Download RBFM v5n1.indb

Editorial
Revista Brasileira de Física Médica.2011;5(1):5-6.
O reconhecimento internacional
da Física Médica
The international acknowledgment of Medical Physics
N
a maior parte dos países desenvolvidos, os físicos médicos são reconhecidos como profissionais imprescindíveis na
área da saúde, recebendo uma compensação salarial condizente com suas atribuições e responsabilidades. No entanto, em alguns países em desenvolvimento, o reconhecimento profissional ainda está em fase de amadurecimento. No Brasil,
os físicos médicos ainda não são considerados profissionais da saúde e costumam ser remunerados inadequadamente
quando comparados aos colegas europeus e norte-americanos.
Os físicos médicos acabaram de ganhar um importante aliado para a valorização de seu trabalho com a inclusão na
Classificação Internacional de Ocupações (ISCO - International Standard Classification of Occupations), uma das principais
classificações internacionais, sob a responsabilidade da Organização Internacional do Trabalho.
Na última versão da classificação da ISCO1, o físico médico está incluído na lista de profissionais do grupo 2111 de físicos
e astrônomos. Na lista de atribuições, um item específico sobre a física médica estabelece que o profissional deve “assegurar
a segurança e a aplicação efetiva da radiação (ionizante e não-ionizante) aos pacientes, de modo a obter um resultado diagnóstico e terapêutico conforme prescrito pelo médico”. Para esclarecer a posição do físico médico, existe uma observação
de que “os físicos médicos são considerados como parte integrante da força de trabalho em saúde, ao lado das ocupações
do subgrupo 22 de profissionais da saúde”. Além disso, no grupo Profissionais da Saúde, há uma nota explícita sobre os
físicos médicos como constituintes da força de trabalho da área de saúde.
A inclusão do físico médico como profissão pela ISCO deve promover a valorização dessa ocupação, sensibilizando
organismos governamentais ao reconhecimento do papel e status desse profissional e contribuindo para a exigência da
presença de físicos médicos qualificados em hospitais e centros de saúde, trabalhando na garantia da qualidade de equipamentos e otimização de procedimentos de diagnóstico e terapia.
O reconhecimento internacional deste profissional é um importante passo para o desenvolvimento da comunidade de
física médica, especialmente nos países em que existe a necessidade de ampliação dos recursos humanos nessa área.
Em 2009, com o intuito de divulgar e ampliar a comunidade de físicos médicos nos países em desenvolvimento, a
International Organization of Medical Physics (IOMP) decidiu sediar sua conferência internacional no Brasil, favorecendo a
participação de profissionais do país e da América Latina.
A 18º edição da International Conference on Medical Physics (ICMP 2011), cujo tema foi “Ciência e Tecnologia para a
Saúde de Todos”, ocorreu de 17 a 20 de abril de 2011, em Porto Alegre, no Rio Grande do Sul, no Centro de Eventos da
Pontifícia Universidade Católica do Rio Grande do Sul (PUCRS). Organizada pela IOMP, pela Asociación Latinamericana de
Física Médica (ALFIM), pela Associação Brasileira de Física Médica (ABFM) e pela Faculdade de Física da PUCRS, a ICMP
2011 foi realizada em conjunto com o 16º Congresso Brasileiro de Física Médica (16º CBFM), que é promovido anualmente
pela ABFM.
A realização da ICMP 2011 em Porto Alegre possibilitou a reunião de físicos médicos, engenheiros e outros profissionais
que atuam nesta área para atividades diversas, como simpósios, workshops, cursos e reuniões que começaram alguns
dias antes da conferência, de 15 a 17 de abril. Um destes eventos foi o 5º Simpósio de Instrumentação e Imagens Médicas
(5º SIIM), promovido em conjunto pela ABFM e pela Sociedade Brasileira de Engenharia Biomédica (SBEB), agregando
físicos, engenheiros biomédicos, tecnólogos e médicos. Neste período também ocorreu o segundo workshop do grupo
HTTG (Health Technology Task Group) da IUPESM (International Union for Physical and Engineering Sciences in Medicine),
que tratou do tema Defining the medical imaging requirements for a health station. Outro evento que reuniu representantes
da América Latina foi a reunião do projeto Regional Meeting to Create a Latin America Network of Medical Professionals on
Radiation Protection of Children (RLA9067/9016/01), financiado pela Agência Internacional de Energia Atômica.
Os participantes da ICMP 2011 tiveram a oportunidade de atualizar-se em seis minicursos nas áreas de radioterapia,
radiologia digital, dosimetria em tomografia computadorizada e quantificação em PET/tomografia computadorizada, com
ministrantes convidados do Brasil (6), Canadá (2), Alemanha (1), Portugal (1) e Estados Unidos da América (13).
A programação da ICMP 2011 incluiu seis sessões plenárias, as quais trataram temáticas inovadoras e de interesse da
comunidade da física médica, tais como:
• Harmful Tissue Effects: Is there always a Dose Threshold?, por Jolyon Hendry, do Gray Institute for Radiation Oncologyand
Biology, University of Oxford (Reino Unido).
Associação Brasileira de Física Médica®
5
Silva AMM
• Novel Dosimetry Concepts based on Nanodosimetry, por Hans Rabus, do Physikalisch-TechnischeBundesanstalt
(Alemanha).
• Clinical Implementation of Volumetric Modulated Arc for Conventionally Fractionated and Stereotactic Body Radiation
Therapy, por Vitali Moiseenko, do Vancouver Cancer Center (Canadá).
• Medical Physicists International Certification: an IOMP Initiative, por Raymond Wu, do Barrow Neurological Institute
(EUA).
• Current Motion Tracking and Motion Correction Technologies for Medical and Preclinical Imaging, por Roger Fulton, da
University of Sydney (Austrália).
• Cell Tracking and In Vivo Single Cell Imaging using MRI and Nanotechnology, por Brian Rutt, da Stanford University
(EUA).
Além das conferências plenárias, as sessões de comunicação científica oral foram coordenadas por 34 palestrantes
convidados, além daquelas compostas por somente comunicações orais, três educacionais e profissionais, duas especiais
da IOMP e cinco mesas-redondas.
Os palestrantes das sessões orais e os moderadores das mesas-redondas eram constituídos de profissionais e pesquisadores convidados da Argentina (3), Austrália (4), Bélgica (1), Brasil (4), Canadá (2), Dinamarca (1), França (1), Alemanha (2),
Malásia (1), México (1), Suécia (1), Suíça (1), Emirados Árabes (1), Inglaterra (3) e Estados Unidos da América (16), e de três
organizações internacionais: a Agência Internacional de Energia Atômica (3), a Organização Panamericana de Saúde (1) e a
Organização Mundial de Saúde (1).
Colaboraram com a ICMP 2011 16 empresas, várias delas expondo materiais e serviços em uma área de 474 m2. Dentre
estas empresas estavam a Varian e a Elekta, que foram também patrocinadoras. O patrocínio das empresas e o apoio de
instituições, como a Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), o Conselho Nacional de
Desenvolvimento Científico e Tecnológico (CNPq), a ABFM, a PUCRS, a IAEA e as Fundações Estaduais de Pesquisa, foram essenciais para a realização da conferência, além de possibilitar uma infraestrutura de excelente qualidade ao público
participante.
A realização da ICMP 2011 foi extremamente produtiva para a comunidade da física médica nacional e internacional, não
apenas sob o aspecto científico, mas também profissional e social. Os participantes puderam encontrar e conversar com
colegas profissionais internacionais, atualizar seus conhecimentos em palestras ministradas pelas maiores autoridades da
física médica mundial e participar de atividades sociais de confraternização.
Indubitavelmente, o encontro teve seu sucesso garantido pela participação do público de mais de 500 estudantes de
graduação e pós-graduação, profissionais e pesquisadores da área de física médica. Além disso, a qualidade da programação e dos trabalhos apresentados deveu-se ao esforço incansável dos membros dos comitês que organizaram a conferência, particularmente do comitê científico, coordenado por Caridad Borrás; do comitê educacional/profissional, coordenado
por Paulo Roberto Costa e do comitê internacional, coordenado por Oswaldo Baffa.
Em relação à produção bibliográfica da conferência, dos quase 500 trabalhos submetidos pelos participantes para apresentação na ICMP 2011, 307 artigos foram aceitos para os diversos temas do evento, sendo 95 para apresentação oral
e 212 como pôsteres. Os temas mais populares, medidos pelo número de trabalhos aceitos, foram: Radiation Dosimetry:
Algorithms, instrumentation and protocols (73), External Beam Radiotherapy (47), Radiation Biology and Radiation Protection
(43) e X-ray Imaging (34).
Esta edição da Revista Brasileira de Física Médica publica uma seleção dos melhores trabalhos apresentados na ICMP
2011, selecionados por um corpo de avaliadores, o qual examinou cuidadosamente os trabalhos submetidos na forma de
artigos completos. A seleção resultou no convite para a publicação de 40 artigos completos em diversas áreas da física
médica, divididos em dois números da Revista Brasileira de Física Médica. Esse seleto grupo de artigos representa apenas
uma pequena parte dos trabalhos apresentados na ICMP 2011. Entretanto, o suplemento com os resumos dos trabalhos
da ICMP 2011 poderá ser acessado pelo site da Revista Brasileira de Física Médica2, fornecendo uma visão mais completa
deste importante evento da física médica recebido no Brasil.
Ana Maria Marques da Silva
Presidente da International Conference on Medical Physics 2011.
Faculdade de Física da Pontifícia Universidade Católica do Rio Grande do Sul. [email protected]
Referências
1
2
6
International Labour Organization. Resolution Concerning Updating the International Standard Classification of Occupations – ISCO-08. [Internet]. [cited 2011 April].
Available at: http://www.ilo.org/public/english/bureau/stat/isco/isco08/index.htm.
Revista Brasileira de Física Médica. 2011. [Internet]. [cited 2011 December 7]. Available at: http://www.abfm.org.br/rbfm.
Revista Brasileira de Física Médica.2011;5(1):5-6.
Editorial
Revista Brasileira de Física Médica.2011;5(1):7-8.
The international acknowledgment
of Medical Physics
O reconhecimento internacional da Física Médica
I
n most developed countries, the medical physicists are acknowledged as essential professionals for the health field, thus
earning a salary that corresponds to their responsibilities and activities. However, in some developing countries, professional acknowledgement is still in progress. In Brazil, medical physicists are not considered as health professionals yet, and
usually do not earn as much as their European and North-American peers.
Medical physicists have just gained an ally with the inclusion of their work in the International Standard Classification of
Occupations (ISCO), one of the main international classifications organized by the International Labor Organization (ILO).
In the last version of ISCO1, the medical physicist is included in the list of professionals of group 2111, comprised
of physicists and astronomers. There is one specific item in the list of responsibilities concerning medical physics
that establishes that the professional should “ensure the safety and the effective application of radiation (ionizing and
non-ionizing) in order to obtain diagnostics and therapy results according to medical prescription”. In order to clarify
the position of the medical physicist, the observation is that “medical physicists are considered as part of the health
workforce, together with the occupations of subgroup 22, comprised of health professionals”. Besides, in the group
of health professionals there is a note about medical physicists as being part of the health workforce.
The inclusion of the medical physicist as a profession by ISCO should value this occupation, thus sensitizing governmental institutions to acknowledge the role and status of this professional, also contributing with the mandatory
presence of qualified medical physicists in hospitals and health centers to ensure the quality of equipment and the
optimization of diagnostic and therapy procedures.
The international acknowledgement of this professional is an important step for the development of the medical
physics community, especially in countries with the need to increase human resources in this field.
In 2009, aiming to publicize and increase the community of medical physicists in developing countries, the
International Organization of Medical Physics (IOMP) decided to hold its international conference in Brazil, thus favoring the participation of local and Latin American professionals
The 18th edition of the International Conference on Medical Physics (ICMP 2011), whose subject was “Science
and Technology of all”, was held from April 17 to 20, 2011, in Porto Alegre, Rio Grande do Sul, in the events center of
Pontifícia Universidade Católica do Rio Grande do Sul (PUCRS). Organized by IOMP, by Asociación Latinamericana
de Física Médica (ALFIM), by the Brazilian Association of Medical Physics (ABMF), and by the School of Physics in
PUCRS, ICMP 2011 was held with the 16th Brazilian Congress of Medical Physics (16th CBFM), which is annually
organized by ABFM.
ICMP 2011, in Porto Alegre, enabled the reunion of medical physicists, engineers and other professionals that work
in this field for different activities, such as symposiums, workshops, courses and meetings that started a few days
before the conference, from April 15 to 17. One of these events was the 5th Symposium of Instrumentation and
Medical Imaging (5th SIIM), promoted by ABDM and the Brazilian Society of Biomedical Engineering (SBEB), which
brought together physicists, biomedical engineers, technologists and doctors. In this period, there was also the second workshop of the Health Technology Task Group (HTTG), which is part of the International Union for Physical and
Engineering Sciences in Medicine; the subject was “defining the medical imaging requirements for a health station”.
Another event that gathered people from Latin America was the Regional Meeting to Create a Latin America Network
of Medical Professionals on Radiation Protection of Children (RLA9067/9016/01), financed by the International Atomic
Energy Agency.
The participants of ICMP 2011 had the opportunity to catch up in six mini courses about radiotherapy, digital
radiology, computed tomography dosimetry and PET quantification/computed tomography, with lecturers from Brazil
(6), Canada (2), Germany (1), Portugal (1) and the United States of America (13).
The schedule of ICMP 2011 included six plenary sessions that discussed innovative subjects that interested the
medical physics community, such as:
• Harmful Tissue Effects: Is there always a Dose Threshold?, by Jolyon Hendry, of the Gray Institute for Radiation
Oncologyand Biology, University of Oxford (the United Kingdom).
Associação Brasileira de Física Médica®
7
Silva AMM
• Novel Dosimetry Concepts based on Nanodosimetry, by Hans Rabus, of the Physikalisch-TechnischeBundesanstalt
(Germany).
• Clinical Implementation of Volumetric Modulated Arc for Conventionally Fractionated and Stereotactic Body Radiation
Therapy, by Vitali Moiseenko, of the Vancouver Cancer Center (Canada.
• Medical Physicists International Certification: an IOMP Initiative, by Raymond Wu, of the Barrow Neurological Institute (USA).
• Current Motion Tracking and Motion Correction Technologies for Medical and Preclinical Imaging, by Roger Fulton, of
the University of Sydney (Australia).
• Cell Tracking and In Vivo Single Cell Imaging using MRI and Nanotechnology, by Brian Rutt, of Stanford University (USA).
Besides the plenary sessions, the oral scientific communication sessions were coordinated by 34 invited lecturers,
besides those comprised only of oral communications; there were three educational and professional sessions, two
special ones for IOMP and five discussion rounds.
The lecturers of the oral sessions and the moderators of the discussion rounds were professionals and researchers from Argentina (3), Australia (4), Belgium (1), Brazil (4), Canada (2), Denmark (1), France (1), Germany (2), Malasia
(1), Mexico (1), Sweden (1), Switzerland (1), Arab Emirates (1), England (3), and the United States of America (16), the
Pan American Health Organization (1) and the World Health Organization (1).
Sixteen companies cooperated with ICMP 2011, and many of them exposed materials and services in a 474 m2
area. Among these companies were Varian and Elekta, which were also sponsors. The sponsorship of companies and
the support of institutions such as the Coordination for the Improvement of Higher Level Education Personnel (CAPES),
the National Council for Scientific and Technological Development (CNPq), ABFM, PUCRS, IAEA and the State Research
Foundations were essential to conduct the conference, besides enabling excellent infrastructure to the participants.
ICMP 2011 was extremely productive for the international and national medical physics community, not only
concerning the scientific aspect, but also the professional and social features. The participants could meet and talk
to foreign colleagues, update their knowledge in lectures given by authorities of the world medical physics and participate in social activities.
There is no doubt that the meeting was successful due to the participation of more than 500 graduate and
postgraduate students, professionals and medical physics researchers. Besides, the quality of the schedule and the
presented papers was due to the restless effort of the members of the committees that organized the conference,
specially the scientific committee, coordinated by Caridad Borrás; the educational/professional committee, coordinated by Paulo Roberto Costa; and the international committee, coordinated by Oswaldo Baffa.
As to the bibliographic production of the conference, out of the 500 papers submitted by the participants to be
presented at ICMP 2011, 307 articles were accepted concerning different subjects: 95 for oral presentation and 212
as posters. The most popular themes, measured by the number of accepted papers, were: Radiation Dosimetry:
Algorithms, instrumentation and protocols (73), External Beam Radiotherapy (47), Radiation Biology and Radiation
Protection (43) and X-ray Imaging (34).
This issue of Revista Brasileira de Física Médica has a selection of the best papers presented at ICMP 2011,
chosen by a group of evaluators that carefully examined the papers submitted as full articles. This selection led to
the invitation to publish 40 full articles of different fields of medical physics, divided in two issues of Revista Brasileira
de Fìsica Médica. This special set of articles represents only a small part of the papers presented at ICMP 2011.
However, the supplement with the abstracts of the papers in ICMP 2011 can be accessed in the website of Revista
Brasileira de Física Médica2, thus providing a wider vision of this important event of medical physics held in Brazil.
Ana Maria Marques da Silva
President of the International Conference on Medical Physics 2011.
School of Physics at Pontifícia Universidade Católica do Rio Grande do Sul. [email protected]
References
1
2
8
International Labour Organization. Resolution Concerning Updating the International Standard Classification of Occupations – ISCO-08. [Internet]. [cited 2011
April]. Available at: http://www.ilo.org/public/english/bureau/stat/isco/isco08/index.htm.
Revista Brasileira de Física Médica. 2011. [Internet]. [cited 2011 December 7]. Available at: http://www.abfm.org.br/rbfm.
Revista Brasileira de Física Médica.2011;5(1):7-8.
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):9-14.
Development of a parallel plate ion
chamber for radiation protection level
Desenvolvimento de uma câmara de ionização
de placas paralelas para proteção radiológica
Márcio Bottaro1, Maurício Moralles2 and Maurício Landi1
1
Instituto de Eletrotécnica e Energia/Seção Técnica de Ensaios em
Equipamentos Eletromédicos (STEEE), Universidade de São Paulo, SP, Brasil.
2
Instituto de Pesquisas Energéticas e Nucleares/Centro do Reator Nuclear de Pesquisas (CRPq),
Comissão Nacional de Energia Nuclear, São Paulo, SP, Brasil.
Abstract
A new parallel plate vented ion chamber is proposed in this paper. The application of this chamber was primarily intended to the measurement of
stray radiation in interventional procedures, but the energy response of about 2.6%, which was obtained in the first prototype, on the range from
40 to 150 kV using ISO 4037-1 narrow qualities, provided the possibility of a wide modality application on radiation protection. Primary studies with
Maxwell 2D electromagnetic field simulator revealed an optimized model regarding effective volume and saturation voltage levels, which conferred
to the ion chamber a dual entrance window feature. The development of this ion chamber has the main contribution of Monte Carlo calculations as a
support tool to the establishment of the effective volume of the chamber and determination of the best materials for housing mounting and conductive
elements, such as guard rings, electrode, and windows. Even the composition of the conductive layers, which would be neglected due to their very
small thicknesses (about 35 µm), had important influence on the results and could be better understood with Monte Carlo N-Particle Transport Code
System (MCNP) simulations.
Keywords: ion chamber, Monte Carlo, energy response, radiation protection.
Resumo
Uma nova câmara de ionização de placas paralelas ventilada é proposta neste trabalho. A aplicação da câmara teve como objetivo principal a
medição da radiação parasita nos procedimentos intervencionistas, porém as variações da resposta em energia de aproximadamente 2,6% na faixa
de 40 a 150 kV, obtida no primeiro protótipo utilizando os feixes padrão estreitos da ISO 4037-1, possibilitou uma ampla aplicação na modalidade
de proteção radiológica. Estudos iniciais feitos com o simulador de campo eletromagnético Maxwell 2D revelaram um modelo otimizado em relação
ao volume efetivo e tensão de saturação, os quais conferiram à câmara de ionização a característica de janela de entrada dupla. O desenvolvimento
desta câmara de ionização teve como principal contribuição as simulações de Monte Carlo como uma ferramenta de suporte para o estabelecimento
do volume efetivo da câmara e para a determinação dos melhores materiais para os elementos de montagem e circuito condutivo, como por
exemplo, os anéis de guarda, eletrodo e as janelas. Até mesmo as composições de camadas condutivas, que seriam negligenciadas devido a sua
pequena espessura (aproximadamente 35 µm), tiveram uma importante influência nos resultados, que foi melhor compreendida com as simulações
realizadas com o Monte Carlo N-Particle Transport Code System (MCNP).
Palavras-chave: câmara de ionização, Monte Carlo, resposta da energia, proteção radiológica.
Introduction
Interventional radiology equipments are extensively used in
medical practice. In the last decades, the minimally invasive
procedures, associated with technological improvements,
resulted in the expansion of the equipment market all over
the world. In Brazil, this panorama is not different and an
increasing number of manufacturers are putting much effort in the development and production of interventional
radiology machines to supply market necessities1.
Medical electrical equipment certification process is
compulsory in Brazil since 1995. In this process, type tests of
these equipments are performed according to international
standards from the International Electrical Commission (IEC)
60601 series. For interventional radiology equipment, such
tests are also performed and some requirements of the applied IEC standards require particular instruments for the X
radiation measurements, especially leakage and stray radiation profiles in significant zones of occupancy. These requirements include special dimensions and volume chambers2.
Corresponding author: Marcio Bottaro – Instituto de Eletrotécnica e Energia da Universidade de São Paulo (USP) – Professor Luciano Gualberto, 1.289 –
São Paulo (SP) – Brazil – E-mail: [email protected]
Associação Brasileira de Física Médica®
9
Bottaro M, Moralles M, Landi M
For leakage radiation measurements, the most important requirements are the entrance window area of 100 cm² and linear dimensions not exceeding 20 cm. In this modality, some
commercial chambers are available and the evaluation of their
time and dose response function is very important. For the
stray radiation profiles in significant zones of occupancy and
also for measurements of isokerma maps of scattered radiation, the most important requirement is related to the volume
of the chamber, 500 cm³, and linear dimensions that cannot
exceed 20 cm. For this modality, no chambers are commercially available, since the old FLUKE 96010A was discontinued. Although FLUKE 96010A regards requirements for both
leakage and stray radiation, it was designed with the specific
purpose of leakage radiation measurement.
This paper presents a new purpose of vented parallel
plate ionization chamber, according to the dimensions and
volume requirements of IEC 60601 international standards
series and regarding all the modalities of diagnostic X-ray
equipment. It was specially designed to present energy response and sensibility necessary to the measurements of
leakage radiation, stray radiation profiles in significant zones
of occupancy, and scattered radiation in isokerma maps.
In the next section, the design method using Maxwell
2D electromagnetic field simulator and the Monte Carlo
calculation code MCNP4C is presented3. Practical measurements and simulations of energy response, based on
normalized X-ray qualities4, are reported in Results and
the fundamental contributions of Monte Carlo calculations
to understand the chamber behavior are discussed in
Discussion and Conclusions sections.
In Figure 2 one observes that there is a large area (orange) corresponding to the effective volume of the chamber, where the charges produced by ionization are collected. There are also two types of regions that determine
dead volumes, where the produced charges are not collected due to different reasons. The first type of dead volume is represented from blue to yellow, where the electric
field is very weak and corresponds to the region outside of
the space between the collecting electrodes. The second
type of dead volume corresponds to the region between
the guard ring and the window, shown in red on the lower
corners, which presents a higher electric field that deviates
the charges to the guard ring, and consequently they are
not collected by the measuring electronics.
Second chamber model
With the purpose to obtain a smaller dead volume, a second model of chamber was designed, which had almost
the same material components, PMMA walls but now with
two polycarbonate windows and a centralized polycarbonate collector electrode. All components were covered with
conductive graphite to provide the chamber polarization
and the charge collection circuit. The guard rings were
placed at a different position, still surrounding the collecting
electrode, but dividing the window circuit and chamber design symmetrically. This confers to the second model other
interesting properties: reduced saturation voltage, reduced
dead volumes and bilateral capacity of measurement. This
new model is shown in Figure 3. Results of Maxwell 2D
calculations are illustrated in Figure 4.
The electromagnetic field calculation for the second
chamber was performed with 250 Vdc, almost half of the
Materials and methods
Electrical field design and simulation
The parallel plate ion chamber was designed with three
fundamental aims: a low cost, robust, and easy manipulation detector. Such objectives lead to a first simple design
with a single entrance window and a collector electrode
surrounded by a guard ring.
First chamber model
The first chamber had polimetilmetacrilate PPMA cylindrical
walls (red), windows and collector electrode (blue) made of
polycarbonate, covered with conductive graphite (gray) in
a way to state a chamber polarization and charge collection circuit. All the internal volume is filled with air (white) as
in a vented chamber. This first model is shown in Figure 1.
It was obtained from Maxwell 2D design tool and indicates
a section of the central axis perpendicular to the window
plane, where all the components can be seen. As with
this free software, we are able only to simulate 2D electric
fields, this section was chosen as a more representative
plane to evaluate electrical field design and behavior.
The electrical filed simulation obtained with the use of
Maxwell 2D software is shown in Figure 2, and it was performed with a chamber voltage of 400 Vdc.
10
Revista Brasileira de Física Médica.2011;5(1):9-14.
D - Air
A - PMMA
C-Graphite
B - Polycarbonate
Figure 1. Section of the first chamber model design used in the
electrical field simulation.
E [V/M]
2.9715e+005
2.7062e+004
2.4647e+003
2.2446e+002
2.0442e+001
1.8617e+000
1.6955e+001
1.5442e+002
1.4063e+003
1.2808e+004
1.1664e+005
1.0623e+006
9.6747e+008
8.8111e+009
8.0245e+010
7.3081e+011
Figure 2. First chamber model electrical field simulation.
Development of a parallel plate ion chamber for radiation protection level
voltage used in the first chamber model. Figure 4 shows
that the regions corresponding to dead volumes and low
electric fields inside the collecting electrodes are significantly smaller than in the first chamber.
Prototype and Monte Carlo model
Prototype
The first prototype model is illustrated in Figure 5.
Water-based graphite was used to avoid corrosion
of the PMMA, since almost all solvents in the most common graphite inks promote PMMA degradation with
time. A painting procedure was developed in order to
guarantee uniform ink distribution and thickness. A
guiding layer was designed to facilitate the painting of
the guard rings. The ion-chamber wiring was performed
by pressing the conductors against the internal side of
the windows and with nylon screws in guard rings and
electrode.
Monte Carlo model
The same ion chamber was also modeled in the MCNP4C3
code to evaluate the energy response of the detector. A
simplified geometry was stated, and details of wiring and
internal nylon screws were neglected. In Monte Carlo calculations, null electron and photon importance was assigned to the dead volume and dead zone, because in
MCNP4C the electric field would not be inserted in the
input code. The components of walls (PMMA), windows
and collector electrode (polycarbonate) were modeled
using standard compositions. The chemical composition
D - Air
C-Graphite
of the graphite cover was modeled using the graphite ink
manufacturer’s specifications for dry ink: 100% graphite.
The graphite thickness was based on the measurements
made in the prototype chamber layers, with and without
graphite coating, using a precise digital Mitutoyo micrometer model 389-251. The graphite mean thickness was
35 µm.
Measurement setup and qualities
As the main application of the detector is related to radiation protection, ISO 4037-1 narrow beams were selected
within diagnostic range4. For the first evaluation, four X-ray
qualities were used (N60, N80, N100 and N150 kV), based
on the available qualities in IEE/USP and in IPEN/CMR,
which are the reference laboratories for ion chamber calibrations in São Paulo. The reference kerma/ion chamber
charge ratio was used to state the energy response evaluation and Monte Carlo comparison parameter.
The same qualities were implemented in STEEE/IEE,
according to ISO 4037-1, where a PTW Freiburg GmbH
300 cm³ TA34055-0 model ion chamber was used to state
the reference doses.
The measurements were then performed in both laboratories and a PTW UNIDOS 457 electrometer was used to collect
charge of the ion chamber prototype during tests. In both laboratories, results were taken using the reference ratio (Eq. 1):
A - PMMA
B - Polycarbonate
Figure 3. Section of the second chamber model design used in
the electrical field simulation.
Figure 5. First chamber prototype.
2.6494e+005
2.5256e+004
2.4075e+003
2.2949e+002
2.1876e+001
1.0854e+000
1.9879e+001
1.8949e+002
1.8063e+003
1.7219e+004
1.6414e+005
1.5646e+006
1.4915e+007
1.4217e+008
1.3553e+009
1.2919e+010
Figure 4. First chamber model design section for electrical field
simulation.
Ratio [C/Gy]
E [V/m]
Reference Ratio Results
3,3E-05
3,1E-05
2,9E-05
2,7E-05
2,5E-05
2,3E-05
2,1E-05
1,9E-05
1,7E-05
1,5E-05
70
170
50
90
110
130
150
ISO 4037-1 narrow quality reference voltage [kV]
IPEN
MCNP
STEEE
Figure 6. First results of Kerma/Charge ratio in reference and
development laboratories and MCNP simulations.
Revista Brasileira de Física Médica.2011;5(1):9-14.
11
Bottaro M, Moralles M, Landi M
Rn =
Cn
Kn
(1)
where:
R is the dose/charge ratio;
K is the laboratory reference air kerma;
and C is the ionization chamber prototype collected charge.
All parameters are related to their ISO narrow quality n.
Ratio [C/Gy]
The same equation was employed for the results of
the simulated data. The setup geometry of each laboratory was implemented in MCNP, and the input spectra
were obtained from earlier simulations in GEANT4 code5.
For the MCNP simulations, the reference air kerma K was
obtained in a prior simulation setup with all narrow qualities over an air volume of 500 cm³ accomplishing the air
kerma definitions. Further simulations with the ion chamber model were performed to collect data of charge C in
the predefined effective volume.
Comparisons between experimental and simulated
data are shown in the next section.
2nd Reference Ratio Results
3,3E-05
3,1E-05
2,9E-05
2,7E-05
2,5E-05
2,3E-05
2,1E-05
1,9E-05
1,7E-05
1,5E-05
70
170
50
90
110
130
150
ISO 4037-1 narrow quality reference voltage [kV]
IPEN
STEEE
MCNP
Ratio [C/Gy]
Final Reference Ratio Results
2,40E-05
2,30E-05
2,20E-05
2,10E-05
2,00E-05
1,90E-05
1,80E-05
1,70E-05
1,60E-05
1,50E-05
50
70
90
110
130 150
170
ISO 4037-1 narrow quality reference voltage [kV]
Ratio [C/Gy]
Figure 7. First results of Kerma/Charge ratio in reference and development laboratories and second results of MCNP simulations.
Weighting Residuals
1,50E+00
1,00E+00
5,00E+01
0,00E+00
70
90
110
130
150
-5,00E+01 50
-1,00E+00
ISO 4037-1 narrow quality reference voltage [kV]
MCNP
STEEE
Residuals
Figure 8. Final results of Kerma/Charge ratio in reference and
development laboratories and MCNP simulations.
12
Revista Brasileira de Física Médica.2011;5(1):9-14.
Results
A summary of results containing charge/kerma ratio is
presented in the graphic of Figure 6. A good agreement
between both laboratories results is clearly seen; however, a large discrepancy with simulated results is also
presented. Both window sides were tested and no relevant differences were found.
Figure 6 clearly presents a detector response with considerable energy dependency. As ISO Narrow 70 kV quality
was not available, an indication of the energy dependency
was based on the c variation coefficient of the results. For
the first results, the energy dependencies were very similar for both laboratories, 13.4% in IPEN/CMR and 10.8%
in STEEE/IEE. However, in MCNP simulations, the result
was 6.9%. All of them were not in agreement with ISO
4037-1 energy dependency requirements of less than 5%.
Such discrepancies were evaluated, and two their possible
sources were studied: MCNP cross-section library used
and composition of the dry graphite ink.
While a new cross-section library was implemented,
a spectroscopic characterization of the graphite ink was
performed and a new chemical composition was found,
which is different from the one specified by the manufacturer. The main components found, other than the
graphite, were: SiO2, CaO, MgO, Al2O3, and Na2O. With
such results, a new chemical model for graphite layer
was implemented in the MCNP code. The new results are
shown in Figure 7.
Both libraries that were used produced almost the
same results, and a new graphite ink was used in a second prototype. The composition of this new ink was also
evaluated by means of spectroscopy procedure. For this
graphite ink, no other elements than carbon (graphite)
were found and a third series of simulations was then performed. The results were very similar to that obtained in
laboratories. Small adjustments in the dead volume were
performed to fit results more adequately. The new radiation
protection ion chamber had its MCNP model defined.
New data are plotted in Figure 8. The results show
good agreement between experimental and simulated results. Unfortunately, no measurements could be performed
in the reference laboratory of IPEN/CMR, but STEEE/IEE
ISO qualities were found to be adequate, as they were validated in earlier measurements. Once again, both window
sides were tested with no significant differences. A good
energy response was also observed for both experimental
and simulated results. In Figure 8 the uncertainties are also
shown within 68% of confidence level, and weighting residuals were calculated in order to compare data.
Residuals are within one standard deviation, and the
new data showed satisfactory results. The energy dependence of experimental and simulated data was also calculated: 2.1% for simulations and 2.6% for STEEE/IEE
laboratory measurements. These results are within ISO
4037-1 specifications of 5%. Further tests and simulations
with other interest spectra and practical application in type
Development of a parallel plate ion chamber for radiation protection level
tests in IEE laboratories are about to be performed, in order to attest compliance of the detector to be applied in
radiation protection measurements.
Discussion
The parallel plate vented chamber presented many successful results and its knowledge could also be improved
with some additional simulations to better understand its
full energy response in the diagnostic range, including
mammographic applications. This is very important as
new IEC standard series are required for all modalities of
diagnostic equipment, including mammographic, dental,
fluoroscopic, conventional, and computed tomography
X-ray generators, and the determination of stray radiation
profiles. The bilateral property was confirmed in the practical tests with no relevant dependence during the measurements. Nevertheless, other practical measurements
and performance tests should be executed to attest its
compliance with international standards and to validate
this detector for use on practical type tests.
Conclusions
Monte Carlo calculation is a very important tool in the development of many radiation detectors. The present study
showed that it can be also applied in the development
of ionization chambers. The parallel plate vented chamber proposed in the present paper corresponds to the
final prototype version. Based on MCNP results, studies
regarding the best materials to be used in the chamber
housing and main parts were performed, and a chamber
model that can be further used to simulate its behaviour on
other X-ray qualities was stated.
A satisfactory energy response was achieved and other performance tests are under execution in the laboratory,
accompanied by MCNP calculations.
Acknowledgment
Authors would like to thank STEEE staff, they were always ready to help and contribute during laboratorial
measurements. Also, thanks to IPEN/CMR for the attention and support.
References
1. Canevaro L (2009) Physical and technical aspects in Interventional
Radiology. Revista Brasileira de Física Médica 3(1):101-15
2. Associação Brasileira de Normas Técnicas. Equipamento eletromédico –
Parte 1: Prescrições gerais de segurança 3. Norma Colateral: Prescrições
gerais para proteção contra radiação de equipamentos de raios X para fins
diagnósticos. ABNT, Rio de Janeiro, 2001 (NBR IEC 60601-1-3)
3. Briesmeister J. MCNP (2000) A general Monte Carlo N-particle transport
code, version 4C. Los Alamos National Laboratory Report, LA-13709-M
4. International Organization for Standardization (1996) X and gamma
reference radiation for calibration dosemeters and doserate meters and for
determining their response as a function of photon energy – Part 1: Radiation
characteristics and production methods. Geneva Switzerland (ISO 4037-1)
5. Guimarães C C, Moralles M, Okuno E (2008) Performance of GEANT4 in
dosimetry applications: Calculation of X-ray spectra and kerma-to-dose
equivalent conversion coefficients. Rad. Meas. 43:1525 DOI 10.1016/j.
radmeas.2008.07.001
Revista Brasileira de Física Médica.2011;5(1):9-14.
13
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):15-20.
Calibration of PKA meters against ion
chambers of two geometries
Calibração de medidores de PKA contra câmaras de
ionização de duas geometrias
José N. Almeida Jr.1, Ricardo A. Terini1, Marco A.G. Pereira2 and Silvio B. Herdade2
2
1
Pontifícia Universidade Católica de São Paulo (PUC-SP); Departamento de Física, São Paulo (SP), Brazil.
Universidade de São Paulo; Instituto de Eletrotécnica e Energia (IEE-USP); Seção Técnica de Desenvolvimento
Tecnológico em Saúde (STDTS), São Paulo (SP), Brazil.
Abstract
Kerma-area product (KAP or PKA) is a quantity that is independent of the distance to the X-ray tube focal spot and that can be used in radiological
exams to assess the effective dose in patients. Clinical KAP meters are generally fixed in tube output and they are usually calibrated on-site by
measuring the air kerma with an ion chamber and by evaluating the irradiated area by means of a radiographic image. Recently, a device was
marketed (PDC, Patient Dose Calibrator, Radcal Co.), which was designed for calibrating clinical KAP meters with traceability to a standard laboratory.
This paper presents a metrological evaluation of two methods that can be used in standard laboratories for the calibration of this device, namely,
against a reference 30 cc ionization chamber or a reference parallel plates monitor chamber. Lower energy dependence was also obtained when the
PDC calibration was made with the monitor chamber. Results are also shown of applying the PDC in hospital environment to the cross calibration of
a clinical KAP meter from a radiology equipment. Results confirm lower energy dependence of the PDC relatively to the tested clinical meter.
Keywords: air kerma-area product, dosimetry, calibration, KAP meters, radiology, ionization chambers.
Resumo
A grandeza produto kerma-área (PKA) independe da distância ao foco do tubo de raios X e pode ser usada nos exames radiológicos para avaliar a
dose efetiva nos pacientes. Medidores clínicos de PKA são geralmente fixados na saída do tubo e usualmente calibrados no local, por meio da medição
do kerma no ar com uma câmara de ionização e da avaliação da área irradiada utilizando uma imagem radiográfica. Recentemente, foi projetado e
comercializado um dispositivo para calibrar medidores clínicos de PKA (PDC, Patient Dose Calibrator – Calibrador da dose do paciente, Radcal Co.),
com rastreabilidade a um laboratório padrão. Este trabalho apresenta uma avaliação metrológica de dois métodos que podem ser utilizados em
laboratórios padrão para calibrar tal dispositivo, ou seja, contra uma câmara de ionização de 30 cc de referência ou uma câmara de monitora de
placas paralelas. Menor dependência energética foi obtida quando a calibração do PDC foi realizada com a câmara de monitora. São mostrados
também resultados do PDC aplicado em um ambiente hospitalar para a calibração cruzada de um medidor clínico de PKA de um equipamento
radiológico. Os resultados confirmam menor dependência energética do PDC em relação ao medidor clínico testado.
Palavras-chave: produto kerma-área de ar, dosimetria, calibração, medidor de produto kerma-área, radiologia, câmaras de ionização.
