Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Personal View Theragnostic imaging for radiation oncology: dose-painting by numbers Lancet Oncol 2005; 6: 112–17 University of Wisconsin Medical School, Department of Human Oncology, K4/316 Clinical Sciences Center, WI, USA (Prof S M Bentzen DSc) Correspondence to: Prof Søren Bentzen, University of Wisconsin Medical School, Department of Human Oncology, K4/316 Clinical Sciences Center, 600 Highland Avenue, Madison, WI 53792, USA [email protected] See page 66 for a Reflection and Reaction article on treatmentplanning approaches in radiation oncology Søren M Bentzen Theragnostic imaging for radiation oncology is the use of molecular and functional imaging to prescribe the distribution of radiation in four dimensions—the three dimensions of space plus time—of radiotherapy alone or combined with other treatment modalities in an individual patient. Several new imaging targets for positron-emission tomography, single-photon-emission CT, and magnetic resonance spectroscopy allow variations in microenvironmental or cellular phenotypes that modulate the effect of radiation to be mapped in three dimensions. Dose-painting by numbers is a strategy by which the dose distribution delivered by inverse planned intensity-modulated radiotherapy is prescribed in four dimensions. This approach will revolutionise the way that radiotherapy is prescribed and planned and, at least in theory, will improve the therapeutic outcome in terms of local tumour control and side-effects to unaffected tissue. Diagnostic imaging, from radiographs to CT, has been used throughout the history of radiotherapy to define the site to which the radiation should be delivered for patients with cancer. Theragnostic imaging is a method by which the radiation dose can be delivered in the four dimensions of space and time to achieve the optimum outcome after radiotherapy. What does theragnostic imaging mean? From the Greek words therapeia (to treat medically) and gnosis Beam 1 5 10 20 25 30 35 40 (knowledge) theragnostic imaging refers to the use of information from medical images to determine how to treat individual patients. Rapid technological and scientific progress in molecular and functional imaging, in radiotherapy planning and delivery, and in clinical radiation biology is making radiotherapy guided by theragnostic imaging a realistic goal. Molecular imaging is providing new research opportunities in the preclinical and clinical development of new therapies and in the study of small-animal models of human disease.1,2 To take full advantage of the curative potential of modern, high-precision radiotherapy, high-quality imaging is needed for identification of target volumes. Conversely, optimum use of the spatial information provided by molecular imaging requires radiotherapy; radiation is a therapeutic agent that can be modulated in the four dimensions of space and time, and the dose can be precisely defined to produce a specified local effect of a given magnitude. This review explores the latter perspective. 45 Intensity-modulated radiotherapy (IMRT) 50 55 60 Courtesy of C Grau, Aarhus University Hospital, Aarhus, Denmark 65 70 Beam 5 75 Beam 2 80 85 Beam 4 Beam 3 50 mm Figure 1: IMRT plan for a patient with oropharyngeal cancer The overlay-colour scale indicates the physically absorbed dose (Gy) of radiation from the whole treatment course. Five non-uniform treatment fields are used. Note the concavity in the high-dose volume, sparing the dose to the spinal cord (located at the intersection point of the five beam axes). This feature allows the risk of radiation myelopathy to be kept very low (<0·1%). Note also the sparing of the left parotid gland, lowering the risk of clinical xerostomia after radiotherapy. 112 One of the key features underpinning this research is the theoretical and technological development of IMRT and inverse treatment planning (figure 1).3 IMRT is the administration of non-uniform intensities of radiation (or photon fluence profiles) to patients as a way to create a specified, non-uniform absorbed dose distribution. This approach has become feasible as a result of the increased computer control of linear accelerators used for radiotherapy during the past few decades, in combination with the introduction and refinement of the multileaf collimator—a computer-controlled device, typically consisting of 20 or more thin collimator plates. The multileaf collimator is used to create irregularly shaped fields that conform to the target volume and modulate the radiation-beam intensity. Traditionally, radiotherapy dose plans have been forward planned—ie, a small number of radiation-beam portals are selected on the basis of the location of the target volume and any healthy structures that need to be avoided. Each beam is assigned a weight that determines