Download The Early Detection Research Network: 10-Year

Mini-Reviews
Clinical Chemistry 59:1
60–67 (2013)
The Early Detection Research Network:
10-Year Outlook
Sudhir Srivastava1*
BACKGROUND:
The National Cancer Institute’s Early
Detection Research Network (EDRN) has made significant progress in developing an organized effort for discovering and validating biomarkers, building resources
to support this effort, demonstrating the capabilities of
several genomic and proteomic platforms, identifying
candidate biomarkers, and undertaking multicenter
validation studies. In its first 10 years, the EDRN went
from a groundbreaking concept to an operational
success.
CONTENTS: The EDRN has established clear milestones
for reaching a decision of “go” or “no go” during the
biomarker development process. Milestones are established on the basis of statistical criteria, performance
characteristics of biomarkers, and anticipated clinical
use. More than 300 biomarkers have been stopped
from further development. To date, the EDRN has prioritized more than 300 biomarkers and has completed
more than 10 validation studies. The US Food and
Drug Administration has now cleared 5 biomarkers for
various clinical endpoints.
SUMMARY: The EDRN today combines numerous collaborative and multidisciplinary investigator-initiated
projects with a strong national administrative and data
infrastructure. The EDRN has created a rigorous peerreview system that ensures that preliminary data—analytical, clinical, and quantitative—are of excellent
quality. The process begins with an internal review with
clinical, biostatistical, and analytical expertise. The
project then receives external peer review and, finally,
National Cancer Institute program staff review, resulting in an exceptionally robust and high-quality validation trial.
© 2012 American Association for Clinical Chemistry
The Early Detection Research Network (EDRN)2 is a
pioneering effort by the National Cancer Institute
(NCI) that is designed to discover and validate biomarkers for assessment of cancer and cancer risk. First
launched in 2000, the EDRN provides a vertically integrated network of academic- and industry-based scientists collaborating to meet the challenge of developing
new cancer-screening and early detection products.
The mission of the EDRN is to implement biomarker
research through systematic evidence-based discovery,
development, and validation of biomarkers for identification of cancer risk, early detection, diagnosis, and
prognosis determination to reduce cancer morbidity
and mortality (Tables 1, 2, and 3).
The identification of biomarkers involves a rigorous process that begins with discovery, which leads to
development, validation, and application. EDRN has
fulfilled these expectations by establishing a process for
biomarker development by using a multidisciplinary
and multiinstitutional approach. This infrastructure,
combined with the development of highly interactive
databases and informatics systems, serves as a model
for the conduct of translational research that is fully
aligned with the goals and objectives of the NCI and the
NIH communities (1 ).
The EDRN has implemented pioneering solutions
to enable data sharing between laboratories using common data elements, thus ensuring and expediting consistent data description across institutions. Research
collaborations occur within an environment of teamwork across different disciplines and laboratories focused on achieving the following common goals (2 ):
• Developing and testing promising biomarkers and
technologies to obtain preliminary information to
guide further testing;
• Evaluating promising, analytically proven biomarkers and technologies, such as measures of accuracy,
sensitivity, and specificity and, when possible, poten-
1
Cancer Biomarkers Research Group, Division of Cancer Prevention, National
Cancer Institute, Rockville, MD.
* Address correspondence to this author at: Cancer Biomarkers Research Group,
Division of Cancer Prevention, National Cancer Institute, 6130 Executive Blvd.,
Suite 3142, MSC 7362, Rockville, MD 20852. E-mail [email protected].
Received July 10, 2012; accepted October 25, 2012.
Previously published online at DOI: 10.1373/clinchem.2012.184697
60
2
Nonstandard abbreviations: EDRN, Early Detection Research Network; NCI, National
Cancer Institute; HUPO, Human Proteome Organisation; MS, mass spectrometry;
TSP1, thrombospondin 1; EPCA-2, early prostate cancer antigen-2; proPSA,
prostate-specific antigen precursor; CA125, cancer antigen 125; DCP, des-␥ carboxyprothrombin; AFP, ␣-fetoprotein; HCC, hepatocellular carcinoma; HE4, human epididymis protein 4.