Introduction
Due to the quantity and frequency with which clinical
examinations are performed, the dose released in diagnostic radiology and interventional procedures should
be accurately determined so as to maintain a reasonable balance between image quality and absorbed dose
to patients.
The more appropriate quantity to express, the levels
of exposure to radiation is the effective dose (E), which
cannot be directly measured1. It can however be obtained
through the quantity air kerma-area product (KAP or PKA),
whose value, by definition, is constant with the distance
between focal spot and patient2.
The issue has special relevance in Brazil, since there
are still few national references about the subject3, reduced
clinical use (yet), despite the recommendation of international standards4, and the need for calibration of PKA meters preferably in Brazil itself, in order to meet the demand
that tends to grow soon.
PKA meters differ from common ionization chambers,
since in those its sensitive volume is only partially irradiated.
Corresponding author: Ricardo A. Terini – Pontifícia Universidade Católica de São Paulo – Rua Marquês de Paranaguá, 111 – Consolação – São Paulo (SP),
Brazil – CEP 01303-050 – E-mail: [email protected]
Associação Brasileira de Física Médica®
15
Almeida Jr. JN, Terini RA, Pereira MAG, Herdade SB
Therefore, in the PKA meters calibration process, which
uses a totally irradiated reference chamber, it is necessary
a method to evaluate the irradiated area which results in an
undesirable increase of the overall uncertainties.
Recently, a device named PDC (Patient Dose
Calibrator, Radcal Co.), which was designed for calibrating clinical PKA meters with traceability to a standard calibration laboratory, was commercialized. This paper had
the aim of investigating two methodologies to calibrate
the PDC PKa meter in laboratory, also analyzing the influence quantities in this process, and minimizing uncertainties in such calibration for direct standard X-ray beams5:
first, against a reference 30 cm3 ion chamber and, second, against a monitor chamber, crossed by the entire
beam, as the PDC. Then, an application was made of
cross calibration of a clinical PKA meter using the previously calibrated PDC.
Materials and methods
Used equipment
All experimental work was made in the Laboratory of
Ionizing Radiations Metrology (LMRI) of the Instituto de
Eletrotécnica e Energia da Universidade de São Paulo
(IEE-USP), using its infrastructure. A constant potential
Philips X-ray equipment was used, which consists of a
bipolar high-voltage generator MG 325 (ripple ≤ 1%), adjustable from 15 to 320 kV, a metal-ceramic tube model
MCN 323 with tungsten anode, large focal spot size 4 mm,
anode angle of 22o, 4 mm Beryllium window, and a control unit model MGC 40. The beam was collimated by a
set of 2 mm thick lead collimators with known apertures,
and filtered by 99.0 to 99.5% purity Al sheets. For the PKA
determination, a reference collimator, 4.5 mm in thickness
and 10.8 cm in aperture diameter, was used near the detector position.
High-voltage waveforms have been invasively acquired by means of the computational acquisition and
A/D channel card of a Tektronix TDS5104 oscilloscope,
and voltage parameters, such as average peak voltage (kVp ave) and practical peak voltage (PPV, as defined
by IEC 61676 standard6), have been calculated by a
LabView (National Instruments) routine developed at
IEE-USP.
Two calibrated detectors have been used for the air
kerma measurements: a PTW TN34014 model monitor
chamber, with carbon coated surfaces, and a reference
PTW 23361 model 30 cm3 cylindrical chamber, both connected to PTW Unidos electrometers.
Setup alignment was made with the help of an optical bench, laser beams and semi-transparent mirrors.
Atmospheric pressure and temperature were measured by
an Oregon Scientific Co. calibrated meter in order to correct all air kerma readings by air density influence.
The investigated PKA meter was a Radcal detector
model PDC, borrowed by Radcal Co.
16
Revista Brasileira de Física Médica.2011;5(1):15-20.
Methodology
X-ray beams characterization
The whole set of conventional radiodiagnostic (RQR) standard X-ray beams from IEC 61267:20057 has been previously characterized in the Philips equipment, using the
reference 30 cm3 ion chamber and 99.9% purity Al filters
for the half value layers (HVL) determinations. First and
second HVL values have been obtained by a logarithmic
interpolation method between two points measured before
and after the HVL corresponding thickness.
kVpave values from Philips equipment, read with the
LabView routine, have been calibrated by means of measured incident beam X-ray spectra obtained with a CdTe
spectrometer (Amptek, Inc.) calibrated, in turn, using
X- and γ-ray known energies from Am-241 and Ba-133
calibrated radioactive sources. The average value of the
maximum spectral energy E (in keV) is numerically equal to
the value of the average peak tube voltage (kVpave, in V), for
low ripple voltage waveforms. In this work8, the endpoint
energy E was determined by a least squares linear regression procedure at the higher energy part of each measured
beam spectrum. Applying this calibration to the whole voltage waveform, PPV values were determined by its definition6 with uncertainties lower than 0.5 kV (k=2), for each
standard beam.
PDC calibration against 30 cm3 reference chamber
In these measurements, monitor chamber has only been
used for correction of X-ray tube output changes. Using the
substitution method, the reference chamber and the PDC
meter were alternatively put in the X-ray beam axis, 100 cm
distant from the focal spot, in each characterized beam.
Each PKA rate value from PDC (PKAPDC) was obtained
as an average of five readings, after correction for air
density influence. For the 30 cm3 chamber, the average
air kerma rate (Kairref) was obtained after correction for air
density and calibration factor. Aperture area of the reference lead collimator (ACOL), placed 8.5 cm in front of the
detector, was determined and, thus, the reference PKA
rate, PKAref, was calculated in each case. Calibration factors for PKA rate readings from PDC detector have been
obtained by Eq. 1:
(1)
In Eq. 1, dFD and dFC are the distances between focal
point and, respectively, reference detector or collimator,
which values are 100.0 and 91.5 cm, both with 0.1 cm
estimated uncertainty (k=1).
PDC calibration against monitor chamber as reference
In these measurements, a monitor chamber has been
moved to a place between detector and reference collimator positions and, firstly, it was calibrated against the
Calibration of PKA meters against ion chambers of two geometries
cylindrical chamber placed in detector position. After this,
the 30 cm3 chamber was substituted by the PDC in such a
way that, in these measurements, the same X-ray beams
crossed both detectors simultaneously (Figures 1 and 2).
In the calibration of the monitor chamber (Figure 1),
each air kerma rate value from reference chamber, Kairref, or
each charge rate value from monitor chamber, Qmon, was
obtained as an average of five readings after correction for
air density. Calibration factors for monitor chamber charge
readings, corrected for distance, in this case, have been
obtained by Eq. 2:
(mGy/nC)
(2)
where:
dFD and dFM are the distances between focal point and,
respectively, reference detector and monitor chamber,
which values were 98.9 and 65.5 cm, both with 0.1 cm
estimated uncertainties (k=1).
Figure 1. Positioning of reference 30 cm3 chamber and graphite
coated monitor chamber in the monitor calibration procedure.
For the calibration of the PDC against the calibrated monitor chamber (Figure 2), each PKA rate value from PDC, PKAPDC,
was also obtained as an average of five readings after correction for air density. The reference values of PKA rate, PKAMon,
have been obtained through the product of average air kerma
obtained with the calibrated monitor chamber, KairMon (= Qmon
. fCAL-Mon_ref ), and the aperture area of the reference collimator,
ACOL, corrected for distance. Both PKA values were corrected
for air density. Calibration factors for PDC PKA readings, in this
case, were obtained by Equation 3:
(3)
Figure 2. Experimental setup showing the relative positions of
PDC and monitor chamber used as air kerma reference, as well
as the reference collimator.
where:
dFC is the distance between focal spot and reference
collimator, which is 60.7±0.1 cm.
Application of calibrated PDC to verify the calibration of
a clinical PKA meter
As application of the PDC detector, after calibration, a
calibration checking of a clinical PKA meter (ScanditronixIBA) coupled to a Philips Omni X-ray equipment, from
Hospital Israelita Albert Einstein (HIAE), in São Paulo, was
carried out. The PDC was supported 17 cm over the exams table, at a distance of 80.5 cm from the X-ray tube
focal spot (Figure 3). The exposure times were 200 ms.
PKA values were measured with both detectors simultaneously irradiated, in a series of measurements with the following conditions: tube voltage varying from 50 to 120 kV,
current-time product fixed as 50 mAs, for three radiation
fields sizes (15x15, 20x20 and 25x25 cm2), adjusted by
means of the tube collimator.
Figure 3. Scheme of setup used for the simultaneous measurements made with the PDC and the Scanditronix-IBA PKA meter
(coupled to a Philips Omni X-ray equipment), in the clinical environment of HIAE (illustration adapted from www.radcal.com ).
Revista Brasileira de Física Médica.2011;5(1):15-20.
17
Almeida Jr. JN, Terini RA, Pereira MAG, Herdade SB
Results
PDC Calibration factors against cylindrical chamber
Figure 4 shows the obtained data of the PDC calibration
factors versus the PPV for air kerma and air kerma-area
product quantities, which were obtained from the measurements against the cylindrical reference chamber.
PDC Calibration factors against monitor chamber
Table 1 shows the obtained results of the calibration factors for the monitor chamber against the cylindrical chamber. In Table 2, we have the results of the PDC calibration made against the calibrated monitor chamber as the
reference detector, for the same characterized standard
beams. Uncertainties appear between brackets to show
only the less significant figure.
Verification of calibration of the clinical PKA meter using
the calibrated PDC
Figure 5 shows the results obtained from measurements
made with PDC and the Scanditronix-IBA PKA meter coupled to Philips Omni X-ray equipment, for some field sizes.
The PKA values showed by the two meters, obtained in a
clinical setting in HIAE, were not corrected for air density
effects, since they were not monitored for temperature and
pressure at the site. Linearity of both meters were checked
all over the investigated intensities range (up to 700 µGy.m2),
and R-coefficient was better than 0.999.
Figure 4. Energy dependence curve of PDC detector versus PPV
values, for PKA and Kair values, measured against the reference
30 cm3 chamber, for RQR standard beams7. All error bars are
shown for k=1.
Table 1. Average air kerma rate values (Kairref) from reference
30 cm3 chamber and charge rate values (QMon) from monitor
chamber (corrected for reference chamber position), as well as
the obtained monitor chamber calibration factors (fCAL-Mon_ref), for
three standard direct beams
Standard
beam
RQR3
RQR6
RQR9
PPV (kV)
49.99(9)
80.00(12)
120.07(17)
Kairref
(mGy/min)
25.5(7)
58(2)
124(3)
QMon
(nC/min)
69.8(2)
159.9(4)
335.3(8)
fCALMon_ref
(mGy/nC)
0.37(1)
0.37(1)
0.37(1)
Conclusions
It is possible to verify that uncertainties in calibration factors are lower for the PDC than for clinical PKA meters like
the investigated ones9. Indeed, uncertainties reached a
maximum of 13% for PDC and 24% for the ScanditronixIBA meter (k=2) (both calibrated against the cylindrical
PTW reference chamber). These results are in accordance
with IEC 60580:2000 standard10, which recommends uncertainties up to 25%.
Also, the energy dependence of PDC was lower than
that of the clinical PKA meter: calibration factors showed
deviations from -10 to -16% for the PDC, and from -1 to
+16% for the other one. For the IBA meter, calibration
factors showed an increasing trend with tube voltage for
all analyzed radiation field sizes9. This difference can be
Table 2. Values of air kerma-area product rate (PKA) (a) and air
kerma rate (Kair) (b) measured with PDC and determined from
calibrated monitor chamber and reference collimator, as well as
the obtained PDC calibration factors (FCALPDC_Mon), for the same
three standard beams above.
a
b
18
Standard
beam
PKAPDC
(mGy.m²/min)
PKAMon
(mGy.m²/min)
FCAL_PKAPDC-Mon
RQR 3
0.63(5)
0.62(4)
0.98(9)
RQR 6
1.41(10)
1.43(8)
1.02(9)
RQR 9
3.11(22)
3.03(17)
0.97(9)
Standard
beam
KairPDC
(mGy/min)
KairMon
(mGy/min)
FCAL_KPDC-Mon
RQR 3
0.064(3)
0.058 (3)
0.91(5)
RQR 6
0.141(7)
0.134(7)
0.95(5)
RQR 9
0.31(2)
0.28(2)
0.91(5)
Revista Brasileira de Física Médica.2011;5(1):15-20.
Figure 5. Energy dependence curve of the Scanditronix – IBA
detector, showing PKA versus nominal kVp values, for three sizes
of radiation field (triangle = 25x25 cm², circle = 20x20 cm² and
square = 15x15 cm2), measured against the calibrated PDC.
Calibration of PKA meters against ion chambers of two geometries
attributed to the lower atomic number of components of
PDC incidence surface, compared to the clinical meters,
which need to be transparent.
Comparing the two methods of calibration of PDC in a
calibration laboratory, although the energy dependence has
remained, the difference among values of the determined
calibration factors is clear: with monitor chamber, PDC calibration factors varied only from 0.97 to 1.02 (Figure 4 and
Table 2). The radiation field that reaches the 30 cm3 chamber, used as reference, is not the same incident onto the
PDC surface, covering only the central portion of the beam.
Unlike, putting the monitor chamber, as reference, after the
reference collimator, the X-ray beam passing through the
chamber will be nearly the same incident on PDC surface.
The beam that reaches the cylindrical chamber is a little
harder than the incident on the PDC surface.
It becomes also evident in the air kerma results. Air
kerma is measured by a 10x10 cm2 chamber located at
the center of PDC. Thus, the beam impinging its surface
is like the one that reaches the cylindrical chamber volume
and so the calibration factor is closer to the unity.
In addition, transmission chambers that are not optically transparent (graphite-coated chambers, for example,
like the used monitor chamber) may be more appropriate
to be used as reference chambers, because they have less
energy dependence.
Acknowledgment
We thank to Fundação de Amparo à Pesquisa do Estado
de São Paulo (FAPESP) and the Brazilian National Counsel
of Technological and Scientific Development agencies
for partially supporting this work; to HIAE and its Medical
Physicist MSc. Marcia C. Silva for allowing the clinical
measurements; to Radcal Co. for lending us the PDC unit
for the tests, as well as to the LMRI of IEE-USP for the
infrastructure use and staff help.
References
1. International Commission on Radiological Protection. Recommendations of
the international commission on radiological protection, ICRP Publication
103. ICRP, Oxford: Pergamon Press; 2007.
2. International Commission on Radiation Units and Measurements. Radiation
quantities and units, ICRU Report 33. ICRU, Bethesda, MD; 1980.
3. Canevaro LV. Aspectos Físicos e Técnicos da Radiologia Intervencionista.
Rev Bras Física Med. 2009;3(1):101-15.
4. European Commission. Council. Directive of June 30, (97/43/Euratom) on
health protection of individuals against the dangers of ionizing radiation in
relation to medical exposure. Official J Eur Commun. No. L180/22; 1997.
5. International Atomic Energy Agency. Dosimetry in Radiology: An International
Code of Practice, Technical Reports Series 457. IAEA, Vienna; 2007.
6. International Electrotechnical Commission. Medical electrical equipment –
Dosimetric instruments used for non-invasive measurements of X-ray tube
voltage in diagnostic radiology, IEC 61676, IEC, Geneva; 2002.
7. International Electrotechnical Commission. Medical Diagnostic X-rays
Equipment – Radiation Conditions for Use in the Determination of
Characteristics, IEC 61267, IEC, Geneve; 2005.
8. Terini RA, Pereira MAG, Künzel R, Costa PR, Herdade SB. Comprehensive
analysis of the spectrometric determination of voltage applied to X-ray
tubes in the radiography and mammography energy ranges using a silicon
PIN photodiode. Brit J Rad. 2004;77:395-404.
9. Toroi P, Kosunen A. The energy dependence of the response of a patient
dose calibrator. Phys Med Biol. 2009;54(N):151-6.
10. International Electrotechnical Commission. Dose Area Product Meters, IEC
60580, 2nd. Ed., IEC, Geneva; 2000.
Revista Brasileira de Física Médica.2011;5(1):15-20.
19
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):21-4.
Sensitivity of film measured off-axis ratios to
film calibration curve using radiochromic film
Sensibilidade das razões fora do eixo central medidas
para a curva de calibração de filmes usando filme
radiocrômico
Diana García-Hernández1 and José M. Lárraga-Gutiérrez2
2
1
Instituto de Física, Universidad Nacional Autónoma de México, Mexico City, Mexico.
Unidad de Radioneurocirugía; Laboratorio de Física Médica, Instituto Nacional de Neurología y Neurocirugía, Mexico City, Mexico.
Abstract
Off-axis ratios of conical beams generated with a stereotactic radiosurgery-dedicated LINAC were measured with EBT2 film and stereotactic diode.
The sensitivity of both full width at half maximum (FWHM) and penumbras (80-20% and 90-10%, respectively), with respect to the characteristics
of the film calibration curve fit, was investigated. In all cases, penumbras resulted to be more sensitive than FWHM. However, these differences
were, in general, smaller than the ones found between EBT2 reference values and the stereotactic diode measurements. The larger variation in OAR
parameters was found to depend on whether the fit intersected or not the origin. A 1D gamma-index analysis showed this difference can be important
in all measured conical beams.
Keywords: small field dosimetry, radiochromic film, penumbra.
Resumo
As razões fora do eixo de feixes cônicos criados com um acelerador de partículas linear (LINAC), dedicado à radiocirurgia estereotáxica, foram medidas
com um filme EBT2 e diodo estereotáxico. A sensibilidade da largura a meia altura (FWHM) e das penumbras (80-20% e 90-10%, respectivamente),
com relação às características da curva de calibração do filme, foi investigada. Em todos os casos, as penumbras mostraram ser mais sensíveis do
que FWHM. Entretanto, estas diferenças foram, em geral, menores do que aquelas encontradas entre os valores de referência do EBT2 e as medidas
do diodo estereotáxico. Encontrou-se que a maior variação nos parâmetros da razão fora do eixo depende se o ajuste se intersectava, ou não, à fonte.
Uma análise do índice de gamma de 1D mostrou que esta diferença pode ser importante em todos os feixes cônicos medidos.
Palavras-chave: dosimetria de campo pequeno, filme radiocrômico, penumbra.
Introduction
Stereotactic radiosurgery (SRS) requires high precision
and accuracy in the calculation of dose distributions, due
to the high dose delivered to the target and the near presence of healthy radiosensitive tissues. One of the major
concerns in SRS is its dosimetry, because of the lack of
lateral electronic equilibrium and steep dose gradients existing in large portions of these fields1.
Along with tissue-maximum-ratios and total output
factors, off-axis ratio (OAR) is one of the most important
dosimetric parameters to be determined during the characterization of small radiation fields. OAR is defined as in
Eq. 1:
OAR(c,r,d)=D(c,r,d)/D(c,0,d)
(1)
where:
c is the diameter of collimator;
r is the off-axis distance perpendicular to central beam
axis, and
d is the depth in water.
The FWHM and the penumbras 80-20% and 90-10%
are relevant information derived from the OAR. The penumbra 90-10% is particularly important in SRS because the
90% isodose curve is commonly used for dose prescription
(instead of the 80% isodose curve used in radiotherapy).
In small field dosimetry, the choice of the suitable detector is another difficulty. In this sense, radiochromic films
are detectors with high-spatial resolution and very interesting properties (tissue-equivalence, dose integration,
self-development, small or null dependence with energy
Corresponding author: José Manuel Lárraga-Gutiérrez – Laboratorio de Física Médica, Instituto Nacional de Neurología y Neurocirugía – Insurgentes Sur 3877,
La Fama, Tlalpan, C.P. 14269 – Mexico City – Mexico – E-mail: [email protected]
Associação Brasileira de Física Médica®
21
García-Hernández D, Lárraga-Gutiérrez JM
of radiation), which make them appropriate for dose distribution measurements in megavoltage radiation fields with
high-dose gradients2,3.
However, as relative dosimetry detectors, radiochromic
films must be calibrated. This paper investigated the sensitivity of measured OAR to the form and characteristics of
the used film calibration curve.
Materials and methods
Radiochromic film EBT2
The recently introduced Gafchromic® EBT2 radiochromic
film was used for all measurements. Sheets were cut in
3x3 cm2 pieces for the calibration curve irradiation, and cut
in 4x5 in2 pieces for the field profiles irradiation. The films
were handled in accordance with the procedures outlined
in the AAPM TG-55 report4.
Irradiation protocol
Beam diameters ranging from 4 to 20 mm were produced by
conical collimators attached to a dedicated SRS 6 MV linear
accelerator (Novalis BrainLAB, Germany). Films were irradiated in liquid water at least 24 hours after they were cut.
The calibration curve was built with 16 equidistant points
covering the dose interval from 0 to 560 cGy, and performing three measurements per dose point. The irradiation was
performed under SAD technique, 5 cm in depth.
The OAR of each conical collimator were obtained irradiating the film pieces under a SAD geometry, 7.5 cm in
depth. The used monitor units were such that the dose delivered was between 400 and 450 cGy for all the cones.
The profile for the cones of 4, 10 and 20 mm were also
measured with a stereotactic diode (SFD, IBA-Dosimetry,
Germany).
Scanning protocol
Film digitization was carried out using a commercial document scanner Epson Perfection V750 Pro, by means of the
SilverFast (LaserSoft Imaging, USA) scanning software, 72
hours after they were immersed in water (Aldelaijan5). The
scanning resolution was 300 dpi, and 48-bits color depth
(RGB mode), although only the red channel was used for
the analysis.
Analysis
Images were first filtered with a Wiener filter (7x7) in a
homemade Matlab (Mathworks Inc., USA) routine. Another
routine gave the fit parameters of the selected calibration model, with its respective χ2. Fifty profiles, 3 cm long
around the centroid of the cone image, were averaged to
obtain a single OD beam profile.
The reference fit was chosen based on the recommendations made by Bouchard et al.6, who listed the minimum
requirements of a suitable fit function:
• the function intersects the origin;
• the function is strictly increasing;
22
Revista Brasileira de Física Médica.2011;5(1):21-4.
•
•
the function has zero or one inflection point in the domain of interest;
if there is an inflection point, it occurs between 0 and
0.5ODmáx.
We analyzed the impact over the OAR of varying the
following parameters in the calibration curve:
Analytical expression for the fit.
Number and distribution of experimental points.
Intersection (or not) with origin.
The behavior of the EBT2 film sensitivity is derived from
multiple hit theory and assumed to be on the form of a
series of dose powers6. Different calibration curves were
used, which did not meet the above criteria and applied to
the measured OAR to evaluate their effect.
Uncertainty analysis
Uncertainty analysis was mainly based on the approach
made by Devic et al.7. The uncertainty in the dose determined from an OD and the calibration curve takes into
account contributions derived from: OD determination;
LINAC output instability and calibration fit uncertainty (χ2).
Variations in film-to-film response, the noise of the ROI used
for the average pixel value determination, and the electronic noise of scanner contribute to the OD uncertainty.
1D Gamma Index
We compared the gamma index (Г) between the profiles obtained with the reference curve, and a profile found with the
same reference curve does not cross the origin. The latter for
the cones of 4 and 20 mm, and allowing a dose difference of
1% in combination with a distance to agreement of 2 mm.
Results and Discussion
Different fits
We chose as a reference calibration fit, the curve with
smaller χ2, and which satisfied the requirements that
have already been described. Table 1 shows the analytical form and parameters of different curves used for the
intercomparison.
All beam profile data were normalized to the central
axis and the beam penumbra was characterized by extracting the beam fall-off widths between 90-10% and 8020% of dose. Table 2 shows the reference values of these
parameters; we also include the values measured using a
stereotactic diode SFD.
Implementing different fits to calibration data, we found
that the variation of FWHM is always smaller than 1.6%
among fits, for all the cone diameters and the penumbras are
even more sensitive (Figure 1). For example, there is a difference of 5% between the 80-20% penumbras of the fits ‘s1’
and ‘log3’ for the cone of 7.5 mm. A similar situation is found
for the penumbra 90-10% (4.6% difference between ‘s2’ and
‘log3’). In general, ‘s1’ underestimates, and ‘s2’ and ‘poli6’
overestimate the values of the penumbras relative to ‘log3’.
Sensitivity of film measured off-axis ratios to film calibration curve using radiochromic film
Intersection with origin
The Figure 1 shows the variation in FWHM and penumbras
after fitting two curves of the form ‘log3’ (passing and not
through origin) to the experimental data set of 16 equidistant points. Again, the FWHM is relatively insensitive to this
variation. However, both penumbras are underestimated
by the noncrossing (0,0) fit, at least in 4%.
Based on the results, it is interesting to see how the
differences among OAR parameters obtained with diverse
film calibration curves are smaller than the ones between
‘log3’-values and the stereotactic diode measurements.
This variation can be as large as 14% for the penumbra
90-10% in the 4 mm cone.
1D Gamma Index
The ‘log3’ is the reference fit curve to the calibration data.
From the previous sections, it can be perceived that the
same form of fit, when it does not cross the origin, offers
one of the largest differences in the values of the OAR parameters. That is the reason why we decided to compare,
through the 1D gamma index, the two beam dose profiles
obtained with the last mentioned fits. Figure 2 shows the
Table 1. Parameters of the different fits used for film calibration
experimental points
‘s1’
‘s2’
‘poli6’
D(OD)=a1Log10[OD+1]
+ a2Log10[OD+1]2 +
a3Log10[OD+1]3
D(OD)=-a1Ln[1-OD/b]
D(OD)=-a1Ln[1-OD/b]
-a2Ln[1-OD/b]2
D(OD)=a1OD+a2OD2+
a3OD3+a4OD4+a5OD5+
a6OD6
Fit parameters
a1=2406.33
a2=-5965.03
a3=175960.50
a1=419.74
b=0.45
a1=506.48
a2=-182.95
b=0.59
a1=933.45
a2=337.55
a3=4669.05
a4=1765.65
a5=4228.97
a6=-6717.24
Cone FWHM FWHM Penumbra Penumbra Penumbra Penumbra
diameter (EBT2) (SFD) 80-20% 80-20% 90-10% 90-10%
(SFD)
(EBT2)
(SFD)
(EBT2)
4
3.87 3.91 1.41
1.35
2.59
2.26
6
5.71
1.58
2.99
7.5
7.54
1.71
3.29
10 10.07 10.12 1.88
1.71
3.84
3.23
12.5 12.41
2.10
4.38
15 14.87
2.11
4.38
17.5 17.16
2.18
4.62
20 19.88 20.00 2.18
1.93
4.70
4.05
1
0
-1
Penumbra 80-20%
Penumbra 90-10%
FWHM
-2
-3
-4
-5
-6
2
4
6
8
12
16
18
20
22
Figure 1. FWHM and penumbras variation depending on the
calibration fit ‘log3’ crossing or not the origin.
Cone of 20mm
Cone of 4mm
1.0
1.0
'log3'
'log3' (x0,y0)
0.8
1.07
0.6
0.4
0.0
20
0.6
0.4
0.2
0
5
10
15
20
Off-axis distance [mm]
25
4mm
30
0.0
30
0
5
Γ
10
10
15
20
Off-axis distance [mm]
25
Mean=8.5
15
Γ
'log3' (x0,y0)
'log3'
0.8
0.2
1.01
14
Collimator diameter [mm]
χ2 [cGy]
0.44
10
OAR
‘log3’
Analytical form
Table 2. Measured OAR parameters [mm] for reference fit
‘log3’
OAR
Fit name
shape of the smallest and biggest collimators’ profiles normalized to the beam axis dose.
The average Г along the profile is 8.5 for the cone
of 4 mm, and 2.4 for the one of 20 mm. The mayor
contribution to these out-of-tolerance values comes
from the low dose regions, where the Г can easily
Percentage difference relative to fit 'log3' that crosses (0,0)
Number and distribution of dose points used for the
calibration curve fit
Here, we fit a calibration curve of the form ‘log3’ for different sets of images: one set – reference set – consisting
of 16 equidistant points in the interval (0,560 cGy); a set
of nine equidistant points; a set of 12 points with detail in
low doses region; a set of 12 points with emphasis in high
doses region. The percentage difference in all of the OAR
parameters determined with the four sets resulted always
less than 2.3% relative to the reference set, being the set
of 12 points (high doses region) that expressed the largest
difference. Table 4 shows that the use of a set with detail in
high doses results in more slightly closer-to-reference set
FWHM and penumbra values than using a set with emphasis in low doses.
25
30
25
30
20mm
Mean=2.4
20
15
10
5
5
0
0
5
10
15
20
Off-axis distance [mm]
0.63
25
30
0
0
5
10
15
20
Off-axis distance [mm]
Figure 2. Gamma index for comparing off-axis ratios between
a beam dose profile obtained with a fit of the form ‘log3’ and
another of the same form, but it does not cross the origin.
Revista Brasileira de Física Médica.2011;5(1):21-4.
23
García-Hernández D, Lárraga-Gutiérrez JM
Table 3. FWHM and penumbras percent difference (relative to FWHM and penumbras obtained with ‘log3’) depending on the kind of
fit selected
Cone diameter
4
6
7.5
10
12.5
15
17.5
20
Penumbra
80-20%
s2
-1.78
0.19
1.15
-0.85
-0.66
-1.16
-1.97
-1.67
FWHM
s1
1.55
0.38
-0.13
0.53
0.44
0.45
0.46
0.36
s2
0.75
0.26
0.01
0.29
0.24
0.24
0.23
0.18
poli6
0.38
0.09
-0.02
0.13
0.11
0.11
0.11
0.09
s1
0.79
3.48
4.92
2.20
2.43
1.69
0.46
1.06
poli6
-0.69
0.30
0.70
-0.15
-0.04
-0.29
-0.70
-0.53
s1
-1.41
0.42
2.19
-0.27
0.04
-0.46
-1.25
-0.82
Penumbra
90-10%
s2
-4.56
-2.87
-1.80
-3.68
-3.23
-3.56
-3.95
-3.70
poli6
-2.51
-1.60
-1.11
-2.01
-1.76
-1.91
-2.09
-1.93
Table 4. FWHM and penumbras percent difference (relative to FWHM and penumbras obtained with ‘log3’) depending on the number
and distribution of points in calibration data
Cone diameter
4
6
7.5
10
12.5
15
17.5
20
Penumbra
80-20%
12 pts (LD)
0.83
0.59
0.49
0.85
0.95
0.01
0.01
0.01
FWHM
9 pts
-0.30
-0.38
-0.36
-0.20
-0.19
-1.4E-03
-9.7E-04
-9.0E-04
12 pts (LD)* 12 pts (HD)**
-0.30
0.07
-0.33
0.08
-0.30
0.07
-0.18
0.04
-0.17
0.04
-1.3E-03
3.1E-04
-9.6E-04
2.2E-04
-8.6E-04
2.0E-04
9 pts
1.52
1.15
0.98
1.52
1.63
0.02
0.02
0.02
12 pts (HD)
-0.22
-0.16
-0.14
-0.22
-0.25
-2.5E-03
-3.0E-03
-2.9E-03
9 pts
2.28
1.81
1.73
2.34
2.20
0.02
0.02
0.02
Penumbra
90-10%
12 pts (LD)
1.29
0.97
0.93
1.34
1.27
0.01
0.01
0.01
12 pts (HD)
-0.33
-0.26
-0.25
-0.35
-0.33
-3.4E-03
-3.7E-03
-3.7E-03
*12 pts (LD), 12 points with detail in low doses; **12 pts (HD), 12 points with detail in high doses.
reach values of 10. Nevertheless, averaging the gamma-index in the profile region, where dose is higher
than 5%, the one in the beam axis, for the 20 mm
cone it results in Г=0.26 (96% of the points satisfy the
acceptance criterion), and for the 4 mm cone it results
in Г=0.53 (with 81% of the points passing the acceptance criterion).
Conclusions
OAR of conical beams generated with a SRS-dedicated
LINAC were measured with EBT2 film. The sensitivity of
the FWHM and penumbras 80-20% and 90-10%, with the
characteristics of the film calibration curve, was investigated. In all the cases, penumbras resulted to be more sensitive than FWHM to the kind of fit. However, these differences were, in general, much smaller than the ones found
between EBT2 reference ‘log3’-values and the stereotactic
diode measurements.
The largest differences in OAR parameters were
found between curves that intersected (and not) the
origin. A 1D gamma analysis showed this difference
can be significant, because, for example, 19% of the
points in the 4 mm-cone profile did not pass the acceptance criteria.
24
Revista Brasileira de Física Médica.2011;5(1):21-4.
Acknowledgment
To Mercedes Rodríguez Villafuerte and Olivia Amanda
García Garduño for their valuable comments to this paper.
References
1. Das IJ, Ding GX, Ahnesjö A. Small fields: Nonequilibrium radiation dosimetry.
Med Phys. 2008;35(1):206-15.
2. Ramani R, Lightstone AW, Mason DL, O’Brien PF. The use of radiochromic
film in treatment verification of dynamic stereotactic radiosurgery. Med
Phys. 1994;21(3):389-92.
3. Wilcox EE, Daskalov GM. Evaluation of Gafchromic EBT film for Cyberknife
dosimetry. Med Phys. 2007;34(6):1967-74.
4. Niroomand-Rad A, Blackwell CR, Coursey BM, Gall KP, Galvin JM,
McLaughlin WL, et al. Radiographic film dosimetry: Recommendations
of AAPM Radiation Therapy Committee Task Group 55. Med Phys.
1998;25(11):2093-115.
5. Aldelaijan S, Devic S, Mohammed H, Tomic H, Liang LH, DeBlois F, et al.
Evaluation of EBT2 model Gafchromic film performance in water. Med Phys.
2010;37(7):3687-93.
6. Bouchard H, Beaudon G, Carrier JF, Kawrakow I. On the characterization
and uncertainty analysis of radiochromic film dosimetry. Med Phys.
2009;36(6):1931-46.
7. Devic S, Seuntjens J, Sham E, Podgorsak EB, Schmidtlein C, Kirov AS, et
al. Precise radiochromic film dosimetry using a flat-bed document scanner.
Med Phys. 2005;32(7):2245-53.
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):25-30.
Dosimetry of cones for radiosurgery system
Dosimetria de cones para sistema radiocirúrgico
Laura Furnari, Camila P. Sales, Gabriela R. Santos, Marco A. Silva and Gisela Menegussi
Department of Radiology, Clinics Hospital, of São Paulo University School of Medicine (FMUSP), São Paulo (SP), Brazil.
Abstract
Dosimetry of small fields, such as cones for radiosurgery, requires a lot of care in its implementation. The acquisition of curves of Percentage depth
dose (PDD) and profiles for nine circular cones with diameters from 4 to 20 mm for 6 MV photons was performed. Measurements with four types
of dosimeters: diode, pinpoint ionization chamber, diamond detector and film were done. A comparison between the data obtained with the several
detectors permitted to conclude that the diode is the detector more reliable. The effect of various methods to “smooth” the curves was studied and
showed that there are methods that change very much the measured data. Interpolations were made in PDD curves in order to eliminate the noise
of small fields due to low signal in the detectors. The most important conclusion refers to the choice of suitable detector, in this case the diode, and
to a careful handling of obtained data to not disturb or modify the results of the measurements.
Keywords: cones, radiosurgery, commissioning, diode detector, diamond detector, pinpoint chamber, dosimetric film.
Resumo
A dosimetria de pequenos campos, como cones para radiocirurgia, requer muito cuidado em sua implementação. A aquisição das curvas de porcentagem
de dose na profundidade (PDP) e perfis para nove cones circulares, com diâmetros de 4 a 20 mm para fótons de 6 MV, foi realizada. Foram realizadas as
medidas com quatro tipos de dosímetros: diodo, câmara de ionização do tipo pinpoint, detector de diamantes e filme. Uma comparação entre os dados
obtidos com diversos detectores permitiu concluir que o diodo é o detector mais confiável. O efeito dos diversos métodos para “atenuar” as curvas foi
estudado e mostrou que existem métodos que realmente mudam os dados medidos. Interpolações foram feitas em curvas de PDP para eliminar os
ruídos de pequenos campos, devido ao baixo ruído nos detectores. A conclusão mais importante refere-se à escolha do detector adequado, neste caso,
o diodo, e ao manuseio cuidadoso dos dados obtidos para não transtornar ou modificar os resultados das medidas.
Palavras-chave: cones, radiocirurgia, comissionamento, detector de diodo, detector de diamantes, câmara detalhada, filme dosimétrico.
Introduction
Materials and methods
The cones are precise accessories used to apply radiosurgery. With them, it is possible to obtain circular fields
with diameter down to 4 mm. At the moment of treatment, the jaws aperture needs to be smaller than the
external circumference of the cone, because this is the
shielded region of this accessory. The recent accidents
related to the use of cones were originated just because
the jaws had been set wrong. As a manufacturer recommendation, the adequated jaw’s aperture must be the
same for all cones size; for our set of cones, it must be
26 mm x 26 mm. So, the use of this accessory demands
a lot of attention and care. This care must be taken in the
entire process of a radiosurgery with cones, beginning by
the data acquisition.
The data needed to feed the planning system are numerous and must be obtained through direct measurements1.This work presents the dosimetry data obtained
with cones and the discussion about how to handle
these data.
The cones, showed in Figure 1, were commissioned in an
accelerator Varian 6EX, 6 MV photon energy, equipped
with micro-multileaf (mMLC) collimator m3 BrainLab.
With an automatic scanning system, it was obtained
the Percentage depth dose (PDD) and the profile curves at
7.5 mm depth in transverse directions for nine cones with
1.a
1.b
Figure 1. Circular cone mounted at 6EX accelerator. 1.a shows
cone 4 mm and 1.b, 20 mm.
Corresponding author: Laura Furnari – Clinics Hospital (USP) – Rua Dr. Eneas de Carvalho Aguiar, 255 – Cerqueira Cesar – São Paulo (SP), Brazil –
CEP 05403-900 – E-mail: [email protected]
Associação Brasileira de Física Médica®
25
Furnari L, Sales CP, Santos GR, Silva MA, Menegussi G
120
cone 20
cone 17.5
cone 15
cone 12.5
cone 10
cone 8
cone 7.5
cone 5
cone 4
100
PDD
80
60
40
20
0
0
50
100
150
200
250
300
350
Depth (mm)
Figure 2. PDD curves for 9 cones measured with diode and not
smoothed.
cone 4
cone 5
cone 7.5
100
80
cone 8
cone 10
60
PDD
diameter from 4 mm to 20 mm. The measurements were
made with diode, pinpoint chamber, diamond detector2
and dosimetric film.
Table 1 presents the type and characteristics of the detectors used.
The diode, pinpoint chamber and diamond detector
were positioned vertically in the water phantom, because
the fields were too small. For the measurements with the
diode and pinpoint chamber, it was used the 3D scanning
Blue Phantom of IBA, and for the diamond detector, it was
used the PTW MP2 phantom, because it has a particular
insulation system.
The EDR2, an enveloped film, was put between slabs
of solid water to obtain the buildup and lateral scatter. For
the measurements of PDD, the film was positioned vertically and this positioning had to be done very carefully, because, as the fields are very small, it is very difficult to put
the film exactly in the direction of the central axis especially
for the cone of 4 mm diameter. For the measurements of
profiles, the film was positioned horizontally.