http://oncology.thelancet.com Vol 6 February 2005 For personal use. Only reproduce with permission from Elsevier Ltd Personal View its relative contribution to the total absorbed dose. A computer program, commonly operating on a set of CT images covering the anatomical region irradiated, is then used to calculate the dose distribution, and the irradiated dose from each beam is then normalised to achieve the prescribed dose in the target volume. Inverse treatment planning, by contrast, takes a prescribed dose distribution as the starting point and uses a mathematical optimisation algorithm to identify a set of beam intensities that approximates the prescribed dose distribution as closely as possible. Typically, the dose distribution is specified as a desired homogeneous dose in the clinical target volume plus several dose-volume constraints for the dose delivered to crucial healthy structures or organs at risk in the surrounding volume of healthy tissue. IMRT and inverse treatment planning have provided new methods to deliver non-uniform or shaped dose distributions that are almost impossible to deliver with conventional static radiation beams. Exploration of the therapeutic potential of non-uniform dose distributions is a challenge for preclinical and clinical radiation research. Theragnostic imaging for radiation oncology Theragnostic imaging for radiation oncology aims to map in three dimensions the distribution of a tumour, tissue, or functional feature, and to provide information about the clinical response of tumours or healthy tissues to radiotherapy. In solid tumours, the aim is to provide images of phenotypic or microenvironmental characteristics known to affect the clinical response. Most research has been based on imaging techniques that use radionuclide-labelled compounds: single-photonemission CT (SPECT) or positron-emission tomography (PET). However, dynamic contrast-enhanced magnetic resonance (MR) imaging and spectroscopy also have much potential. Research has also focused on radiobiological mechanisms that affect—or are suspected to affect—the outcome after radiotherapy and for which an appropriate imaging surrogate exists. Tumour burden and clonogen density At the most fundamental level, molecular imaging has the potential to define the real target volume—the volume consisting of malignant cells that need irradiation to a therapeutic dose to control the disease.4 One example is the use of proton MR spectroscopy to discriminate between benign prostate hyperplasia and malignant tissue,5 and the application of this information in the planning of conformal transperineal implantation of therapeutic iodine-125 seeds in the prostate.6 PET scanning with 18-fluorodeoxyglucose is commonly used as a proxy for tumour burden and has been intensively investigated for staging of various solid cancers. However, [18F]FDG is a hexokinase substrate indicative of glucose metabolism and is an indirect http://oncology.thelancet.com Vol 6 February 2005 probe that reflects the activity of target enzymes rather than their concentration.2 Bos and colleagues7 investigated the correlation between [18F]FDG activity (assessed visually on an arbitrary four-point scale by three observers) and various histopathological and immunohistochemical markers in breast tumours.7 They concluded that [18F]FDG uptake is a function of: the microvasculature for delivery of nutrients; SLC2A1 (GLUT1) for transport of [18F]FDG into cells; hexokinase for entry of [18F]FDG into glycolysis; and the number of tumour cells in the volume, the proliferation rate, number of lymphocytes, and hypoxia-inducible factor 1 (HIF1) for upregulation of SLC2A1. Tumour hypoxia Low partial oxygen pressure in human tumours as measured with polarographic microelectrodes is associated with poor outcome in patients receiving radiotherapy.8 Although tumour hypoxia is associated with malignant progression and a poor prognosis after other treatment modalities as well as after radiotherapy,9 the idea that hypoxic regions in solid tumours constitute a resistance problem for radiotherapy seems reasonable. One strategy to resolve this problem is to boost the radiation dose to hypoxic regions of the tumour, provided that these regions can be imaged in three dimensions. Several radionuclide-labelled compounds have shown preferential uptake or long-lasting retention in hypoxic tumour regions in clinical and preclinical studies. One class of these agents is compounds that contain a 2-nitroimidazole group (eg, fluoride-18misonidazole, iodide-123-iodoazomycin arabinoside, and D-125I-iodoazomycin galactopyranoside. Another class of compounds includes copper-62-labelled diacetylbis(N(4)-methylthiosemicarba-zone) and technetium99m-labelled 4,9-diaza-3,3,10,10-tetramethyldodecan2,11-dione dioxime. The mechanism of action of these agents is not completely understood, but