Early Detection Research Network: 10-Year Outlook
Mini-Reviews
Table 1. List of standard and protocols developed by EDRN.a
Assays
Validation of bleomycin-induced chromosomal breakage
in lymphocytes
Biomarker of lung cancer susceptibility
Validation of 3.4-kb mitochondrial DNA deletion
Biomarker for prostate cancer risk
Development of high-density breast and prostate tissue
microarrays
Testing of in situ hybridization and other
molecular probes
Validation of SOPsb for microsatellite instability and
DNA methylation assays
a
b
Applications
Biomarkers for bladder cancer
Validation of saliva-based mRNA assay
Biomarkers of oral cancer
Validation of proteomic prostate-specific biomarkers,
including percent proPSA and other PSA isoforms
Biomarkers for improving PSA screening
of prostate cancer
Urine PCA3 assay
Detection of prostate cancer
Urine/TMA assay for T2S:ERG fusion
Detection of prostate cancer
FISH assay for T2S:ERG fusion
Detection of prostate cancer
Aptamer-based assay
Detection of lung cancer
Proteomic panel for Lung Cancer
Detection of lung cancer
OVA1TM test for ovarian cancer
Differential diagnosis of benign pelvic
mass from ovarian cancer
ROMA for ovarian cancer
Differential diagnosis of benign pelvic
mass from ovarian cancer
Vimentin methylation in stool
Detection of colon cancer
SOPs for blood, sera, plasma, urine, stool
Standard reference sample sets
8-oxyguanine DNA glycosylase (8OGG); alkyl-adenine
DNA gycosylase; APE1 endonuclease
Measuring DNA repair capacity for lung
cancer risk
Retrieved from the EDRN 5th report [NCI (3 )].
SOP, standard operating procedure; PCA, prostate cancer; TMA, transcription-mediated amplification; ERG, v-ets erythroblastosis Virus E26 oncogene homolog;
FISH, fluorescence in-situ hybridization; ROMA, risk of ovarian malignancy algorithm; APE1, apurinic endonuclease 1.
tial predictors of outcomes or surrogate endpoints
for clinical trials;
• Analyzing biomarkers and their expression patterns to
serve as a background for large, definitive validation
studies;
• Conducting early phases of clinical and epidemiological biomarker studies;
• Encouraging collaboration and dissemination of information to ensure progress and avoid fragmentation of effort.
In its first 10 years, the EDRN evolved from a
groundbreaking concept to an operational success.
With the primary mission to discover and scrupulously validate biological markers that signal the earliest stages of cancer (e.g., premalignant lesions, genetic variations, and risk indicators) and reagents for
risk assessment in primary organ systems (e.g., prostate, breast, ovary, lung, colon, pancreas, and liver).
The EDRN combines collaborative, multidisciplinary, investigator-initiated projects with a strong
administrative and data infrastructure (3 ).
1998 –2000: Inception and Inauguration
In 1997, a 20-member Cancer Prevention Program Review Group, seeking a means to revitalize the National
Cancer Prevention and Control Program, recommended the concept of the EDRN. The NCI’s Board of
Scientific Advisors and the National Cancer Advisory
Board approved the concept on November 13, 1998 (4 ).
The EDRN structure emerged (Fig. 1) with working components that comprised laboratories, validation centers, a data management center, and 2 oversight components including a steering committee and a
network consulting team.
2001–2003: Meeting the Scientific Challenges
The network developed systematic, comprehensive
guidelines to develop, evaluate, and validate biomarkers. This 5-phase approach established both a scientific
standard and a roadmap for successfully translating biomarker research from the laboratory to the clinic (5 ).