The curves of PDD and profile obtained were smoothed.
Absolute measurements of output factor for a field
10 cm x 10 cm and for each cone were done, and the
scatter factor was calculated.
The treatment planning system iPlan for radiosurgery
with cones needs beam data used for the Clarkson dose
calculation methods. The major beam data required for
the commissioning of this unit include depth dose data
measured at source-to-surface distance (SSD) of 98.5 cm,
beam profiles measured at the depth of 7.5 cm with SSD
of 92.5 cm and relative scatter factor data measured at the
depth of 1.5 cm with SSD of 98.5 cm.
cone 12.5
cone 15
cone 17.5
40
cone 20
20
0
0
50
100
150
200
250
300
Depth (mm)
Figure 3. PDD curves for 9 cones measured with diode and
smoothed.
120
Results and discussion
100
Diode results
PDD
80
PDD curves
The curves presented in Figure 2 are the PDD for 9
cones obtained with diode, and Figure 3 is the same
curves smoothed.
The Figure 4 shows curves, the original and the smoothed
manually for cone of 4 mm. The original curve presents a lot
of noise that may be perceived because, as the electric signal is small, the size of noise is proportionally big. We tried
to smooth the curves using several methods presented as
tools in the software OmniPro-Accept, but finally it was chosen to do manual interpolations and corrections. The same
procedure was applied to all other PDD curves.
Edited
60
Original
40
20
0
0
50
100
150
200
250
300
350
Depth (mm)
Figure 4. PDD curves for cone with 4 mm diameter: original
and edited.
Table 1. Characteristics of the detectors
Detector type
Diode
Pinpoint chamber
Diamond detector
Film
26
Model
Stereotactic SFD
CC01
60003
EDR2
Revista Brasileira de Física Médica.2011;5(1):25-30.
Brand
IBA
IBA
PTW
Kodak
Effective measurement point
<0.9 mm
2.3 mm from tip
0.5 mm
Thick ness (mm)
0.06
3.6
0.3
Diameter of active area
0.6 mm
2 mm
3 mm
Dosimetry of cones for radiosurgery system
Profile curves
The curves presented in Figure 5 are halves of the profiles for
9 cones obtained with diode. These curves furnished the off
axis factor for each cone inserted in the planning system.
Scatter factor
The relative scatter data were measured with diode at the
depth of 1.5 cm with SSD of 98.5 cm. The relative scatter factor was calculated by dividing the measured data
for each collimator setting, by the measured data at the
same depth and same SSD for a field size of 10 cm x10
cm. Figure 6 shows the relative scatter factors for different
circular cones with collimator jaws of 2.6 cm x 2.6 cm.
Film results
Profile curves
The Figure 9 is the image obtained irradiating the film perpendicularly to the central axis at the depth of 7.5 cm with
SSD of 92.5 cm.
This was scanned in an Epson Expression 10000XL
device to construct the profile curves presented in the
Figure 10.
120
100
80
PDD curves
The same results of PDD were obtained with pinpoint
chamber, but with much more noise because of the size of
the chamber. (Figure 7)
Profile curves
The profile curves obtained with pinpoint chamber, showed
in Figure 8, are like that obtained with diode, but also present a lot of noise.
100,00
PDD
Pinpoint results
60
40
20
0
0
50
100
60,00
40,00
10
15
50
cone 8 mm
cone 10 mm
40
cone 12.5 mm
cone 15 mm
20
20
300
350
60
cone 20
cone17.5
cone15
cone12.5
cone10
cone 8
cone 7.5
cone5
cone 4
30
10
-40
20,00
5
250
70
cone 17.5 mm
cone 20 mm
0
150 200
Depth (mm)
Figure 7. PDD curves for 9 cones measured with pinpoint
chamber.
cone 4 mm
cone 5 mm
cone 7.5 mm
80,00
0,00
cone 4
cone 5
cone 7.5
cone 8
cone 10
cone 12.5
cone 15
cone 17.5
cone 20
-30
-20
0
0
10
-10
Off axis distance (mm)
-10
20
30
40
Figure 8. Profile curves for 9 cones measured with pinpoint
chamber.
25
Distance off center (mm)
Figure 5. Profile curves for 9 cones measured with diode.
Scatter factor
1,3
1,1
0,9
0,7
0,5
0
5
10
15
cone diameter (mm)
20
25
Figure 6. Scatter factor relative of 9 cones measured with diode.
Figure 9. Image film irradiated to obtain profile curves for the
cones 4, 5, 7.5, 8, 10, 12.5, 15, 17.5 and 20mm.
Revista Brasileira de Física Médica.2011;5(1):25-30.
27
Furnari L, Sales CP, Santos GR, Silva MA, Menegussi G
Diamond results
PDD curves
The Figure 11 represents the PDD curves obtained with
diamond detector. In the enlarged detail, it is possible to
see the noise presented by this detector.
Profile curves
In the Figure 12, we see the profile curves obtained with
diamond detector.
-40
-30
-20
350
300
250
200
150
100
50
0
-10
0
-50
10
20
30
40
off axis distance (mm)
Figure 10. Profile curves for 9 cones obtained from the film.
110
100
cone 20
cone 17.5
cone 15
cone 12.5
cone 10
cone 8
cone 7.5
cone 5
cone 4
90
80
PDD
70
60
50
Comparison of PDD curves for different detectors
The Figure 13 is the representation of the PDD obtained
with diode, pinpoint and film for cone of 4 mm. We may
observe that they present different results: the measurements with pinpoint show more collection of charge and
the film less than diode. This effect is important only in
small cones. To establish a criterion to know what curve
was the correct, we did measurements with diode, moving
it manually and verifying after each movement if the diode
was well positioned.
The comparison of the data obtained with automatic
and manual movement is in the Figure 14. It is possible
to see that the curves are very similar, almost coincident,
what indicates that the measurements done with automatic scanning of the diode were correct. Otherwise, that
data were compared with gold standards BrainLab library
and were approved. So, we concluded that the diode is
the best detector for small cones and we used this data to
feed the planning system.
Effect of smooth curves
The smooth of the curves offered by the software is an interesting procedure, because even if the curves have some
noise they are usable. However, it is necessary to make it
carefully as each smooth is going to change the parameters of the curve. As an example, the Figure 15 presents
100
80
40
pinpoint
diode
PDD
60
30
film
40
20
10
0
010 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250
20
Depth (mm)
0
Figure 11. PDD curves for 9 cones measured with diamond detector.
0
50
100
150
200
250
300
350
Depth (mm)
Figure 13. PDD curves for cone 4 mm obtained with pinpoint
chamber, diode and film.
120
100
80
PDD
cone 20
cone 17.5
cone 15
cone 12.5
cone 10
cone 8
cone 7.5
cone 5
cone 4
cone 4 automatic
cone 4 manual
60
40
20
-30
-20
-10
0
10
Off axis distance
20
30
Figure 12. Profiles curves for 9 cones measured with diamond
detector.
28
Revista Brasileira de Física Médica.2011;5(1):25-30.
0
0
50
100
150
200
Depth (mm)
250
300
Figure 14. PDD curves for cone 4mm obtained with automatic
and manual movement of the diode.
Dosimetry of cones for radiosurgery system
Quantity:
R100:
D50:
D100:
D200:
Dmax:
TPR200/100:
Dose
9 mm
74.3 %
50.8 %
23.1 %
132.5 %
0.516
Dose
9 mm
74.4 %
50.3 %
22.9 %
132.7 %
0.516
Dose
9.8 mm
75.2 %
51.1 %
23.2 %
132.2 %
0.516
Figure 15. Results of maximum dose depth (R100), depth maximum dose (R50), percentage of dose in 100 mm depth (D100),
percentage of dose in 200 mm depth (D200), percentage of maximum dose (Dmax), ratio of Tissue-phantom ratio depth 200 mm
to 100 mm (TPR) for three types of smooth: colunm 1 - without
smooth, column 2 - least square with mean value region of 5
mm, and column 3 - envelope with mean value region of 5 mm.
80
70
60
50
40
30
20
10
(G)
0
-30
-20
-10
0
10
Off-axis distance (Inline) [mm]
20
(T)
Figure 16. Profile curve for cone 20 mm obtained with pinpoint
chamber in the horizontal position.
Orientation of the detector
The Figure 16 represents a profile curve obtained with
pinpoint chamber in the horizontal position for the cone
20 mm. A comparison with the corresponding data of
the Figure 9 shows that, if the chamber is put in the
horizontal orientation, it cannot do the collection of
charge properly.
cones (4 to 7.5 mm). The best positioning of chambers for
these measurements is vertical orientation, and the speed
to acquisition is important to charge collection; it is convenient to use step by step instead of continuous measurements. It is also necessary a carefully handling of obtained
data to not disturb or modify the results.
The dosimeter size and its orientation can increase the
signal’s noise and disturb the results, and, as high speed,
may result in wrong values. Another important concern is
the evaluation of the data, because the smoothing tools
can produce big distortions to the final results.
Conclusions
References
The decision about the best dosimeter to data acquisition
in small fields is a serious task. It involves each center’s
needs and availability of the equipments. For our purpose,
the diode was the best detector for measures with small
1.
the effect of different type of smooth and it is possible to
see a difference of almost 10% in the position of the maximum (R100) with the smooth envelope.
2.
Fang-Fang Y. Dosimetric characteristics of Novalis shaped beam surgery
unit. Med Phys. 2002;29(8):1729-38.
Rustgi NS, Frye DM. Dosimetric characterization of radiosurgical beams
with diamond detector. Med Phys. 1995;22(12):2117-21.
Revista Brasileira de Física Médica.2011;5(1):25-30.
29
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):31-4.
Implementation of intraoperative
radiotherapy in a linear accelerator
VARIAN 21EX
Implementação da radioterapia intraoperatória em um
acelerador linear VARIAN 21EX
Gustavo H. Píriz1, Enrique Lozano1,2, Yolma Banguero1, Carlos Fernando Varón1, Claudio S.
Mancilla1,2, Cristian Parra1,2 and P. Pacheco3
1
Instituto Nacional del Cáncer/Radioterapia, Santiago, Chile.
2
Universidad de la Frontera, Temuco, Chile.
3
Universidad Nacional Mayor de San Marcos, Lima, Perú.
Abstract
The aim of this paper is to present the experience on intraoperative radiotherapy, which has as the reference center the network of radiotherapy in
Chile. It is detailed the construction of a system of applicators with an easy coupling on a linear accelerator collimator. It is also detailed the cost and
the measurements set up with their corresponding percentage depth dose and isodose curves. This technique was implemented in a Varian Clinac
21EX for beams with 6, 9 and 12 MeV electron energy. The coupling system provides a good dose distribution both laterally and in depth for different
energies. This provides a good coverage of treatment planning volume.
Keywords: intraoperative radiotherapy, dosimetry, electron beam.
Resumo
O objetivo deste estudo é apresentar a experiência com a radioterapia intraoperatória, que tem como centro de referência a rede de radioterapia no
Chile. Detalha-se a construção do sistema de aplicadores de fácil ajuste em um acelerador linear. Também detalhou-se os custos e as medidas em
relação ao PDD correspondente e às curvas de isodose. Esta técnica foi implementada em um Varian Clinac 21EX para feixes com 6, 9 and 12 MeV.
O sistema de acoplamento fornece uma boa distribuição da dose lateralmente e em profundidade para diferentes energias. Com isso, é possível
planejar o volume do tratamento.
Palavras-chave: radioterapia intraoperatória, dosimetria, feixe de elétrons.
Introduction
The current radiotherapy requires high-tech equipment
and multidisciplinary professionals in order to comply with
the requirements in implementation of special techniques1.
Intraoperative radiotherapy (IORT) is an area of interest for
the treatment of certain cancers; a single electron dose
is given intraoperatively on the tumor bed2-3. IORT avoids
conventional radiotherapy after surgery and improved tumor control4-5. It also allows direct visualization of the tumor precisely defined. This allows full or partial protection
of normal tissues through the organs mobilization and/or
energy selection6.
This paper shows the implementation of IORT in a linear accelerator Varian 21EX for beams with 6, 9 and 12
MeV electron energy. It was based on calibration protocols: TRS 3987, ICRU 718 and TG72 (report 92)9.
Materials and methods
Equipment description
We worked in a dual accelerator VARIAN 21EX, which has
6 and 18 MV nominal photon energies and 6, 9, 12, 15 and
18 MeV electrons energies. Measuring was made with a
relative dosimetry phantom PTW MP3, a Markus TN34045
Corresponding author: Gustavo Hector Píriz Monti – Instituto Nacional del Cáncer – Profesor Zañartu, 1010 – Independencia – Santiago – Chile –
E-mail: [email protected]
Associação Brasileira de Física Médica®
31
Piriz GH, Lozano E, Banguero Y, Varón CF, Mancilla CS, Parra C, Pacheco P
parallel plane camera, a Pinpoint TN31014 and a semiflexible TN 31002 with a PTW Freiburg electrometer
(Unidos E).
In order to complete the system, five acrylic cylindrical applicators were designed and built with 4.4 cm
internal diameter, 0.5 cm thick, and 22 cm in length.
They have bevelled in 0°, 15º, 30º, 45º and 50º in the
extreme (Figure 1).
For the location of these applicators in the cone of 10
x 10 cm, it was necessary to build a cerrobend block of
equal area, as depicted in Figure 2. For the construction
of these acrylic applicators over the coupling system, it
was not necessary a large investment as it is required
when the institutions buy the standard equipment generally used for this technique. In our case, the cost was
under US$ 100.00.
Figure 3 shows the system developed at the Instituto
Nacional Del Cancer. Measurements were made in it
with the source-skin distance (SSD) equals to 112 cm
and non gap is left between the water surface and the
applicator.
Results
Figures 4, 5 and 6 show the curves of isodose distributions for the energies of 6, 9 and 12 MeV obtained for electron beams, to beveled applicators for
0, 15, 30 and 45 degrees. These curves were taken
with the semi-flexible ionization chamber TN 31002
(0.125 cm 3).
Figure 7 shows the coverage of the isodose curves
using the applicator without bevel to 6 MeV.
Figures 8, 9, 10 and 11 show some of the measured
curves of percentage depth dose using different applicators on axes: these curves were taken with the semiflexible ionization chamber TN 31002 (0.125 cm3).
Conclusions
One of the most important aspects in this paper for the
implementation of the technique is significantly lower
costs for the construction of the applicators and the
coupling system.
The data obtained are similar to the values published
in the TG 72 report.
As shown in Figures 4, 5 and 6, for small cones, a good
homogeneity of the isodose distribution was obtained and
confirmed with the coverage curve in Figure 7.
Fifty per cent of the isodose curves have a diameter
close to the applicator.
The measured of R50 in the different axes (bisector,
beam and clinical) for each applicator was specific for
each of the used energies (6, 9 and 12 MeV), but it is
dependent on both the angle and radius of the acrylic
cone.
32
Revista Brasileira de Física Médica.2011;5(1):31-4.
Figure 1. Different acrylic cylindrical applicators.
a
b
Figure 2. a) cylindrical applicator inserted into the cerrobend
block; b) set-up for measurements.
Aplicator without oblique
Accesory
Accesory
Aplicator with
oblique
15 cm
15 cm
Phamton
Phamton
Beam axis
=
Clinical Axis
Clinical axis
Beam axis
Bisectrix axis
Figure 3. Definition of “bisectrix axis” for intraoperative radiotherapy
electron applicator with oblique incidence of the beam axis (angle θ).
a)
b)
c)
d)
Figure 4. Typical isodose distributions measured from the accelerator VARIAN for 6 MeV beams using the applicator: a) 0 degree bevel;
b) 15 degrees bevel; c) 30 degrees bevel; and d) 45 degrees bevel.
a)
Percentage depht dose (PDD%)
Implementation of intraoperative radiotherapy in a linear accelerator VARIAN 21EX
b)
c)
d)
a)
b)
d)
Figure 6. Typical isodose distributions measured from the accelerator VARIAN for the 12 MeV beams using the applicator a)
0 degree bevel; b) 15 degree bevel; c) 30 degree bevel; and d)
45 degrees bevel.
45
40
wide (mm)
35
30
25
90%
20
85%
15
70%
10
50%
5
0
8
12
20
16
Depth (mm)
24
28
Figure 7. Coverage of the isodose curves of 90, 85, 70 and
50%. Width versus depth, using applicator from 0 degree with
6 MeV.
80.00
a
b
c
d
60.00
40.00
20.00
0.00
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50
Depth (cm)
PDD
120.00
100.00
80.00
a
b
c
d
60.00
40.00
20.00
0.00
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50
Depth (cm)
Figure 9. a) The beam axis percentage depth doses using the
0 degree bevel applicator. Percentage depth dose using the 30
degree bevel applicator: b) In the beam axis, c) In the bisectrix
axis, d) In the clinical axis for a 6 MeV electron beam.
Percentage depht dose (PDD%)
c)
100.00
Figure 8. a) The depth doses percentage of the beam axis using
the 0 degree bevel applicator. Percentage depth dose using the
15 degree bevel applicator: b) In the beam axis, c) In the bisectrix axis, d) In the clinical axis. For a 6 MeV electron beam.
Percentage depht dose (PDD%)
Figure 5. Typical isodose distributions measured from the accelerator VARIAN for 9 MeV beams using the applicator: a) 0
degree bevel; b) 15 degrees bevel; c) 30 degrees bevel; and d)
45 degrees bevel.
PDD
120.00
120.00
PDD
100.00
80.00
a
b
c
d
60.00
40.00
20.00
0.00
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00
Depth (cm)
Figure 10. a) The beam axis percentage depth doses using the
0 degree bevel applicator. Percentage depth dose using the 30
degree bevel applicator: b) In the beam axis, c) In the bisectrix
axis, d) In the clinical axis for a 12 MeV electron beam.
Revista Brasileira de Física Médica.2011;5(1):31-4.
33
Percentage depht dose (PDD%)
Piriz GH, Lozano E, Banguero Y, Varón CF, Mancilla CS, Parra C, Pacheco P
PDD
3.
120.00
100.00
4.
80.00
60.00
40.00
20.00
a
b
c
d
0.00
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00
Depth (cm)
Figure 11. a) The beam axis percentage depth doses using the
0 degree bevel applicator. Percentage depth dose using the 30
degree bevel applicator: b) In the beam axis, c) In the bisectrix
axis, d) In the clinical axis for a 12 MeV electron beam.
References
1.
2.
34
DeLaney TF, Trofimov AV, Engelsman M, Suit HD. Advanced technology
radiation therapy in management of bone and soft tissue sarcomas. Cancer
Control. 2005;12(1):27-35.
Nemoto K, Ogawa Y, Matsushita H, Takeda K, Takai Y, et al. Intraoperative
Revista Brasileira de Física Médica.2011;5(1):31-4.
5.
6.
7.
8.
9.
radiation therapy (IORT) for previously untreated malignant gliomas. BMC
Cancer. 2002;2:1.
Zachario Z, Sieverts H, Eble MJ, Gfrörer S, Zavitzanakis A. IORT (Intraoperative
Radiotherapy) in neuroblastoma: experience and first results. Pediatr Surg.
2002;12(4):251-4.
Ronsivalle C, Picardi L, Vignati A. Accelerators development for intraoperative
radiation therapy. Proceedings of the 2001 particle accelerator conference,
Chicago; 2001.
Chu SS, Kim GE, Loh JL. Design and dose distribution of docking applicator for an
intraoperative radiation therapy. Korean Soc Ther Radiol. 1991;9(1):123-30.
Ronsivalle C, Picardi L, Iacoboni V, Teodoli S, Barca G, Sicilianno R. Technical
features and experimental characterization of the IORT-1 system, a new
IORT dedicated accelerator. Nuclear instruments & methods in physics
research. Section A, Accelerators, spectrometers, detectors and associated
equipment. 2006;565(2):1042-5.
International Atomic Energy Agency. Adsorbed dose determination in
external beam radiotherapy: an international code of practice for dosimetry
based on standards of absorbed dose to water. Technical Report Series No.
398. Vienna: IAEA, 2000.
International Commission on Radiation Units and Measurements,..
Prescribing, recording, and reporting electron beam therapy : ICRU Report,
No. 71. Series: Journal of the ICRU, Vol. 4, No. 1, 2004.
Beddar AS, Biggs PJ, Chang S, Ezzell GA, Faddegon BA, Hensley FW, et al.
Intraoperative radiation therapy using mobile electron linear accelerator:
report of AAPM radiation therapy committee task group no. 72. Med Phys.
2006;33(5):1476-89.
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):35-40.
A comparative study using both coded
excitation and conventional pulses in the
evaluation of signal to noise ratio sensitivity
and axial resolution in ultrasonic A-mode scan
Estudo comparativo entre pulso de excitação codificada
e pulso convencional na avaliação da relação sinal-ruído
e resolução axial em ultrassom por inspeção modo-A
Tiago M. Machado1 and Eduardo T. Costa1,2
1
University of Campinas/Department of Biomedical Engineering, School of Electrical and Computer Engineering, Campinas (SP), Brazil.
2
University of Campinas/Center for Biomedical Engineering, Campinas (SP), Brazil.
Abstract
In this paper, we have made a comparative study of backscattering of ultrasound conventional and chirp codified pulses. We simulated the interaction of
these two different pulses with a computational phantom constructed with variable amplitude and phase scatterers following a Gamma distribution. We
have used the echo signal-to-noise ratio (eSNR) metric of the backscattered signals from both coded excitation pulse (CEP) and conventional pulse (CP)
for various scenarios, as well as the evaluation of the axial resolution (AR) of the system, using both pulses. The computational phantoms were created
with regular and variable scatterers spacing with amplitude and phase variation for three transducers: 2.25, 5.0 and 7.5 MHz center frequencies. The
duration of the excitation CEP was 18 μs with chirp frequency bandwidth varying from a multiplying factor of 3.7, 2.0 and 1.2 times the transducer
bandwidth, respectively. The pulse compression was performed using matched (MF) and mismatched (MMF) filters. The results for different transducers
and phantoms are in accordance to the literature, and they have given an improvement of the SNR for coded pulse above 20 dB (in average) over
conventional pulse excitation. In addition, the axial resolution for both codified and conventional pulses are in the same range. For a 2.25 MHz transducer,
ARs were 1.33, 1.18 and 1.38 λ for CP, CEP/MF and CEP/MMF filters. Similarly, ARs for 5 MHz for all above three conditions were 1.34, 1.14 and 1.29
λ, and for the 7.5 MHz transducer 1.31, 1.23 and 1.38 λ. Our results have confirmed the increase in gain and very close agreement of the AR. Further
research and development should be carried out to use the potentialities of CEP techniques in medical ultrasound imaging equipment.
Keywords: signal-to-noise ratio, coded excitation pulse, pulse compression, axial resolution, gamma distribution.
Resumo
Neste artigo realizou-se um estudo comparativo do retroespalhamento ultrassônico por inspeção modo-A, obtidos pela utilização de pulsos codificados
e convencionais. Foi simulada a interação destes dois diferentes pulsos em phantoms computacionais construídos com espalhadores de fase variável
e amplitude seguindo a distribuição estatística gama. Mediu-se a relação sinal-ruído (eSNR) para os sinais de eco obtidos tanto por pulso de excitação
chirp codificada (CEP) quanto por pulso convencional (CP) em diversos cenários, bem como a resolução axial do sistema (AR) por sua métrica usual, para
ambos os pulsos. Os espalhadores foram distribuídos espacialmente de forma regular e variável, variando-se a amplitude e fase para três transdutores
operando em 2,25, 5,0 e 7,5 MHz. A duração da excitação chirp utilizada foi de 18 μs, varrendo uma largura de banda de 3,7, 2,0 e 1,2 vezes maior do
que a largura de banda de cada transdutor descrito, respectivamente. A compressão de pulso foi realizada usando-se filtros casado (MF) e descasado
(MMF). Os resultados para os diferentes transdutores e phantoms estudados apresentaram boa concordância com a literatura e indicam uma melhora
da SNR para o CEP em torno de 20 dB (em média) quando comparados com o CP. Além disso, as ARs comparadas tanto para o CEP quanto para o CP
estiveram dentro da mesma faixa de valores. Para o transdutor de 2,25 MHz, os valores das ARs foram 1,33, 1,18 e 1,38λ para CP, CEP/MF e CEP/MMF,
respectivamente. Semelhantemente, para o transdutor de 5 MHz, nas mesmas condições acima foram 1,34, 1,14 e 1,29λ, e para o transdutor de 7,5
MHz: 1,31, 1,23 e 1,38λ. Os resultados confirmaram o aumento no ganho da SNR e uma concordância próxima em relação à resolução axial. Contudo,
novos estudos e pesquisas devem continuar a ser realizados sobre a potencialidade de uso da técnica CEP em sistemas de ultrassom médico.
Palavras-chave: razão sinal e ruído, pulso de excitação codificada, compressão do pulso, resolução axial, distribuição gama.
Corresponding author: Tiago de Moraes Machado – Departamento de Engenharia Biomédica; Faculdade de Engenharia Elétrica e de
Computação da Universidade Estadual de Campinas (DEB/FEEC/UNICAMP) – Rua Alexander Fleming, 181 – Campinas (SP), Brazil –
E-mail: [email protected]
Associação Brasileira de Física Médica®
35
Machado TM, Costa ET
Introduction
Ultrasound imaging is one of the most important medical
imaging procedures mainly due to the possibility of getting
real-time images, be a noninvasive and ionizing radiation
free technique and low cost equipment compared to those
of other imaging modalities (computed tomography – CT,
X-Ray, magnetic resonance image – MRI, etc)1.
However, poor image quality, due to the ultrasound
wave attenuation frequency-dependence with speckle
artifacts, poses a constant challenge to overcome these
limitations.
Conventional pulse (CP) imaging technique has a
peak power limitation imposed by the safety limits for human body to avoid, for instance, cavitations and internal
heating2.
Thus, adaptation of coded excitation pulse (CEP) from
radar and sonar theory has been implemented with success in medical ultrasound. This technique comprises the
application of long pulse with frequency modulation as excitation in the transmission, distributing the energy by its
frequency components without increasing the peak power
limits as would do in the CP. In the reception, the echo
signal obtained with CEP is compressed by matched filter,
restoring the possibility of detection of medium targets, improving the echo signal to noise ratio (eSNR) and retaining
the axial resolution (AR), even though this filter presents
sidelobe artifacts adjacent to the mainlobe, which degrade
the image quality2,3.
As our research group is working on the development
of low cost ultrasound equipment and studying different
approaches to obtain good quality images, we carried out
a comparative study between both CEP and CP techniques interacting on a scattering medium obeying statistical gamma distribution by simulations with computational
phantoms in one dimension (A-line simulation), for three
different transducer frequencies, highlighting the feasibility and validity of the techniques for several distribution
scenarios. Our results are confronted against those of the
literature.
Theory
Linear frequency modulation
Linear frequency modulation (chirp) is one among several
possible coded excitations and is the more usual, because
of its ease generation and unique properties in both time
and frequency domains. Mathematically, a common definition is denoted by Eq. 12,4:
B 2¹¼ T
T
¬ ©
s(t) = a(t).cos ­2p ª f0 t+
t º , f t f
2T » ½¾ 2
2
® «
where:
a(t) is the amplitude modulation function;
36
Revista Brasileira de Física Médica.2011;5(1):35-40.
(1)
f0 is the start frequency;
T is the chirp time duration; and
B is the bandwidth swept.
The core of signal modulation is the distribution of
the energy over all frequency components during time
T, allowing the increase of the time-bandwidth product
(TBP), which is nearly one for mono-frequency signals.
Therefore, to get TBP >1 is the key for modulation and
pulse compression4.
Ultrasonic pulse compression
The matched filter is a common filter to perform ultrasound echo pulse compression, because it maximizes
the eSNR in the presence of white noise. The complex
conjugate of the excitation signal used to excite the
transducer is the transfer function in the frequency domain of the matched filter4.
The main purpose of this filter is to concentrate the
energy distribution performed by the pulse modulation for
a single instant, bringing back the TBP to approximately
one again.
Nevertheless, the transducer bandwidth has a bandpass behavior, which poses limitations on the use of the
chirp bandwidth as well in the gain of the eSNR2,4.
The disadvantage in the compression process is the
generation of adjacent sidelobes around the mainlobe,
affecting the resolution and contrast of the image2,4.
Therefore, invariably, a new requirement is to apply a
weighted tapering of the excitation signal in the transmission and also move the matched filter (MF) to a mismatched condition for the sidelobe reduction below -45
dB, according to Haider et al.5.
To achieve this attenuation level, the mismatched filter
(MMF) is done by applying a window function on the transfer function of the MF in the reception4.
Signal-to-noise ratio and axial resolution
The ultrasonic echo detection sensitivity is measured by
SNR value. Therefore, eSNR is a good metric to evaluate
a pulse-echo system because this relation helps to determine the contrast resolution of the system6,7. The eSNR for
CEP can be written as in Eq. 2:
eSNR(χ)(dB) = 10.log(TBP) + eSNRCONV (χ)
(2)
where:
eSNRCONV is the known metric for conventional pulse.
To evaluate the axial resolution, the common metric is defined as in Eq. 32:
AR =
ct
2 where:
c is the sound speed,
t is the pulse length.
(3)
A comparative study using both coded excitation and conventional pulses in the evaluation of signal to noise ratio sensitivity and axial resolution in ultrasonic A-mode scan
Materials and Methods
Results
All algorithms used in both eSNR and AR test validation
were developed using MATLAB® (MathWorks Inc., EUA)
software.
We present the main aspects of the methodology adopted as follows:
Ultrasonicpulse: a Gaussian pulse was generated with
a Gaussian envelope modulating a sine wave and we assumed a circular transducer.
Amplitude tapering and mismatched filter: the chirp
was tapered by 0.15 ratio Tukey window implemented
with the MATLAB tukeywin function. For MMF, the -60 dB
Chebyshev window was chosen.
Computational phantoms: one-dimension structures
were built and the scatterers were modeled as complex
variables containing magnitude and phase. The phase
was distributed between 0 and 2 π, randomly varying. The
amplitude obeys the statistical gamma distribution and
we used the MATLAB gamrnd function1. Figure 1 shows
an example of the arrangement of scatterers in one of the
constructed phantoms.
A set of ten simulations was performed for several distributions and conditions of scattering: regular, regular plus
random, and random. For each condition we varied amplitude and phase in a combined way. Both CP and CEP
interacted with the scatters by convolution generating RF
A-lines, which were then processed.
We used transducers operating at 2.25, 5.0, and 7.5
MHz. The chirps sweeping bandwidths (B) were 3.7, 2.0
and 1.2 greater than the respective transducer bandwidth
(65% of transducer center frequency). We then compute
AR, SNR average values and their associated standard
deviation for each case.
The global parameters used were:
• Soundspeed(c):1540m/s.
• Phantomdimension(axialdirection):60mm.
• Relativetransducerbandwidth(BWR):65%.
• Timedurationofthechirp(T):18μs.
In Figure 2 it is shown the A-line obtained by CEP interaction with the phantom of Figure 1. As one can see, the
target information is not resolved and pulse compression
mechanism must be applied to solve the problem of resolution. In Figure 3, we present the output profile of both MF
and MMF filters on CEP.
Applying the filters of Figure 3 to the A-line shown in
Figure 2, the target information (position along the transducer axis) is obtained (Figure 4), where the RF signal and
its respective envelop are shown.
In Tables 1, 2 and 3, we show eSNR values for each
transducer after processing the echo signal obtained with
CP and CEP after application of both MF and MMF filters
for several scenarios. In all tables, we have: the amplitude
variation (AV); random phase (RP); σ as the standard deviation. In Table 4 it is shown the AR for the transducers
excitation pulses (CP, CEP+MF, CEP+MMF).
0.8
Normalized Amplitude
0.6
0.2
0
-0,2
-0,4
-0,8
0
10
20
30
40
Depth [mm]
Matched Filter (MF)
Mismatched Filter (MMF)
-10
Normalized Amplitude [dB]
0.8
0.6
0.4
60
Pulse compression outputs
0
1
50
Figure 2. Echo signal with unresolved targets due long length of
the CEP convoluted with phantom of the Figure 1.
Scatterers
Amplitude
0.4
-0,6
Computational phantom
-20
-30
-40
-50
-60
-70
-80
-90
0.2
0
Echo returned from scatterers
1
-100
0
0
10
20
30
40
Depth [mm]
Figure 1. Spatial distribution of the scatterers.
50
60
2
4
6
8
10
Time, [μs]
12
14
16
18
Figure 3. Profile of the pulse compression outputs for 2.25 MHz
transducer.
Revista Brasileira de Física Médica.2011;5(1):35-40.
37
Machado TM, Costa ET
Table 4. AR values obtained for both 2.25, 5.0 and 7.5 MHz
transducers (ultrasound wave propagating in water)
RF and envelope of A-line echo signal
1
RF signal
Envelope
0.8
Spatial resolution [mm and λ]
Transducer
frequency [MHz]
CP
CEP+MF CEP+MMF
2.25 (λ=0,7 mm)
0.91 (1.33 λ) 0.80 (1.18 λ) 0.93 (1.38 λ)
5.0 (λ=0,3 mm)
0.41 (1.34 λ) 0.35 (1.14 λ) 0.40 (1.29 λ)
7.5 (λ=0,2 mm)
0.26 (1.31 λ) 0.25 (1.23 λ) 0.28 (1.38 λ)
Normalized Amplitude
0.6
0.4
0.2
0
-0,2
-0,4
Discussion
-0,6
-0,8
0
0
10
20
30
40
Depth [mm]
50
60
Figure 4. Resolved target of structured phantom of Figure1 after pulse compression applied to signal shown in Figure 2. We
show RF signal and its envelope.
Table 1. SNR values obtained for the 2.25 MHz transducer
Spatial distribution
of the scatterers
CP
eSNR
σ
(dB)
(dB)
CEP+MF
eSNR
σ
(dB)
(dB)
CEP+MMF
eSNR
σ
(dB)
(dB)
Regular
25.15 1.41 47.52 0.95 45.87 1.06
(AV + RP)
Regular + Random
30.85 1.42 46.45 0.77 44.66 0.70
(AV + RP)
Random
31.68 0.83 48.16 0.49 46.47 0.51
(AV + RP)
Table 2. SNR values obtained for the 5.0 MHz transducer
Spatial distribution
of the scatterers
CP
eSNR
σ
(dB)
(dB)
CEP+MF
eSNR
σ
(dB)
(dB)
CEP+MMF
eSNR
σ
(dB)
(dB)
Regular
26.48 0.89 46.43 0.98 45.88 0.94
(AV + RP)
Regular + Random
32.00 1.33 45.76 1.29 45.23 1.29
(AV + RP)
Random
31.24 1.37 47.08 0.88 46.58 0.90
(AV + RP)
Table 3. SNR values obtained for the 7.5 MHz transducer
Spatial distribution
of the scatterers
CP
eSNR
σ
(dB)
(dB)
CEP+MF
eSNR
σ
(dB)
(dB)
CEP+MMF
eSNR
σ
(dB)
(dB)
Regular
27.09 1.13 46.94 0.73 45.98 0.79
(AV + RP)
Regular + Random
31.78 1.29 46.91 1.15 45.87 1.17
(AV + RP)
Random
31.40 1.49 47.30 0.76 46.30 0.71
(AV + RP)
38
Revista Brasileira de Física Médica.2011;5(1):35-40.
O’Donnell8 and Misaridis4 discuss the potential gain factor of the eSNR, indicating a numerical range between 15
and 20 dB. Misaridis4 also claims that the use of MMF in a
medium without attenuation leads to a loss of eSNR in the
range of 1 to 2 dB.
It is important to note that the application of MMF
leads to a slight decrease in eSNR, because it invariably
broadens the mainlobe. Moreover, some authors4,7 indicate that the gain factor of the eSNR is given by the TBP.
The results summarized in Tables 1, 2 and 3 show good
agreement with those obtained in the literature, taking
into account that in those situations, tests were mainly
focused on B-mode images while in our case we used
only A-line mode.
In Table 1 we see that the gain obtained with CEP varies between 15.60 to 22.37 dB. When MMF was applied
over MF, there was a loss in the eSNR gain between 1.65
and 1.79 dB. The eSNR theoretical gain expected in this
case (TBP=97) is 19.85 dB. For data in Table 2, the predicted theoretical gain is 20.68 dB with TBP=117. As one
can see, the CEP gain ranges between 13.76 and 19.95
dB and after MMF application, the loss varied from 0.50
to 0.55 dB. In Table 3, the SNR with CEP ranged from
15.13 to 19.85 dB with TBP=105 (eSNR theoretical gain
is 20.21 dB).
For spatial resolution, we compare the ordinary values
for CP with those obtained by CEP+MF and CEP+MMF
(pulse compression). According to Behar and Adam2 and
Misaridis4, the AR produced by these filters is calculated
at -20 dB of the maximum mainlobe. We can notice that,
if we only work with CEP+MF, the AR shows better results
than after application of the MMF (CEP+MMF). However,
in the first case the sidelobes levels were in the dynamic
range of the ultrasound system (range between -40 and
-50 dB), which is unacceptable. Therefore, after applying MMF over the CEP+MF, the attenuation comes to acceptable levels at the expense of a slight broadening of
the mainlobe, but not compromising too much the AR
system.
Conclusions
We have presented a comparative study of the SNR and
AR obtained when processing ultrasound A-line echo
A comparative study using both coded excitation and conventional pulses in the evaluation of signal to noise ratio sensitivity and axial resolution in ultrasonic A-mode scan
signals obtained with conventional and with codified pulse
transducer excitation. The present results are in accordance with literature data and have also shown that it is
possible to improve SNR with codified pulse, when proper
procedures are applied to regain system resolution. These
are important results for our research group and shall be
used in our ultrasound system development, although further studies are required.
References
1. Vivas GC. Estudos de modelos estatísticos utilizados na caracterização de
tecidos por ultra-som. [dissertação]. Campinas: Faculdade de Engenharia
Elétrica e de Computação, Universidade Estadual de Campinas, 2006.
2. Behar V, Adam D. Parameter optimization of pulse compression in ultrasound
imaging systems with coded excitation. Ultrasonics. 2004;42(10):1101-9.
3. Pedersen MH, Misaridis TX, Jensen JA. Clinical evaluation of chirpcoded excitation in medical ultrasound. Ultrasound Med Biol.
2003;29(6):895-905.
4. Misaridis T. Ultrasound imaging using coded signals. [thesis]. Denmark:
Technical University of Denmark, 2001.
5. Haider B, Lewin PA, Thomenius KE. Pulse elongation and deconvolution
filtering for medical ultrasonic imaging, IEEE Trans Ultrason Ferroelect Freq
Contr. 1998;45(1):98-113.
6. Liu J, Kim KS, Insana MF. SNR comparisons of beamforming strategies.
IEEE Trans Ultrason Ferroelectr Freq Contr. 2007;54(5):1010-7.
7. Oelze ML. Bandwidth and resolution enhancement through
pulse compression. IEEE Trans Ultrason Ferroelectr Freq Contr.