they seem to be bioreducible compounds retained in regions with hypoxic cells with intact mitochondria. One difficulty with all of these agents is that the signal-to-background ratio is poor. The search for better markers continues.10 Dynamic contrast-enhanced MR imaging visualises differences between tissues in the behaviour of a gadolinium-based contrast medium.11 One example of this technique is BOLD (blood-oxygen-level dependent) imaging, which is being investigated as an indirect method for mapping of hypoxic regions in tumours.12 Work in progress has shown a significant correlation between pimonidazole immunostaining of coregistered histological sections and the value of R2*, a relaxation rate that is sensitive to deoxyhaemoglobin concentration, derived from BOLD. Hypoxia causes activation of several cellular response pathways,9 and one of the most important of these is regulated by HIF1. Collaborative work by Harris’s group at Oxford University and our group at the Gray 113 For personal use. Only reproduce with permission from Elsevier Ltd Personal View Cancer Institute, suggests that the outcome after definitive radiotherapy in patients with squamous-cell carcinoma of the head and neck depends on the exact hypoxic pathways activated in a tumour. If this finding is correct, hypoxic radioresistance is much more than just the absence of oxygen. Again, the requirements for molecular imaging would be changed. Serganova and colleagues’ study13 on imaging of HIF1 transcriptional activity in tumours in mice has also shown that it is feasible to map the activation of a specific pathway by use of PET. Tumour proliferation Several randomised controlled trials14–16 have provided evidence on the importance of overall treatment time, especially in squamous-cell carcinoma of the head and neck, with evidence from other tumour sites being less strong. This feature has been linked to rapid tumour-cell proliferation during radiotherapy as a resistance mechanism in fractionated radiotherapy.16 Thus, there is a rationale for theragnostic imaging of tumour-cell proliferation, and research on this topic has focused on radionuclide-labelled nucleosides or aminoacids as possible imaging agents.17 Several studies18 have looked at radiolabelled deoxyuridines, but the signal-to-noise ratio is poor, owing to the rapid degradation of these compounds in vivo. This problem was largely solved by the development and clinical testing of 18F-3-deoxy-3fluorothymidine (FLT). Two clinical studies19,20 found encouraging correlations between local FLT-specific uptake values and the Ki-67 labelling index detected in biopsy samples by immunohistochemistry. Ki-67 is a protein expressed in all phases of the cell cycle but not in resting cells; it is therefore used widely to measure cell proliferation. Despite these encouraging results, a study in animals21 showed limited incorporation of 18F-FLT into DNA relative to background and further validation studies seem warranted. New imaging targets Many of the new imaging targets under investigation2 are of immediate interest for theragnostic imaging for radiation oncology. Examples are cyclin D (which is overexpressed in many cases of squamous-cell carcinoma of the head and neck, for example, for assessment of tumour-cell proliferation), and mitogen-activated protein kinase for measurement of activation of the RASsignalling pathway. Epidermal growth factor receptor (EGFR) is expressed in many tumours of epithelial origin and is implicated in several processes associated with the malignant phenotype. Clinical studies22,23 have shown a relation between EGFR expression and the outcome of definitive radiotherapy in cancer of the head and neck. Our work has suggested that EGFR expression is directly linked to the time factor in radiotherapy, making this an interesting target for theragnostic imaging.24–26 Regions with high EGFR expression have a large capacity for 114 accelerated repopulation, and these regions would probably benefit from accelerated radiotherapy (ie, delivered in a shorter overall time). With the current advances in molecular risk profiling and the search for fingerprints of malignant phenotypes that are sensitive to a specific type of modified radiotherapy,27 many new theragnostic imaging targets are likely to be identified in the coming years. Functional imaging of crucial healthy tissues Optimisation of the therapeutic ratio of IMRT must take into account the dose distribution in healthy tissue as well as tumours. The dose distribution to healthy tissues can be determined through placement of simple constraints on the clinically acceptable maximum dose to an organ at risk or by specification of a single variable of the dose-volume histogram, such as the fraction of the total organ volume receiving a dose that exceeds a threshold value. However, just as with