Clinical Chemistry 59:1 (2013) 61
Mini-Reviews
Table 2. Adoption of EDRN-supported assays.a
Detection/biomarker assay
Clinical translation validation
Blood
FDAb IVD pending review
proPSA
Urine
a
b
PCA3
FDA cleared
Urine/TMA assay for T2S:ERG fusion for PCA
CLIA in process
FISH for T2S:ERG fusion for prostate cancer
In CLIA Lab
Aptamer-based markers for lung cancer
In CLIA Lab
Proteomic panel for lung cancer
In CLIA Lab
OVA1TM for ovarian cancer
FDA cleared
SOPs for blood (serum, plasma), urine, stool
Frequently used by biomarker research community
Vimentin methylation marker for colon cancer
In CLIA Lab
ROMA algorithm for CA125 and HE4 tests for pelvic
mass malignancies
FDA cleared
Blood/DCP and AFP-L3 for HCC
FDA cleared
Blood GP73 for HCC
Together with AFP-L3 used in China for monitoring/
risk assessment of cirrhotic patients for HCC
Retrieved from the EDRN 5th report [NCI (3 )].
FDA, US Food and Drug Administration; IVD, in vitro diagnostic. For other abbreviations see the text and the Table 1 footnote.
Phase 1: Discovery, involving exploratory study, to
identify potentially useful biomarkers.
Phase 2: Validation, in which biomarkers are studied to determine their capacity for distinguishing people with cancer from those without.
Phase 3: Determination of the capacity of a biomarker to detect preclinical disease by testing the
marker against tissues collected longitudinally from research cohorts.
Phase 4: Prospective screening studies on biomarker performance in large populations, and determination of false-referral rates.
Phase 5: The penultimate period, in which largescale population studies are performed to evaluate the
biomarker both for its role in cancer detection and its
overall screening impact.
Although the network’s focus is mainly on phases
1 through 3, researchers have welcomed the 5-phase
structure because it provides an orderly succession of
studies that build upon each other to yield an efficient
and thorough approach to biomarker development.
2003–2005: The Network Surges Ahead in Real Time
In late 2006, the EDRN’s Program for Rapid Independent Diagnostic Evaluation was established as an administrative means to assist extramural investigators in
successfully conducting cross-laboratory validation of
biomarkers (6 ). The network implemented the follow62 Clinical Chemistry 59:1 (2013)
ing guiding principles for biomarker validation and
used criteria for the prioritization of collaborative
projects:
1.
2.
3.
4.
5.
6.
Biologic rationale/strength of hypothesis;
Strength of study design;
Technical parameters;
Clinical or scientific impact;
Practicality;
Collaborative strength/team effort.
To broaden the opportunities for scientific interactions and coordinated research, collaborative groups
were formed. These organ-specific research groups
were structured to promote information exchange on
organ-related biomarkers and to identify research priorities within the EDRN. One major role of the collaborative groups was to serve as advisors and liaisons with
associate members.
The associate membership component was designed for investigators not affiliated with the EDRN
but who wished to join the network by proposing
collaborative studies within its scope and objectives.
More than 200 collaborative projects spanned the
various organ sites. These projects are monitored
through the EDRN’s electronic Study Information
System.
Useful bioinformatics tools were developed to
maximize standardization of information and data
management across multiple activities and research
Early Detection Research Network: 10-Year Outlook
Table 3. Candidate biomarkers tested for early
detection of ovarian cancer.