2007;54(4):768-81.
8. O’Donnell M. Coded excitation system for improving the penetration of realtime phased-array imaging systems. IEEE Trans Ultrason Ferroelectr Freq
Contr. 1992;39(3):341-51.
Revista Brasileira de Física Médica.2011;5(1):35-40.
39
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):41-6.
Neutron stimulated emission computed
tomography applied to the assessment
of calcium deposition due to the presence
of microcalcifications associated
with breast cancer
Tomografia computadorizada de emissão estimulada
por nêutrons aplicada para avaliar a deposição de cálcio
devido à presença de microcalcificações associadas ao
câncer de mama
Rodrigo S. S. Viana1 and Hélio Yoriyaz1
1
Instituto de Pesquisas Energéticas e Nucleares, Comissão Nacional de Energia Nuclear (IPEN/CNEN), São Paulo (SP), Brazil.
Abstract
In this paper we presented an application of the Neutron Stimulated Emission Computed Tomography (NSECT), which uses a thin beam of fast neutrons
to stimulate stable nuclei in a sample, emitting characteristic gamma radiation. The photon energy is unique and it is used to identify the emitting
nuclei. This technique was applied for evaluating the calcium isotopic composition changing due to the development of breast microcalcifications.
A particular situation was simulated in which clustered microcalcifications were modeled with diameters less than 1.40 mm. In this case, neutron
beam breast spectroscopy was successful in detecting the counting changes in the photon emission spectra for energies, which are characteristics
of 40Ca isotope in a low deposited dose rate.
Keywords: NSECT, microcalcifications, breast cancer, spectroscopy, diagnosis.
Resumo
Neste trabalho, apresentou-se a aplicação da tomografia computadorizada de emissão estimulada por nêutrons (TCEN), que usa um feixe de
nêutrons rápidos para estimular os núcleos estáveis em uma amostra, emitindo radiação gama característica. A energia do fóton é única e é utilizada
para identificar os núcleos emissores. Esta técnica foi aplicada para avaliar a mudança da composição isotópica de cálcio devido ao desenvolvimento
de microcalcificações mamárias. Uma situação em particular foi simulada, na qual microcalcificações agrupadas foram modeladas com diâmetros
inferiores a 1,40 mm. Neste caso, a espectroscopia da mama do feixe de nêutrons obteve sucesso ao detectar as mudanças da contagem nos
espectros de emissão de fótons para energias, que são características do isótopo de 40Ca em uma razão de baixa taxa de dose depositada.
Palavras-chave: TCEN, microcalcificações, câncer de mama, espectroscopia, diagnóstico.
Introduction
Breast cancer is the second most common cancer worldwide and the leading cause of death among women in
Brazil. According to estimates for 2010, it is expected approximately 49,000 new diagnosed cases1.
One of the main signs of breast cancer at an early diagnosis is the development of microcalcifications. Because
of calcium radiological properties, microcalcifications are
associated with nonpalpable lesions that can be visualized on mammography, which makes it the primary mode
of breast cancer diagnosis2. The importance of detecting
microcalcification formations in their early stages is a wellknown fact and, according to the literature, the survival
rate of patients who developed breast cancer is inversely
proportional to the lesion size. Regardless of prognosis,
Corresponding author: Rodrigo Sartorelo Salemi Viana – Instituto de Pesquisas Energéticas e Nucleares – IPEN – Avenida Lineu Prestes, 2.242, Cidade
Universitária, P.O. Box 11049, São Paulo (SP), Brazil – CEP 05508-000 – E-mail: [email protected]
Associação Brasileira de Física Médica®
41
SSV Rodrigo, Yoriyaz H
women with invasive breast tumors with a diameter of 10
mm or even smaller die due to complications at diagnosis3. However, in early stages, those microcalcifications
are very small, which become difficult to be detected by
mammography.
The presence of isolated breast microcalcifications is
not a decisive factor in the diagnosis of breast cancer, but
it is one of the first signs of metabolism disorder. In addition, the morphological changes caused by microcalcifications and visible on mammography screening occurs
later on physiological changes due to increased calcium
deposition.
In recent years, a new technique for in vivo spectrographic imaging of stable isotopes was presented as
Neutron Stimulated Emission Computed Tomography
(NSECT)4. In this technique, which uses multiple projections, a fast neutron beam interacts with the stable isotopes of the irradiated tissue through inelastic scatterings,
making them jump into an excited state. When they return
to their ground state, they emit photons, which energies
are intrinsic to the emitting nuclei. The emitted gamma
energy spectra can be used for two purposes: to reconstruct the target tissue image and; to determine the tissue
elemental composition. Considering a clinical application,
the spectroscopy of elements distribution in the body may
be used in the study of the tissues metabolism. As the
development of calcium deposits in the form of microcalcifications alter the abundance of this element in the breast,
this spectrographic technique may be used to evaluate the
calcium isotopic composition changing due to the development of microcalcifications.
In the present work, the energy spectrum data obtained from the simulated spectroscopy of a healthy breast
have been compared to those obtained from the simulated
spectroscopy of a breast model with inserted microcalcifications and different diameters. Simulations have been
done using the Monte Carlo code MCNP5. From these
comparisons, it was possible to establish a relationship between the microcalcification sizes and the calcium emission
photopeak intensities. A particular situation represented
by clustered microcalcifications has also been analyzed. In
this approach, a mammography unit was simulated in order to relate the variation of calcium isotopic composition
with the spatial distribution of microcalcifications.
Methodology
Monte Carlo code MCNP5
The Monte Carlo method can be described as a statistical one, which uses a sequence of random numbers to
perform a simulation. In terms of radiation transport, the
stochastic process can be seen as a family of particles
moving randomly in each individual collision as they travel
through matter. The average behavior of these particles is
described in terms of macroscopic quantities, such as flux
or particle density. The expected value of these quantities
42
Revista Brasileira de Física Médica.2011;5(1):41-6.
corresponds to the deterministic solution of the Boltzman
equation. Specific quantities, such as deposited energy or
dose, are derived from these quantities.
The MCNP code is a well-known and widely used
Monte Carlo code for neutron, photon, and electron transport simulations5. The first MCNP version was released in
the mid-1970s for neutron and photon transport, and it was
enhanced over the years to include generalized sources
and tallies, electron physics and coupled electron-photon
calculations, macrobody geometry, statistical convergence
tests and other features. The present work used the last
MCNP released version, which is the 5th. The MCNP5 particle transport simulation requires an input file (inp), which
allows the user to specify all the information about geometry modeling, source specifications, material compositions,
and specific quantities to be estimated (tallies).
Simulations
The simulated breast was modeled as a half of an ellipsoid
placed in the x-y plane. The breast composition was taken
from the literature6. Microcalcifications assume two distinct
chemical compositions: calcium oxalate (CaC2O4.2H2O)
and hydroxyapatite (Ca10(PO4).6H2O). Some morphological
characteristics revealed that benign tumors have microcalcifications predominantly composed by calcium oxalate,
while microcalcifications composed by hydroxyapatite can
be associated with both benign and malignant tumors2.
The microcalcifications simulated in this paper were modeled considering these two chemical compositions.
The NSECT is a spectrographic technique and the
analysis of any sample is understood by both the spatial
distribution of stable isotopes and the photon emission
spectrum, which characterizes the isotopic composition
of irradiated medium. However, the approach proposed in
this paper uses only the spectroscopic analysis of tissues
under investigation.
First, the photon emission spectrum of the healthy
breast was obtained and used as a reference assuming the existence of a normal6 calcium concentration.
Subsequently, the breast was modeled with the inclusion
of microcalcifications with different diameters (1-14 mm)
and using both chemical compositions already described.
The resultant spectra obtained from the simulations were
compared with the reference spectrum, with the aim of
establishing a relationship between the diameters of the
microcalcifications and the calcium emission photopeak
intensities. Since the background is a common factor in all
obtained spectra, it was not necessary to adopt any suppression or background extraction procedure.
Two hyper-pure germanium (HPGe) detectors were
modeled as cylinders of 5.32 g/cm3 density, with 12 cm
diameter and 15 cm height. The detectors were separated
90° from each other and both forms 45° with the neutron
beam axis. The neutron source was modeled in MCNP5 as
a monoenergetic energy beam of 7 MeV and with a square
section of 1 cm2. 5x108 incident neutrons have been simulated and photons, whose emission was stimulated by
Neutron stimulated emission computed tomography applied to the assessment of calcium deposition due to the presence of microcalcifications associated with breast cancer
inelastic scattering of fast neutron beam, were recorded
on the surface of the detectors using the F2 superficial flux
tally. This tally estimates the average particle scalar flux on
a user-specified surface and it was associated with the
En card that allows separating the counting photons according to energy bins of interest. Using this MCNP5 tally
resource, it was possible to build the energy spectrum of
the scattered photons arriving in the detectors. The configuration adopted of the spectrometric system is shown
in Figure 1.
Since 1913, when the first description of microcalcifications in a mammography was reported, many studies
were conducted to characterize and classify the types
of microcalcifications7. Because microcalcifications are
radiopaque structures, some researches show the importance of monitoring the development of calcifications
mainly in the early diagnosis through mammography
screening. Additionally, as hydroxyapatite can be found
in both malignant and benign tumors, other parameters
associated with microcalcifications should be evaluated,
such as shape, composition, quantity, and distribution.
However, the probability of malignancy is proportional to
the number of calcifications8. Based on this fact, in order
to simulate a more realistic case, using the spectrometric
system setup showed in Figure 1, randomized clustered
hydroxyapatite microcalcifications with different diameters (0.15-1.40 mm) have been modeled for analysis.
The spectrum obtained was compared with the reference
spectrum of the healthy breast.
As already described, mammography is the primary
mode of diagnosis of breast cancer and currently it is
in constant technological development. To confirm the
change in the calcium abundance, due to clustered microcalcifications development, a mammography unit was
modeled considering a 23 keV photon beam focusing on
a molybdenum target at a distance of 15 cm from the rhodium filter 25 μm thick and 45 cm from the compressor
plate. A 10 cm compression thick and a decrease in breast
volume by 10% were considered. Using the resources
available in MCNP5, the mammography screening was
simulated using the flux image radiograph (FIR) tally, which
property reproduces a radiographic image of the photon
flux that goes through a user-specified image grid.
In any diagnostic techniques with ionizing radiation, the
absorbed dose in patients during the procedure requires
special attention, and all intrinsic parameters to the diagnosis should ensure that the ALARA principle be satisfied.
Therefore, the absorbed dose rate on neutron spectroscopy of the breast, with clustered microcalcifications, was
estimated and compared with the allowed limits for the average glandular dose for mammography.
and 6 GB RAM desktop, in an average CPU time of five
days. All results presented in this section were obtained
considering a maximum standard deviation of 3%.
As described in the methodology section, in the first
approach, the emission spectrum of the healthy breast
and the spectrum of the breast with the inclusion of microcalcifications of different diameters have been simulated.
Comparing both calculated spectra, it was possible to
observe the change of the breast isotopic composition in
function of breast calcium concentration. Figure 2 shows
the behavior of the normalized counts by emission spectrum of the healthy breast for different photopeak energies
characteristics of 40Ca isotope, due to the increase in the
microcalcifications diameter.
The first feature that can be observed is the difference
in the range of normalized counts relative to differences in
the microcalcification compositions. This behavior is justified by the number of calcium atoms in hydroxyapatite
and calcium oxalate molecules, which has a ratio of 10-1.
According to the literature, the recognition of lesion malignancy by invasive methods is determined by physical,
chemical, and morphological characteristics of the lesion
sample, and the most used noninvasive procedure is the
analysis of mammographic findings. However, noninvasive
methods like mammography for diagnosis of breast cancer are limited to a minimum detectable size of the microcalcification itself. As an example, we can mention the use
of high-frequency ultrasound9.
With the obtained spectra, it is possible to verify the
sensitivity of the presented spectrometric technique to distinguish the composition of microcalcifications as a function of the amplitude of the normalized counts once calcium oxalate microcalcifications are strictly associated with
benign tumors. Another favorable factor for the diagnosis
is associated with the possibility of evaluating 2 mm diameter microcalcifications and even smaller.
The second approach proposed is to model a breast
with clustered microcalcifications and obtain spectroscopy
and mammography to verify the change in the isotopic
composition of calcium through the presence of microcalcifications. To achieve this purpose, 16 hydroxyapatite
Results and Discussion
Simulations were performed on a Linux environment at
Ubuntu operating system on a 2.67 GHz Intel® CoreTM i7
Figure 1. The MCNP5 spectrometric system setup: Breast (1);
HPGe detectors (2), and neutron beam (3).
Revista Brasileira de Física Médica.2011;5(1):41-6.
43
SSV Rodrigo, Yoriyaz H
Figure 2. Change of normalized counts in accordance with the increase in diameter of microcalcifications of different chemical compositions: hydroxyapatite (A) and calcium oxalate (B).
Figure 3. Photon emission spectra of the healthy breast and the one with clustered microcalcifications.
microcalcifications with diameters ranging from 0.15
to 1.40 mm were simulated. Figure 3 shows the energy
spectrum of the healthy breast and breast with clustered
microcalcifications in a range from 50 to 4,000 keV with
the detail of the energy range of interest.
Even for clustered microcalcifications with diameters
smaller than 1.40 mm, it is possible to observe the increased calcium isotopic composition, altering the normalized counts of the emission spectrum for the 3,736
keV and 3,904 keV energies. In a clinical application, this
result confirms the ability of NSECT spectroscopy mode
in detecting the change of calcium abundance due to the
development of microcalcifications or even other conditions that would be associated with the disorder in the
calcium production.
Using the same microcalcifications arrangement,
the mammography screening was performed using the
MCNP5 FIR tally to simulate the radiographic image of the
44
Revista Brasileira de Física Médica.2011;5(1):41-6.
photon on a high-resolution matrix array over a 16 x 16
cm field, with 0.1 mm resolution discrimination per pixel.
Figure 4 shows the radiographic image obtained, it is possible to visualize some clustered radiopaque structures. In
a hypothetical clinical case, this image could represent the
first indication of breast physiology disorder.
Neutron radiation dose
Using the resources available in MCNP5, the average energy deposited by each 7 MeV neutron that interacts in
breast was estimated to be 2.92 MeV. As the results provided by MCNP5 are normalized by the number of simulated particles, to obtain the absolute absorbed dose is
necessary to assume that the neutron source intensity is
known. For the calculation purpose, the standard intensity
of an Am-Be sealed natural source, with the intensity of
107 neutrons emitted per second on a single projection,
was adopted. Assuming that the fast neutron beam has
Neutron stimulated emission computed tomography applied to the assessment of calcium deposition due to the presence of microcalcifications associated with breast cancer
been simulated with this intensity, the absorbed dose rate
obtained in the breast spectroscopy is 0.0074 mGy/s.
According to the American College of Radiology10, the
tolerated limit for the breast average glandular dose is 3
mGy in a single exposure. Whereas the exposure times are
low, the absorbed dose rate in mammography could be
superior to that of neutron spectroscopy.
Conclusion
It was demonstrated the ability of NSECT spectroscopy
mode in detecting the change in calcium deposition due
to the development of hydroxyapatite and calcium oxalate
microcalcifications. The results obtained, where microcalcifications with different diameters were inserted on the
healthy breast, revealed the change of the breast isotopic
composition in function of the increasing calcium abundance through normalized counts of photopeak energies
of this element.
In a clustered microcalcifications situation, even when
considering microcalcifications with diameters less than
1.40 mm, the breast spectroscopy was able to detect the
isotopic composition changing, and this task was achieved
under a low deposited dose rate if compared to the average glandular dose limit for mammography.
Considering a compromise between deposited dose
and the counting efficiency of the detectors, the factors
that could have prevented the application of NSECT spectroscopy mode on breast isotopic composition analysis
are exposure time and neutron source intensity.
Figure 4. Mammography screening performed with 0.1 mm
resolution discrimination per pixel of breast with clustered microcalcifications.
3.
4.
5.
Acknowledgment
This paper was supported by Fundação de Amparo à
Pesquisa do Estado de São Paulo (FAPESP), grant number 2010/04206-4.
References
1. Brasil. Ministério da Saúde. Instituto Nacional de Câncer. Estimativa 2010.
Incidência de câncer no Brasil. Available at: http://www.inca.gov.br/
estimativa/2010/estimativa20091201.pdf.
2. Morgan M, Cooke M, McCarthy M. Microcalcifications associated with
6.
7.
8.
9.
10.
breast cancer: an epiphenomenon or biologically significant feature of
selected tumors? J Mammary Gland Biol Neoplasia. 2005;10:181-7.
Tabár L, Chen H, Duffy S, Yen M, Chiang C, Dean P, et al. A novel method for
prediction of long-term outcome of women with T1a, T1b, and 10-14 mm
invasive breast cancers: a prospective study. Lancet. 2000;355:429-33.
Floyd C Jr, Bender J, Sharma A, Kapadia A, Xia J, Harrawood B, et al.
Introduction to neutron stimulated emission computed tomography. Phys
Med Biol. 2006;51:3375-90.
Browm F, Barrett R, Booth T, Bull J, Cox L, Forster R, et al. MCNP Version
5. Applied Physics Division - Los Alamos National Laboratory, LA-UR-023935; 2002.
Bender J, Kapadia A, Sharma A, Tourassi G, Harrawood B, Floyd C Jr. Breast
cancer detection using neutron stimulated emission computed tomography:
prominent elements and dose requirements. Med Phys. 2007.34:
3866-71.
Salomom A, Beiträge ZU. Pathologie der mamacarcinome. Arch Klin Chir.
1913;101:572-668.
Hallgrimsson P, Kåresen R, Artun K, Skjennald A. Non-palpable breast
lesions. Diagnostic criteria and preoperative localization. Acta Radiol.
1988;29:285-88.
Woo M, Jung-Gi I, Young H, Dong-Young N, In A. US of Mammographically
Detected Clustered Microcalcifications. Radiology. 2000;217:849-54.
American College of Radiology. Mammography Quality Control Manual.
ACR Committee on Quality Assurance in Mammography; 1999.
Revista Brasileira de Física Médica.2011;5(1):41-6.
45
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):47-52.
Image processing techniques to evaluate
mammography screening quality
Técnicas de processamento de imagem para avaliar a
qualidade de exames de mamografia
Clara Quintana1,2, Germán Tirao1,2 and Mauro Valente1,2
1
Consejo Nacional de Investigaciones Científicas y Técnicas, Ciudad Autónoma de Buenos Aires, Argentina.
2
Facultad de Matemática, Astronomía y Física, Universidad Nacional de Córdoba, Córdoba, Argentina.
Abstract
Mammography imaging has proved to be the best noninvasive method for breast cancer diagnosis, but it requires that irradiation parameters are
set within Protocols recommendations (minimal dose delivering). This work presents an investigation on mammography image formation by means
of validated Monte Carlo simulations along with further image analysis and mathematical processing. Several image processing methods have
been suitably introduced and investigated according to their capability for micro-calcification detection and quality evaluation. The obtained results
suggest the feasibility of all the proposed methods. Furthermore, it was possible to characterize the reliability of each and to infer the corresponding
advantages or disadvantages, obtaining an image quality evaluation as a function of several parameters configurations.
Keywords: X-ray imaging, image processing, quality evaluation.
Resumo
A imagem por mamografia é comprovadamente o melhor método não-invasivo para o diagnóstico do câncer de mama, mas requer que os parâmetros
de irradiação sejam estabelecidos de acordo com as recomendações de protocolos (dose mínima). Este trabalho apresenta uma investigação
completa sobre a formação da imagem pela mamografia por meio de simulações Monte Carlo validadas juntamente com uma análise da imagem
e o processamento matemático. Diversos métodos de processamento da imagem foram apropriadamente introduzidos e investigados de acordo
com a capacidade deles em detectar microcalcificações e avaliar a qualidade. Os resultados obtidos sugerem a facilidade de todos os métodos
propostos. Além disso, foi possível caracterizar a confiabilidade de cada método e deduzir as vantagens ou desvantagens correspondentes, obtendo
uma avaliação de qualidade da imagem como uma função de várias configurações dos parâmetros.
Palavras-chave: imagem por raios X, processamento de imagem, avaliação da qualidade.
Introduction
Breast cancer is the most frequent cancer in women all
over the world. The main established strategies for breast
cancer control are based on primary prevention along
with early diagnosis. In this sense, breast imaging plays
an outstanding role for the screening and diagnosis of
symptomatic women. Mammography units may differ by
X-ray beam characteristics, breast compressing plate system, and X-ray detector1. Breast and micro-calcification
(μCa) composition material absorption properties show a
strong dependence on X-ray spectrum. Furthermore, specific patient breast characteristics, lesion shape and size
as well as examination exposure may influence the final
image quality. This paper presents different mathematical
algorithms with the aim of evaluating the mammographic
image quality, based on dedicated image processing techniques and devoted to µCa detection. In order to assess its
feasibility, the integral system was applied to a wide range
of clinical situations including different spectral characteristics, breast composition and thickness, µCa shape, size,
and composition.
Materials and Methods
Mammography image formation
A typical mammography facility has been modeled by
means of dedicated and validated Monte Carlo sub-routines2, which are capable of performing absorption contrast
images. This simulation toolkit, based on the PENELOPE
v. 2008 main code3, has been applied to investigate the
Corresponding author: Prof. Germán Tirao – CONICET & FaMAF University of Córdoba – Medina Allende y Haya de la Torre, Ciudad Universitaria – Córdoba –
Argentina – E-mail: [email protected]
Associação Brasileira de Física Médica®
47
Quintana C, Tirao G, Valente M
dependence of image quality upon irradiation configurations, according to different parameters: anode and accelerating voltage, breast tissue composition and thickness,
and compositions, shapes and sizes of µCa, taking into
account typical situations. Fifty-four different combinations
have been studied, each containing nine µCas of different
shapes and sizes, in different positions within the breast.
Image processing
Different image processing techniques along with the corresponding numerical algorithms have been developed
with the aim of assessing suitable and almost automatic
methods for µCa detection. The first approach consists on
true-false threshold segmentation methods, which assigns
1 or 0 value to each pixel when its intensity is greater or
lower than a certain preestablished threshold value. The
second approach implements the Canny algorithm for
edge detection, which finds edges by looking for gradient
(as the Gaussian filter derivative) local maxima within the
original image. The method incorporated two thresholds,
devoted to detect strong (T1) and weak (T2) edges, including weak edges in the output only if they were connected to the strong ones. The third approach consists on
a template matching routine, designed for searching a user
predefined pattern within the original image. The matching
process moves the template image to all possible positions
in the source image and computes a numerical index (correlation), which indicates how well the template matches
the image in that position, resulting in a correlation map
with peaks located at positions of µCa.
In order to assess the quality of the image as a function
of physical parameters involved in its formation and the ability to µCa detection, several mathematical processing techniques were applied to the entire set of simulated images.
These techniques were based on intrinsic properties calculations and comparison with ideal binary pattern images,
defined with the same geometry details (positions and sizes
of µCa used in the simulation). In additions, typical techniques based on intrinsic properties were implemented for
this paper, such as: Dynamic Range (DR), Signal-to-noise
radio (SNR)4, Contrast-to-noise radio (CNR)4 and Entropy
(H)5. The implemented comparative techniques were developed based on concepts of information theory, namely:
Joint Entropy (JH)5, Mutual Information (MI)5, Normalized
Cross Correlation (NCC)6, and index Q (Q)7.
The DR is the number of gray levels in which the information is distributed. It is expected that the quality of
the image may preserve a direct relation with the DR. It
was calculated from the differences between maximum
and minimum gray levels within the image. The SNR value of image was calculated as the average SNR value of
each pixel, defined as the ratio of pixel intensity to its noise
standard deviation. Similarly, CNR value of image can be
defined as the ratio of differences between DR and noise
standard deviation. Therefore, it may be expected that
both SNR and CNR values should increase as image quality gets improved.
48
Revista Brasileira de Física Médica.2011;5(1):47-52.
The entropy can be seen as a measure of uncertainty,
since its maximum occurs when all symbols have equal
probability of occurrence5. Images may also be considered
as carriers of information, where instead of probabilities of
symbol occurrence, one has the distribution of gray values.
It seems intuitive to infer that an image with a nearly uniform distribution of gray tones may have very little information. The image quality and entropy relationship depends
on image noise. Since noise can be interpreted as information, if the noise is low, it is expected higher quality image
with more information. The entropy can be calculated using the Eq. 1:
(1)
where:
p(i )
is the probability distribution of intensity i.
From the entropy, joint entropy is defined as Eq. 2:
(2)
where:
p (i, j )
is the joint probability defined from the joint
histogram;
A represents a simulated image and
B is the ideal binary image.
Then, images of better quality should minimize the JH.
The MI can be defined in different ways and each of these
definitions leads to different interpretations5. In this paper,
the one used is as in Eq. 3:
MI ( A, B)  H ( A)  H ( B)  H ( A, B)
(3)
where:
A represents a simulated image and
B is the ideal binary image.
In this sense, MI can be interpreted as the amount of
information that is not exclusive to any of the two images. It
is remarkable that the maximization of MI is closely related
to minimizing the JH.
In signal processing theory, cross-correlation (or sometimes called “cross-covariance”) is a measure of similarity
between two signals, often used to find relevant features
in an unknown signal by comparison with another that is
known. NCC has application in pattern recognition and
cryptanalysis. NCC of simulated image A and the ideal binary image B were calculated as in Eq. 4:
NCC ( A, B) 
 A(i, j )  A  B(i, j )  B 
i, j
med
med
 A(i, j )  A  B(i, j )  B 
i, j
med
2
2
med
(4)
Image processing techniques to evaluate mammography screening quality
where:
Amed (Bmed) is the average value of the image A (B).
Wang and Bovik7 proposed a mathematically defined
universal image quality index Q. By “universal”, the authors
meant that the measured quality approach did not depend
on the images being tested, the viewing conditions, and individual observers. It should be applicable to several image
processing applications and provide meaningful comparisons across different types of image distortions. The index
Q was calculated using the Eq. 5:
Q

4 AB Amed Bmed
2
A
  B2
 A
med
2
 Bmed
2
(5)

values, in order to achieve a good performance. As expected, greater values for the T2 threshold parameter (focused
on weak edges detection) resulted in best performances
for detecting image details, but becoming also greater the
possibility of “false true” due to the detection of noise as
image detail. Therefore, it may be advisable to overcome
this risk by means of considering high T1 threshold values (dedicated to strong edges detection), as reported in
Figures 2a and 2b. However, for lower T2 values, it may be
advisable to employ also lower T1 in order to be able to
detect µCa edges, as reported in Figures 2c and 2d.
The template matching approach has been applied,
obtaining a very good performance. In fact, this was the
method that best worked, detecting all µCa (9 of 9), while
the Canny algorithm only detected 8 µCa. The true-false
threshold segmentation method detected 7 µCa, but with
where:
 X2 
1
2
 X (i, j )  X med 
N  1 i, j
 AB 
and X = A or B,
1
 A(i, j )  Amed B(i, j )  Bmed 
N  1 i, j
The dynamic range of Q is [-1,1]. The value of 1 is
achieved when the original image and test image are equal
and the worst value when the test image is twice the mean
of original image subtracted by the original image.
Results and Discussion
Due to the fact that the considered breasts have different absorption paths associated with their ellipsoidal shapes, the
threshold segmentation method could not provide a satisfactory performance because of the nonuniform background.
Therefore, its suitability is strongly dependent on the image
position, as indicated in Figure 1. It was found that the performance of this method showed significant dependence upon
the threshold value. However, it should be mentioned that in
some cases this drawback may overcome if a suitable background (BG) is first subtracted from the original image.
The subtracted BG was calculated from the original
image following these steps: suitable smoothing process
is applied to original image, for each point in the longitudinal central axis (around which μCas where positioned,
as shown in Figure 1). It has been calculated the average
along the transverse profile obtaining therefore a suitable
BG for each point on the longitudinal central axis, which
has been used for μCa detection by means of the 1D algorithms. A whole 2D BG can be directly and straightforwardly calculated generalizing the proposed method to 2D
dimensions by means of considering repeating the mechanism for all non-central longitudinal axes (image rows).
The implemented edge detection method is required to
assess suitable combination of the user-defined threshold
Figure 1. µCa detection using image segmentation algorithms.
From left to right – Central axis profiles: original (top curve) and
BG subtracted (bottom curve) image; Original image, segmented image (threshold B), segmented image (threshold A), BG
subtracted image segmented (threshold C). Image parameters:
glandular breast tissue with 30 mm thick and 40 kV incident
spectrum, calcium oxalate µCa composition.
Figure 2. µCa detection using edge filter algorithms with different tolerance values. From left to right: a: Original image; b:
T1=0.05, T2=0.1; c: T1=0.1, T2=0.1; d: T1=0.05, T2=0.8; e:
T1=0.1, T2=0.8. Mammographic parameters: glandular breast
tissue with 30 mm thick for 40 kV incident spectrums, calcium
oxalate µCa composition.
Revista Brasileira de Física Médica.2011;5(1):47-52.
49
Quintana C, Tirao G, Valente M
a suitable BG subtraction it can detect all µCa. Another important advantage of this method is that the detection efficiency depends sensitively on the details size of the template, and not its form. This fact is clearly shown in Figure
3, where correlation maps for different templates sizes can
be observed, and also how a large template distinguishes
more noticeable large (small) µCa.
Regarding mathematical processing techniques to
evaluate image quality and µCa detection, which are both
the intrinsic properties calculations as parameters for comparison, it was noticed that almost all the studied parameters, except for JH, showed the expected tendencies with
respect to the physical parameters involved in the image
formation: accelerating voltage, breast tissue composition
and thickness, and µCa composition. The JH parameter
did not reflect the expected behavior, inferred according
to image formation processes and physical parameters
dependence. This fact may be explained because very
different images were compared. Despite the functional
relationship of MI and JH, MI showed the correct behav-
Figure 3. Correlation profiles for different templates sizes.
iors because JH is several orders of magnitude (about 3)
smaller than H.
Besides, the studied parameters were very useful to
establish a ranking of the images and thus to assess a
quality criterion. The ranking is set taking into account the
relationship of the parameter with the image quality, i.e.,
a maximum, minimum, or an appropriated value. In this
sense, the DR, CNR, H, and Q establish almost the same
ranking. These parameters measure the overall image contrast compared to the other studied parameters. The obtained ranking does not reflect the feasibility of µCa detection and therefore the image quality. For example, Figure
4 shows the central axis profiles for three different simulated images, the black, red and green lines correspond
to 30, 70 and 30 mm thickness of glandular breast tissue,
and 34, 40 and 24 kV incident spectrums, respectively.
For these cases, µCa composition was calcium oxalate. In
Figure 4, the µCa-9 was detected only for the black and
green axis profiles. Then, the image corresponding to the
red line presents a lower quality than the one corresponding to the green line, and therefore should be in a worse
position in the ranking. But the obtained ranking with this
parameter was sixth, seventh, and eight for the black, red
and green lines, respectively.
However, the ranking obtained using SNR values was
very similar to those obtained from DR, CNR, H, and Q
presenting only some minor differences. This parameter
weights the average value of intensity, resulting in a nonsuitable ranking, which does not reflect the ability to µCa
detection. As example, it can be seen from Figure 4 that
central axis profiles for a 24 kV incident spectrum and 30
mm thickness of glandular breast tissue (green line) should
not be better ranked than the image corresponding to red
line. However, the obtained ranking predicts the opposite.
The ranking using SNR values for these three examples
was fifth, sixth and eight for the red, black and green lines,
respectively.
Finally, the NCC parameter created a classification
quite different from those presented earlier, but much more
representative of the image quality. For the same examples
used in Figure 4, the corresponding ranking of this parameter was fifth, sixth and tenth for the black, green and
red lines, corresponding to 30, 70 and 30 mm thickness
of glandular breast tissue, and 34, 40 and 24 kV incident
spectrums, respectively. It can be observed that although
the red line has greater contrast than the others, it fails to
detect the µCa-9. This µCa is clearly detected by the other
profiles. The image quality defined by these parameters
represents the feasibility of µCa detection.
Conclusions
Figure 4. Central axis profiles for simulated images. The black,
red and green lines correspond to 30, 70 and 30 mm thickness
of glandular breast tissue, and 34, 40 and 24 kV incident spectrums, respectively. μCa composition was calcium oxalate.
50
Revista Brasileira de Física Médica.2011;5(1):47-52.
Different mammography image processing techniques
have been proposed and investigated. A suitable simulation
toolkit has been satisfactory implemented for this investigation. Image processing techniques reliability and suitability
Image processing techniques to evaluate mammography screening quality
for automatic detail detection have been carefully studied.
Image segmentation and edge filtering approaches showed
good performance for µCa detection. The template matching approach and its correlation values were excellent and
helpful tools for µCa detection, and they could also be used
as a quality parameter for different irradiation set-ups. These
algorithms proved to be a valuable and promising tool for
mammography images processing. The DR, CNR, H, and Q
algorithms establish almost the same ranking. Such parameters measure the overall image contrast. But, SNR values
weights average value of intensity, resulting in a nonsuitable
ranking, which does not reflect the ability to µCa detection.
Finally, NCC was the only one capable of satisfactory reflecting image quality as well as providing feasible µCa detection. NCC makes few requirements on the image sequence
and has no parameters to be searched by the user.
Acknowledgment
This paper has been partially supported by grants from
research Projects PIP 11420090100398, PICT 2008-243
along SeCyT from Consejo Nacional de Investigaciones
Cientificas y Tecnicas (CONICET), Agencia Nacional
de Promoción Científica y Tecnológica (ANPCyT) and
Universidade Nacional de Córdoba (UNC) of Argentina.
References
1. Berns EA, Hendrick R, Solari M, Barke L, Reddy D, Wolfman J, et al. Digital
and screen-film mammography: comparison of image acquisition and
interpretation times. Am. J. Roentgenol. 2006;187(1):38-41.
2. Tirao G, Quintana C, Malano F, Valente M. X–ray spectra by means of
Monte Carlo simulations for imaging applications. X Ray Spectrom.
2010;39(6):376-83. DOI: 10.1002/xrs.1279.
3. Salvat F, Fernández-Varea J, Sempau J. PENELOPE – Version 2008, NEA:
France; 2008.
4. Bushberg JT, Seibert JA, Leidholdt Jr EM, Boone JM. The essential physics
of medical imaging. California: Lippincott Williams & Wilkins; 2001.
5. Pluim JPW, Maintz JBA, Viergever MA. Mutual-information-based
registration of medical images: a survey. IEEE Transactions on Medical
Imaging. 2003;22(8):986-1004.
6. Lewis JP. Fast Template Matching, Vision Interface 95, Canadian Image
Processing and Pattern Recognition Society, Quebec, Canada, 1995; p.
120-3.
7. Wang Z, Bovik AC. A Universal Image Quality Index, IEEE Signal processing
Letters. 2002;9:81-4.
Revista Brasileira de Física Médica.2011;5(1):47-52.
51
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):53-6.
A pilot study – acute exposure to a lowintensity, low-frequency oscillating magnetic
field: effects on carrageenan-induced paw
edema in mice
Estudo piloto – exposição aguda a campo magnético
oscilante de baixa intensidade e baixa frequência: efeitos
sobre edema de pata induzido por carragenina, em
camundongos
Tania M. Yoshimura1, Daiane T. Meneguzzo2 and Rodrigo A.B. Lopes-Martins3
1
Linfospin, São Paulo (SP), Brazil.
Nuclear and Energy Research Institute, São Paulo (SP), Brazil.
3
Laboratory of Pharmacology and Experimental Therapeutics, Department of Pharmacology, Institute of Biomedical
Sciences, University of São Paulo, São Paulo (SP), Brazil.
2
Abstract
The purpose of this paper was to evaluate the effects of an oscillating magnetic field (MF) on edema evolution in an animal model. Paw edema was
induced in 32 female Swiss mice by injecting 50 µL of 1.0% carrageenan, diluted in saline solution in the left hind footpad. Animals were randomly
assigned into four Experimental (exposed to different field frequencies) and two Control Groups. Groups 1 (0 Hz), 2 (3 Hz), 3 (9 Hz) and 4 (15 Hz) were
exposed for 60 seconds to an oscillating MF (300 mT) in the first, second, and third hour after the injection. Control Groups (CGN and DCL) were not
exposed to the MF and diclofenac was administered to DCL Group one hour after the edema induction. Paw volumes were determined every hour
using a water plethysmometer. The results were graphed against time and, to evaluate the edema, the area under the curve (AUC) was measured.
All groups receiving some form of intervention (1, 2, 3, 4 and DCL) revealed AUC values that were substantially lower than those of the CGN Group.
DCL had the lowest reduction percentage (25.0±6.1%) and Group 3, the highest (46.9±4.0%). Compared to the results of DCL, only Groups 2 and
3 showed significantly lower AUC values. Also, with statistical relevance, Group 3 showed lower AUC values than Groups 1 and 4. According to this
experiment, acute exposure to oscillating MF yields positive results in the regression of carrageenan-induced edema in mice, with indications that
such effect depends on the field frequency.
Keywords: magnetic field, edema, carrageenan, inflammation, mice.
Resumo
O objetivo desse trabalho foi avaliar os efeitos da exposição aguda a um campo magnético (CM) oscilante sobre a evolução do edema em um modelo
animal. O edema foi induzido através de injeção subcutânea de 50 μL de carragenina (diluída a 1,0 % em solução salina) na região subplantar da
pata traseira esquerda de 32 camundongos fêmeas da linhagem Swiss. Os animais foram distribuídos de forma aleatória em 4 grupos experimentais
(expostos a diferentes frequências deCM) e 2 grupos controle. Os Grupos 1 (0 Hz), 2 (3 Hz), 3 (9 Hz) e 4 (15 Hz) foram expostos a CM oscilante (300
mT) durante 60s na primeira, segunda e terceira horas após a injeção de carragenina. Os Grupos Controle (CGN e DCL) não foram expostos ao CM, e
o Grupo DCL recebeu diclofenaco sódico 1h após a indução do edema. O volume das patas foi determinado a cada hora com auxílio de pletismógrafo
digital. Os resultados foram dispostos em gráfico contra o tempo, e foi calculada a área sob a curva (AUC) para avaliação do edema. Todos s grupos
que receberam algum tipo de intervenção (1, 2, 3, 4 e DCL) apresentaram valores de AUC significativamente menores do que CGN. DCL apresentou a
menor porcentagem de redução (25,0±6,1%) e o grupo 3, a maior (46,9±4,0%). Em relação ao DCL, somente os grupos 2 e 3 apresentaram valores
de AUC menores e significativos. O grupo 3 apresentou, com relevância estatística, valores de AUC menores do que os grupos 1 e 4. De acordo com
esse experimento, a exposição aguda a CM oscilante promove efeitos positivos na regressão de edema induzido por carragenina em camundongos,
com indicações de que tais efeitos são dependentes da frequência do CM.
Palavras-chave: campo magnético, edema, carragenina, inflamação, camundongos.