tumours, the full four-dimensional dose distribution should ideally be constrained. One interesting idea is to lower the risk of side-effects by reducing the dose absorbed in regions of an organ that have special functional importance. These regions could be identified by use of functional imaging. Seppenwoolde and colleagues28 incorporated lung-perfusion information from SPECT directly into the radiotherapy-plan optimisation for patients with medically inoperable non-small-cell lung cancer. The researchers showed that this procedure increased the relative weights of radiation beams passing through hypoperfused lung regions and reduced the predicted risk of side-effects to healthy tissue compared with conventionally optimised treatment plans. From dose-painting to dose-painting by numbers Dose-painting was the term coined by Ling and colleagues29 in their review of image-guided radiotherapy. The idea was to visualise tumour subvolumes with a potential resistance problem and to paint some additional dose onto that volume. This notion was applied in a study by Chao and co-workers,30 who identified regions with pronounced retention of 62 Cu(II)-diacetyl-bis(N(4)-methylthiosemicarba-zone) on a PET scan of a patient with a squamous-cell carcinoma of the head and neck. In a planning study, they showed that 80 Gy could be delivered in 35 fractions to the hypoxic target volume, with 70 Gy in 35 fractions delivered to the rest of the clinical target volume. By use of standard bioeffect modelling31 to adjust for the slightly higher dose per fraction in the hypoxic target volume, the biological equivalent dose to this volume is about 82 Gy in 2 Gy per fraction. The researchers used dosevolume histograms to show that this dose distribution could be planned and delivered by means of inverseplanned IMRT. http://oncology.thelancet.com Vol 6 February 2005 For personal use. Only reproduce with permission from Elsevier Ltd Personal View Dosimetric features of IMRT delivery Although a review of the advantages and drawbacks of IMRT compared with conventional radiotherapy is beyond the scope of this report, a brief summary of some of these from the viewpoint of dose-painting by numbers is relevant. Three-dimensional conformal radiotherapy (3D-CRT) is a broad class of radiotherapy techniques that aim to improve the match between the clinical target volume and the volume irradiated to a high radiation dose. This aim is achieved without intensity modulation by use of a limited number of static beams. 3D-CRT plans typically reduce the volume of healthy tissue that receives a high dose, while increasing the dose to the clinical target volume. The two most frequently aired arguments against IMRT are that this method of planning and delivery makes quality assurance very difficult and that there might be an increased risk of radiation-induced secondary cancers. Certainly, quality assurance is much more demanding with IMRT techniques than with 3DCRT or with conventional radiotherapy.34 However, more and more reports have been published from clinical departments reporting their experience with the introduction of clinical and dosimetric quality-assurance programmes for IMRT.35,36 The quality assurance needed does not differ substantially between a dose plan prescribed on the basis of theragnostic imaging and that prescribed on the basis of a limited number of specified point dose constraints. The possible increase in radiation-induced second cancer in long-term survivors after IMRT remains controversial. In a review, Hall and Wuu37 pointed out that the change from conventional radiotherapy techniques to 3D-CRT is expected to reduce the incidence of radiation-induced cancers as a result of the relative decrease in the healthy-tissue volume receiving high doses of radiation. However, a change from 3D-CRT to IMRT typically involves an increase in the number of fields used. Furthermore, because a large proportion of the radiation field is shielded by the collimator during irradiation, the time for which the accelerator beam is switched on will increase by a factor of two or three. This feature will increase the total body Prostate tumour T2* relaxivity 120 Frequency 100 80 60 40 20 0 10 13 16 19 22 25 28 31 34 37 40 R2* (s–1) Normo-oxic 70 Anoxic 75 80 85 85 Dose in 35 factions (Gy) Courtesy of N J Taylor, D Carnell, and A Padhani, Mount Vernon Hospital, UK Although much research is needed before dosepainting can be introduced into clinical practice, the issue is whether this step is far enough. The problem with discrete volumes is that they are binary: voxels are either inside or outside the volume. However, the clinical and biological reality is that there are gradients of oxygen tension or three-dimensional distributions of the density of radioresistant cellular phenotypes. This discrepancy led me to propose a more radical