Marker
a
Diagnostic sensitivity at 95%,
diagnostic specificity (95% CI)
CA125
0.73 (0.64–0.84)
HE4
0.57 (0.50–0.70)
Transthyretin
0.47 (0.38–0.56)
CA15.3
0.46 (0.39–0.54)
CA72.4
0.40 (0.33–0.52)
IGFBP2a
0.38 (0.30–0.45)
Cytokeratin 19
0.37 (0.25–0.45)
IGF2
0.36 (0.28–0.49)
Kallikrein 6
0.36 (0.26–0.43)
Mesothelin
0.35 (0.27–0.46)
B7-H4
0.35 (0.27–0.42)
Apolipoprotein A1
0.34 (0.26–0.43)
sV-CAM
0.34 (0.22–0.42)
Prolactin
0.34 (0.15–0.49)
Chitinase
0.31 (0.23–0.38)
Spondin-2
0.28 (0.18–0.36)
MIC1
0.24 (0.14–0.30)
CA19.9
0.23 (0.18–0.30)
Transferrin
0.23 (0.15–0.32)
IGFBP1
0.21 (0.15–0.31)
Hepcidin
0.21 (0.12–0.28)
TSH
0.19 (0.13–0.25)
CTAP-III
0.19 (0.12–0.24)
MMP3
0.17 (0.11–0.27)
MMP2
0.17 (0.08–0.28)
GH
0.16 (0.10–0.23)
MIF
0.15 (0.09–0.24)
Eotaxin
0.15 (0.07–0.23)
MPO
0.14 (0.09–0.22)
EGFR
0.10 (0.05–0.17)
ITIH4
0.09 (0.05–0.16)
B2M
0.05 (0.02–0.12)
IGFBP2, insulin-like growth factor binding protein-2; IGF2, insulin-like
growth factor-2; sV-CAM, Soluble vascular cell adhesion molecule; MIC1,
macrophage inhibitory cytokine 1; TSH, thyroid stimulating hormone;
CTAP-III, connective tissue activating peptide-III; MMP3, matrix metalloproteinase-3; GH, growth hormone; MIF, macrophage migration inhibitory
factor; MPO, myeloperoxidase; EGFR, epidermal growth factor receptor;
ITIH4, inter-alpha-trypsin inhibitor heavy chain family, member 4; B2M,
␤-2 microglobulin.
sites. These tools include the EDRN Virtual Specimen
Bank, also known as the ERNE Knowledge System,
which was deployed in early 2003 to allow a common
web-based query to search for available samples across
Mini-Reviews
the EDRN clinical validation centers. The Validation
Study Information Management System, another of
these tools, encompasses all the security features of the
auditing systems required by the US Food and Drug
Administration.
A number of public–private partnerships were established since the inauguration of the EDRN. The
most notable partnerships are with the Canary Foundation, the Lustgarten Foundation, and the international Human Proteome Organisation (HUPO). The
premise of the HUPO-sponsored study was an initiative to evaluate multiple proteomic technology platforms, develop bioinformatics tools and standards for
protein identification, and to create a database of the
plasma proteome. The entire study was published in
the August 2005 issue of the Journal of Proteomics (7 ).
2006 –2008: An Investment in Prevention
The EDRN is one of the first organizations to recognize
the importance of standardized, prospective collection
of biologic samples in the context of clinically relevant
circumstances to discover and accelerate the validation
of biomarkers for cancer detection and prognosis.
Samples from cancer cases and controls are collected
prospectively for each major cancer organ site, including cancers of the ovary, breast, lung, colon, liver, prostate, and pancreas (2 ).
The sample sets contain comprehensive clinical,
demographic, and epidemiologic information that
help in the rapid evaluation of technologies and biomarkers in prevalidation before the initiation of prospective, large, and expensive validation trials. Over the
past 10 years, the EDRN has collected a wide range of
biological samples with a primary focus on noninvasively available samples. These samples include serum,
plasma, urine, DNA, and sputum from major epithelial
cancers, e.g., colon, breast, prostate, pancreas, bladder,
and lung. All samples were collected using the NCI Biospecimens First Generation Guidelines. The samples
are clinically annotated with EDRN-developed common data elements and accessible through an online
query using the EDRN Resource Network Exchange
system. Samples collected by the EDRN are provided to
EDRN and non-EDRN principal investigators.
EDRN investigators at the Data Management and
Coordinating Center have developed a study design for
phase 2 and 3 biomarker validation trials. They have
termed this the PRoBE (Prospective Specimen Collection Retrospective Blinded Evaluation) design (5 ). A
key feature of this design is the avoidance of bias by
collecting all samples before diagnosis, so that cases
and controls are enrolled under the same conditions
and all samples are collected and processed identically.
This design can require large numbers of study particiClinical Chemistry 59:1 (2013) 63
Mini-Reviews
Fig. 1. Structural organization of the EDRN.
pants because the incidence of most cancers are low;
most of those study participants enrolled will not have
cancer.