Corresponding author: Tania Mateus Yoshimura – IPEN - Instituto de Pesquisas Energéticas e Nucleares – Centro de Lasers e Aplicações – Av. Lineu Prestes
2242 - Cidade Universitária – São Paulo (SP), Brazil – CEP 05508-000 – E-mail: [email protected]
Associação Brasileira de Física Médica®
53
Yoshimura TM, Meneguzzo DT, Lopes-Martins RA
Introduction
In 1979, the Food and Drug Administration (FDA, USA) approved, for the first time, the use of pulsed electromagnetic
field (PEMF) for the treatment of non-united fractures and
failed arthrodesis1. Since then, several studies have been
conducted to demonstrate the benefits of electromagnetic
fields (EMFs) for the treatment of various conditions, such
as psoriasis2, edema3-5, osteoarthritis6, and difficult-to-heal
wounds7-9. The positive effects of EMFs in the modulation
of hemodynamics10, and pain relief caused by carpal tunnel syndrome11, besides the established use for bone formation, were also reported. The mechanisms that explain
the occurrence of these observed effects, however, are yet
to be completely understood.
Considering the principles that govern the electromagnetic phenomena, it is possible to infer that, in biological
systems and living organisms, EMFs are ubiquitous and
constantly created by physiological processes (cell movements, ionic fluxes, fluids flow in the circulatory systems,
mitochondrial electron transport chain, action potentials,
and so on). Therefore, biological effects can result from the
interaction of these fields with exogenous EMFs.
A recent review12 points out how electrical oscillations may play important physiological roles in the living
organism. Previous studies have reported that right after
the occurrence of a wound, an electric field is created at
the injured tissue, and this may be a primary stimulus perceived by epithelial cells to start the repair process of the
tissue. This “electrical disturbance” caused by the injury
may remain for several hours, and it only ceases when the
tissue is re-epithelialized. These observations indicate that
the cells are sensitive to electrical changes, being able to
respond specifically to them.
The major purposes of this paper were to evaluate the
effects of acute exposure to a low-intensity, low-frequency
oscillating magnetic field on paw edema induced by carrageenan in mice, and to verify if these effects depend on
the magnetic field (MF) frequency. It is also our objective to
compare the possible effects of the MF exposure to those
promoted by a renowned anti-inflammatory drug (sodium
diclofenac).
Materials and methods
Animals
All experiments were conducted in accordance with the
rules and regulations for animal care and use set forth
by the Institute of Biomedical Sciences (University of São
Paulo). Since this pilot study was based on classically established protocols, submission to the Institute’s Ethics
Committee was not obligatory.
Thirty-two female Swiss mice, 44 days-old, weighting
between 21 and 28 g, kept under ideal temperature conditions, with water and food ad libitum, in 12 hours/ 12 hours
dark/light cycles, were used for this study. At the day of
54
Revista Brasileira de Física Médica.2011;5(1):53-6.
the experiment, animals were randomly allocated in groups
of five or six inside PVC cages and were kept conscious
throughout the experimental procedure.
Paw edema induction
To induce the paw edema, the experiment followed the
classical model proposed by Winter et al., in 196213, 50 µL
of 1.0% carrageenan (Sigma Chemical Co., St. Louis, MO,
USA) diluted in a 0.9% saline solution were subcutaneously
injected in the left hind footpad of all mice. Carrageenan, a
sulfated polysaccharide that turns into gel in room temperature, is widely applied in the promotion of acute inflammatory response in animal models, as it induces the release of
inflammatory mediators, such as histamine, bradykinin and
prostaglandins, among others.
Edema evaluation
The volume of each animal’s left hind paw (submersed until the knee articulation) was determined using a water plethysmometer (Plethysmometer 7150 – Ugo Basile®, Italy,
0.01 mL precision). Evaluations were made immediately
before (T0) and one, two, three and four hours (T1, T2, T3
and T4) after the carrageenan injections. Each measurement was repeated three times for every animal; therefore,
the paws’ volumes are expressed as averages.
In order to evaluate the edema evolution, relative volumes (%) of the animal’s paws were calculated in relation
to their basal volumes, for each of the four hours in the
experiment, using the formula:
Relative Volume =
(VFm - VBm )
VBm x 100
Where,
VBm refers to the average basal volume at T0;
VFm is the average final volume at T1, T2, T3 and T4.
With the values obtained for each animal, the averages of their relative volumes were calculated for each
group. The relative volume averages were graphed
against time, and the area under the curve (AUC) was
measured. With the AUC values, the percentages of
edema reduction for each group in relation to the CGN
group were calculated.
Exposure to MF
The equipment (NVL70, Linfospin, Brazil) used for the
present study induces an oscillating MF, with adjustable
frequencies (0 to 24 Hz). Each variation cycle induces a
polar inversion, which switches back and forth between the
North-South and South-North orientation. The field’s intensity, measured with a gaussmeter (TMAG-1T, Globalmag,
Brazil, 1 mT precision), was of 3,000 G (300 mT) on the
surface of the equipment.
Animals were placed in the orthostatic position on the
equipment, with the edema-inflicted paw immobilized and
directly touching its surface. Even though there was a full
body exposure to the MF, the injured paws were exposed
to the highest field intensity. During the exposure (with 60
A pilot study – acute exposure to a low-intensity, low-frequency oscillating magnetic field: effects on carrageenan-induced paw edema in mice
Experimental design
Experimental and Control Groups are described in Table 1.
Groups 1, 2, 3 and 4 were exposed to MF of the same
intensity, during the same time period, with variations only
regarding the field’s frequency (0, 3, 9 and 15 Hz, respectively). The exposures were carried out during the first,
second, and third hours, starting from the carrageenan
injection.
Two Control Groups with edemas induced through the
same protocol were included, though without exposure to
the MF: CRG and DCL – the animals in this last group were
administered an intramuscular injection of sodium diclofenac (1 mL/ kg) one hour after the carrageenan injection.
Relative Volume (%)
seconds of duration)¸ the mice were immobilized through
the torso-cervical region.
160%
140%
120%
CGN
DCL
1 (0Hz)
2 (3Hz)
3 (9Hz)
4 (15Hz)
100%
80%
60%
40%
20%
0%
MI
1h
2h
3h
4h
Time
MI: initial measurement moment.
Figure 1. Relative volume averages (%) and their respective standard errors (groups DCL and 3 (9 Hz), n=6; other groups, n=5).
Statistical analysis
A Z-test was conducted to compare the AUC values obtained from different groups throughout the experiment,
with p≤0.05 values being considered statistically relevant.
Results
The relative volume averages for each group, along with
the AUC values, can be found in Figures 1 and 2, respectively. The edema reduction percentages are detailed in
Table 2.
All groups receiving some form of intervention (1, 2, 3,
4 and DCL) revealed AUC values that were substantially
lower than the ones of CGN Group (p≤0.05). DCL had the
lowest reduction percentage (25.0±6.1%) and Group 3,
the highest (46.9±4.0 %). Compared to the results of DCL,
only Groups 2 (3 Hz) and 3 (9 Hz) showed significantly
lower AUC values (p≤0.05). Also, with statistical relevance
(p≤0.05), Group 3 (9 Hz) showed lower AUC values than
Groups 1 (0 Hz) and 4 (15 Hz).
Conclusion
Acute exposure to a 300 mT oscillating MF yields positive
results in the regression of carrageenan-induced edema
in mice, with indications that such effect depends on the
field frequency.
Figure 2. AUC values and their respective standard errors calculated throughout the four hours of experiment (groups DCL and
3 (9 Hz), n=6; other groups, n=5). Different symbols represent
statistical difference between the respective groups (p≤0.05).
Table 2. Edema percentage reduction (reduction%) and the respective standard errors (±SE% reduction) calculated in relation
to the CGN Group (Groups DCL and 3 (9 Hz), n=6; other groups,
n=5).
Groups
CGN
DCL
1 (0 Hz)
2 (3 Hz)
3 (9 Hz)
4 (15 Hz)
Reduction %
±SE% reduction
25,0
35,3
44,5
46,9
34,9
6,1
4,6
4,2
4,0
4,9
Table 1. Experimental and Control groups.
1
2
3
4
CGN
5
5
6
5
5
Average weight
(g)
24,6
25,0
24,8
24,2
25,0
DCL
6
24,5
Groups
Number of subjects
Frequency
(Hz)
0
3
9
15
Intensity
Exposure duration
Exposure moment
(mT)
(seconds)
300
60
T1, T2 & T3
300
60
T1, T2 & T3
300
60
T1, T2 & T3
300
60
T1, T2 & T3
Animals with induced paw edema not exposed to the MF
1 hour after the carrageenan injection, the animals received a high anti-inflammatory dose (sodium
diclofenac 1 mL/ kg) of intramuscular injection
Revista Brasileira de Física Médica.2011;5(1):53-6.
55
Yoshimura TM, Meneguzzo DT, Lopes-Martins RA
References
1. Yan QC, Tomita N, Ikada Y. Effects of static magnetic field on bone formation
of rat femurs. Med Eng Phys. 1998;20(6):397-402.
2. Castelpietra R, Dal Conte G. First experiments in the treatment of psoriasis
by pulsating magnetic fields. Bioelectrochem Bioenerg. 1985;14(1-3):22533.
3. Curri SB. Morphohistochemical changes in rat paw carrageenin oedema
induced by pulsed magnetic fields. Bioelectrochem Bioenerg. 1985;14(13):57-61.
4. Zecca L, Dal Conte G, Furia G, Ferrario P. The effect of alternating magnetic
fields on experimental inflammation in the rat. Bioelectrochem Bioenerg.
1985;14(1-3):39-43.
5. Morris CE, Skalak TC. Acute exposure to a moderate strength static
magnetic field reduces edema formation in rats. Am J Physiol Heart Circ
Physiol. 2008;294(1):H50-7.
6. Ciombor DM, Aaron RK, Wang S, Simon B. Modification of osteoarthritis
by pulsed electromagnetic field: a morphological study. Osteoarthritis
Cartilage. 2003;11:455-62. In: Morris CE, Skalak TC. Acute exposure to a
moderate strength static magnetic field reduces edema formation in rats.
Am J Physiol Heart Circ Physiol. 2008;294(1):H50-7.
7. Canedo-Dorantes L, Garcia-Cantu R, Barrera R, Mendez-Ramirez I, Navarro
VH, Serrano G. Healing of chronic arterial and venous leg ulcers through
systemic effects of electromagnetic fields. Arch Med Res. 2002;33:281-9.
In: Morris CE, Skalak TC. Acute exposure to a moderate strength static
56
Revista Brasileira de Física Médica.2011;5(1):53-6.
8.
9.
10.
11.
12.
13.
magnetic field reduces edema formation in rats. Am J Physiol Heart Circ
Physiol. 2008;294(1):H50-7.
Patino O, Grana D, Bolgiani A, Prezzavento G, Mino J, Merlo A, et al. Pulsed
electromagnetic fields in experimental cutaneous wound healing in rats.
J Burn Care Rehabil. 1996;17:528-31. In: Morris CE, Skalak TC. Acute
exposure to a moderate strength static magnetic field reduces edema
formation in rats. Am J Physiol Heart Circ Physiol. 2008;294(1):H50-7.
Stiller MJ, Pak GH, Shupack JL, Thaler S, Kenny C, Jondreau L. A
portable pulsed electromagnetic field (PEMF) device to enhance
healing of recalcitrant venous ulcers: a double-blind, placebocontrolled clinical trial. Br J Dermatol. 1992;127:147-54. In: Morris
CE, Skalak TC. Acute exposure to a moderate strength static magnetic
field reduces edema formation in rats. Am J Physiol Heart Circ Physiol.
2008;294(1):H50-7.
Xu S, Okano H, Ohkubo C. Acute effects of whole-body exposure to static
magnetic fields and 50-Hz electromagnetic fields on muscle microcirculation
in anesthetized mice. Bioelectrochemistry. 2001;53(1):127-35.
Weintraub MI, Cole SP. A randomized controlled trial of the effects of a
combination of static and dynamic magnetic fields on carpal tunnel
syndrome. Pain Med. 2008;9(5):493-504.
Funk RH, Monsees T, Ozkucur N. Electromagnetic effects - From cell biology
to medicine. Prog Histochem Cytochem. 2009;43(4):177-264.
Winter CA, Risley EA, Nuss GM. Carrageenin-induced oedema in the hind
paw of the rat as an assay for anti-inflammatory drugs. Proc Soc Exp Biol.
1962;111:544-7.
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):57-62.
Reproducibility of radiant energy
measurements inside light scattering liquids
Reprodutibilidade de medidas de energia radiante dentro
de líquidos espalhadores de luz
Ana C. de Magalhães and Elisabeth M. Yoshimura
Institute of Physics of São Paulo University/Nuclear Physics Department, São Paulo (SP), Brazil.
Abstract
The aim of this project is to evaluate the uncertainty associated with measurements performed with laser beams in scattering liquids with optical
fibers. Two lasers with different wavelengths were used, 632.8 nm and 820 nm, to illuminate a cuvette with Lipovenos PLR, the scattering liquid. A
mask at the top of the cuvette was used to control the positioning of 250 µm optical fiber and the laser entrance point. The light energy was measured
with an optical power meter, and the integration time was 60 s, measured with a chronometer. There were no systematic errors associated with the
integration time. Two tests were done to evaluate the uncertainty associated with the positioning of the components of the experimental arrangement.
To evaluate the uncertainty of the positioning of the cuvette, seven series of measurements with the optical fiber 5 mm far from the beam were
performed. Between two series, the cuvette was removed from the holder, the liquid was mixed and the cuvette was put back in the same position.
A second test was done to evaluate the reproducibility of fiber positioning and the relative positioning of mask and cuvette. Three distances between
the beam and the fiber were used: 4 mm, 5 mm and 6 mm, through the positioning of the fiber at three different holes in the mask. Six series of
measurements were performed. Between two series, the cuvette was removed from the holder, the liquid was mixed and the cuvette was put back
in the same position. Each series was done in a different order, permuting the three positions. The uncertainty associated with the positioning of the
experimental elements was of the order of 7%. The variation of incident laser energy was also evaluated, resulting in 6.5% and 4.0% for the red and
infrared lasers respectively.
Keywords: laser, scattering liquids, light energy.
Resumo
O objetivo deste projeto é avaliar a incerteza associada às medidas realizadas com feixes de laser em líquidos espalhadores com fibras ópticas. Dois
lasers com diferentes comprimentos de onda foram utilizados, 632,8 e 820 nm, para iluminar uma cubeta com Lipovenos PLR, o líquido espalhador.
Uma máscara no topo da cubeta foi utilizada para controlar o posicionamento da fibra óptica de 250 µm e o ponto de entrada do laser. A energia
da luz foi medida com um medidor de potência óptica (Optical Meter), o tempo de integração foi de 60 segundos, medido com um cronômetro.
Não houve erros sistemáticos associados com o tempo de integração. Dois testes foram realizados para avaliar a incerteza associada com o
posicionamento dos componentes do arranjo experimental. Para avaliar a incerteza do posicionamento da cubeta, foram realizadas sete séries de
medidas com fibra óptica a 5 mm de distância do feixe. Entre duas séries, a cubeta foi removida do recipiente, o líquido foi misturado e a cubeta
foi colocada na mesma posição novamente. Um segundo teste foi realizado para avaliar a reprodutibilidade do posicionamento da fibra e o relativo
posicionamento da máscara e da cubeta. Três distâncias entre o feixe e a fibra foram utilizadas: 4, 5 e 6 mm, por meio do posicionamento da fibra
em três diferentes orifícios na máscara. Seis séries de medidas foram realizadas. Entre duas séries, a cubeta foi removida do recipiente, o líquido
foi misturado e a cubeta foi posta novamente na mesma posição. Cada série foi realizada em uma ordem diferente, mudando as três posições. A
incerteza associada com o posicionamento dos elementos experimentais foi da ordem de 7%. A variação da energia de laser incidente também foi
avaliada, resultando em 6,5 e 4,0% para os lasers vermelho e infravermelho, respectivamente.
Palavras-chave: laser, líquidos espalhadores, energia da luz.
Introduction
The determination of the uncertainty associated with an experimental arrangement is very important. It determines the
reliability of the experiment. In our experiment, with the uncertainty, is possible to distinguish measurements and to investigate variations of light energy inside an illuminated volume.
The aim of this project was to evaluate the reproducibility of light energy measurements on scattering liquids with
the use of an optical fiber coupled in an optical meter, in
order to determine the uncertainty associated to the positioning of the experimental components and analyze the
main uncertainty components.
Corresponding author: Ana Carolina de Magalhães – Instituto de Física – Rua do Matão, Travessa R, 187 – Cidade Universitária – São Paulo (SP), Brazil –
CEP: 05508-090 – E-mail: [email protected]
Associação Brasileira de Física Médica®
57
Magalhães AC, Yoshimura EM
Materials and methods
A cylindrical cuvette was used, with 25 mm diameter, with
a plane window with 10 mm of width, parallel to the cylinder axis, as it is possible to see at Figure 1-A and 1-B.
Lipovenos PLR, which is a lipid emulsion for intravenous use and has scattering behavior1, was used as scattering liquid.
An optical fiber with a diameter of 250 μm was used
to collect the scattered light. The fiber was guided by a
metallic tube (internal diameter of 0.85 mm, external diameter of 1.5 mm and 47.65 mm length) in order to make
sure that the fiber was vertical and that its entering position in the liquid could be measured. The fiber depth in
the liquid was 5 mm and this portion of fiber was without
protective cover.
A mask, detailed in Figure 1-C, was used in order to
allow the controlled positioning of the optical fiber in relation to the laser beam. The central hole, for the laser beam,
has a diameter of 3 mm, and each one of the other holes,
for the optical fiber, has a diameter of 1.5 mm. The first
hole is at a distance of 4 mm of the center, the second is
at 5 mm and so on, until the last one that is at 11 mm of
the center.
The final experimental arrangement is represented at
Figure 2.
Lasers with two different wavelengths were used, one
in red range (He-Ne, 632.8 nm) and the other in infrared
range, 820 nm. The circular beams reached the center of
the cuvette, passing through the mask central hole. In the
case of infrared laser, an aperture was used to ensure that
the laser beam reaching the liquid was circular.
An optical meter (Newport Hand-Held Optical meter,
model 1918-C) with a connector to optical fibers was used
in integration mode, and the signal was integrated in 60 s.
This measurement was taken with a chronometer, as the
optical meter doesn’t have this function.
In order to determine the uncertainty associated with
the positioning of the cuvette and the optical fiber, two
reproducibility tests were carried out. The objective of the
first one (R1) was to evaluate the reproducibility of the
positioning of the cuvette in the holder for illumination.
Seven series of seven measurements with the optical fiber 5 mm far from the beam were performed. Between
two series, the cuvette was removed from the holder, the
liquid was mixed and the cuvette was put back in the
same position.
The objective of the second test (R2) was to evaluate
the reproducibility of the optical fiber positioning and the
relative positioning of mask and cuvette. Three distances between the beam and the fiber were used: 4 mm,
5 mm and 6 mm, through the positioning of the fiber at
three different holes in the mask. Six series of seven measurements with the fiber at each of the three positions
were performed. Between two series, the cuvette was
removed from the holder, the liquid was mixed and the
cuvette was put back in the same position. Each series
was composed by a different permutation of the three
positions.
Data of the incident energy from the lasers were collected in order to determine the stability of this value. These
measurements were done with the same optical meter, but
without the optical fiber attachment, in order to determine
how the lasers, the mask and the aperture affect the results
obtained. For red laser, measurements were taken at two
different positions: position 1 - before the mask; and position 2 - behind the mask. For infrared laser, measurements
at three different positions were taken: position 0 - before
the aperture; position 1 - between the aperture and the
mask; and position 2 - behind the mask. The positions 1
and 2 are the same for both lasers.
50 mm
B
C
A
Figure 1. Figure of the cuvette used at the experiment. A) Threedimensional view; and B) cross-section view. At C, is shown the
mask used at the experiment to position the optical fiber and
the laser beam.
58
Revista Brasileira de Física Médica.2011;5(1):57-62.
Figure 2. Figure of the experimental arrangement used. The
laser is represented by the red line. The optical fiber is represented by the green line, the fiber depth in the liquid is 5 mm. The
mask used is represented in A, it has a central hole for the laser
beam and eight radial holes for the fiber.
Reproducibility of radiant energy measurements inside light scattering liquids
Results
He-Ne laser
The results obtained for the wavelength 632.8 nm for the
test R1 are represented at the graphs of Figures 3 and 4.
The data sequence for test R2 is shown at the graph of
Figure 5, in which it is possible to see a change at the data
for all distances, because of the same reasons of test R1.
At Table 1 are the values for average, standard deviation and coefficient of variation for each test.
The results for the measurements of laser incident
energy are shown at Figure 6. The energy is relative to the
average of the measurements of each position (position 1
and 2). There are no irregularities at the data, showing that
the laser output is stable. Results for average, standard
deviation, coefficient of variation and sample size are shown at Table 1.
Figure 5. Graph with data for test R2, red laser. The graph shows data in chronological order for each distance from the fiber
to the center of the cuvette.
380
Energy (nJ)
370
360
350
340
330
59,85 59,90 59,95 60,00 60,05 60,10 60,15 60,20
Integration time (s)
Figure 3. Graph with data for test R1, red laser. The graph
shows integration time dispersion data, measured with manual
chronometer.
Figure 6. Graph with data for laser incident energy, red laser.
The energy is relative to the average of the measurements of
each position.
380
Energy (nJ)
370
360
Table 1. Average, standard deviation and coefficient of variation
of the radiant energy for each test (R1 and R2) realized with red
laser and data for laser incident energy at positions 1 and 2
(lines Pos 1 and Pos 2, respectively).
350
340
330
0
7
14
21
28
35
42
49
Data number
Figure 4. Graph with data for test R1, red laser. The graph shows data in chronological order. The arrows correspond to the
change of measure series.
Test/
distance
R1
R2/4 mm
R2/5 mm
R2/6 mm
Pos 1
Pos 2
Average
(nJ)
353
475
400
229
333×106
362×106
Standard
deviation (nJ)
10
24
30
9
7×106
23×106
Coefficient of
variation (%)
2.9
5.0
7.6
3.9
2.2
6.5
Sample
size
49
42
42
42
11
10
Revista Brasileira de Física Médica.2011;5(1):57-62.
59
Magalhães AC, Yoshimura EM
Infrared laser
The data obtained for test R1, in chronological order, are
shown at the graph of Figure 7.
It is possible to note, at graph of Figure 7, that data oscillate a lot, without any clear change, neither tendencies.
For test R2, the results are shown at the graph of
Figure 8, in which is possible to note that there is a big
change at energy for all distances after the seventh measurement. The reason for this alteration is a variation of the
incident laser energy, which is changed by a factor of 2.6.
After this big change, other variations with this magnitude
did not occur.
Table 2. Average, standard deviation and coefficient of variation
of the radiant energy for each test (R1 and R2) realized with
infrared laser and data for laser incident energy at positions 0, 1
and 2 (lines Pos 0, Pos 1 and Pos 2, respectively).
Test/
distance
R1
R2/4 mm
R2/5 mm
R2/6 mm
Pos 0
Pos 1
Pos 2
Average (nJ)
88.4
196.8
152.5
113.9
262.2×106
89.3×106
81.3×106
Standard
Coefficient of
deviation (nJ) variation (%)
2.1
2.3
3.4
1.8
8.0
5.2
3.6
3.2
6.1×106
2.3
3.4×106
4.0
0.61×106
0.75
Sample
Size
49
21
35
35
11
11
11
The results for the measurements of laser incident
energy are shown at Figure 9. The energy is relative to
the average of the measurements of each position (position 0, 1 and 2). There are no irregularities at the data,
what means that there are not big variations of incident
energy and there is not any important effect of the aperture and the mask. Results for average, standard deviation, coefficient of variation and sample size are shown
at Table 2.
Figure 7. Graph with data for test R1, infrared laser. The graph
shows data in chronological order.
Figure 9. Graph with data for laser incident energy, infrared
laser. The energy is relative to the average of the measurements
of each position.
Figure 8. Graph with data for test R2, infrared laser. The graph
shows data in chronological order for each distance measured.
At Table 2 are the values for average, standard deviation and coefficient of variation for each test. Probably, the
low standard deviation obtained for test R2, at distance of
4 mm, occurred because three series of measurements at
this position were lost.
60
Revista Brasileira de Física Médica.2011;5(1):57-62.
Conclusions
Through the analysis of the data, it was concluded that there are no tendencies related to the integration time, what is
seen at integration time dispersion graphs. Therefore, it is
possible to say that this source of error is considered when
the energy analysis is done.
There is an uncertainty associated to the optical fiber and cuvette positioning, shown by the standard
Reproducibility of radiant energy measurements inside light scattering liquids
deviation obtained for each group. As they have similar
magnitudes, varying roughly between 2 and 8%, we will
adopt 7% as a value of the uncertainty associated with
the data collection. It is equivalent to the largest standard deviation found for the series data. A large portion
of the radiant energy data variation is due to the laser
stability, as the coefficients of variation of the incident
energy are similar to the standard deviation of the reproducibility experiments.
Acknowledgment
The authors thank the personnel of the Optical Laboratory of
Institute of Physics, where the measurements were done.
References
1.
Pogue BW, Patterson MS. Review of tissue simulating phantoms
for optical spectroscopy, imaging and dosimetry. J Biomed Opt.
2006;11(4):041102.
Revista Brasileira de Física Médica.2011;5(1):57-62.
61
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):63-6.
Optimization of the scan protocol in the
measurements of coronary artery calcium
Otimização do protocolo de exame nas medidas do
cálcio arterial coronário
Larissa C. G. Oliveira1, Ilan Gottlieb2, Fabrício M. de Carvalho2, Larissa C. Pinheiro3,
Simone Kodlulovich3, Fernando A. Mecca4 and Ricardo T. Lopes1
Laboratório de Instrumentação Nuclear (LIN/COPPE/UFRJ) - Centro de Tecnologia, Rio de Janeiro (RJ), Brazil.
2
Clínica de Diagnóstico por Imagem , Rio de Janeiro (RJ), Brazil.
3
Instituto de Radioproteção e Dosimetria (IRD/CNEN), Rio de Janeiro (RJ), Brazil.
4
Instituto Nacional do Câncer, Rio de Janeiro (RJ), Brazil.
1
Abstract
The aim of this study was to evaluate the influence of the tube current applied for studies of calcium score. The research was carried out in a private
clinic of Rio de Janeiro, using a 64-slice MDCT scanner and an anthropomorphic cardiac computed tomography phantom. In all images, the Agatston
score, the volume and mass of the calcifications, and the noise for each current tube were determined. The average computed tomography attenuation
number obtained for all tube currents was 261.6±3.2 HU for the CaHA density insert and -0.2 HU±2.0 for the water insert. The images obtained
at lower tube currents were noisier and grainier than those obtained at higher tube currents. However, no significant differences were found in the
calcium measurements, which suggest a high potential of patient dose reduction, around 50%, without compromising diagnostic information.
Keywords: cardiovascular disease, multi-detector computed tomography and quantification of coronary artery calcium.
Resumo
O objetivo deste estudo foi avaliar a influência da corrente de tubo aplicada aos estudos do escore do cálcio. A pesquisa foi realizada em uma
clínica particular no Rio de Janeiro, utilizando um tomógrafo computadorizado de 64 cortes e fantoma antropomórfico cardíaco de tomografia
computadorizada. Em todas as imagens, o escore de Agatston, o volume e a massa das calcificações e o ruído para cada corrente de tubo foram
determinados. O número de tomografia computadorizada médio , em unidades de Hounsfield (HU), obtido para todas as correntes de tubo foi de
261,6±3,2 HU, para a inserção da densidade CaHA, e -0,2 HU±2,0, para a água. As imagens obtidas em correntes de tubo baixas foram mais
ruidosas e granuladas do que aquelas obtidas em correntes de tubo elevadas. No entanto, não foram encontradas diferenças significativas nas
medidas do cálcio, o que sugere um grande potencial de redução da dose ao paciente, em torno de 50%, sem comprometer as informações para
o diagnóstico.
Palavras-chave: doença cardiovascular, tomografia computadorizada multidetectores e quantificação de cálcio arterial coronário.
Introduction
The amount of calcium deposits in the coronary arteries
is an indirect marker of total atherosclerotic burden, and it
has been strongly associated with future cardiac events in
asymptomatic patients1,2. As calcium has high X-ray attenuation, its detection can be easily performed with a gated
noncontrasted computed tomography (CT) of the heart3.
The Agatston score method was first used in 1990
and is based on the area and density of the calcified
plaques4. High reproducibility of this scoring method is
desirable, since it is widely used both in clinical and in
research settings5.
The introduction of the multidetector CT (MDCT) improved the detection of lesions or obstructions (stenosis) of coronary arteries, especially in contrasted cardiac
studies6,7. The combination of higher rotation speeds
and larger coverage per rotation has allowed a more
accurate depiction of the coronary artery tree. As a result, the usage of this procedure significantly increased
in recent years, providing great clinical benefit in terms
of incipient and obstructive atherosclerosis detection.
However, the progressive increase of the collective dose
became an important concern8. The aim of this study
was to evaluate the influence of the tube current in measured calcium score.
Corresponding author: Larissa C. G. Oliveira – Laboratório de Instrumentação Nuclear – Rua Pajurá, 95 / 501 Taquara / Jacarepaguá – Rio de Janeiro (RJ),
Brazil – E-mail: [email protected]
Associação Brasileira de Física Médica®
63
Oliveira LCG, Gottlieb I, Carvalho FM, Pinheiro LC, Kodlulovich S, Mecca FA, Lopes RT
Materials and methods
The quantification of coronary calcium was performed
in a private clinic of Rio de Janeiro, using a 64-slice
MDCT scanner (Briliance 64, Philips Medical Systems,
the Netherlands) and an anthropomorphic cardiac CT
phantom (QRM, Moehrendorf, Germany). The cardiac
phantom was positioned on the patient’s couch, and its
rear edge was aligned with the laser beam of the gantry
(Figure 1).
The cardiac CT contains nine calcified cylinders and
two large calibrations inserts. The nine calcified cylinders
are divided into three sets, each with calcium hydroxyapatite (CaHA) densities of 200, 400, and 800 mg/cm3
and diameters of 5, 3, and 1 mm, respectively. The two
large calibration inserts are made of water and spongy
bone (200 mg/cm3 CaHA density), which are equivalent
materials.
The scan protocol used was the standard spiral CT
protocol for chest examination: 120 kVp, 64 x 0.625 mm
collimation and 0.5 seconds, per rotation. The effective tube
current levels increased from 80 to 180 mAs with an interval
of 20 mAs.
After data acquisition, all images were transferred to a
dedicated Philips workstation, where the Agatston, volume and mass scores were determined according to guidelines, but, in short, the Agatston score was determined
by setting a threshold of 130 HU and ignoring structures
smaller than 1 mm2 to exclude noise from the calculation6.
Depending on the peak attenuation of the calcified cylinder, the calcified area was multiplied by one of the following
factors (F): 130-199-HU: F=1; 200-299 HU: F=2; 300-399
HU: F=3 and for higher than 400 HU: F=4.
The calcified cylinders volume was determined as the
number of voxels Nvoxel in the volume data set, which belong to the calcification multiplied by the number of one
voxel Vvoxel, according to the following equation:
V = Nvoxel . Vvoxel
(1)
To obtain the CaHA mass, a calibration measurement
of a calcified cylinder with known CaHA density (ρCaHA) was
performed, and a calibration factor c was determined for
each current level, according to the Eq. 2:
(2)
c = ρCaHA /(CTcylinder - CTwater)
The calibration factor c is, therefore, given by the CaHA
density (ρCaHA) of the known calcified cylinder divided by the
mean difference in CT numbers of the calcified cylinder and
one of the two large inserts made of water-equivalent material (CTcylinder - CTwater), in the calibration measurement. The
measured CaHA mass multiplied by the respective calibration factor corresponds to the value of the CaHA mass.
Image noise measurements were assessed in all images, which were obtained for different values of tube current. For each tube current, three region-of-interests (ROIs)
were evaluated. A circular ROI (200 mm2 approximately)
was placed in each image. The CT number and the standard deviation (SD) were determined in the homogeneous
large insert (CaHA density of 200 mg/ cm3). The CT number and SD in the water insert were similarly measured to
calculate the calibration factor.
Results and Discussion
The averaged CT number attenuation obtained for all tube
currents was 261.6 HU±3.2 for the CaHA density insert
and -0.2 HU±2.0, for the water insert.
The values for Agatston score, the volume and mass
measurements of the individual calcified cylinders, and
their corresponding tube current are presented in Table 1.
Of the nine calcified cylinder presented in the phantom,
only six cylinders were visible on the image and measurable at the 130 HU threshold in each tube current. The
three cylinders with 1 mm diameter were excluded in this
survey due to their size limitation. The Agatston, volume
and mass values, when compared, the values measured
by the manufacture, our values were lower.
Image noise expressed as the SD of the CT number of
the CaHA insert ranged from 13.5 HU, at 80 mAs, to 9.1
HU at 160 mAs. Figure 2 presents the values found in this
survey and by Cheng et al.1. As expected, images obtained
at lower tube currents were noisier and grainier than those
Table 1. Measurements for calcified cylinders in calibration insert at different tube current.
HA density Size
Mean CT Agatston
(mg/cm3) (mm) number (HU) score
5
197.0
39.2
200
3
146.5
7.9
1
-*
5
747.7
78.4
400
3
473.6
31.6
1
5
398.3
58.8
800
3
239.8
15.8
1
-
Figure 1. Anthropomorphic cardiac CT phantom in the gantry.
64
Revista Brasileira de Física Médica.2011;5(1):63-6.
*Nonmeasurable.
Volume
Mean
(mm3) mass (mg)
58.7
9.7
2.6
2.4
1.4
58.7
37.3
22.6
8.9
6.8
0.3
58.7
19.8
22.6
4.6
6.8
0.4
Optimization of the scan protocol in the measurements of coronary artery calcium
50
14
This survey
Cheng (2002)
200 mg/cm3
400 mg/cm3
800 mg/cm3
40
CAHA mass (mg)
Image Noise (HU)
12
10
8
6
30
20
10
80
100
120
140
Tube current (mAs)
160
180
Figure 2. Relationship between tube current and image noise.
0
80
100
120
140
160
180
Tube current (mAs)
Figure 4. Relationship between tube current and CaHa mass.
Acknowledgments
A
B
The authors thank the financial support of Conselho Nacional
de Pesquisa (CNPq) and the Radiology department.
References
Figure 3. Images obtained at 80 (A) and 160 mAs (B) demonstrating image noise.
obtained at higher tube currents. However, in the Figure 3,
it is possible to observe that it is feasible to obtain an image
with an optimized current value adequate to the diagnostic
and, consequently, to reduce the patient’s dose.
Although a reduction in the tube current from 160 to 80
mAs resulted in a noisier image, no significant differences
were found in the calcium measurements obtained with
the CaHa mass (Figure 4). The same behavior was observed for the calcium volume.
Conclusions
The results obtained in this research of calcium quantification deposits on coronary artery the quantitative scoring methods, such as Agatston, calcium volume, and mass scoring suggests a high potential of patient’s dose reduction,
around 50%, without compromising diagnostic information.
1. Hong C, Bae KT, Pilgram TK, Suh J, Bradley D.. Coronary Artery Calcium
measurements with multi-detector row CT: In vitro assessment of
effect of radiation dose. Radiology. 2002;225:901-6.
2. Greuter MJW, Dijkstra H, Groen JM, Vliegenthart R. 64 slice MDCT
generally underestimates coronary calcium scores as compared to
EBT: A phantom study. Med Physics. 2007;34:3510-9.
3. Wexler L, Brundage B, Crouse J, Detrano R, Fuster V, Maddahi J,
et al. Coronary artery calcification: pathophysiology, epidemiology,
imaging methods, and clinical implications. A statement for health
professionals from the American Heart Association. Circulation. 1996;
94:1175–92.
4. Raggi P. Coronary Calcium on Electron Beam Tomography Imaging
as a Surrogate Marker of Coronary Artery Disease. Am J Cardiol.
2001;87:27-34A.
5. Hausleiter J, Meyer T, Hermann F, Hadamitzky M, Krebs M, Gerber T, et
al. Estimated Radiation Dose Associated with Cardiac CT Angiography.
JAMA. 2009;301:500-7.
6. Agatston A, Janowitz F, Hildner N, Zusmer M, Viamonte R, Detrano
R. Quantification of coronary artery calcium using ultrafast computed
tomography. JACC. 1990;15:827-32.
7. Miller JM, Rochitte CE, Dewey M, Arbab-Zadeh A, Niinuma H, Gottlieb
I, et al. Diagnostic performance of coronary angiography by 64-row CT.
N Engl J Med. 2008;359:2324-36.
8. Kalra MK, Maher M, Toth T, Hamberg LM, Blake MA, Shepard JA,
et al. Strategies for CT Radiation dose Optimization. Radiology.
2004;230(30):619-28.
Revista Brasileira de Física Médica.2011;5(1):63-6.
65
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):67-72.
Evaluation of the image quality in
computed tomography: different phantoms
Avaliação da qualidade de imagem na tomografia
computadorizada: diferentes fantomas
Vinicius C. Silveira1, Larissa C. Oliveira2, Rômulo S. Delduck1,
Simone Kodlulovich1, Fernando A. Mecca³ and Humberto O. Silva4
1
Instituto de Radioproteção e Dosimetria, Comissão Nacional de Energia Nuclear (CNEN), Rio de Janeiro (RJ), Brazil.
2
Nuclear Instrumentation Laboratory / Instituto Alberto Luiz Coimbra de Pós-graduação e pesquisa de Engenharia
(COPPE), Universidade Federal do Rio de Janeiro (UFRJ), Rio de Janeiro (RJ), Brazil.
3
National Institute of Cancer (INCa), Rio de Janeiro (RJ), Brazil.
4
Copa D’Or Hospital, Rede Labs D’Or, Rio de Janeiro (RJ), Brazil.
Abstract
The aim of this paper was to compare the simulators provided by the CT manufactures and Catphan’s Phantom with the American College of
Radiology (ACR) computed tomography phantom. The image evaluation followed the protocols established by the manufactures of the phantoms. For
slice thickness evaluation, the maximum percentage difference was 9% between the phantoms ACR and Siemens. In CT number accuracy test, the
measurements of CT number of water showed a difference of 10 HU between the CT simulators. Comparing the uniformity results, the discrepancy
was 11% and 55% for Siemens and Philips respectively in relation to the result obtained with the ACR phantom. The result of low contrast was
the same for all phantoms. The MTF50 and MTF10 obtained with Siemens phantom was 4 and 8 pl/mm. For Catphan, 6 and 7 pl/mm. Results
demonstrate that the ACR simulator was the most comprehensive and flexible to be used in several scanner models. Some simulators did not present
all image quality indicators to perform a complete test.
Keywords: computed tomography, image quality, phantoms.