change in the way theragnostic imaging can be used in prescribing four-dimensional dose distributions—dose-painting by numbers. Named after the painting-by-numbers activities in children’s activity books, the principle is illustrated in figure 2 in a patient with localised prostate cancer, a tumour type characterised by regions with substantial hypoxia.32 The clinical target volume, typically defined on a CT or MR scan, is transferred to the BOLD MR scan, and an intensity histogram is calculated for the marker of hypoxia. These intensities will correspond to voxel oxygenation status covering the full range from normal to anoxic. The non-linear scale indicates the prescribed dose as a function of marker intensity. A lower boundary of 70 Gy ensures that no part of the tumour will have a prescribed dose lower than the standard prescription. An upper boundary of 85 Gy is defined to limit the dose to any structure in the patient, and this value is repeated in figure 2 to indicate that it is a ceiling on the prescribed dose. Then, all of the prescribed doses on a pixel-by-pixel basis are used to specify the desired dose distribution in inversetreatment-planning software. The technical feasibility of this procedure was shown by Alber and colleagues,33 who modified their inverse-planning software to implement dose-painting by numbers on the basis of 18 F-misonidazole imaging data. Figure 2: BOLD MR image of a patient with prostate cancer The region of interest was defined from haematoxylin-eosin staining of the prostate sections and transferred to the MR image. The distribution of R2* values in the region of interest covers the range of oxygenation status from normal to anoxic. Dose-painting by numbers involves the prescription of a specific, physically absorbed dose of radiation on a voxel-by-voxel basis according to a defined prescription function as indicated by the non-linear scale below the histogram. http://oncology.thelancet.com Vol 6 February 2005 115 For personal use. Only reproduce with permission from Elsevier Ltd Personal View exposure owing to leakage radiation. Hall and Wuu estimate that IMRT will increase the frequency of second cancers in patients surviving for 10 years to 1·75% compared with 1·00% after conventional radiotherapy. 10-year survival in most solid cancers is below 50%, so this higher rate corresponds to an increase of less than 0·5% in all cases treated. Clearly, this effect will be outweighed by any clinically relevant gain in control of the primary tumour resulting from IMRT. Hall and Wuu underlined the uncertainties in these estimates, but they must be regarded as current best values. The issue of radiation-induced second cancers will be most important in children and other patients with long life expectancy. Again, there is no major difference between standard IMRT and IMRT prescribed by dose-painting by numbers. Are we ready for dose-painting by numbers? Validation of the imaging target is the first objective in building the case for a new theragnostic imaging procedure. This aim generally involves two steps. The first is to show that the imaging variable correlates with a local biological property, such as a local microelectrode measurement of oxygen tension for hypoxic imaging or a mitotic index or a Ki-67 labelling index for imaging of cell proliferation. The second step is to show the clinical importance of the validation marker for the radiobiological characteristic in question; for example, that the Ki-67 labelling index actually selects for a benefit from accelerated radiotherapy. This step might at first seem superfluous, but it is needed to justify the prescription of a specific temporal modulation of radiotherapy based on the imaging information. Even with a validated imaging target, temporal stability is a concern. The usefulness of theragnostic imaging in prescribing four-dimensional dose distributions obviously depends on the short-term and long-term stability of the three-dimensional map of density of specific cellular phenotypes or microenvironmental variables. Oxygenation is one example: intermittent closing and opening of vessels can cause microscopic changes in oxygenation, so-called acute hypoxia, on a typical time-scale of minutes.38,39 The extent to which this process affects hypoxia as estimated with radionuclidelabelled compounds is not clear. In addition, there is reoxygenation of hypoxic regions after the radiation cell killing a few hours after each dose fraction and throughout the full 6–7 weeks of fractionated radiotherapy. Even if reoxygenation does occur, boosting of the radiation dose to the region that was hypoxic at the start of therapy can still be worthwhile. Studies of the temporal stability of hypoxia maps are in progress in several centres. Spatial resolution is poorer with PET, SPECT, and MR spectroscopy than with MR