2008 to the Present: Delivery of Clinically Useful
Biomarkers
The EDRN plays very important roles as a “brake” and
“accelerator.” A number of biomarkers undergo rigorous validation tests before they are adopted for larger
validation studies. Biomarkers that are not effective for
the intended clinical objective are not considered further. Each biomarker is tested against the intended goal
and then considered for a validation study. A few examples with the intended clinical objectives are illustrated below.
• Is the biomarker assay reproducible in an independent laboratory? If not, then it is a “no go.”
Example: the EDRN pioneered a systematic evaluation of protein profiles as a candidate biomarker using SELDI-TOF-mass spectrometry (MS) and MALDI
platforms for the diagnosis of prostate cancer. The
study was initiated in 2003 to test published claims that
64 Clinical Chemistry 59:1 (2013)
MS-based protein profiling in ovarian, prostate, lung,
and other cancers could serve as a source of reliable
biomarkers. Network investigators designed a 3-stage
study to validate the reproducibility of the platform
(stage 1), validate the diagnostic use of protein patterns
(stage 2), and conduct a clinical validation (stage 3)
using well-annotated, prospective samples from stratified risk groups and prostate cancer cases and controls.
The stage 1 study confirmed the portability and reproducibility of the SELDI-TOF-MS platforms
among the 8 participating institutes. The stage 2
study could not substantiate previous claims that
protein profiling significantly discriminates between
prostate cancer and noncancer. Thus, the approach
that used protein expression profiling based on
SELDI/MALDI-TOF-MS did not perform well
enough to warrant the launch of the stage 3 prospective study. It was a no go. (8 –9 ).
• Is the biomarker’s performance reproducible when
checked by use of an independent set?
Example: Thrombospondin 1 (TSP1); kallikreins
2, 3, 5, and 11; and early prostate cancer antigen-2
(EPCA-2) were no go for further validation studies af-
Mini-Reviews
Early Detection Research Network: 10-Year Outlook
ter they failed in a blinded test on an EDRN prostate
cancer reference set. Prostate-specific antigen precursor (proPSA), %[⫺2]proPSA, held its performance
[area under the concentration curve (AUC) ⫽ 0.69] on
the same set of samples (10 ). For TSP1 and EPCA-2,
the AUC dropped from above 0.95, observed in preliminary data, to below 0.55 on the reference set.
• Does the biomarker outperform currently used
markers, or add significant value to it? If not, then it is
a no go.
Example: In the phase 3 validation study, 5 predictive models, each containing 6 – 8 biomarkers, were
evaluated according to a predetermined analysis plan.
The diagnostic sensitivity of each predictive model to
distinguish ovarian cancer from controls at 98% specificity was first determined for the entire set of samples.
In this study design, 1 model showed performance
comparable with that of cancer antigen 125 (CA125) in
terms of diagnostic sensitivity and specificity, whereas
the remaining models had poorer performance than
CA125. A second component of this study involved
unblinding half of the samples to perform training with
each model, followed by testing on the remaining samples. In this case again, only 1 model showed performance that was merely comparable with CA125 alone.
Thus, all tested markers failed to show improvement
over CA125 (11–12 ).
• Does the biomarker have a clear potential clinical
use? If not, then it is a no go.
Example: des-␥ carboxyprothrombin (DCP) indicated better diagnostic sensitivity and specificity in preliminary data than did ␣-fetoprotein (AFP), the conventional clinical biomarker for hepatocellular
carcinoma (HCC). Therefore, if the performance is
true, it will replace or add to AFP as a better noninvasive blood test. A phase 2 validation study has just been
completed, which found that that better performance
holds only on later stage HCC but does not hold for
early stage HCC. However, for early stage HCC, in particular HCC with viral etiology, DCP, when combined
with AFP, increases the diagnostic specificity by 50%
(from 24% to 36%) and holds diagnostic sensitivity at
95% (13 ).