Resumo
O objetivo deste trabalho foi comparar os simuladores fornecidos pelos fabricantes de tomógrafos e o fantoma Catphan com o fantoma de
tomografia computadorizada do Colégio Americano de Radiologia (ACR). A avaliação da imagem seguiu os protocolos estabelecidos pelos fabricantes
dos fantomas. Para a avaliação da espessura de corte, a maior diferença foi de 9% entre os fantomas ACR e Siemens. No teste de exatidão do
número de CT, as medidas do número de CT da água mostraram uma diferença de 10 HU entre os fantomas de CT. Comparando os resultados de
uniformidade, a discrepância foi de 11% e 55 % para os fantomas da Siemens e da Philips em relação ao valor obtido com o fantoma do ACR. O
resultado de baixo contraste foi o mesmo para todos os fantomas. Os valores de MTF50 e MTF10 para a resolução de alto contraste do Siemens
foram 4,2 e 7,6 pl/mm e para o Catphan, 6 e 7 pl/mm. Os resultados demonstraram que o simulador do ACR foi o mais compreensivo e flexível a
ser usado em diversos modelos de tomógrafos. Alguns simuladores não apresentaram dados suficientes para realizar o teste completo.
Palavras-chave: tomografia computadorizada, qualidade da imagem, fantomas.
Introduction
The optimization program includes an evaluation of the image quality. Each manufacture has developed a specific
simulator for their computed tomography (CT) scanner.
These phantoms present differences in the physical indicators to evaluate image quality, values of tolerance, and especially the procedure to carry out the tests. The American
College of Radiology (ACR) CT phantoms have been used
in an accreditation program in the USA. By applying a
standard methodology, it is possible to evaluate and to
compare scanners from all manufactures or models.
The minimum physical indicators recommended for
the evaluation of image quality are: positioning of coach,
CT number accuracy, slice width, low contrast resolution,
high contrast (spatial) resolution, CT number uniformity,
and noise. However, several simulators do not have these
indicators to conduct a full assessment.
Despite of the increase of multi-slice scanners in
Latin America, hospitals do not have personnel trained,
Corresponding author: Vinicius da Costa Silveira – Institute of Radioprotection and Dosimetry – Av. Salvador Allende, s/n – Recreio dos Bandeirantes –
Rio de Janeiro (RJ), Brazil – E-mail: [email protected]
Associação Brasileira de Física Médica®
67
Silveira VC, Oliveira LC, Delduck RS, Kodlulovich S, Mecca FA, Silva HO
instrumentation, and phantoms to implement the quality assurance program. Nowadays, the regulatory authority does not have any information about the performance of scanners of the services. Besides, countries of
Latin America do not have a CT accreditation program.
Consequently, there is no information about patient dose
and image quality.
The aim of this paper was to compare the simulators provided by the manufactures and the Catphan’s Phantom with
the ACR CT Phantom. The image evaluation followed the
protocol established by the manufacture of each phantom.
Materials and methods
The phantoms evaluated were provided by General
Electric (GE), Siemens, Philips, Catphan 500, and ACR CT
Phantom.
Image quality tests were performed in three scanners:
Philips Brilliance 40, GE Light speed and Siemens Somaton,
with their respective simulators and the CT ACR Phantom.
Additional tests were carried out in the public hospital with
Catphan 500 and ACR phantom on the scanner Picker.
A
B
Figure 1. (a) Catphan 500; (b) ACR.
A
B
Figure 2. (a) Phantom Philips / Brilliance 40; (b) Phantom GE/
Light speed).
Table 1. Quality assurance of each manufacturer
Test QA
Collimation
Accuracy of # CT
Pixel Size
High Contrast
Low Contrast
Noise and Uniformity
Contrast scale
Alignment
Accuracy laser light
CAT¹
ACR
GE
x
x
x
x
x
x*
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x* - only water and air
QA: quality assurance; 1CAT: Catphan; 2PHI: Philips; 3SI: Siemens.
68
Revista Brasileira de Física Médica.2011;5(1):67-72.
PHI²
x
x
x
x
x
x
-
SI³
x
x*
x
x
x
x
x
-
Simulators characteristics
The simulators have distinct characteristics, according to
the specificity of the test performed.
The Catphan 5001 (Figure 1a) is made in The Phantom
Laboratory Incorporated, in New York. It is a solid Phantom
containing four modules: CTP528, 21 line-pair high resolutions; CTP 515, sub-slice and supra-slice low contrast;
CTP404, position verification, slice width, sensitometry
and pixel size; CTP486, solid image uniformity module.
The ACR2 CT accreditation phantom (Figure 1b) is a solid
phantom containing four modules, constructed primarily from
solid water. There are external markings (BBs) on the first and
last module to allow the alignment of the phantom in the axial,
coronal, and sagittal directions. Using this phantom, it is possible to evaluate alignment, CT number accuracy, slice width,
low and high contrast resolution, uniformity, and noise.
The Philips Phantom3 (Figure 2a) has two parts: head
and body. The part of the head contain: physical layer, impulse response and slice width; water layer, noise and uniformity; multi-strip, contrast scale and sensitometry. The
part of the body contain: a Teflon strip and water hole.
GE’s Phantom (Figure 2b) for scanners of light speed
series can evaluate six quality image criteria: contrast
scale, resolution of high and low contrast, noise, uniformity,
slice width, accuracy of laser, and linearity of CT number4.
It is divided into three parts: resolution’s block, contrast
membrane, and water hole.
The Siemens’ Phantom for Somaton scanners contains
a number of modules suitable for testing different CT image
quality characteristics, such as: slice thickness, impulse
response, CT number accuracy (water and air), high and
low contrast, noise and uniformity, and alignment. Table 1
presents the proposed tests by each manufacturer.
For this study, the following were compared: collimation, accuracy and linearity of CT number, evaluation of
high and low contrast, noise and uniformity, and contrast
scale. Therefore, they demonstrate the adequacy to a
proper image quality evaluation of each phantom.
Experimental setup
Simulators were placed and aligned using light beams of
the scanners. All phantoms are cylindrical; the alignment in
the gantry was performed considering sagittal and coronal projections. The position was determined by specifics
marks of each simulator. The evaluation followed the respective manual of the manufacture. To compare the ACR
and Catphan phantoms, the head routine protocol on axial
acquisition was used.
Results and discussion
Slice thickness
Siemens Somaton Phantom
Table 2 shows the results of slice thickness to ACR and
Siemens phantoms on Siemens/Somaton.
Evaluation of the image quality in tomography computed: different phantoms
In Table 2, the maximum difference between the results
obtained with the phantoms was 9% for the slice width
with 10 mm. In relation to nominal slice, differences were
lower than 3% for Siemens and 10% for ACR.
Philips
In this study, it was possible to compare the phantoms
results only for 5 mm slice thickness (Table 3), in which the
result obtained was equal for both phantoms and nominal
slice thickness. For 10 mm, the percentage difference between the nominal and ACR was of 46%.
CT number accuracy
Siemens phantoms
Table 6 presents the accuracy values of the CT number carried out with ACR and Siemens Phantoms.
The Siemens Phantom only has water and air inserts.
Important structures, like soft tissue and bone, are not
available. For all cases, the values are in the tolerance
range.
CT GE Hi-Speed
In the Table 4, for nominal 3 mm slice thickness, the Phantoms
GE and ACR presented the same values. For the other thicknesses (5, 7 and 10), we could only obtain the comparison with
nominal values. For ACR measurements, the values for 3 and
7 mm were the same of the selected. For GE, the difference
between the measured and the nominal values was 5%.
CT Brillance, Philips
Table 7 presents the results of CT number accuracy with ACR and Philips Phantoms. For polyethylene
and acrylic, the results showed a difference of 14 and
2%, respectively. For water, although the values were in
accordance to the tolerance, the value measured with
ACR phantom was approximately 7 HU. Philips Phantom
does not have a material similar to air. Therefore, with
the exception of acrylic, the CT number accuracy was
adequate.
CT Picker
Table 5 presents the slice thickness measurements using
Catphan and ACR phantoms. For all nominal slice thickness, the differences between the phantoms results were
5%. Comparing with the nominal value of 3 mm, the percentage difference was approximately 40%.
CT Hi-Speed, GE
In Table 8, the results of the CT number accuracy for
CT Hi-Speed from GE are presented. In this case,
water is the only common material in the phantoms.
The result to the discrepancy was 43% between the
simulators.
Table 2. Comparison of slice thickness measured with the CT
Somaton phantom and ACR
Table 6. Accuracy of CT number for CT Somaton, Siemens
Nominal slice
thickness (mm)
2
3
10
Slice thickness measured (mm)
Phantom ACR
Phantom Siemens
2.5
2.3
3.0
3.2
9.0
9.8
D (%)
8
7
9
Material
Polyethylene
Water
Acrylic
Bone
Air
Average number of CT (HU)
Phantom ACR Phantom Siemens Reference (HU)
-90.5
-107 and 87
0
-1.0
-7 and +7
126.7
+ 110 and 130
894.0
+ 850 and 970
- 983.1
-999.0
- 1,005 and 970
Table 3. Slice thickness for CT Brilliance, Philips
Slice thickness
selected (mm)
5
7
Slice thickness selected measure (mm)
Phantom ACR
Phantom Philips
5
5
10
-
Table 4. Slice thickness for CT Hi-Speed, GE
Slice thickness
selected (mm)
3
5
7
10
Slice thickness selected measure (mm)
Phantom ACR
Phantom GE
3.0
3.0
5.0
7.0
9.5
Table 7. Accuracy of CT number for CT Brillance, Philips
Material
Polyethylene
Water
Acrylic
Bone
Air
Average number of CT (HU)
Phantom ACR Phantom Philips Reference (HU)
-81.2
- 70
- 107 and 87
5.7
0
-7 and +7
136.6
140
+ 110 and 130
893.7
+ 850 and 970
-973.9
-1,005 and 970
Table 8. Accuracy of CT number for CT Hi-Speed, GE
Average number of CT (HU)
Phantom ACR
Phantom GE
Reference (HU)
Polyethylene
-90.4
- 107 and 87
water
-0.7
-0.4
- 7 and +7
Acrylic
124.9
+ 110 and 130
Bone
917.2
+ 850 and 970
Air
-985.1
- 1,005 and 970
Polyethyrene
-1,1
-
Material
Table 5. Slice thickness for CT Picker
Slice thickness
selected (mm)
3
5
10
Slice thickness selected measure (mm)
Phantom ACR
Phantom Catphan
2,2
2,1
5
5,1
10
10
Revista Brasileira de Física Médica.2011;5(1):67-72.
69
Silveira VC, Oliveira LC, Delduck RS, Kodlulovich S, Mecca FA, Silva HO
Picker
The CT numbers obtained with ACR and Catphan Phantoms
are presented in Table 9. For polyethylene and air, the discrepancy between results was of 4 and 1%, respectively. The reading of water had a very high discrepancy between phantoms.
Low contrast resolution
The Siemens and ACR Phantoms contain low contrast
groups of objects inside a similar background with different sizes. For both phantoms, groups with 5 and 2 mm
diameter were identified applying the manuals.
For Philips Brilliance CT, we could not visualize the
group of 5 mm as the manufacture in ACR and Philips.
The GE Phantoms contain a polystyrene membrane
suspended in water with holes with diameters of 10, 7.5,
5, 3 and 1 mm. It was observed holes of 3 mm. The differences between CT number of the membrane and water
are equal to 10 (contrast level).
In the Catphan 500, the contrast levels are measured
marking region of interest (ROIs) over the largest target visualised in supra-slice, sub-slice, and in the background
(Table 10). With the ACR’s Phantom, inserts of 6 mm were
visualized inserts by applying head’s protocols. With the
Catphan, the smallest diameter discernible was 5 mm for
supra (0.3% contrast level) and 5 mm for sub-slice (1%
contrast level).
For Siemens CT scanner using its own phantom, it was
possible to evaluate only uniformity. ACR showed a uniformity of -0.6 HU and Siemens, 7.3 HU. The results were
satisfactory.
The results to the Hi-Speed GE Scanner and Philips
Brilliance by uniformity and noise using their own simulator were satisfactory. Nevertheless, when comparing
to the ACR, the values for uniformity and noise were respectively 3.4 HU and 2.4 to GE and -1.4 HU and 4.8 to
ACR. For Philips, the noise showed a difference of 16%
in relation to the manufactures tolerance. Compared
with the ACR, the # CT was 55% superior and the standard deviation was 20% lower.
For the Picker, the value of uniformity obtained by
Catphan was 8.7 HU and 10.4 with the ACR. The maximum values of noise were 9.1 with the Catphan and 7 with
ACR phantom.
Uniformity and noise
Table 10 presents the results of uniformity and noise according to ACR manual for all scanners.
High contrast resolution
Table 11 shows the results to the test of high contrast
carried out with the ACR Phantom on scanners: Philips,
Siemens, and Picker.1
Results to the CT GE using the own phantom of manufacturer presented a difference between the measured and
reference value (18%) equal to 17%.
To the Siemens and Catphan phantom, the MTF
method was used to quantify the values of high resolution. The results obtained for Siemens and Picker scanners were according to the manufacturer’s tolerances
(Tables 12 and 13).
Table 9. Accuracy of CT number for CT Picker
Table 11. High contrast resolution - Phantom ACR
Material
Average number of CT (HU)
Phantom ACR Phantom Catpham
-98,2
-94,4
-107 and 87
Water
10,1
0
-7 and +7
Acrylic
133
120
+110 and 130
Bone
978
-
+850 and 970
-974,6
-980
-1005 and 970
Table 10. Uniformity and noise – Phantom ACR
Average number of CT (HU) ± standard deviation
Parameter /
Position
CT Siemens CT Philips
CT GE
CT Picker
Center (C)
-0.4±5.2
3.8±6.9
-1.4±4.8
10.4±5.2
3h
-1.0±4.3
3.3±5.2
-0.7±4.2
10.7±6.3
6h
-2.3±4.7
4.1±5.5
-0.1±4.9
11,0±7,0
9h
-1.0±4.2
3.6±5.8
-0.1±4.3
10,8±4,7
12h
0.3±4.2
4.1±5.4
0.4±4.2
11.6±4.8
Revista Brasileira de Física Médica.2011;5(1):67-72.
CT Picker
Reference
Abdomen adult
6
6
-
5
Chest
Hi-Resolution
8
7
-
6
Head
-
-
7
-
Table 12. High contrast resolution – CT Siemens
Nominal Value
(lp/cm)
Tolerance
(lp/cm)
50%
4,50
0,45
4,2
y
10%
8,00
0,80
7,6
y
2%
10,00
1,00
9,6
y
MTF (u)
70
CT Siemens CT Philips
Reference (HU)
Polyethylene
Air
Spatial frequency (pl/mm)
Technique
Measured
Conform
value (lp/cm)
Table 13. High contrast resolution – CT Picker
MTF (u) (%)
Value (lp/cm)
Tolerance (lp/cm)
60
5
±50 %
50
6
±50 %
8
7
±50 %
Evaluation of the image quality in tomography computed: different phantoms
Conclusions
Evaluation of slice thickness showed similar results for all
phantoms. For the accuracy of CT number, the water’s
CT number showed a very large discrepancy for all simulators. The GE and Siemens Phantoms do not have the
structures to simulate soft and high materials, which are
necessary to image quality evaluation. For low contrast
resolution, all phantoms showed equivalent results.
For uniformity and noise, GE and Siemens phantoms
presented results with a very large discrepancy in relation
to ACR. However, Philips and Cathan Phantoms showed
equivalent results.
Results demonstrate that the ACR simulator was the
most comprehensive and flexible for use in several scan-
ner models. It also had all the tests recommend by the
International Image Quality Assurance5.
References
1. Goodenough DJ. CatPhan® 504 Manual Laboratories Incorporated,
Copyright 2009.
2. American College of Radiology. Instruction Manual for testing the ACR CT
Phantom. Preston White Drive: Reston VA, 1891.
3. Philips Medical System 4535 673 86351_D – Volume 1.
4. Manual manufacture GE 2211214-127 Rev.
5. European Commission. Quality Criteria for Computed Tomography. Working
Document EUR 16262 (Brussels: EC) (1997).
Revista Brasileira de Física Médica.2011;5(1):67-72.
71
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):73-8.
Estimation of patient dose in computed
tomography: an extension of IAEA
Project in Brazil
Estimativa de doses em pacientes submetidos a exames
de tomografia computadorizada: uma extensão do
Projeto IAEA no Brasil
Romulo S. Delduck1, Simone Kodlulovich1, Larissa C. Oliveira2, Vinicius C. Silveira1,
Humberto O. Silva3, Helen Khoury4 and Alejandro Nader5
2
1
Institute of Radiation Protection and Dosimetry (CNEN), Rio de Janeiro (RJ), Brazil.
Nuclear Instrumentation Laboratory / Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa de
Engenharia (COPPE), Universidade Federal do Rio de Janeiro (UFRJ), Rio de Janeiro (RJ), Brazil.
3
Hospital Copa D’Or, Rio de Janeiro (RJ), Brazil.
4
Department of Nuclear Energy of Universidade Federal de Pernambuco (UFPE), Recife (PE), Brazil.
5
International Agency of Nuclear Energy, Vienna, Austria.
Abstract
The aim of this paper was to estimate the patient’s dose in routine procedures in Brazil and to identify the potential of optimization in adults and
pediatric procedures. The sample included ten hospitals distributed in different states of the country. In each hospital, the routine protocols of head,
chest, high resolution chest, abdomen, and pelvis were recorded. The values of Cw, Cvol and Pkl were estimated based on the nCw values provided by
IMPACT. For the same procedure, significant differences in patient’s doses were verified between the hospitals and also in the same department.
In some cases, the technical factors are so low that suggest a rigorous evaluation of the image quality. Problems were also observed regarding
procedures records, and the information about the procedure is insufficient. This study indicated the necessity of an implementation of an action
plan that includes training program to operate the scanner in an optimized mode, to carry out the dosimetry, and to evaluate the image quality. The
large range of patient’s doses indicated that there is an expressive potential of patient’s dose reduction and optimization maintaining the diagnostic
information.
Keywords: computed tomography, patient dose, optimization.
Resumo
O objetivo deste trabalho foi estimar a dose dos pacientes em procedimentos de rotina no Brasil e identificar o potencial de otimização em
procedimentos adultos e pediátricos. A amostra incluía dez hospitais distribuídos em diferentes estados do país. Em cada hospital, os protocolos de
rotina da cabeça, tórax, alta resolução do tórax, abdome e pélvis foram registrados. Os valores de Cw, Cvol e Pkl foram estimados com base nos valores
de nCw, fornecidos pela IMPACT. Para o mesmo procedimento, diferenças significativas foram encontradas nas doses de acordo com os hospitais,
também nos mesmos departamentos. Em alguns casos, os fatores técnicos são tão baixos que sugerem uma avaliação rigorosa da qualidade de
imagem. Problemas também foram observados em relação aos registros dos procedimentos, já que a informação sobre eles é insuficiente. Este
estudo indicou a necessidade de implementar um plano de ação que incluísse um programa de treinamento para operar o equipamento de forma
otimizada, fazer a dosimetria e avaliar a qualidade da imagem. A grande variedade de doses indicou que existe um grande potencial de redução das
doses e otimização, mantendo a informação do diagnóstico.
Palavras-chave: tomografia computadorizada, dose do paciente, otimização.
Corresponding author: Rômulo de Sena Delduck Pinto Filho – Institute of Radioproteccion and Dosimetry – Avenida Salvador Allende, s/n – Recreio dos
Bandeirantes – Rio de Janeiro (RJ), Brazil – E-mail: [email protected]
Associação Brasileira de Física Médica®
73
Delduck RS, Kodlulovich S, Oliveira LC, Silveira VC, Silva HO, Khoury H, Nader A
In Brazil, the number of computed tomography (CT) scanners increased exponentially in the last five years. Most of
them are multi-slice CT scanners (MDCT), enabling new
clinical application such as CT angiography and virtual
endoscopy1. However, this new technology has been introduced before a preparation of the diagnostic radiology
departments. There is a lack of health professionals trained
in CT and in the instrumentation necessary to carry out the
dosimetry and the image quality evaluation.
The advances in MDCT increased the resources to
obtain an image of the patient with more diagnostic information. However, these scanners introduce new concepts
to be understood and tradeoffs to be made. CT scanners
have been recognized as a high radiation modality, when
compared to other diagnostic X-ray techniques.
Even though the recognized benefits derived from CT
procedures, the high frequency and the magnitude of the
patient doses from these examinations have drawn the attention to potential risk from this practice. The amount of
radiation dose received from a CT scan depends on many
factors, including the design, maintenance and the operation mode. Especially the MDCT can potentially result in
higher radiation risk to the patient due to the tendency to
perform long scan lengths at high tube currents, fast acquisition times, and multi-phase contrast studies.
The protocols established by the manufacturer for the
routine procedures should be evaluated to achieve lower
patient dose and diagnostic quality image. It is also fundamental to adequate the protocol for the size and characteristics of the patients and the clinical indication2. While
the determination of patient dose is a common practice in
Europe, few surveys have been carried out in Brazil. This
paper, with the support of the International Atomic Energy
Agency (IAEA), is the first step to establish Brazilian reference levels for CT.
Materials and methods
Selection of the hospitals
Ten hospitals, public and private, distributed in four states,
participated of the survey, they were: Rio de Janeiro, Minas
Gerais, Paraná, and Pernambuco. The year of fabrication
of the CT scanner varied between 2000 and 2008. The
scanners were single and multi-sliced.
Data collected
The participant hospitals voluntarily were in the survey.
Data were collected with the technologist of the service
according to a standard questionnaire prepared during
the IAEA regional projects (RLA067 and 057). In the first
part, the technologist provided to the physicist general information: state, manufacturer, model, age of the scanner
and number of examinations per year for each procedure.
In the second part, scan parameters related to standard
74
Revista Brasileira de Física Médica.2011;5(1):73-8.
protocols conducted on typical adult (average-size) and
pediatric (<1 year, 7 years) were required. The selected
procedures included: head, chest, high-resolution chest,
abdomen and pelvis.
CT dose index
The diagnostic reference level (DRL) is a fundamental tool
for the optimization process. The quantities used in the CT
were: weighted CT kerma index (Cw), Kerma-volume product (Cvol) and kerma-length product (Pkl). Values of Cvol and
Pkl were calculated for each procedure based on nCw value
from IMPACT3 for the respective scanner model. When the
tube current modulation was used, the doses were calculated using reported values of average mAs.
Results and Discussions
Data collected
For each hospital, the patient number varied significantly.
Of the ten participant hospitals, only five forms were complete. These five hospitals correspond to two states only
(Figure 1). The total number of patients per year in these
five hospitals was 50.220. Hospitals C and B showed the
highest and lowest patient number per year, respectively.
Distribution of the exams
Figure 2 shows the frequency of CT procedures performed annually in the different centres. CT of abdomen
and pelvis are the most frequent procedures, which rep-
Patients for year
Patients
Introduction
16.000
14.000
12.000
10.000
8.000
6.000
4.000
2.000
0
Total
B
C
D
H
J
Hospitals
Figure 1. Number of patients per year in each hospital.
16000
14000
12000
10000
8000
6000
4000
2000
0
Pelvis
Abdomen
High Resolution
Chest
Chest
B
C
D
H
J
Head
Figure 2. CT procedures distribution for each hospital per year.
Estimation of patient dose in computed tomography: an extension of IAEA Project in Brazil
Comparison of Pkl and Cw values for each procedure according to the age group
The distribution of Pkl values for different adults CT procedures are presented in Figure 4.
Table 1. Comparison of technical factors for routine adult procedures.
CT exam
Head
Chest
Abdomen
CT exam
Head
Chest
Abdomen
CT e xam
Head
Chest
Abdomen
Present survey
kVp
mAs
Cw (mGy) Pkl (mGy.cm)
120-130 120-500
13.4-115
175-3744
120-130
70-285
7.7-16.3
94.1-684
120-130
80-350
6.2-20
162-767.3
Literature adult
Other countries
European DRL
kVp
mAs
Cw (mGy) Pkl (mGy.cm)
120
250-270
60
1050
120-140 120-267
30
650
120
120-267
35
780
IRPA 2008 Adult
kVp
mAs
Cw (mGy) Pkl (mGy.cm)
90-140
20-600
4-77
62-1773
90-130
40-440
3-50
2.7-999
90-140
40-457
3-66
40-21.9
lvi
s
st
Ch
res es
olu t hi
tio gh
n
ab
do
me
n
97,39
ch
e
Cvol máx / Cvol min
Br
ai n
sk
ull
ba
se
Adult
Children <1y
Children 7y
89,60
Procedure
Figure 3. Rate of the maximum and minimum values of Cvol for
each procedure and age group.
Table 2. Technical factors for the pediatric (<1 year-old) routine
procedures used in this survey.
Exam
kVp
mAs
Head
80-130
80-200
Chest
120-130
41-80
Abdomen
120-130
41-150
IRPA 2008
Exam
Cw (mGy)
Pkl (mGy.cm)
Head
43
376
Chest
31
617
Abdomen
31
226
Table 3. Technical factors for the pediatric (seven years-old)
routine procedures used in this survey.
CT exam
Head
Chest
Abdomen
Present survey
mAs
Cw (mGy)
80-350
9-80,5
70-100
9,0-23
41-150
9-27,6
IRPA 2008
Pkl (mGy.cm)
546
738
498
kVp
80-130
120-130
120-130
CT exam
Head
Chest
Abdomen
PKL (mGy.cm)
Technical factors
The routine protocols in most of the countries of Latin America
and IRPA 2008 showed a very large range of technical factors for all procedures. The results obtained for adults are presented in Table 1.
The kVp values obtained in this study were similar to the
ones in literature. However, the range of mAs was excessively
large for all procedures, increasing consequently the Cw and
Pkl values. This result indicated that some actions should be
done in these hospitals, where the patient doses are unnecessary high.
The variation of Cvol values obtained for each procedure
can be observed in Figure 3.
Comparing high resolution chest with other procedures,
the technical factors used are much higher and in approximately 71% of the procedures the acquisition mode was
axial. In Table 2 the doses values for children under one year
are presented. Also, for this age, the variation of mAs is very
expressive. For head, the range was from 80 to 200 mAs.
The Pkl varied from 59 to 820 mGy.cm. Comparing the average values of Pkl for different procedures obtained in this
study with the ones obtained in IRPA 2008, it is possible to
observe that the Pkl values were: 16.5% above for head exams, 77.8% below for chest and 28.1% lower for abdomen.
The range of doses values for seven-year-old patients
are presented in Table 3. Similar results obtained for under one year-old patients were observed including large
ranges of mAs and Pkl. For head, the Pkl maximum was
2,146 mGy.cm.
177,78
pe
resent aproximately 45% of the total number. Chest high
resolution procedure is the least performed.
4000
3500
3000
2500
2000
1500
1000
500
0
Pkl (mGy.cm)
58,5-2146
120-337,3
113-441,6
Adult
3RD Quartile
=1892,6
3RD Quartile 3RD Quartile
=417,7
=193,4
H
C
CHR
3RD Quartile
=469,8 3RD Quartile
=380,4
A
P
Figure 4. Pkl for adult patients. The procedures are represented
as: head (H), chest (C), chest high resolution (CHR), abdomen (A)
and pelvis (P).
Revista Brasileira de Física Médica.2011;5(1):73-8.
75
Delduck RS, Kodlulovich S, Oliveira LC, Silveira VC, Silva HO, Khoury H, Nader A
Conclusions
The Cw values both for adults and pediatric patients
were lower than the European DRL and IRPA 2008. Due
to the large differences in the scan length, pitch, table
increment and mAs used in the protocols for patients
with the same characteristics, the result of Pkl was not
consistent. The most critical procedure was the high
resolution chest.
The large range of technical factors is a concern
and should be investigated. The difficulty to answer the
form also indicates that the professionals do not have
the training necessary to understand the scanner, which
would be essential to the optimization program. In many
76
Revista Brasileira de Física Médica.2011;5(1):73-8.
PKL (mGy.cm)
procedures, the pediatric doses were higher than the
adults’ and much higher than the European DRL’s. This
suggests that a survey specific for children should be
carried out in order to as soon as possible the hospitals
implement optimization programs. It is also important to
expand this survey in the other states of the country.
900,0
800,0
700,0
600,0
500,0
400,0
300,0
200,0
100,0
0,0
Children <1y
3RD Quartile
=341,3
3RD Quartile
=130,0
C
H
3RD Quartile 3RD Quartile
=195,6
=174,4 3RD Quartile
=124,9
CHR
A
P
Figure 5. Pkl for pediatric patients (<1 year-old). The procedures
are represented as: head (h), chest (c), chest high resolution
(chr), abdomen (a) and pelvis (P).
2.500,0
PKL (mGy.cm)
Comparing the values of Pkl for various adult procedures, the largest range of Pkl values and the higher
third quartile are observed. The third quartile of Pkl for
head examinations was 80.3% higher compared to the
European DRL and 106% higher than the IRPA 2008
value. For chest, the third quartile of Pkl was 35.7% and
16.5% lower than the European DRL and IRPA 2008 value. For abdomen, the Pkl was 39.8% and 47.5% lower
than the DRL European and IRPA 2008.
The distribution of Pkl values for children under one
year-old is presented in Figure 5. Also, for this age
group, the range of Pkl values for the head CT was very
large and the values higher than the ones obtained for
the other procedures. For head and abdomen, the third
quartiles were 10 and 30%, lower than the DRLs obtained in IRPA 2008. For chest, the Pkl value for pediatric
patient was 374.6% lower than the IRPA 2008 value.
The distribution of Pkl values and the respective
third quartile values for pediatric patient (seven yearsold) submitted to routine procedures are presented in
Figure 6.
Comparing the third quartiles of Pkl obtained in this
study with the Pkl presented in IRPA 2008, our results
indicated that all values of this survey were lower than
these references: 6.0% for head, 183.3% for chest and
31.4% for abdomen .
As can be observed in Table 4, the third quartiles of
Cw obtained in this study were lower than the IRPA 2008
value and European DRL.
Comparing the third quartiles of Cw presented in
Table 5 for children under one year-old, with the Pkl presented in IRPA 2008, it is possible to observe that all
values of this survey were lower than these references:
13% for head, 63% for chest and 55% for abdomen.
Table 6 presents the Cw values for children of seven
years-old in surveys of head, chest, and abdomen. For
this age, the difference between the third quartile values
of Cw obtained in this study and the IRPA values were
much lower than the results obtained for adults and children under one year-old: 12.9% for head, 24.4 % for
chest, and 26.1% for abdomen.
Children 7y
2.000,0
1.500,0
3RD Quartile
=512,0
3RD Quartile
3RD Quartile RD
=378,9
Quartile
3
=260,4
=117,0
1.000,0
500,0
3RD Quartile
=307,1
0,0
H
C
CHR
A
P
Figure 6. Pkl for pediatric patients (seven years-old). The procedures are represented as: head (h), chest (c), chest high resolution (chr), abdomen (a) and pelvis (P).
Table 4. Third quartile of Cw (mGy) for routine procedures: adults.
Study
IRPA 2008
DRL European
Present survey
Head
56
60
55
Chest
19
30
10
Abdomen
20
35
14.6
Table 5. Third quartile of Cw(mGy) for routine procedures: children under one year-old.
Study
IRPA 2008
this survey
Head
43,00
37,50
Chest
31.00
11.50
Abdomen
31.00
13.80
Table 6. Cw(mGy) values for the routine procedures: seven-yearold children.
Study
IRPA 2008
Present survey
Head
44.00
38.30
Chest
25.00
18.90
Abdomen
26.00
19.20
Estimation of patient dose in computed tomography: an extension of IAEA Project in Brazil
References
1. Kodlulovich SD, Khoury H, Blanco S. Survey of radiation exposure for adult
and pediatric patients undergoing CT procedures in Latin America – IRPA
2008. Prot Dos. 2005;114(1-3):303-7.
2. Tsapaki V, Aldrich JE, Sharma R, Staniszewska MA, Krisanachinda A,
Rehani M, et al. International Atomic Energy Agency. Dose reduction in
CT while maintaining diagnostic confidence: diagnostic reference levels at
routine head, chest, and abdominal CT-IAEA-coordinated research project.
Radiology. 2006;240(3):828-34.
3. ImPACT (Imaging Performance Assessment of CT scanners). CT patient
dosimetry Excel spreadsheet. Available from home webpage of the ImPACT
evaluation centre of the DH Medicines and Healthcare products Regulatory
Agency (MHRA). [cited 2011 Jan]. Available from: http://www.impactscan.org.
Revista Brasileira de Física Médica.2011;5(1):73-8.
77
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):79-84.
Evaluation of patients’ skin dose undergoing
interventional cardiology procedure using
radiochromic films
Avaliação da dose na pele de pacientes submetidos a
procedimentos de cardiologia intervencionista usando
filmes radiocrômicos
Mauro W. Oliveira da Silva1, Bárbara B. Dias Rodrigues1,2 and Lucía V. Canevaro1
2
1
Instituto de Radioproteção e Dosimetria (IRD/CNEN); Serviço de Física Médica, Rio de Janeiro (RJ), Brazil.
Universidade Federal do Rio de Janeiro (UFRJ); Programa de Engenharia Nuclear (PEN-COPPE), Rio de Janeiro (RJ), Brazil.
Abstract
In interventional cardiology (IC), coronary angiography (CA) and percutaneous transluminal coronary angioplasty (PTCA) procedures are the most
frequent ones. Since the 1990s, the number of IC procedures has increased rapidly. It is also known that these procedures are associated with high
radiation doses due to long fluoroscopy time (FT) and large number of cine-frames (CF) acquired to document the procedure. Mapping skin doses
in IC is useful to find the probability of skin injuries, to detect areas of overlapping field, and to get a permanent record of the most exposed areas of
skin. The purpose of this study was to estimate the maximum skin dose (MSD) in patients undergoing CA and PTCA, and to compare these values
with the reference levels proposed in the literature. Patients’ dose measurements were carried out on a sample of 38 patients at the hemodynamic
department, in four local hospitals in Rio de Janeiro, Brazil, using Gafchromic© XR-RV2 films. In PTCA procedures, the median and third quartile
values of MSD were estimated at 2.5 and 5.3 Gy, respectively. For the CA procedures, the median and third quartile values of MSD were estimated at
0.5 and 0.7 Gy, respectively. In this paper, we used the Pearson’s correlation coefficient (r), and we found a fairly strong correlation between FT and
MSD (r=0.8334, p<0.0001), for CA procedures. The 1 Gy threshold for deterministic effects was exceeded in nine patients. The use of Gafchromic©
XR-RV2 films was shown to be an effective method to measure MSD and the dose distribution map. The method is effective to identify the distribution
of radiation fields, thus allowing the follow-up of the patient to investigate the appearance of skin injuries.
Keywords: interventional cardiology, radiation protection, patient dose, skin dose, reference levels.
Resumo
Em cardiologia intervencionista (CI), os procedimentos de angiografia coronária (AC) e angioplastia coronária transluminal percutânea (ACTP) são os
mais frequentes. Desde os anos 1990, o número de procedimentos de CI tem crescido rapidamente. Sabe-se, também, que estes procedimentos
estão associados às altas doses de radiação, devido ao logo tempo de fluoroscopia (TF) e ao grande número de imagens (FC) adquiridas para
documentar o procedimento. Mapear as doses na pele em CI é útil para estimar a probabilidade de lesões cutâneas, para detectar as áreas dos campos
sobrepostos e registrar as áreas mais expostas da pele. O objetivo deste estudo foi estimar a dose máxima na pele (DMP) em pacientes submetidos
a AC e ACTP, e compará-la com os níveis de referência propostos na literatura. As medições das doses dos pacientes foram realizadas em uma
amostra de 38 pacientes no departamento de hemodinâmica, em quatro hospitais locais no Rio de Janeiro, Brasil, utilizando os filmes Gafchromic®
XR-RV2. Nos procedimentos de ACTP, os valores da mediana e do terceiro quartil da DCM foram estimados em 2,5 e 5,3 Gy, respectivamente. Para
os procedimentos de AC, os valores da mediana e do terceiro quartil da DCM foram estimados em 0,5 e 0,7 Gy, respectivamente. Neste trabalho,
utilizou-se o coeficiente de correlação de Pearson (r) e encontrou-se uma correlação razoavelmente forte entre o TF e a DCM (r=0,8334, p<0,0001),
para os procedimentos de AC. O limiar de 1 Gy para efeitos determinísticos se excedeu em nove pacientes. O uso dos filmes Gafchromic© XR-RV2
se mostrou um método eficaz para medir a DCM e o mapa de distribuição da dose. O método é eficaz para identificar a distribuição dos campos de
radiação, permitindo o acompanhamento do paciente de forma a investigar o aparecimento de lesões cutâneas.
Palavras-chave: cardiologia intervencionista, proteção radiológica, dose na pele, níveis de referência.
Corresponding author: Mauro Wilson Oliveira da Silva – Instituto de Radioproteção e Dosimetria – IRD/CNEN – Avenida Salvador Allende, s/n – Recreio dos
Bandeirantes – Rio de Janeiro (RJ), Brazil – CEP 22780-160 – E-mail: [email protected]
Associação Brasileira de Física Médica®
79
Silva MWO, Rodrigues BBD, Canevaro LV
Introduction
The growing use of interventional cardiology procedures offers enormous benefits to patients and contributes significantly to the radiation exposure of patients1-3.
Interventional cardiology procedures can involve high
doses to patients and, in particular, to patients’ skin,
the tissue at greatest risk of deterministic injuries. The
evaluation of skin dose from interventional cardiology
procedures is recommended, but it is difficult to perform
due to the different X-ray fields and projections used in
the procedure4-6. Many studies have investigated the radiation dose to patients during interventional cardiology
procedures7-13.
The main task of radiation protection is not only to
minimize the stochastic risks, but also to avoid deterministic injuries. The International Commission on Radiological
Protection (ICRP) recommends the establishment of reference levels as a method of optimizing the radiation
exposure14-17.
Patient dosimetry in interventional cardiology procedures is complex due to difficulties in the identification of
irradiated areas of the skin, as well as the several projections used, different field sizes, radiation qualities, focusskin distance, and focus-image intensifier distance. These
complex procedures require long fluoroscopy times and
the acquisition of a large number of pictures to record the
injury and its results. Mapping skin doses in interventional
cardiology procedure is useful to find the probability of
any skin injury, to detect areas of overlapping field and
to get a permanent record of the most exposed areas of
skin11,13,18-22.
The purpose of this study was to evaluate the maximum skin dose (MSD) received by patients undergoing interventional cardiology procedures, especially because the
number of procedures performed annually has increased
over the past 20 years23. An additional purpose was to
compare these values with the reference levels proposed
in the literature1,13,24-27.
Radiochromic films
In this study, the patients’ back MSD was evaluated using 35 x 43 cm Gafchromic XR-RV2 radiochromic films
(International Specialty Products, Wayne, NJ, USA)28,29.