imaging and CT and there can be difficulties with partial volume artefacts. At present, commercial high-resolution clinical PET 116 Search strategy and selection criteria Data were identified by searches of PubMed using the search terms “radiotherapy or radiation oncology”, “functional imaging”, “molecular imaging”, “positron emission tomography”, “magnetic resonance imaging”, and “magnetic resonance spectroscopy”. scanners have a full width at half-maximum resolution of 3·5–5·0 mm in both the axial and transverse planes. The development of CT-PET and MR-PET scanners might help in reducing the partial volume artefacts and could improve the accuracy of image coregistration. By contrast, the achievable resolution with current PET scanners should be seen on the scale of achievable spatial precision in delivering 35 dose fractions in the clinic. This requirement defines an effective limit on the steepness of gradients that are deliverable in the clinic, which, with current radiotherapy technology, is probably not much better than the above 3–4 mm. Clearly, dosepainting by numbers will need optimum precision of the planning dosimetry and the multileaf-collimator technology as well as immobilisation of the patient and reproducibility. Physiological organ movement is another possible limitation that is being addressed by several research groups. The final, hugely challenging, topic of research is to establish the prescription function that is the mathematical link between a specific value of an imaging variable and the optimum clinical dose to be prescribed to the corresponding voxel. This function is unlikely to be defined from radiobiological measurements made in vitro but will have to be derived from outcome data in human beings or in animals. Ideally, evidence-based prescription rules should be derived from controlled clinical trials. Simple linear interpolations between a clinically justified minimum and maximum dose over the range of values covered by the imaging variable might suffice as a first approximation, but strategies are under investigation that could further refine this prescription function. In conclusion, theragnostic imaging for radiation oncology will revolutionise the whole process of radiotherapy prescription and planning. At the present rate of progress, a realistic prophecy is that this therapeutic principle will be in early clinical testing within the next 5 years. Conflict of interest I declare no conflicts of interest. References 1 Weissleder R. Molecular imaging: exploring the next frontier. Radiology 1999; 212: 609–14. 2 Herschman HR. Molecular imaging: looking at problems, seeing solutions. Science 2003; 302: 605–08. 3 Intensity Modulated Radiation Therapy Collaborative Working Group. Intensity-modulated radiotherapy: current status and issues of interest. Int J Radiat Oncol Biol Phys 2001; 51: 880–914. http://oncology.thelancet.com Vol 6 February 2005 For personal use. Only reproduce with permission from Elsevier Ltd Personal View 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Bentzen SM. High-tech in radiation oncology: should there be a ceiling? Int J Radiat Oncol Biol Phys 2004; 58: 320–30. Garcia-Segura JM, Sanchez-Chapado M, Ibarburen C, et al. In vivo proton magnetic resonance spectroscopy of diseased prostate: spectroscopic features of malignant versus benign pathology. Magn Reson Imaging 1999; 17: 755–65. Zelefsky MJ, Cohen G, Zakian KL, et al. Intraoperative conformal optimization for transperineal prostate implantation using magnetic resonance spectroscopic imaging. Cancer J 2000; 6: 249–55. Bos R, Der Hoeven JJ, van Der WE, et al. Biologic correlates of (18)fluorodeoxyglucose uptake in human breast cancer measured by positron emission tomography. J Clin Oncol 2002; 20: 379–87. Nordsmark M, Overgaard J. A confirmatory prognostic study on oxygenation status and loco-regional control in advanced head and neck squamous cell carcinoma treated by radiation therapy. Radiother Oncol 2000; 57: 39–43. Harris AL. Hypoxia: a key regulatory factor in tumour growth. Nat Rev Cancer 2002; 2: 38–47. Chapman JD, Bradley JD, Eary JF, et al. Molecular (functional) imaging for radiotherapy applications: an RTOG symposium. Int J Radiat Oncol Biol Phys 2003; 55: 294–301. Padhani AR, Husband JE. Dynamic contrast-enhanced MRI studies in oncology with an emphasis on quantification, validation and human studies. Clin Radiol 2001; 56: 607–20. Taylor NJ, Baddeley H, Goodchild KA, et al. BOLD MRI of human tumor oxygenation during carbogen breathing. J Magn Reson Imaging 2001; 14: 156–63. Serganova I, Doubrovin M, Vider J, et al. Molecular imaging of temporal dynamics and spatial heterogeneity of hypoxia-inducible factor-1 signal transduction activity in tumors in living mice. Cancer Res 2004; 64: 6101–08. Bernier J, Bentzen SM. Altered fractionation and combined radiochemotherapy