Four validation studies have already led to successful submission and clearance of clinically useful biomarkers by the US Food and Drug Administration: (a)
validation of AFP and DCP for differentiating hepatocellular cancer from nonmalignant liver diseases; (b)
validation of human epididymis protein 4 (HE4) in
conjunction with CA125 for measuring the risk of
ovarian cancer in high-risk patients; (c) a panel of 5
proteomics-based markers for the detection of ovarian
cancer; and (d) a noncoding RNA test for measuring
the risk of second biopsy in prostate cancer. These
clearances support EDRN-based validation studies that
are rapidly moving basic science discoveries to the bedside (3 ).
As an example, we tested more than 68 biomarkers
for the detection of ovarian cancer as part of the verification process before launching a large validation
study. EDRN uses a verification process for biomarker
performance by measuring the sensitivity, specificity,
and ROC of each biomarker individually and as part of
a panel. As previously mentioned, the EDRN ovarian
cancer study tested more than 68 biomarkers in samples collected under stringent protocol and operating
procedures. Individual biomarkers were then evaluated for their performance to determine if they warranted a full validation trial to determine the clinical
utility. Except for CA125 and HE4, none of the biomarkers listed in Table 3 performed with adequate sensitivity and specificity for preclinical detection of
ovarian cancer (11 ). The US Food and Drug Administration has now cleared a panel of CA125 and HE4 for
monitoring women for early recurrence of ovarian
cancer.
Future Outlook
Genome-wide global profiling is expected to lead to a
molecular taxonomy of cancer that goes beyond organ
and tissue types. These technologies will supersede and
enhance the classifications based on histopathology or
on the patterns of expression of genes of unknown biological significance uncovered by global transcriptomic profiling. Such an integrated systems approach
may help us to find molecular features of cancer that
are manifested at the earliest stages, leading to effective
target-based screening and early intervention. Strategies must be put to use in the most appropriate manner
for integrating clinical and genomics data with other
omics and imaging data for an effective tailored therapy in the future.
Recent advances in clinical diagnostic imaging and
screening have increased our ability to detect lesions,
but our inability to differentiate aggressive vs indolent
cancers has led to overdiagnosis and overtreatment.
Moreover, overdiagnosis has also led to greatly increased healthcare costs due to the technologies used,
the procedures performed, and the number of patients
examined. To address these problems, the EDRN has
begun discussing the challenges and the scientific opportunities involved in lowering costs, minimizing
overdiagnosis, and improving cancer detection, diagnosis, and staging through the combined use of biomarkers and imaging.
Clinical Chemistry 59:1 (2013) 65
Mini-Reviews
EDRN investigators are proactively analyzing the data
along with their own expression data to triage biomarker information relevant to early detection of cancer. An example is the use of data generated by the
NCI’s The Cancer Genome Atlas to guide the discovery
of ovarian cancer–related secretome genes that can be
further prioritized and subjected to scrutiny for clinical
validation.
including tumor grade, migration, invasion, and apoptosis, may or may not predict the clinical outcomes
under assessment, such as metastasis and death. Unfortunately, current studies often ignore the potential role
of a tumor’s local environment. This is a major limitation. How can researchers design relevant studies or
establish implementable strategies to determine which
cancers will likely lead to an undesirable clinical
outcome?
INTEGRATION OF IMAGING WITH BIOMARKERS
Coping with Challenges
NEXT GENERATION SEQUENCING AND BIOMARKER DISCOVERY
Moving promising molecular diagnostics into the
clinic subsequent to validation is a major undertaking
in terms of expense, hours, and resources required. In
this regard, the merging of ex vivo and in vivo imaging
methodologies with biomarker information offers:
1. Noninvasive near-simultaneous detection of multiple targets and surrogate endpoints in an efficient, robust, and cost-effective manner;
2. Noninvasive assessment of dynamic physiologic
processes in near real time, such as rate constants of
cellular processes, blood flow, oxygenation, pH, temperature, and diffusion;
3. Ability to locate biomarker origins in 3-dimensional
space during a single exam.
It is anticipated that imaging-based clinical validation studies could be used to significantly shorten the
current timeline required for clinical validation and
subsequent regulatory clearance by the US Food and
Drug Administration and recommendations by the
Centers for Medicare and Medicaid Services.