Gafchromic XR-RV2 film has a higher sensitive dose
range (1 cGy to 50 Gy). This film has been developed to
specifically measure absorbed dose at both low and high
energy photons, in which the energies are between 30 keV
and 30 MeV29. The active layer of Gafchromic XR-RV2 is
approximately 17 ?. It is sandwiched between two sheets
of polyester: one transparent film substrate with thickness
of 97 ? and one opaque, white film substrate with thickness of 97 ?. The transparent polyester substrate used in
the film contains a yellow dye. It enhances the visual contrast of the chromatic changes when the film is exposed
to radiation29.
Each batch of films comes with a specific lot number.
Therefore, each batch has a tape calibration30.
The Gafchromic XR-RV2 film was placed on the table
of procedures, under the mattress where the patient was
positioned during the interventions (Figure 1).
The methodology is applied to quantify and map the
dose on the patient’s back. If there is a careful study of
the images, one can evaluate some parameters such as
geometry and irradiation conditions, distribution and intensity fields, the possible overlap of radiation fields, etc. The
exposure time of each procedure was also recorded.
Statistical analysis
The Pearson correlation test was applied to assess if the
MSD is linearly related to the fluoroscopy time. A p-value
lower than 0.05 was considered statistically significant. All
calculations were performed by using R-program statistical analysis software31.
Materials and methods
Results and discussion
Institutions
Measurements were performed in four cardiac catheterization laboratories in Rio de Janeiro, Brazil: two public hospitals (A and B) and two private hospitals (C and D). The
hospital A is a reference hospital in interventional cardiology procedures.
Table 1 shows the data related to the 26 patients who underwent CA procedures. Table 2 shows the data related to
the PTCA procedures.
Patients
Data were obtained from a sample of 38 patients undergoing interventional procedures during coronary angiography (CA) and percutaneous transluminal coronary angioplasty (PTCA). Twenty-six CA and 12 PTCA procedures
were studied. In CA, the mean patient weight was 78.3
kg (range was from 50 to 159 kg) and in PTCA the mean
80
patient weight was 77.5 kg (range was from 58 to 120 kg).
Patients were previously prepared for the procedures, according to the clinical practice of the institution and were
aware of the risks and complexity of the procedure.
Revista Brasileira de Física Médica.2011;5(1):79-84.
Figure 1. Radiochromic film position.
Evaluation of patients’ skin dose undergoing interventional cardiology procedure using radiochromic films
Table 1. Exposure parameters in patients undergoing CA procedures
Statistical analysis
Minimum
1st quartile
Median
Mean
3rd quartile
Maximum
Standard
deviation
Total Fluoroscopy
time [minutes]
2
3.08
5.24
5.98
6.91
16.5
Number of
images
398
677
1,127
1,297
1,818
2,602
Number of
series
6
7
10
10
12
17
4.09
709
3
Table 2. Exposure parameters in patients undergoing PTCA procedures
Statistical analysis
Minimum
1st quartile
Median
Mean
3rd quartile
Maximum
Standard
deviation
Total Fluoroscopy
time [minutes]
3.6
16.6
21.2
30.1
33.8
89.7
Number of
images
932
1,076
1,330
1,605
1,625
3,363
Number of
series
14
16
19
26
33
52
24
820
13
12
CA
10
Frequency
The frequency distribution of the MSD measured with
Gafchromic® XR-RV2 radiochromic films over the population of patients is shown in Figure 2.
We investigated the correlation between the various
parameters (total fluoroscopy time and MSD, number of
images and MSD, number of series and MSD, and patient weight and MSD) separately for the CA and PTCA to
examine whether these factors could be useful in estimating the MSD during the procedures.
In the case of CA procedures, we have found a strong
correlation between total fluoroscopy time and MSD
(r=0.8334, p<0.0001, r2=0.694). Conversely, there is a
poor statistically correlation between number of series
and MSD (r=0.3573, p=0.07, r2=0.128), number of images and MSD (r=0.0746, p=0.72, r2<0.01).
In the case of PTCA procedures, we have found a
strong correlation between total fluoroscopy time and
MSD (r=0.8755, p=0.0009, r2=0.77). The correlation between number of series and MSD and between number
of images and MSD was analyzed and a moderate correlation was found (r=0.7525, p=0.0193, r2=0.5662, and
r=0.5428, p=0.1310, r2=0.2946, respectively).
Tables 3 and 4 show the mean values of fluoroscopy
time, number of images and MSD and published data, for
patients undergoing CA and PTCA, respectively.
Figure 3 shows a digital image of the Gafchromic®
XR-RV2 radiochromic film used to evaluate the skin
dose distribution and MSD during PTCA. In this procedure, the patients’ weight was 120 kg; fluoroscopy time
was 25 minutes, number of images was 1,330, and the
MSD was 3 Gy.
Analyzing the correlation between fluoroscopy time
and MSD, the fluoroscopy time only gives an approximate indicator of the dose to the skin. For the evaluation
of MSD, it is important to consider the complexity of the
procedure.
The reference levels concern: fluoroscopy time, number of images and MSD. The number of images in this
study is comparable with that in other published studies
for the CA procedures, but considerably higher for the
PTCA procedures. The fluoroscopy time is comparable
for CA procedures, but higher for PTCA. In our study, the
results show an excess time spent in fluoroscopy during
PTCA. Consequently, a high MSD was registered.
PTCA
8
6
-.+,
4
2
0
0,25 0,5 0,75
1
1,5 2
3
4
!
5
6
MSD (Gy)
Figure 2. Frequency distribution of maximum skin dose measured by radiochromic films.
Table 3. Mean values for CA procedure
Conclusions
The method to identify the distribution of the radiation fields
in the patient’s back is effective and safe without interfering
in the procedures, thus allowing the patient’s follow-up in
order to investigate the occurrence of skin injuries.
Although these procedures are widely justified, it is
necessary that the practices are optimized to avoid unnecessary exposure of the patient. It is more important to
consider that some measured values are above the threshold dose for the manifestation of deterministic effects.
Published studies
Fluoroscopy
time [minutes]
Number of
images
MSD [Gy]
Neofotistou et al.1
6
1270
-
Giordano et al.13
3.80
562.5
0.09
Padovani et al.24
6.5
700
0.65
IAEA 25
7.1
867.7
-
9.9
1079
0.27
6.2
-
0.28
Hansson et al.
Trianni et al.
26
27
This study
0.57
Revista Brasileira de Física Médica.2011;5(1):79-84.
81
Silva MWO, Rodrigues BBD, Canevaro LV
Table 4. Mean values for PTCA procedure
Published studies
Fluoroscopy
time [minutes]
Number of
images
MSD [Gy]
Neofotistou et al.1
16
1355
-
Giordano et al.13
16.20
963,5
0.49
Padovani et al.24
15.5
1000
1.5
15.2
1100
1.4
Hansson et al.26
20.3
1190
0.98
Trianni et al.27
13.4
-
1.03
This study
30.1
1605
3.04
IAEA
25
Figure 3. Gafchromic® XR-RV2 radiochromic film.
If we compare our results with those proposed by IAEA
research group25, we conclude that it is imperative to carry
on a trial on local practices.
Acknowledgment
This work was partially supported by the Comissão
Nacional de Energia Nuclear (CNEN), Conselho Nacional
de Desenvolvimento Científico e Tecnológico (CNPq), and
International Atomic Energy Agency (IAEA).
References
1. Neofotistou V, Vano E, Padovani R, Kotre J, Dowling A, Toivonen M, et al.
Preliminary reference levels in interventional cardiology. Eur Radiology.
2003;13(10):2259-63.
2. Benini A, Pedersen F, Jorgensen E. Doses to patients in interventional
cardiology. IFMBE Proceedings; 2009.
3. Herrmann HC, Baxter S, Ruiz CE, Feldman TE, Hijazi ZM; SCAI Council on
Structural Heart Disease. Results of the society of cardiac angiography and
interventions survey of physicians and training directors on procedures
for structural and valvular heart disease. Catheter Cardiovasc Interv.
2010;76(4):E106-10.
4. Chida K, Kato M, Kagaya Y, Zuguchi M, Saito H, Ishibashi T, et al.
Radiation dose and radiation protection for patients and physicians during
82
Revista Brasileira de Física Médica.2011;5(1):79-84.
interventional procedure. J Radiat Res (Tokyo). 2010;51(2):97-105.
5. Den Boer A, de Feijter PJ, Serruys PW, Roelandt JR. Real-time quantification
and display of skin radiation during coronary angiography and intervention.
Circulation. 2001;104(15):1779-84.
6. Jaco JW, Miller DL. Measuring and monitoring radiation dose
during fluoroscopically guided procedures. Tech Vasc Interv Radiol.
2010;13(3):188-93.
7. Zontar D, Kuhelj D, Skrk D, Zdesar U. Patient peak skin doses from cardiac
interventional procedures. Radiat Prot Dosimetry. 2010;139(1-3):262-5.
8. Khodadadegan Y, Zhang M, Pavlicek W, Paden RG, Chong B, Schueler BA,
et al. Automatic Monitoring of Localized Skin Dose with Fluoroscopic and
Interventional Procedures. J Digit Imaging. 2010;24(4):626-39.
9. Sarycheva SV, Golikov, Kalnicky S. Studies of patient doses in interventional
radiological examinations. Radiat Prot Dosimetry. 2010;139(1-3):258-61.
10. Zaman A, Ahmed A, Naseer H, Yunous N, Ali M, Zaman M. Estimation of
patient doses from interventional radiology procedures in Pakistan result of
IAEA project RAS /9/047. IFMBE Proceedings; 2009.
11. Tsapaki V, Ahmed NA, AlSuwaidi JS, Beganovic A, Benider A, BenOmrane L,
et al. Radiation exposure to patients during interventional procedures in 20
countries: initial IAEA project results. Am J Roentgenol. 2009;193(2):559-69.
12. Giordano S. Radiation-Induced Skin Injuries During Interventional
Radiography Procedures. J Radiol Nursing. 2010;29(2):37-47.
13. Giordano C, D’Ercole L, Gobbi R, Bocchiola M, Passerini F. Coronary
angiography and percutaneous transluminal coronary angioplasty
procedures: Evaluation of patients’ maximum skin dose using Gafchromic
films and a comparison of local levels with reference levels proposed in the
literature. Phys Med. 2010;26(4):224-32.
14. Clarke R, Valentin J. A history of the International Commission on
Radiological Protection. Health Physics. 2005;88(6):717-32.
15. Clarke R, Valentin J. A history of the international commission on radiological
protection. Health Physics. 2005;88(5):407-22.
16. Clarke RH, Valentin J. The History of ICRP and the Evolution of its Policies.
Annals of the ICRP. 2009;39(1):75-110.
17. ICRP International Commission on Radiological Protection Avoidance of
radiation incurie from medical interventional procedures. Annals of the
ICRP. 2000;30:7-67.
18. Vañó E, Guibelalde E, Fernández JM, González L, Ten JI. Patient
dosimetry in interventional radiology using slow films. Br J Radiol.
1997;70:195-200.
19. Guibelalde E, Vano E, Gonzalez L, Prieto C, Fernandez JM, Ten JI. Practical
aspects for the evaluation of skin doses in interventional cardiology using a
new slow film. Br J Radiol. 2003;76(905):332-6.
20. Vano E, Gonzalez L, Guibelalde E, Aviles P, Fernandez JM, Prieto C, et
al. Evaluation of risk of deterministic effects in fluoroscopically guided
procedures. Radiat Prot Dosimetry. 2006;117(1-3):190-4.
21. Tsapaki V, Patsilinakos S, Voudris V, Magginas A, Pavlidis S, Maounis T, et al.
Level of patient and operator dose in the largest cardiac centre in Greece.
Radiat Prot Dosimetry. 2008;129(1-3):71-3.
22. Tsapaki, V., Radiation dose in interventional cardiology. Imaging in Medicine,
2010. 2(3): p. 303-312.
23. Miller DL, Balter S, Schueler BA, Wagner LK, Strauss KJ, Vañó E. Clinical
radiation management for fluoroscopically guided interventional procedures.
Radiology. 2010;257(2):321-32.
24. Padovani R, Vano E, Trianni A, Bokou C, Bosmans H, Bor D, et al. Reference
levels at European level for cardiac interventional procedures. Radiat Prot
Dosimetry. 2008;129(1-3):104-7.
25. IAEA. In: Safety Report Series No. 59. Establishing guidance levels in X ray
guided medical interventional procedures: a pilot study. Viena; 2009.
26. Hansson B, Karambatsakidou A. Relationships between entrance skin
dose, effective dose and dose area product for patients in diagnostic and
interventional cardiac procedures. Radiat Prot Dosimetry. 2000;90(12):141-4.
Evaluation of patients’ skin dose undergoing interventional cardiology procedure using radiochromic films
27. Trianni A, Chizzola G, Toh H, Quai E, Cragnolini E, Bernardi G, et
al. Patient skin dosimetry in haemodynamic and electrophysiology
interventional cardiology. Radiat Prot Dosimetry. 2006;117(13):241-6.
28. Blair A, Meyer J. Characteristics of Gafchromic® XR-RV2 radiochromic
film. Med Physics. 2009;36(7):3050-8.
29. Ying CK, Kandaiya S. Patient skin dose measurements during coronary
interventional procedures using Gafchromic film. J Radiol Prot.
2010;30(3):585-96.
30. Gafchromic radiochromic dosimetry film background information
and characteristic performance data. [2008 nov]. Available at http://
online1.ispcorp.com/_layouts/Gafchromic/index.html
31. The R project for statistical computing. [2011 jan]. Available at http://
www.r-project.org.
Revista Brasileira de Física Médica.2011;5(1):79-84.
83
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):85-8.
EMITEL e-Encyclopaedia of Medical Physics
and Dictionary of Terms
Enciclopédia eletrônica de Física Médica e
Dicionário de Termos - EMITEL
Slavik Tabakov1, Peter Smith2, Franco Milano3, Sven-Erik Strand4, Cornelius Lewis5,
Magdalena Stoeva6 and Vassilka Tabakova1
2
1
King’s College London, UK.
International Organization for Medical Physics (IOMP).
3
University of Florence, Italy;
4
University of Lund, Sweden.
5
King’s College Hospital, UK.
6
AM Studio Plovdiv, Bulgaria.
Abstract
EMITEL, the e-Encyclopaedia of Medical Physics and its Multilingual Translator (dictionary) have been launched at WC2009 (www.emitel2.eu). This
international project attracted more than 300 specialists from 36 countries and grew to be the largest international project in the profession. This
paper describes the development of EMITEL, its effective use, and its planned future development.
Keywords: education, training, encyclopedia.
Resumo
A EMITEL, a enciclopédia eletrônica de Física Médica e seu tradutor multilíngue (dicionário) foram lançados no WC2009 (www.emitel2.eu). Esse
projeto internacional atraiu mais de 300 especialistas de 36 países e é o mais amplo projeto internacional desta profissão. O presente trabalho
descreve o desenvolvimento da EMITEL, seu uso eficaz e o desenvolvimento planejado para o futuro.
Palavras-chave: educação, treinamento, enciclopédia.
Introduction
EMITEL project initial partners and phases
The EMITEL project was a consequence of the first
International Conference on Medical Physics Education
and Training (in 1998 and 2003), which was organized by
King’s College London in connection with the e-learning
projects EMERALD and EMIT1. After this, the European
Medical Imaging Technology e-Encyclopaedia for Lifelong
Learning (EMITEL) was funded with the help of the EU
Leonardo da Vinci programme. The result was an original
e-learning tool used by a wide spectrum of specialists in
Medical Physics and Engineering. The tool was merged
with a dedicated translator of terms (dictionary).
The dedicated EMITEL web site (www.emitel2.eu), which was built by AM Studio, has more than 6,000 users per
month. This paper describes the main features of EMITEL
(Encyclopaedia and Dictionary) and the plans for its future
development.
The idea for EMITEL appeared around 2001 and was initiated as a dictionary (translator) of terms. Initially, it had five
languages, now this number was increased to 29. Later,
in 2005, the EU project was developed2, it started in 2006
and its main phase was completed by the end of 2009.
The project partnership included the core of the
Institutional partners from previous projects (EMERALD
and EMIT) – King’s College London (Contractor) and
King’s College Hospital, University of Lund and Lund
University Hospital, University of Florence, AM Studio
Plovdiv and the International Organization for Medical
Physics (IOMP). Then, additional specialists volunteered
as contributors. Thus, an EMITEL Network was formed
(300+ specialists).
Although the project’s name was specified, Medical
Imaging Technology was specially underlined in the name
Corresponding author: Slavik Tabakov, Dept. Medical Engineering and Physics – King’s College London – Denmark Hill, London SE5 9RS, UK – E-mail:
[email protected]
Associação Brasileira de Física Médica®
85
Tabakov S, Smith P, Milano F, Strand S-E, Lewis C, Stoeva M, Tabakova V
of the project. Radiotherapy and Radiation Protection
were also added. Therefore, the main areas of the
Encyclopaedia are: X-ray Diagnostic Radiology, Nuclear
Medicine; Magnetic Resonance Imaging, Ultrasound
Imaging, Radiotherapy, Radiation Protection, General
terms linked to Medical Physics. Special care was taken
for covering the aspects of Medical Engineering related
to Imaging.
EMITEL Encyclopaedia and Dictionary
EMITEL developed an expandable database of specific terms (4,000+). The terms have one to three or
more words. These terms were translated into many
languages by working groups of national specialists.
Thus, the dictionary includes: English, Swedish, Italian,
French, German, Portuguese, Spanish, Bulgarian,
Czech, Croatian, Japanese, Estonian, Finnish, Greek,
Hungarian, Latvian, Lithuanian, Polish, Romanian,
Russian, Slovenian, Bengal, Chinese, Iranian, Arabic,
Malaysian, Thai, and Turkish.
The dictionary uses tables of terms, and thus crosstranslates terms between any two languages. The dictionary database is expandable to allow the addition of
new languages (with different alphabets). It was coordinated by S Tabakov and its software was made by
AM Studio. The same software company developed
the whole web database and search engines for the
e-Encyclopaedia.
To build the Encyclopaedia, each term from the dictionary was covered by an explanatory article (entry) in
English. The entries were aimed at MSc-level and above.
Their volume varies from approximately 50 to 500 words.
The model of the Encyclopaedia was built around a larger number of specific entries, rather than small number of
multi-page articles. This model allows an easy and effective search for information. About 3,400 articles were developed with an overall volume of 2,100 A4 pages. To avoid
problems related to complexity of the web site, the articles
are not internally hyperlinked, instead most of them include
list of related articles.
More than 2,000 images, graphs etc. were included in the articles to enhance the educational value
of the reference material. The articles were grouped
in seven categories – Physics of: X-ray Diagnostic
Radiology, Nuclear Medicine; Radiotherapy; Magnetic
Resonance Imaging; Ultrasound Imaging; Radiation
Protection; General terms. Each article includes contribution from three specialists – author, referee, and
group coordinator.
The EMITEL web site combines the Dictionary and the
Encyclopaedia and it uses the ability of the current Internet
browsers to operate with all languages. So, each translated term comes with an area-specific hyperlink displaying
the corresponding article for this term.
86
Revista Brasileira de Física Médica.2011;5(1):85-8.
How to use EMITEL web site
To use the Encyclopaedia (in English only): select
Encyclopaedia; write the term you want to see at the window; and click Enter. A list with terms is displayed – against
each one is a blue hyperlink related to the area of the term,
so click the hyperlink to read the article.
EMITEL can also search inside the text of the articles.
To do so, select Search in Full Text; specify the area and
proceed as above. In case of UK or American English differences (i.e. colour>color; optimise>optimize), try both
spellings or search only part of the term (e.g. colo, optim).
To use the Dictionary: select Dictionary; choose the
Input and Output languages; write the term you want to
see at the window; and click Enter. A list with terms is displayed, where the terms are found either single, or in combination with other words (the e-Dictionary assumes that
the user’s Internet browser already supports languages).
To use both the Encyclopaedia and Dictionary: select
Combined and proceed as above (this search is limited
only to the title of the article, not inside its text).
The web site was built with two Search Engines – one
searching into the Lists of terms (in all languages) and another one searching inside the text of the articles. The latter
allows significant increase of the potential of the e-Encyclopaedia, including search for related terms, acronyms, and
synonyms. To use this facility, the user has to select Search
in Full Text and specify the category/area of the search (as
described above).
EMITEL future development
EMITEL Network continues its activities related to the support and update of the Dictionary and Encyclopaedia. To
allow this, an additional web site was developed to handle the updates. This web Content Management System
(CMS, also developed by AM Studio) allows not only online
editing of the materials, but also adding new terms/entries
and including new languages. EMITEL Consortium and
Network have editorial control over the online material.
The e-Encyclopaedia attracted more than 300 specialists from 36 countries (United Kingdom, Sweden,
Italy, Bulgaria, Austria, Croatia, Cyprus, Czech Republic,
Denmark, Estonia, Finland, France, Germany, Greece,
Hungary, Ireland, Japan Latvia, Lithuania, Poland, Portugal,
Romania, Slovenia, Spain; Australia, Bangladesh, Canada,
China, Croatia, Iran, Libya, Malaysia, Morocco, Russia,
Thailand, Turkey, USA) – the number of contributors expands rapidly. The network has a dedicated Administrator
at King’s College London –who is also link of contact for
new contributors.
Alongside the development of the digital content of
EMITEL, an Agreement is made with a Publishing company to allow the paper print of the Encyclopaedia (expected by the WC2012).
EMITEL e-Encyclopaedia of Medical Physics and Dictionary of Terms
Acknowledgment
References
EMITEL gratefully acknowledges the support of the EU
1.
Leonardo Programme, the Partner Institutions and its
many Contributors (EMITEL Network), all listed in the web
site www.emitel2.eu.
2.
Tabakov S, Roberts C, Jonsson B, Ljungberg M, Lewis C, Strand S, et al.
Development of Educational Image Databases and e-books for Medical
Physics Training. J Med Eng Physics. 2005;27(7):591-9.
EC project 162-504 EMITEL. Available from: http://ec.europa.eu/research/
index.cfm.
Revista Brasileira de Física Médica.2011;5(1):85-8.
87
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):89-92.
Influence of brain region of interest location
for apparent diffusion coefficient maps
calculation for reference values to be used
in the in vivo characterization of brain
tumors in magnetic resonance images
Influência da localização da região de interesse cerebral
para cálculo dos mapas do coeficiente de difusão
aparente para valores de referência a serem utilizados
na caracterização in vivo de imagens de ressonância
magnética de tumores cerebrais
Edna M. Souza1,2,3, Gabriela Castellano3,4 and Eduardo T. Costa1,2
1
Center of Biomedical Engineering of the State University of Campinas (UNICAMP) - Campinas (SP), Brazil.
Biomedical Engineering Department of Electrical and Computational Engineering School of UNICAMP - Campinas (SP), Brazil.
3
Neurophysics Group, Cosmic Rays and Cronology Department, Gleb Wataghin Physics Institute, UNICAMP - Campinas (SP), Brazil.
4
CInAPCe Program (Cooperação Interinstitucional de Apoio a Pesquisas sobre o Cérebro) - São Paulo (SP), Brazil.
2
Abstract
In general, pathologic processes, such as neoplasic cell changes, tend to alter the magnitude of structural organization by destruction or reorganization of membranous
elements or by a change in cellularity. These changes will also have an impact on proton mobility, which can be followed up by DWI (diffusion weighted magnetic
resonance imaging). From DWI is obtained the ADC (apparent diffusion coefficient) map, which is a representation of the magnitude of water diffusion at the points
of a given region of interest (ROI). The purpose of this study was to assess the variation of ADC values in different brain ROIs of normal subjects, using a computer
tool previously developed. The aim of this assessment was to verify whether ADC values could be used to differentiate between normal subjects and patients with
multiform glioblastoma (a high-grade glioma) and meningioma. ADC maps were calculated for 10 controls, 10 patients with glioblastoma and 10 with meningioma.
For controls, mean ADC values were calculated for 10 different ROIs, located in the same places where the tumors were present in the patients. These values were
then averaged over ROIs and over subjects, giving a mean ADC value of (8.65±0.98)x10-4 mm2/s. The mean ADC values found for brain tumors were (5.03±0.67)
x10-4 mm2/s for meningioma and (2.83±0.45)x10-4 mm2/s for glioblastoma. We concluded that the ROIs used for computing brain ADC values for controls were not
essential for the estimation of normal reference ADC values to be used in the differentiation between these types of tumors and healthy brain tissue.
Keywords: magnetic resonance, diffusion, brain tumors, computer-assisted image processing.
Resumo
Em geral, os processos patológicos, como as alterações celulares neoplásicas, tendem a alterar a magnitude da organização estrutural pela destruição ou
reorganização dos elementos membranosos, ou pela mudança na celularidade. Tais mudanças também terão um impacto na mobilidade do próton, que pode ser
acompanhada pela imagem ponderada de difusão. Pela imagem ponderada de difusão, pode-se obter o mapeamento do coeficiente de difusão aparente, que é a
representação da magnitude da difusão da água nos pontos de certa região de interesse. O objetivo deste estudo foi avaliar a variação dos valores do coeficiente de
difusão aparente em diferentes regiões de interesse cerebral de indivíduos normais, utilizando uma ferramenta computacional que foi previamente desenvolvida. O
objetivo desta avaliação foi verificar se os valores do coeficiente de difusão aparente poderiam ser utilizados para diferenciar indivíduos normais de pacientes com
glioblastoma multiforme (um glioma de alto grau) e meningioma. Os mapeamentos do coeficiente de difusão aparente foram calculados para dez controles, dez
pacientes com glioblastoma e dez com meningioma. Para os controles, os valores do coeficiente de difusão aparente médio foram calculados para dez diferentes
regiões de interesse, localizadas nos mesmos lugares onde os tumores estavam presentes nos pacientes. Esses valores foram, em seguida, calculados sobre
as regiões de interesse e sobre os sujeitos, fornecendo um valor do coeficiente de difusão aparente médio de (8,65±0,98)x10-4 mm2/s. Os valores médios do
coeficiente de difusão aparente encontrados para tumores cerebrais foram de (5,03±0,67)x10-4 mm2/s, para o meningioma, e (2,83±0,45)x10-4 mm2/s, para o
glioblastoma. Concluiu-se que as regiões de interesse utilizadas para se computar os valores do coeficiente de difusão aparente cerebral para os controles não foram
essenciais para estimar os valores de referência normal, que deverão ser usados na diferenciação entre esses tipos de tumores e tecido cerebral saudável.
Palavras-chave: ressonância magnética, difusão, neoplasias encefálicas, processamento de imagem assistida por computador.
Corresponding author: Edna Marina de Souza – Cidade Universitária Zeferino Vaz, Center of Biomedical Engineering, UNICAMP – Barão Geraldo –
CEP: 13.083-970 – Campinas (SP), Brazil – E-mail: [email protected]
Associação Brasileira de Física Médica®
89
Souza EM, Castellano G, Costa ET
Introduction
In diffusion-weighted magnetic resonance imaging (DWI),
the contrast is determined by the microscopic and random
motion of water protons. In general, pathologic processes,
such as neoplasic cell changes, tend to alter the structural
organization of membranous elements through changes
in cellularity1. Such changes affect the average trajectory
of water molecules through tissue, which can be analyzed qualitatively and quantitatively using DWI. Based on
these images and on T2 weighted images, ADC (Apparent
Diffusion Coefficient) maps are calculated, whose values
can be used to distinguish between normal and pathological brain tissue.
DWI can be obtained by pulse sequences commonly used for the acquisition of structural images with
the insertion, in these sequences, of two gradients of
equal magnitude and opposite orientations (or same
orientation, but separated by a radiofrequency pulse of
180º), as shown as in Figure 12. Thus, water protons
that moved between the applications of both diffusion
gradients will generate signals of different magnitudes,
being of lower amplitude the signal from the instant after
the last application of diffusion gradient. Figure 1 shows
a Spin-Echo (SE) pulse sequence, commonly used to
acquire diffusion images.
49.6 ± 4.5 years, 40% women), and 10 patients with
meningioma aged between 36 and 54 years (mean =
42.5 ± 2.8 years, 60% women). All tumor cases were
confirmed by histopathological analysis performed after
images acquisitions. The study was approved by the
Ethics Review Board of UNICAMP Medical Sciences
School. DWI and T2-weighted images were acquired in
DICOM format in a Prestige 2T scanner, manufactured
by Elscint (Haifa, Israel). Diffusion images were registered
on T2 images using Mutual Information Maximization
(MIM) and Affine Transformations (AT)3. This step was
aimed at aligning DWI and T2 images, since despite
these are acquired one after another, small head displacements along the scan result in voxel shift between
the images. The ADC maps are calculated using a computational tool developed previously in Matlab®. From
the DWI acquired in the x, y and z directions, a mean
DW image (SI) was calculated, containing information
about water diffusion. The SI image is given by:
SI = (SIxSIySIz)1/3
SIx, SIy and SIz are the DW images acquired along the
x, y and z directions. The calculation of ADC values is performed using the following equation:
SI = SI0 × e-bADC
90º
180º
RF
GDiff
GDiff
GM
GS
GP
t1
GDiff
GDiff
GDiff
GDiff
I
)
TE
Figure 1. Spin-Echo (SE) pulse sequence for acquisition of
DWI. (RF: radiofrequency pulse; GS: slice selection direction;
GP: phase codification direction; GM: frequency codification
direction. Gdiff: diffusion gradient; t1: time between application
of first RF pulse and first diffusion gradient; Δ: time between
two diffusion gradients; δ: application time of diffusion gradient; TE: echo time).
Materials and methods
To develop the present study, DWI and T2 images present in a database of the Neuroimaging Laboratory, in
UNICAMP hospital, were used. We analyzed 10 control
subjects aged between 22 and 48 years (mean = 33.5 ±
3.8 years, 40% women), 10 patients with multiform
glioblastoma aged between 42 and 64 years (mean =
90
Revista Brasileira de Física Médica.2011;5(1):89-92.
(1)
(2)
SI0 is the intensity of the T2 image, SI is the intensity
of the diffusion-weighted image, b is the coefficient of diffusion sensitization in s/mm2 and ADC is the ADC value,
in mm2/s. In the MRI scanner used, the parameter b was
fixed to a value of 700 s/mm2.
In order to facilitate the calculations and minimize
noise in the ADC maps, a mask was developed in
Matlab, using the Mathematical Morphology operations
of dilation and closing, with a structuring element type
diamond4. This mask allowed the application of the presented equations only on the places of the image corresponding to brain. ADC maps obtained were converted
into DICOM images, and the ADC mean value in regions
of interest (ROIs) corresponding to normal brain tissue and tumors were calculated and compared among
themselves and with literature values. The ROIs were
drawn using the software ImageJ® with guidance of a
neurosurgeon, based on visual aspects of the tumor on
the ADC map and considering its possible proliferation
pathways.
To evaluate the possible existence of dependence
between the ADC values and the brain region, mean
ADC values were calculated on controls for 20 different
ROIs located in the same places where the tumors (10
meningiomas and 10 glioblastomas) were present in the
patients. These values were then averaged over ROIs
and over subjects, giving a mean ADC value used as
reference.
Influence of brain region of interest location for apparent diffusion coefficient maps calculation for reference values to be used in the in vivo characterization of brain tumors magnetic resonance images
Results
MEAN ADC VALUES - COMPARISON
12
ADC
-4
9
2
ADC (mm /s) x10
Figure 2 shows examples of ADC maps calculated for
a control subject and a patient with glioblastoma. The
white ROI corresponds to the area of mean ADC calculation in tumor. Figure 3 shows examples of ADC maps
calculated for a control subject and a patient with meningioma. The white ROI corresponds to the area of mean
ADC calculation in tumor. Figure 4 shows a plot of mean
ADC values for control subjects, patients with glioblastoma and patients with meningioma. Figure 5 shows the
distribution of mean ADC values for control subjects obtained in ROIs of glioblastomas. Figure 6 shows mean
ADC values for the control group in ROIs of glioblastoma
excluding lateral ventricles. Figure 7 shows the distribution of mean ADC values for control subjects obtained in
ROIs of meningioma.
6
3
0
MENINGIOMA
GLIOBLASTOMA
CONTROL
Figure 4. Mean ADC values for control group ((8.65±0.98) ×10-4
mm2/s), glioblastoma patients ((2.83±0.45) ×10-4 mm2/s), and
meningioma patients ((5.03±0.67) ×10-4 mm2/s).
6
-4
2
Peak=8,88x10 mm /s
2
-7
2
=3,6x10 mm /s
Frequency count
Gaussian fit
Frequency
4
2
Figure 2. (A) ADC map for a patient with glioblastoma. (B) ROI
for mean ADC calculation in tumor. (C) ADC map for a control
subject. (D) Same (B) ROI applied to (C) ADC map for calculation
of mean ADC values in healthy brain tissue located in the same
place where the tumor was present in the patient.
0
8,6
8,8
-4
9,0
2
9,2
9,4
ADC (x10 mm /s)
Figure 5. Distribution of mean ADC values for control group
healthy brain tissue using the same ROIs used for mean ADC
calculation in glioblastoma.
5
-4
2
Peak=8,25x10 mm /s
-5
2
=3,7x10 mm /s
Frequency counts
Gaussian fit
Frequency
4
3
2
1
0
Figure 3. (A) ADC map for a patient with meningioma. (B) ROI
for mean ADC calculation in tumor. (C) ADC map for a control
subject. (D) Same (B) ROI applied to (C) ADC map for calculation
of mean ADC values in healthy brain tissue located in the same
place where the tumor was present in the patient.
7
8
-4
2
ADC ( x10 mm /s)
9
10
Figure 6. Distribution of mean ADC values for control group
healthy brain tissue using the same ROIs used for mean ADC
calculation in glioblastoma excluding the ROI portion that corresponds to the lateral ventricles.
Revista Brasileira de Física Médica.2011;5(1):89-92.
91
Souza EM, Castellano G, Costa ET
4
-4
2
Peak=8,9x10 mm /s
-5
2
=1,2x10 mm /s
Frequency counts
Gaussian fit
Frequency
3
2
1
0
8,4
8,6
8,8
-4
9,0
2
9,2
9,4
ADC ( x10 mm /s )
Figure 7. Distribution of mean ADC values for control group
healthy brain tissue using the same ROIs used for mean ADC
calculation in meningioma.
Discussion
For control subjects, patients with glioblastoma and patients with meningioma, the mean ADC values were (8.65
± 0.98)×10-4 mm2/s, (2.83 ± 0.45)×10-4 mm2/s and (5.03 ±
0.67)×10-4 respectively, as seen in the graph of Figure 4.
A t-test applied to the ADC values showed that they were
significantly different (p < 0.001) between the groups of
patients compared to healthy subjects. For glioblastoma,
the values obtained agree with information found in the
literature5. For control subjects, the value corresponds to
regions containing normal white and gray matter6.
The results found show that there is no significant influence of ROI location in the determination of ADC values
for normal brain tissue in the control group. Moreover, it is
possible to differentiate between healthy and tumoral brain
tissue using ADC values. The protocol developed in this
work should be further associated with other techniques
of image processing, among which texture analysis tools
that apply second-order statistics, such as co-occurrence
and run length matrices. The calculation of ADC values for
normal brain tissue using the same ROIs as in the group of
tumors showed no significant dependency of ADC values
of normal tissues with the brain region, as shown by the
92
Revista Brasileira de Física Médica.2011;5(1):89-92.
graphs of Figures 5-7. However, ROIs portions that overlap
the lateral ventricles should be excluded from the calculation, since in those regions ADC values are much higher
due to the flow of cerebrospinal fluid (CSF). The non-exclusion of the lateral ventricles from these calculations makes
the distribution of ADC values show a tendency to higher
values than those found in areas where there is only gray
and white matter (Figure 5). In Figure 6, there is not this
tendency. For meningioma, it is not necessary to exclude
the lateral ventricles from the ADC calculation.
Glioblastoma are tumors of high cellularity6. The large
concentration of tumor cells in a given region hinder the
flow of water molecules through it, resulting in lower ADC
values compared to the control condition.
Conclusion
Based on these results, it appears that the calculation of
ADC values can be a useful tool for distinguishing between
normal brain tissue and tumor.
Moreover, this tool should be associated with other
techniques of image processing, such as co-occurrence
and run-length matrices7, taking into account the neighborhood relations between pixels in a given ROI in an attempt to obtain a computational resource that allows the
characterization of healthy and pathological brain tissues
noninvasively and in vivo in routine clinical practice.
References
1. Weiss TF. Cellular Biophysics, v. 1: Transport. Cambridge: Oxford Publishing
Group; 2004.
2. Mansfield P. Multiplanar imaging formation using NMR spin-echoes. Solid
State Physics. 1977;10:55-8.
3. Quasi AA. Image Registration Toolkit. Department of Computer Science,
Copenhagen University; 2008.
4. Lotufo RA, Dougherty ER. Hands-on morphological image processing.
Bellingham: SPIE Press; 2003.
5. Norris DG. The effects of microscopic tissue parameters on the diffusion weighted
magnetic resonance imaging experiment. NMR Biomed. 2001;14(2):77-93.
6. Beaulileu C, Allen PS. Determinants of anisotropic water diffusion in nerves.
Magn Reson Med. 1994;31(4):394-400.
7. Haralick RM. Statistical and structural approaches to texture. Proceedings
of IEEE. 1979;67:786 -804.
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):93-8.
Effect of the scanner background noise
on the resting brain networks detected by
functional magnetic resonance imaging
Efeito do ruído de fundo do tomógrafo nas redes cerebrais
de repouso detectado pela imagem por ressonância
magnética funcional
Carlo Rondinoni1,2, Antônio Carlos dos Santos2 and Carlos Ernesto G. Salmon1
1
Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto - Department of Physics, University of São Paulo, Ribeirão Preto (SP), Brazil.
2
Faculdade de Medicina de Ribeirão Preto - Department of Medical Clinics, University of São Paulo, Ribeirão Preto (SP), Brazil.
Abstract
Resting state studies by fMRI are carried out in order to identify the brain networks responsible for their basal functioning, which are known as resting state
networks. Although considered to be in rest, subjects are unavoidably under a massive charge of environmental acoustic noise produced by the magnetic
resonance imaging equipment. Our aim was to verify if the massive auditory information input could mask the “real” resting state networks. The functional
volumes were acquired when seven naïve subjects (four women) had their eyes opened under default echo planar imaging (EPI) sequences or during soft-tone
sequences (slew-rate reduction), as allowed by a Philips Achieva 3T magnetic resonance imaging scanner. The sound pressure level difference between the
default and soft sequences reached 12 dB. Experimental sessions consisted of two runs of seven minutes each under different levels of noise. The sequence
of conditions was counterbalanced between subjects. The functional volumes were pre-processed in BrainVoyager and submitted to self-organizing group
Independent Component Analysis (sogICA). The influence of the higher noise level was evaluated by identifying the BOLD components and by comparing the
functional volumes of the five representative resting state networks under each condition (random effects – Independent Component Analysis). The results
show that a lower level of noise may uncover functionally wider components. A t-test showed that the high noise condition induced significantly higher BOLD
signal in the posterior cingulate cortex only. However, lower noise levels induced higher BOLD activity in the bilateral parietal lobule, bilateral superior frontal
gyrus, and insula. Yet, the motor resting state network seems to be wider under low noise, reaching auditory areas in the temporal cortices, and an oscillatory
component on the thalamus was identified in the low noise condition. The results indicate that a compromise should be taken into account when studying
rest, balancing between noise reduction, and speed of acquisition.