approaches: pioneering new opportunities in head and neck oncology. Eur J Cancer 2003; 39: 560–71. Overgaard J, Hansen HS, Specht L, et al. Five compared with six fractions per week of conventional radiotherapy of squamous-cell carcinoma of head and neck: DAHANCA 6 and 7 randomised controlled trial. Lancet 2003; 362: 933–40. Bentzen SM. Repopulation in radiation oncology: perspectives of clinical research. Int J Radiat Biol 2003; 79: 581–85. Van de Wiele C, Lahorte C, Oyen W, et al. Nuclear medicine imaging to predict response to radiotherapy: a review. Int J Radiat Oncol Biol Phys 2003; 55: 5–15. Shields AF, Grierson JR, Dohmen BM, et al. Imaging proliferation in vivo with [F-18]FLT and positron emission tomography. Nat Med 1998; 4: 1334–36. Buck AK, Schirrmeister H, Hetzel M, et al. 3-deoxy-3-[(18)F]fluorothymidine-positron emission tomography for noninvasive assessment of proliferation in pulmonary nodules. Cancer Res 2002; 62: 3331–34. Vesselle H, Grierson J, Muzi M, et al. In vivo validation of 3deoxy3-[(18)F]-fluorothymidine ([(18)F]FLT) as a proliferation imaging tracer in humans: correlation of [(18)F]FLT uptake by positron emission tomography with Ki-67 immunohistochemistry and flow cytometry in human lung tumors. Clin Cancer Res 2002; 8: 3315–23. Lu L, Samuelsson L, Bergstrom M, et al. Rat studies comparing 11C-FMAU, 18F-FLT, and 76Br-BFU as proliferation markers. J Nucl Med 2002; 43: 1688–98. http://oncology.thelancet.com Vol 6 February 2005 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 Ang KK, Berkey BA, Tu X, et al. Impact of epidermal growth factor receptor expression on survival and pattern of relapse in patients with advanced head and neck carcinoma. Cancer Res 2002; 62: 7350–56. Eriksen JG, Steiniche T, Askaa J, et al. The prognostic value of epidermal growth factor receptor is related to tumor differentiation and the overall treatment time of radiotherapy in squamous cell carcinomas of the head and neck. Int J Radiat Oncol Biol Phys 2004; 58: 561–66. Ramos-Suzarte M, Rodriguez N, Oliva JP, et al. 99mTc-labeled antihuman epidermal growth factor receptor antibody in patients with tumors of epithelial origin, part III: clinical trials safety and diagnostic efficacy. J Nucl Med 1999; 40: 768–75. Mishani E, Abourbeh G, Rozen Y, et al. Novel carbon-11 labeled 4-dimethylamino-but-2-enoic acid [4-(phenylamino)-quinazoline-6yl]-amides: potential PET bioprobes for molecular imaging of EGFR-positive tumors. Nucl Med Biol 2004; 31: 469–76. Divgi CR, Welt S, Kris M, et al. Phase I and imaging trial of indium 111-labeled anti-epidermal growth factor receptor monoclonal antibody 225 in patients with squamous cell lung carcinoma. J Natl Cancer Inst 1991; 83: 97–104. Buffa FM, Bentzen SM, Daley FM, et al. Molecular marker profiles predict locoregional control of head and neck squamous cell carcinoma in a randomized trial of continuous hyperfractionated accelerated radiotherapy. Clin Cancer Res 2004; 10: 3745–54. Seppenwoolde Y, Engelsman M, De Jaeger K, et al. Optimizing radiation treatment plans for lung cancer using lung perfusion information. Radiother Oncol 2002; 63: 165–77. Ling CC, Humm J, Larson S, et al. Towards multidimensional radiotherapy (MD-CRT): biological imaging and biological conformality. Int J Radiat Oncol Biol Phys 2000; 47: 551–60. Chao KS, Bosch WR, Mutic S, et al. A novel approach to overcome hypoxic tumor resistance: Cu-ATSM-guided intensity-modulated radiation therapy. Int J Radiat Oncol Biol Phys 2001; 49: 1171–82. Bentzen SM, Baumann M. The linear-quadratic model in clinical practice. In: Steel GG, ed. Basic clinical radiobiology, 3rd edn. London: Arnold, 2002: 134–46. Parker C, Milosevic M, Toi A, et al. Polarographic electrode study of tumor oxygenation in clinically localized prostate cancer. Int J Radiat Oncol Biol Phys 2004; 58: 750–57. Alber M, Paulsen F, Eschmann SM, Machulla HJ. On biologically conformal boost dose optimization. Phys Med Biol 2003; 48: N31–35. Low DA. Quality assurance of intensity-modulated radiotherapy. Semin Radiat Oncol 2002; 12: 219–28. McNair HA, Adams EJ, Clark CH, et al. Implementation of IMRT in the radiotherapy department. Br J Radiol 2003; 76: 850–56. Boehmer D, Bohsung J, Eichwurzel I, et al. Clinical and physical quality assurance for intensity modulated radiotherapy of prostate cancer. Radiother Oncol 2004; 71: 319–25. Hall EJ, Wuu CS. Radiation-induced second cancers: the impact of 3D-CRT and IMRT. Int J Radiat Oncol Biol Phys 2003; 56: 83–88. Chaplin DJ, Durand RE, Olive PL. Acute hypoxia in tumors: implications for modifiers of radiation effects. Int J Radiat Oncol Biol Phys 1986; 12: 1279–82. Dewhirst MW, Kimura H, Rehmus SW, et al. Microvascular studies on the origins of perfusion-limited hypoxia. Br J Cancer Suppl 1996; 27: S247–51. 117 For personal use. Only reproduce with permission from Elsevier Ltd