INDOLENT VS AGGRESSIVE TUMORS: OVERDIAGNOSIS
Many detectable lesions and cancers are asymptomatic
and not life threatening. However, the current inability
to discern whether lesions are aggressive cancers that
will lead to clinically significant morbidity and death or
benign, slow-growing cancers that will lead to excessive
testing leads to a phenomenon known as overdiagnosis. Because overdiagnosis leads to unnecessary medical
procedures and treatment and increased risk of adverse
side effects, it is important to develop strategies to
quantify and manage its occurrence. About 25% of
breast cancers detected on mammograms and about
60% of prostate cancers detected with PSA tests could
represent overdiagnosis. A lung cancer screening trial
of chest x-rays and sputum tests led to the estimation
that 50% of the suspicious abnormalities detected represented overdiagnosis.
In future directions the EDRN’s aims to develop
strategies for studying the natural history of cancer to
aid in developing better tools for determining which
cancers are clinically important. The surrogate markers
of cancer aggression often used as evaluative criteria,
66 Clinical Chemistry 59:1 (2013)
The EDRN has admirably overcome many obstacles
and challenges, particularly in its vehement approach
to reducing silo-oriented research structures. In overcoming these obstacles, the EDRN has created a growing army of crusading researchers who promote the
collaboration that has essentially changed the EDRN’s
overarching philosophy and guiding principles.
A consensus of collaboration and credit sharing
within and between organ-based collaborative groups
appears to be emerging. Substantial attention is now
paid to coping with problems that previously were considered mundane (for example, QC of sample collection and management). The EDRN has built a set of
data elements for translational research that includes
demographic, clinical, biosample, research, and clinical data. Development and use of these data elements
have been critical factors in creating group cohesion
and in linking resources from diverse units throughout
the EDRN. Leveraging other government resources,
such as the Jet Propulsion Laboratory, the EDRN has
developed a novel informatics infrastructure that
permits interrogation of diverse databases at long
distances.
The maturing culture of the network has engaged
its participants in a unifying effort to leverage resources, maximize returns on investments, and crossfertilize and synthesize new ideas by exposing previously unrelated concepts to one another. Through
these efforts the levels of some challenges have diminished considerably since the emergence of the EDRN.
The culture of working with industry has presented challenging working relationships. Small companies are attracted to the EDRN because of its rich
resource and expertise. However, there continues to be
a lack of investment from industrial partners in biomarker discovery research, and occasionally there is
resistance to disclosure of biomarkers to be tested using
EDRN reference samples.
Communication among scientists with diverse
backgrounds (e.g., biochemists, epidemiologists, and
statisticians) challenges individual views of scientific
inquiry that narrowly focus on respective disciplines
without efforts on a larger scale to move discovery to
Mini-Reviews
Early Detection Research Network: 10-Year Outlook
translational research. The struggle to enable effective
communication and mutual respect is a major challenge for any vertically integrated unit such as the
EDRN.
The EDRN has learned from these problems,
grown scientifically and culturally, and transmitted lessons learned to other parts of the NCI and the broader
scientific community. Obstacles should not be interpreted as weakness of the concept or the model, but
rather should be interpreted as a process of evaluation and
change as the commitment of the stakeholders and the
scientific community alike continues to grow collectively.
Conclusions
The EDRN has made significant progress in developing
an organized effort for biomarker discovery and validation, building resources to support this effort, demonstrating the capabilities of several genomic and proteomic platforms, identifying candidate biomarkers,
and undertaking multicenter validation studies.
In its first 10 years, the EDRN has gone from a
groundbreaking concept to an operational success. To
date, the EDRN has developed a rich pipeline of more
than 300 biomarkers that have been subjected to rigorous
phase 2 validation and are ready for large validation studies. The EDRN has generated more than 1450 publications, more than 28 patents, and more than 14 licenses (2).
Without the EDRN, research into new biomarkers
of early cancer detection and risk would have remained
on the periphery of research, with a strong but fragmented laboratory presence and little translational interest in the academic scientific community. However,
with the network, a new translational paradigm has
defined organizations, approaches, and standards by
which biomarkers are developed and assessed. The network has created substantial focus, energy, and new
research in the field of early detection.