Keywords: fMRI, default-mode network, resting state, acoustic noise, ICA.
Resumo
Os estudos relativos ao estado de repouso pela imagem por ressonância magnética funcional (fMRI) são realizados para identificar as redes cerebrais
responsáveis pelas função basal das redes neurais, conhecidas como redes em estado de repouso. Embora considerados em repouso, os indivíduos
inevitavelmente recebem uma alta intensidade de sons do ambiente produzidos pelo equipamento de ressonância magnética. O objetivo deste estudo foi
verificar se a informação auditiva recebida poderia mascarar as “verdadeiras” redes em estado de repouso. Os volumes funcionais foram obtidos quando
sete indivíduos (quatro mulheres) tiveram seus olhos abertos em sequências pulso ecoplanares (EPI) ou durante sequências silenciosas (redução da taxa de
variação), utilizando um scanner de ressonância magnética Philips Achieva 3T. A diferença no nível de pressão do som entre o EPI padrão e a silenciosa
chegou a 12 dB. As sessões experimentais consistiam de duas etapas de sete minutos com diferentes níveis de ruído. A sequência das condições foi
contrabalanceada entre os indivíduos. Os volumes funcionais foram pré-processados no programa BrainVoyager e submetidos a análise de componentes
independentes em grupos auto-organizados (sogICA). A influência do nível mais alto de ruído foi analisada pela identificação dos componentes BOLD e pela
comparação dos volumes funcionais das cinco redes cerebrais em estado de repouso representativas para cada condição (efeitos aleatórios da análise
independente dos componentes). Os resultados mostram que um nível menor de ruído pode revelar componentes funcionalmente mais amplos. Um teste
t mostrou que o ruído intenso induziu sinais BOLD mais intensos somente no córtex cingulado posterior. Porém, níveis menores de ruído induziram maior
atividade BOLD no lobo parietal bilateral, no giro frontal superior bilateral e na insula. Além disso, a rede motora de repouso parece ser mais ampla sob ruídos
baixos, alcançando áreas auditivas no córtex temporal, e um componente oscilatório no tálamo foi identificado sob ruído baixo. Os resultados indicam que um
compromisso deve ser considerado ao estudar o repouso, ponderando-se a redução do ruído e a velocidade da aquisição.
Palavras-chave: fMRI, rede de modo padrão, estado de repouso, ruído acústico, ICA.
Corresponding author: Carlos Ernesto Garrido Salmon – Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto – University of São Paulo – Avenida
Bandeirantes, 3.900 – Ribeirão Preto (SP), Brazil – E-mail: [email protected]
Associação Brasileira de Física Médica®
93
Rondinoni C, Santos ACD, Salmon CEG
Introduction
It was observed that, when the voluntary action is required,
some areas of the brain are paradoxically deactivated. This
finding led the researchers to suggest that there is an organized mode of brain function, known as “default mode”, as
a baseline that is partially suspended during one specific
behavior of directed action1. Two explanations were raised
to account for these deactivations during targeted actions.
Firstly, these brain areas could be deactivated following the
reallocation of attentive resources for tasks that require extreme focus of attention. Secondly, areas may be deactivated because they are related to the basal monitoring of the
external environment or linked to free-thought processes,
such as: mind wandering or sensory-motor awareness2.
The subtleness in the functioning of the brain basal
state and its relationship to the external world raised a concern about the influence of the environmental factors, once
they could mask deactivations related to basal monitoring
of the external environment or free-thought. The effect of
the scanner background noise (SBN) may be intertwined
in the functional data, avoiding the identification of the real
resting state networks (RSN). It has been shown that the
brain activity was differently modulated depending on the
type of acquisition. Continuous sampling, that is, image
acquisition under continuous noise showed that the brain
activity was different from sparse sampling, when echo
planar imaging (EPI) volumes were acquired after periods
of silence. The SBN seems to have suppressed the default
network components, like the medial prefrontal cortex,
posterior cingulated, and precuneus. The authors state
that the noise does not alter the spatial distribution of the
default network, but it influences its magnitudes in a nonlinear fashion3. This is also true for the working memory
functioning4, as shown by the higher recruitment of attentive resources under high noise conditions. BOLD activation was higher in the cerebellum, frontal cortex, fusiform
cortex and lingual gyrus, and lower in the anterior cingulated and putamen. As found in another study assessing
SBN5, the higher attentive demand to hear sentences under noise resulted in a higher activation in the left temporal
and inferior parietal cortices. The authors conclude that
silent EPI sequences would be more adequate for auditory perception studies and its applicability depends on the
regions of interest in the brain.
If from one side the RSNs seem to be influenced by
environmental factors, the same is not true if the subjects
rest with their eyes opened, open with fixation or with the
eyes closed6. The most relevant finding in this study is the
observed behavioral competition between focused attention and free-thought processes. While posterior cingulate
and medial prefrontal cortices oscillate in phase, the intraparietal sulcus shows out of phase activity with the former
areas, which evidences a functioning pattern of anti-correlated fluctuations.
Along with direct correlation, another common way of
analyzing data for intrinsic patterns, without assuming any
94
Revista Brasileira de Física Médica.2011;5(1):93-8.
a priori condition, is the Independent Component Analysis
(ICA)7,8. The algorithm is based on an adaptive filter that
maximizes the independence of the temporal series components by the progressive decreasing of mutual information. The ICA algorithm is applied on all voxels in order to
separate the set of information in networks with coherent
and maximally independent fluctuations9. Group results
can be achieved by applying a clustering method on ICA
maps of different subjects, which allows to identify the
common activity across individuals10.
Materials and methods
Our approach consists in using the fMRI acquisition to verify if the massive input of auditory information shall mask the
true brain RSNs. Data acquired with EPI sequences producing two different levels of SBN were compared. First,
data were submitted to the ICA. Then, the components
under each noise level were grouped by a self-organizing
clustering algorithm. Finally, the RSN maps under each
noise level were compared in order to show which areas
presented significant differences, indicating if the sound
pressure produced by the different EPI sequences were
influencing resting state results.
fMRI acquisition
The functional volumes were acquired when seven naïve
subjects (four women) had their eyes opened under default
EPI sequences or during soft-tone sequences (slew-rate reduction), as allowed by a Philips Achieva 3T magnetic resonance imaging (MRI) scanner. The acquisition of functional
images was accomplished with a standard eight-channel
head coil. Echo-planar images had the following parameters: 200 volumes, 29 slices in ascending order, 4 mm
slice thickness, voxel size 1.83x1.83 mm, slice time 66 ms,
FOV=240x240 mm, FH=95 mm, and TR/TE=2000/30 ms.
The silent sequence was designed by setting to maximum
level in maximum the “soft tone” parameter offered by the
MRI tomograph, which decreases the gradient slew rate
leading to lower coil vibration levels during acquisition. The
high noise condition was done with “soft tone” parameter
turned off. The difference between the sound pressures in
each scanner setting reached the order of 12 dB, as evidenced by a recording done with a microphone connected
to the inbuilt MRI apparatus communication device. The
only difference between the two acquisitions was related
to the slew rate and consequent lower noise, while repetition time and other parameters were kept constant. After
functional scans, each subject was scanned for the acquisition of anatomical 3D T1-weighted images (TR=9.7 ms;
TE=4 ms; flip angle 12°; matrix 256x256; FOV=256 mm;
1 mm slice thickness; voxel size 1x1x1 mm).
Experimental procedure
Experimental sessions consisted of two runs of seven minutes each under two different levels of noise. The sequence
Effect of the scanner background noise on the resting brain networks detected by functional magnetic resonance imaging
of conditions was counter-balanced between subjects.
Voluntaries were instructed to keep their eyes opened while
looking steadily through the head coil mirror. The field of vision included the outside of the scanner bore and a curtain
over the control room window. The functional volumes were
pre-processed and submitted to self-organizing group ICA
(random effects sogICA) in BrainVoyager. The difference of
the noise levels was evaluated by comparing the maps of
five categories of RSNs under each sound pressure level.
fMRI data analysis
Data were processed in BrainVoyager (Brain Innovations, The
Netherlands). Functional volumes were corrected for 3D motion with reference to the first volume. Subjects that showed
movements larger than 2 mm were excluded. Slice order correction and co-registration to the anatomical volume were done
before standardization into the Tailarach space. After linear drift
filtering, functional data entered the ICA algorithm. BOLD components were identified depending on the fingerprint of each
map11, leading to a classification as published in a previous
research12. Components, which showed high spectral densities between 0.02 and 0.05 Hz, high values of clustering and
skewness, high temporal and spatial entropy, high lag-1 autocorrelation and low kurtosis, were selected and considered as
a component related to BOLD signal. The number of voxels in
each region of interest indicated by the ICA under each noise
level was compared in a bi-caudal t-test. Voxels with significant
differences were identified in the Tailarach atlas and they are
presented in the next session.
Results
The RSNs were identified here as the following, as categorized in a previous paper11:
• default-modenetwork(RSN1);
• bilateralvisualcortices(RSN3);
• fronto-parietalnetwork(RSN2);
• bilateralmotorandauditorycortices(RSN4and5);
• anteriorcingulatecortex(RSN6).
Figures 1 and 2 present the networks found in our study,
pointing out each RSN category. A pair-wise comparison between the five maps under each noise condition showed that
the high noise condition induced significantly higher BOLD
Figure 1. Independent Component Analysis maps calculated
on the data during standard acquisition (loud noise, soft-tone
parameter off). Numbers indicate the resting state networks as
defined in a previous paper11: 1. default-mode network, internal
processing; 2. retinotopic occipital cortex, visual processing; 3.
dorsal attentive network, goal-directed action; 4. sensory motor cortices, motor control; 5. pre-frontal cortex, self-referential
mental activity. The color scales on the Independent Component
Analysis maps are arbitrary and they are not related to positive
or negative levels of activation. Maps show t-values between
3.0 and 10.
Figure 2. Independent Component Analysis maps calculated on
the data during soft-tone acquisition (low noise level). Numbers
indicate the resting state networks as defined elsewhere11. The
color scales on the Independent Component Analysis maps are
arbitrary and they are not related to positive or negative levels of
activation. Maps show t-values between 3.0 and 10.
Revista Brasileira de Física Médica.2011;5(1):93-8.
95
Rondinoni C, Santos ACD, Salmon CEG
signal in the posterior cingulate cortex, while the soft-tone sequence induced higher BOLD activity in the bilateral parietal
lobule, bilateral superior frontal gyrus, and insula (Table 1).
Visual inspection of each RSN under different sound pressures indicates that, apart from RSN 1 (default-mode network
- DMN), lower level of noise has uncovered components that
are spread over a larger area of the cerebral cortex.
As shown in Table 2, the number of voxels in each
group component is higher only for the classical defaultmode network. The remaining networks show wider areas
of activity under soft-tone acquisition or show the same
extension of activation, as it is the case of the dorsal attentive network (Table 2).
Three aspects should be noted in these results. First,
the motor RSN shows much wider extensions of activity under lower levels of noise, as denoted by the disparate number of voxels in each environmental condition
(Table 2). Second, a component with activity in the thalamus (picture not shown) was identified in the low noise
group analysis, while no such activity was evidenced for
the standard MRI sequence. Third, the salience network
component has significant activity widespread across the
cerebral hemispheres during soft-tone acquisition, as it
can be seen on the comparison between the maps of
Figures 1 and 2 (RSN 5).
Table 1. Brain areas and statistical significance found in the
gross pair-wise comparison between resting state independent components under standard and soft-tone acquisitions
(Student’s t-test, bi-caudal).
Higher for soft-tone acquisition (p<0.01)
Brodmann’s
Topographic region Hemisphere Voxels t-value
area
40
Inferior parietal lobule
Left
528
3.63
40
Inferior parietal lobule
Right
1,163 3.91
13, 22
Insula
Right
1,403 3.92
9
Superior frontal gyrus
Left
473
3.78
6
Superior frontal gyrus
Right
448
3.85
Higher for standard tone acquisition (p<0.01)
Brodmann’s
Topographic region Hemisphere Voxels t-value
area
30
Posterior cingulate
Right
1,168 4.02
Table 2. Extension (in voxels) of group independent components, depending on the level of acoustic noise produced by the
magnetic resonance imaging scanner. Resting state networks
(RSNs) on the list are named as defined elsewhere11.
Number of voxels (relative proportion)
Resting state
networks category
1
DMN
2
VISUAL
3
MOTOR
4
DORSAL
5
SALIENCE
96
HARD noise
SOFT noise
55913 (0.55)
41822 (0.43)
8210 (0.13)
36940 (0.50)
12170 (0.43)
45273 (0.45)
54396 (0.57)
55318 (0.87)
37273 (0.50)
16021 (0.57)
Revista Brasileira de Física Médica.2011;5(1):93-8.
Discussion
The motivation of this research is the concern about
the possible influence of the environmental noise on
the organized brain function at rest, mainly represented by the DMN. During the last years, resting state
paradigms became the approach of choice of many
investigators, given their simplicity and reliability, being applicable on a wide variety of subjects. Thus, the
concern about the influence of the acoustic noise on
the RSNs comes from the fact that the whole body of
research could be looking at no rest at all, but into networks that, in fact, are defending the sensory system
from an annoying massive input of the acoustic noise.
A previous research was carried out using the “nearrest” approach, when periods of rest intertwined with
directed behavior were scrutinized for the active brain
areas. The present study attempts for the first time to
investigate the brain dynamics during pure rest conditions. In a scenario of high acoustic noise, reallocation
of attentive resources may be difficult since the basal
monitoring of the external environment is saturated.
That is, both functions that are suggested for the DMN
are under stress.
Our results seem to be in complete accordance with
this rationale and as suggested by others. The aversive
nature of the loud sound seems to damp the sensorial processes, which shall induce a reduction in the activity in the
brain areas related to sensory-motor, auditory, and salience
processing. On the other hand, lower levels of sound may
favor the free dynamics among the totality of the RSNs,
which has a higher place for large oscillatory activity in the
motor, auditory and salience cortices with the contribution
of the thalamus.
Conclusions
Our concern now, along with the future refinement of the
data analysis, is to bring into discussion the implications
of applying the soft-tone sequence in the all-day of the
research, as a way to permit the appreciation of the “real”
RSNs. Our results suggest that the standard level of noise
in the all-day resting brain research must be taken into account, once different arrangements of active areas may be
found when the scanner room gets less loud and more
comfortable. Further research is needed to address the
subtleness of such differences, gathering clues from a
greater number of subjects or from a group of neurological
patients.
Acknowledgments
We thank João Pereira Leite and Jaime Shinsuke Ide
for the valuable comments about the approach of our
research.
Effect of the scanner background noise on the resting brain networks detected by functional magnetic resonance imaging
References
1. Raichle ME, MacLeod AM, Snyder AZ, Powers WJ, Gusnard DA, Shulman
GL. A default mode of brain function. PNAS. 2001;98(2):676-82.
2. Shulman GL, Fiez JA, Corbetta M, Buckner RL, Miezin FM, Raichle ME, et al.
Common blood flow changes across visual tasks: II. Decreases in cerebral
cortex. J Cognit Neuroscience. 1997;95:648-63.
3. Gaab N, Gabrielli JDE, Glover GH. Resting in peace or noise: scanner
background noise suppresses default-mode network. Human Brain
Mapping. 2008;29:858-67.
4. Tomasi DG, Caparelli EC, Chang L, Ernst T. (2005). fMRI-acoustic noise
alters brain activation during working memory tasks. NeuroImage
27(2):377-386.
5. Peelle JE, Eason RJ, Schmitter S, Schwarzbauer C, Davis MH. Evaluating
an acoustically quiet EPI sequence for use in fMRI studies of speech and
auditory processing. NeuroImage. 2010;52(4):1410-9.
6. Fox MD, Snyder AZ, Vincent JL, Corbetta M, Van Essen DC, Raichle ME.
From the Cover: the human brain is intrinsically organized into dynamic,
anticorrelated functional networks. PNAS. 2005;102(27):9673-8.
7. Fox MD, Raichle ME. Spontaneous fluctuations in brain activity observed
with functional magnetic resonance imaging. Nature Rev Neuroscience.
2007;8:700-11.
8. Fox MD, Greicius M. Clinical applications of resting state functional
connectivity. Front. Syst. Neurosciences. 2010;4:19.
9. Formisano E, Esposito F, Di Salle F, Goebel R. Cortex-based independent
component analysis of fMRI time series. Magnet Reson Imaging.
2004;22(10):1493-504.
10. Esposito F, Scarabino T, Hyvarinen A, Himberg J, Formisano E, Comani
S, et al. Independent component analysis of fMRI group studies by selforganizing clustering. NeuroImage. 2005;25(1):193-205.
11. Mantini D, Perrucci MG, Del Gratta CM, Romani GL, Corbetta M.
Electrophysiological signatures of resting state networks in the human
brain. PNAS. 2007;104(32):13170-5.
12. De Martino F, Gentile F, Esposito F, Balsi M, Di Salle F, Goebel R, et al.
Classification of fMRI independent components using IC-fingerprints and
support vector machine classifiers. NeuroImage. 2007;34:177-94.
Revista Brasileira de Física Médica.2011;5(1):93-8.
97
Artigo Original
Revista Brasileira de Física Médica.2011;5(1):99-104.
Fluorescein isothiocyanate labeled,
magnetic nanoparticles conjugated
D-penicillamine-anti-metadherin and in vitro
evaluation on breast cancer cells
Avaliação do isotiocianato de fluoresceína marcado, das
nanopartículas magnéticas conjugadas da D-penicilamina
antimetaderina e in vitro nas células do câncer de mama
Özlet Akça1, Perihan Ünak1, E.İlker Medine1, Çağlar Özdemir3, Serhan Sakarya2 and Suna Timur3
Institute of Nuclear Sciences, Department of Nuclear Applications, Ege University, Turkey.
ADUBILTEM Science and Technology Research and Development Center, Adnan Menderes University, Turkey.
3
Science Faculty, Department of Biochemistry, Ege University, Turkey.
1
2
Abstract
Silane modified magnetic nanoparticles were prepared after capped with silica generated from the hydrolyzation of tetraethyl orthosilicate (TEOS).
Amino silane (SG-Si900) was added to this solution for surface modification of silica coated magnetic particles. Finally, D-penicillamine (D-PA)-antimetadherin (anti-MTDH) was covalently linked to the amine group using glutaraldehyde as cross-linker. Magnetic nanoparticles were characterized by
scanning electron microscopy (SEM), X-ray diffraction (XRD), vibrating sample magnetometer (VSM), and atomic force microscopy (AFM). AFM results
showed that particles are nearly monodisperse, and the average size of particles was 40 to 50 nm. An amino acid derivative D-PA was conjugated
anti-MTDH, which results the increase of uptaking potential of a conjugated agent, labelled fluorescein isothiocyanate (FITC) and then conjugated
to the magnetic nanoparticles. In vitro evaluation of the conjugated D-PA-anti-MTDH-FITC to magnetic nanoparticle was studied on MCF-7 breast
cancer cell lines. Fluorescence microscopy images of cells after incubation of the sample were obtained to monitor the interaction of the sample
with the cancerous cells. Incorporation on cells of FITC labeled and magnetic nanoparticles conjugated D-PA-anti-MTDH was found higher than FITC
labeled D-PA-anti-MTDH. The results show that magnetic properties and application of magnetic field increased incorporation rates. The obtained
D-PA-anti-MTDH-magnetic nanoparticles-FITC complex has been used for in vitro imaging of breast cancer cells. FITC labeled and magnetic
nanoparticles conjugated D-PA-anti-MTDH may be useful as a new class of scintigraphic agents. Results of this study are sufficiently encouraging
to bring about further evaluation of this and related compounds for ultraviolet magnetic resonance (UV-MR) dual imaging.
Keywords: Fe3O4 magnetic nanoparticles, D-penicillamine, Anti-Metadherin, fluorescein isothiocyanate (FITC), MCF-7.
Resumo
Nanopartículas magnéticas modificadas de silano foram preparadas após serem tampadas com sílica criada da hidrolização do ortossilicato de
tetraetilo (TEOS). Aminossilano (SG-Si900) foi adicionado à solução para modificação da superfície da sílica revestida por partículas magnéticas. Por
fim, a D-penicilamina (D-PA)-antimetaderina (anti-MTDH) foi covalentemente ligada ao Grupo da Amina, utilizando glutaraldeído como ligante cruzado.
As nanopartículas magnéticas foram caracterizadas pela microscopia eletrônica de varredura (MEV), difração de raios X (DRX), magnetômetro de
amostra vibrante (MAV) e microscopia de força atômica (MFA). Os resultados da MFA mostraram que as partículas estão quase monodispersas e
o tamanho médio das partículas era de 40 a 50 nm. Um derivado aminoácido da D-PA foi conjugado como anti-MTDH, que resulta no aumento do
potencial de absorção de um agente conjugado, isotiocianato de fluoresceína marcado (FITC) e depois conjugado às nanopartículas magnéticas.
A avaliação in vitro da D-PA-anti-MTDH-FITC conjugada à nanopartícula magnética foi estudada em linhagens das células do câncer de mama
MCF-7. As imagens da microscopia de fluorescência das células após a incubação da amostra foram obtidas para monitorar a interação da amostra
com as células cancerígenas. A incorporação nas células do FITC marcado e das nanopartículas magnéticas conjugadas de D-PA-anti-MTDH foi
encontrada superior ao FITC marcado D-PA-anti-MTDH. Os resultados mostram que as propriedades magnéticas e a aplicação do campo magnético
aumentaram as taxas de incorporação. O complexo D-PA-anti-MTDH das nanopartículas magnéticas do FITC foi utilizado para a visualização in
vitro das células de câncer de mama. O FITC marcado e as nanopartículas magnéticas conjugadas em D-PA-anti-MTDH podem ser úteis como
uma nova classe de agentes cintilográficos. Os resultados deste estudo favorecem a realização de futuras avaliações para este e outros compostos
relacionados para a visualização com técnicas de dupla imagem da ressonância magnética ultravioleta.
Palavras-chave: nanopartículas magnéticas Fe3O4, D-penicilamina, antimetaderina, isotiocianato de fluoresceína, MCF-7.
Corresponding author: Ünak Perihan – Institute of Nuclear Sciences, Department of Nuclear Applications, Ege University – 35100 Bornova Izmir Turkey –
E-mail: [email protected]
Associação Brasileira de Física Médica®
99
Özlet A, Perihan Ü, E.İlker M, Çağlar Ö, Serhan S, Suna T
Introduction
In recent years, magnetic nanoparticles have attracted
much attention due to their unique magnetic properties and
widespread application in cell separation1,2, drug delivery3,4,
magnetic resonance image (MRI) techniques5, cancer diagnosis and treatment6-9. These ferrofluids can be directed
to magnetic area due to their magnetic properties10-12.
Biomedical applications have increased the interest of magnetic nanoparticles into silica. The nontoxic
silica is an ideal coating material because of its capability form extensive cross-linking, which leads to an inert
outer shield. Silanized nanocomposites are stable in a
wide range of biological environments. They are biocompatible and can also be easily activated to provide
new functional group.
D-Penicillamine (D-PA) is an aminothiol and a powerful chelating agent13. Penicillamine is largely used in
medicine in rheumatoid arthritis, Wilson’s disease for
the removal of copper, and in heavy metal poisoning16,17.
Penicillamine is a pharmaceutical of chelator class.
The pharmaceutical form is D-penicillamine. Like
L-penicillamine, it is highly toxic18.
Metadherin is a type 2 transmembrane protein, in
which its overexpression was first described in breast cancers. It plays an important role in the metastasis of breast
cancers into the lungs as a secondary site of development.
Metadherin is located in a small region of human chromosome 8, and it seems to be crucial to cancer’s spread or
metastasis since it helps tumor cells to tightly stick to blood
vessels in distant organs. The gene also makes tumors
more resistant to the powerful chemotherapeutic agents
normally used to wipe out the deadly cells.
Antibodies reactive to the lung-homing domain of
metadherin and siRNA-mediated knockdown of metadherin expression in breast cancer cells inhibited experimental
lung metastasis, indicating that tumor cell metadherin mediates localization at the metastatic site19,20.
Fluorescein isothiocyanate (FITC) is the original fluorescein molecule functionalized with an isothiocyanate reactive group (-N=C=S). The isothiocyanate group reacts
with amino terminal and primary amines in proteins. It has
been used for labeling proteins including antibodies and
lectins21,22. Anti-MTDH conjugated D-penicillamine was labeled with FITC using the amine group.
A based-novel antibody and D-penicillamine, magnetic
nanoparticle conjugated fluorescent complex for in vitro
imaging of breast cancer cells was reported here.
Material and Methods
Materials
All reagents were commercially available and analytical
grade. Anti-metadherin (100 mg/400mL) was purchased
from Zymed. D-PA and FITC were purchased from Aldrich
Chemical Co., and other chemicals were supplied from
100
Revista Brasileira de Física Médica.2011;5(1):99-104.
Merck Chemical Co. The MCF-7 human breast cancer
cell line was obtained from the American Type Culture
Collection.
Formation and Surface modification of Magnetic
Nanoparticles
Synthesis of core-shell (Fe3O4–SiO2) magnetic nanoparticles
Silica coated magnetic nanoparticles have been prepared
by partial reduction of Fe3+ ion with sodium sulfide under
nitrogen atmosphere. While Fe3O4 magnetite nanoparticles
have been formed at the first step of the reaction, they
were coated with silica at the second one.
6Fe3+ + SO32- + 18NH3.H2O → 2Fe3O4 + SO42- + 18NH4+
+ 9H2O
Surface modification of silica coated magnetic particles
with amino silane
Trialkoxysiylalkil substitute polymethylene diamine compounds, like SG-Si900 (N-[3-(trimethyoxysiyl)propyl]-ethylenediamine), are used for surface coating of inorganic
materials.
Glutaraldehyde Conjugation of Magnetic Nanoparticles
Determination of magnetic particles properties
The Scanning Electron Microscope (SEM) (Phillips XL-30
S FEG) was taken to determine surface morphology and
size of magnetic particles. Since the samples should be
dry for imaging, sample was dispersed in an evaporating
solvent (methanol) after being washed with ethanol-water
mixture for three times in this study. Samples were taken
with micropipette and put on the steel plates for SEM measurements. They were followed five minutes after methanol treated samples on steel plates were dried and images
were taken.
X-Ray diffractometer (XRD) (Phillips X’Pert Pro) analyses of magnetic particles were studied. Elemental analyses
of particles were made by X-ray diffraction. They were irradiated with collimated monochromatic X-rays. The diffraction angle and intensity of diffracted X-rays gave the known
several comparable pattern of samples.
Magnetic properties of magnetic particles were examined with the Vibrating Sample Magnetometer (VSM)
(LakeShore 7407) at Izmir Institute of Technology .
Atomic Force Microscope (AFM) (Q-Scope 250
Scanning Probe Microscope Ambios. Tech.) analyses of
magnetic particles were carried out at the Institute of Solar
Energy, Ege University.
Anti-Metadherin (Anti-MTDH) conjugation of D- Penicillamine
Five mg of D-penicillamine was dispersed in 1470 µL of
0.1 M sodium carbonate buffer (pH 9.0). Then, 30 µL of
glutaraldehyde was added and mixture was stirred at 4°
C for 24 hours. After that, 4 µL of anti-MTDH was added
to the mixture, and it allowed to stand overnight at 4 °C.
Fluorescein isothiocyanate labeled, magnetic nanoparticles conjugated D-penicillamine-anti-metadherin and in vitro evaluation on breast cancer cells
The solution was centrifugated for five minutes at 10,000 X
g by using centrifugal filter units (50,000 NMWL) to remove
unbound anti-MTDH .
labeling of anti-MTDH conjugated D- penicillamine
D-PA-anti-MTDH solution: The D-PA-anti-MTDH solution
was prepared by dissolving 5 mg of D-penicillamine in freshly
prepared 1 mL of 0.1 M sodium carbonate buffer (pH=9.0).
FITC solution: 1 mg FITC was used as labeling
agent, and it was dissolved in 1 mL of dimethyl sulfoxide
(DMSO).
FITC-D-PA-anti-MTDH: 100 μL of FITC solution was
added into D-PA-anti-MTDH solution at dark condition.
The mixture stayed at 4 °C during eight hours in order
to label D-PA-anti-MTDH with FITC.
500 μL of NH4Cl buffer (50 mM) and FITC-D-PA-antiMTDH were mixed and stayed at 4 °C, during two hours.
Then, 100 μL of glycerine was added. The mixture passed
through the column to leave unbound FITC.
nanoparticles-FITC, and magnetic field applying D-PA-antiMTDH-magnetic nanoparticles-FITC incorporation to cells.
9 cm2 tissue culture Petri dishes were visualized with 100 X
magnification and photographed through epi-fluorescence
microscopy (Olympus, Tokyo, Japan). Besides, the magnetic field effect was determined for the several cellular incorporations of the ligands conjugated magnetic particles.
Results and discussion
Structural Properties of Magnetic Particles
SEM Analyses Results
The SEM imaging of magnetic particles is as depicted in
Figures 1 to 3. SEM results showed that particles are nearly monodisperse. The average particle size is found to be
from 40 to 50 nm. Sizes of the silica coated particles did
not change after surface modification with silane.
Magnetic Nanoparticles Conjugation of FITC Labelled
D-Penicillamine-Anti-MTDH
Approximately 1.5 mL of FITC-D-PA-anti-MTDH solution
was obtained. 250 μL of 0.1 M PBS, containing 0.15 M
NaCl, 0.005 M EDTA, was added to 1 mL of FITC-D-PAanti-MTDH solution. After the addition of 2μl (150 mg/mL)
of magnetic nanoparticles, the mixture was kept at room
temperature during 12 hours.
Incorporation Rates of FITC-D-PA-Antibody Conjugated Magnetic Nanoparticles with MCF-7 Cells
MCF-7 breast cancer cell lines were used for this study.
The cells were cultured and seeded into the wells of a 24
well culture plate and 9 cm2 tissue culture petri dishes for
fluorescence microscopy, after enough had been produced. In the study of biological activity detection of FITC
labelled, magnetic nanoparticles conjugated D-PA-antiMTDH in vitro, exactly 105 MCF-7 cells were implanted on
petri dishes. Cells were cultured to confluence at 37 °C
and 5.0% CO2 D-PA-anti-MTDH-FITC, D-PA-anti-MTDHmagnetic nanoparticles-FITC, control solution and magnetic field applying D-PA-anti-MTDH-magnetic nanoparticles-FITC were used in the study.
Medium over the cells was removed and the cells were
washed with PBS for three times. 250 µL of these samples
were put into the wells of a 24 well culture plate and 500
µL of these samples were put into 9 cm2 tissue culture petri
dishes after washing. NdFeB magnets were placed each
well of the plate and magnetic field was applied to each
well of these plates, while other well was not under magnetic field. Optimum incubation time was defined for two
hours in the study. Culture medium was discarded from
the wells at the optimum time (two hours) and washed with
PBS for three times. After incubation time, 24 well culture
plates were fluorometrically assayed in a multiwell fluorescence plate reader (Thermo, Milford, MA) to determine
the %D-PA-anti-MTDH-FITC, D-PA-anti-MTDH-magnetic
Figure 1. SEM Images of Silica Coated Magnetite.
Figure 2. SEM Images of Silanated Magnetite. Nanoparticles
with 50000 X magnification.
Revista Brasileira de Física Médica.2011;5(1):99-104.
101
Özlet A, Perihan Ü, E.İlker M, Çağlar Ö, Serhan S, Suna T
Figure 3. SEM Images of Fe3O4 Magnetite. Nanoparticles with
100000 X magnification.
XRD Analyses Results
XRD analyses of magnetic particles, after surface modification, show the X-ray diffraction pattern of the samples
paired with Fe3O4 diffraction pattern (Figure 4).
VSM Analyses Results
Magnetic properties of magnetic particles were determined with VSM (LakeShore 7407). Magnetization value
versus applied magnetic field for Fe3O4 magnetic particles
was 16.28 emu/g.
AFM Analyses Results
Incorporation Rates of FITC-D-PA-anti-MTDH Conjugated Magnetic Nanoparticles with MCF-7
The obtained D-PA-anti-MTDH-magnetic nanoparticlesFITC complex have been used for in vitro imaging of breast
cancer cells. The cellular binding efficiency of D-PA-antiMTDH-FITC and D-PA-anti-MTDH-magnetic nanoparticles-FITC was calculated using the fluorescence signals as
a result of the targeted-cell interaction.
Figure 4. XRD Images of Fe3O4 Magnetic Nanoparticles.
102
Revista Brasileira de Física Médica.2011;5(1):99-104.
Figure 5. Atomic force microscopy images.
Incorporation on cells of FITC labeled and magnetic
nanoparticles conjugated D-PA-anti-MTDH was found
higher than FITC labeled D-penicillamine. The results
show that magnetic properties and applying magnetic
field increased incorporation rates. On the other hand,
D-penicillamine throat cancer cells (Detroid) using the
same method on cytotoxic effects were shown in another
study23.
D-penicillamine is a good chelating agent for antibody conjugating to nanoparticles. These nanoparticles
may be useful as a new class of agents to target antibody or biomolecules for imaging and targeted therapy
of cancer. Results of this study are sufficiently encouraging to bring about further evaluation of this and related
compounds.
Fluorescein isothiocyanate labeled, magnetic nanoparticles conjugated D-penicillamine-anti-metadherin and in vitro evaluation on breast cancer cells
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Figure 6. Fluorescence microscopy images of D-penicillamine-(Anti-MTDH)-FITC (A), D-penicillamine-(Anti-MTDH)-FITCMagnetic nanoparticles (B) and magnetic field applying D-penicillamine-(Anti-MTDH)-FITC-Magnetite nanoparticles (C), after a
two-hour incubation with MCF-7 cells with 100 magnifications.
References
1. Pope NM, Alsop RC, Chang YA, Smith AK. Evaluation of magnetic
alginate beads as a solid support for positive selection of CD34+
cells. J Biomed Mater Res. 1994;28:449-57.
2. Liu ZL, Ding ZH, Yao KL, Tao J, Du GH, Lu QH, Wang X, Gong FL, Chen X.
Preparation and characterization of polymer-coated core-shell structured
magnetic microbeads. J Magn Magn Mater. 2003;265:98-105.
3. Zhou J, Wu W, Caruntu D, Yu MH, Martin A, Chen JF, et al. Synthesis of
18.
19.
20.
21.
22.
23.
porous magnetic hollow silica nanospheres for nanomedicine application.
J Phys Chem. 2007;111:17473-7.
Pieters BR, Williams RA, Webb C. Magnetic Carrier Technology. Oxford,
England: Butterworth-Heinemann; 1992.
Abo M, Chen Z, Lai LJ, Reese T, Bjelke B. Functional recovery after brain
lesion - contralateral neuromodulation: an FMRI study. Neuroreport.
2001;12:1543-7.
Edelstein RL, Tamanaha CR, Sheehan PE, Miller MM, Baselt DR, Whitman
LJ, et al. The BARC biosensor applied to the detection of biological warfare
agents. Biosensors Bioelect. 2000;14:805-13.
Kim CK, Lim SJ. Recent progress in drug delivery systems for anticancer
agents. Pharm Soc Korea. 2002;25(3):229-39.
Goodwin S, Peterson C, Hoh C, Bittner C. Targeting and retention of
magnetic targeted carriers (MTCs) enhancing intra-arterial chemotherapy.
J Magn Mater. 1999;194:132-9.
Liu JW, Zhang Y, Chen D, Yang T, Chen ZP, Pan SY, et al. Facile synthesis
of high-magnetization-Fe2O3/alginate/silica microspheres for isolation
of plasma DNA Colloids and Surfaces A. Physicochem Eng Aspects.
2009;341:33-9.
Yee Mak S, Hwang Chen D. Binding and sulfonation of poly (acrylic acid)
on Iron oxide nanoparticles: a Novel, Magnetic, Strong Acid cation NanoAdsorbent. Macromol Rapid Commun. 2005;26:1567-71.
Lee SY, Harris MT. Surface modification of magnetic nanoparticles capped
by oleic acids: characterization and colloidal stability in polar solvents. J
Colloid Interface Sci. 2006;293:401-8.
Haddad PS, Martins TM, Souza-Li LD, Li LM, Metze K, Adam RL, et al.
Structural and morphological investigation of magnetic nanoparticles based
on iron oxides for biomedical applications. Mater Sci Eng. 2008;28:489-94.
Vande Stat RJ, Muijsers AO, Henrichs AMA, VanderKorst JK. D-Penicillamine:
biochemical, metabolic and pharmacological aspects. Scand J Rheumatol.
1979;28:13-20.
Schilsky ML. Wilson disease: genetic basis of copper toxicity and natural
history. Semin Liver Dis. 1996;16:83-95.
Acar Ç, Teksöz S, Ünak P, Biber Müftüler FZ. Investigation Of New
Bifunctional Agents: D-Penicillamine. J Radioanalytical Nuclear Chem.
2007;273:641-7.
Horiuchi K, Yokoyama A, Tanaka H, Saji H, Odori T, Morita R, et al.
Technetium coordination state as a factor of stability in 99mTc-complexes
used in hepatobiliary system: comparative studies on 99mTc-complexes of
prididoxial with glutamate (Tc-PG) and isoleucine (Tc-Pl). Eur J Nucl Med.
1981;6:573-9.
Unak P, Tunç M, Duman Y. Labeling of Penicillamine di sulfide with
technetium-99m Appl Rad Isot. 1998;49(7):805-9.
Tröger W, The Isolde Collaboration. Hg(II) Coordination Studies in Penicillamine
Enantiomers by 199mHg-TDPAC. Hyperf Int. 2001;136/137:673-80.
Brown DM, Ruoslahti E. Metadherin, a cell surface protein in breast tumors
that mediates lung metastasis. Cancer Cell. 2004;5(4):365-74.
Sutherland HG, Lam YW, Briers S, Lamond AI, Bickmore WA. 3D3/lyric: a novel
transmembrane protein of the endoplasmic reticulum and nuclear envelope,
which is also present in the nucleolus. Exp Cell Res. 2004;294(1):94-105.
Hafeli UO, Sweeney SM, Beresford BA, Humm JL, Macklis RM. Effective
Targeting of Magnetic Radioactive 90-Y-Microspheres to Tumor Cells by an
Externally Applied Magnetic Field. Preliminary In Vitro and In Vivo Results.
Nucl Med Biol. 1995;22:147-55.
Mohapatra S, Mallick SK, Skghosh T, Pramanik P. Synthesis Of Highly Stable
Folic Acid Conjugated Magnetite Nanoparticles For Targeting Cancer Cells.
Nanotechnology. 2007;18(38):385102.
Gülcüler G, Yıldır A, Hepsöğütlü B. İNEPO 17. Uluslararası Çevre Proje
Olimpiyatı; 2009.
Revista Brasileira de Física Médica.2011;5(1):99-104.
103