Author Contributions: All authors confirmed they have contributed to
the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design,
acquisition of data, or analysis and interpretation of data; (b) drafting
or revising the article for intellectual content; and (c) final approval of
the published article.
Authors’ Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form.
Disclosures and/or potential conflicts of interest:
Employment or Leadership: S. Srivastava, EDRN.
Consultant or Advisory Role: None declared.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: None declared.
Expert Testimony: S. Srivastava, EDRN.
Patents: None declared.
References
1. Wagner PD, Srivastava S. New paradigms in
translational science research in cancer biomarkers. Transl Res 2012;159:343–53.
2. Crichton DJ, Mattman CA, Tornquist IM, Anton K,
Hughes JS. Bioinformatics: biomarkers of early
detection. Cancer Biomark 2010;9:511–30.
3. NCI. The Early Detection Research Network. November 2011, 5th report, NIH publication number
10-7696. http://edrn.nci.nih.gov/docs (Accessed
March 2012).
4. NCI. EDRN initial report: translational research to identify early cancer and cancer risk,
October 2000, publication number 01-4852.
http://edrn.nci.nih.gov/docs (Accessed March
2012).
5. Pepe MS, Feng Z, Janes H, Bossuyt, PM, Potter
JD. Pivotal evaluation of the accuracy of a biomarker used for classification or prediction: standards for study design. J Natl Cancer Inst 2008;
100:1432– 8.
6. NIH. Announcement of the NCI Early Detection
Research Network (EDRN) Program for Rapid,
Independent Diagnostic Evaluation (PRIDE) for
cancer biomarkers. http://grants.nih.gov/grants/
guide/notice-files/NOT-CA-07-003.html (Accessed
December 2012).
7. Omenn GS, States DJ, Adamski M, Blackwell
TW, Menon R, Hermjakob H, et al. Overview of
the HUPO Plasma Proteome Project: results
from the pilot phase with 35 collaborating
laboratories and multiple analytical groups,
generating a core dataset of 3020 proteins and
a publicly-available database. Proteomics
2005;5:3226 – 45.
8. McLerran D, Grizzle WE, Feng Z, Bigbee WL,
Banez LL, Cazares LH, et al. Analytical validation
of serum proteomic profiling for diagnosis of
prostate cancer: sources of sample bias. Clin
Chem 2008;54:44 –52.
9. McLerran D, Grizzle WE, Feng Z, Thompson IM,
Bigbee WL, Cazares LH, et al. SELDI-TOF MS
whole serum proteomic profiling with IMAC surface does not reliably detect prostate cancer. Clin
Chem 2008;54:53– 60.
10. Sokoll LJ, Sanda M, Feng Z, Kagan J, Mizrahi IA,
Broyles DL, et al. A prospective, multicenter, NCI
EDRN study of [⫺2]proPSA: Improving prostate
cancer detection and correlating with cancer aggressiveness. Cancer Epidemiol Biomarkers Prev
2010;19:1193–200.
11. Daniel CW, Bast RC Jr, Berg CD, Diamandis EP,
Godwin AK, Hartge P, et al. Ovarian cancer biomarker performance in prostate, lung, colorectal,
and ovarian cancer screening trial specimens.
Cancer Prev Res 2011;4:365–74.
12. Zhu CS, Pinsky PF, Cramer DW, Ransohoff DF,
Hartge P, Pfeiffer RM, et al. A framework for
evaluating biomarkers for early detection: validation of biomarker panels for ovarian cancer. Cancer Prev Res 2011;4:375– 82.
13. Marrero JA, Feng Z, Wang Y, Nguyen MH, Befeler
AS, Roberts LR. Alpha-fetoprotein, des-gamma
carboxyprothrombin, and lectin-bound alphafetoprotein in early hepatocellular carcinoma.
Gastroenterology 2009;137:110 – 8.
Clinical Chemistry 59:1 (2013) 67