Utilizing Existing Clinical and Population Biospecimen Resources for Discovery or Validation of Markers for Early Cancer Detection


Despite the large investment in biomarker discovery and validation, very few major cancer biomarkers have been approved for clinical use over the last 25 years. In addition, many initially promising biomarkers have not been validated. One major problem is that certain study designs result in specimens that are not ideal for biomarker discovery or validation. For example, if there is a systematic difference between cases and controls or in handling of their specimens, a biased comparison occurs.

To accurately discover and/or validate cancer biomarkers, access to resources where serial collections, epidemiological data, clinical data, and outcomes were assessed over time is required. Specimens collected just-proximal to the diagnosis of cancer are ideal because they are typically not biased by knowledge of the diagnosis or treatment of the disease. These include serial blood samples (including plasma or serum) collected in a systematic and equivalent manner, with appropriate epidemiological and clinical annotation; such samples are ideal for discovery and/or validation of biomarkers for early detection.

There are several NCI- and NIH-sponsored cohort studies and clinical trials with high-quality, well-annotated specimens that have been utilized to assess biomarker validity. Important characteristics of these cohorts and trials needed to perform such assessments are the availability of serial blood draws and the collection of outcomes data (e.g., development of cancer) over time. However, much larger numbers and disease-free populations are needed. Consequently, it may be useful to aggressively leverage existing clinical and epidemiology cohort resources since new collections are so expensive and time-consuming.

Return to Top


The aim of the meeting was to learn whether ongoing NCI-funded projects, e.g. cohort studies, health maintenance organizations (HMOs), or clinical trials, might be leveraged for unbiased studies of biomarker study and validation. Cancer researchers from these diverse studies met to discuss the challenges and opportunities for using existing infrastructures. The workshop had three main goals:

  1. Facilitate sharing of methods, approaches, and lessons learned in performing this type of research using existing clinical and epidemiology infrastructures, mainly for studies of cancer early detection.
  2. Identify obstacles or challenges that need to be addressed to perform this type of research.
  3. Develop best practices and contrasting alternative approaches to be documented in a white paper.

Return to Top


View agenda for Wednesday, August 28
Wednesday, August 28 Topic
12:00 p.m. - 1:00 p.m. Registration
1:00 p.m. - 1:10 p.m. Welcome
Muin Khoury
Epidemiology and Genomics Research Program (EGRP), Division of Cancer Control and Population Sciences (DCCPS), NCI
1:10 p.m. - 1:30 p.m. Charge to Participants
Barry Kramer
Division of Cancer Prevention (DCP), NCI
Session 1: Leveraging Cohort Studies and RCTs for Discovery and Validation of Biomarkers for Diagnosis; Part 1 – Theory and Considerations Regarding Design
Moderator: Muin Khoury, EGRP, DCCPS, NCI
1:30 p.m. - 2:00 p.m. Description and Background of the Problem
David F. Ransohoff
University of North Carolina at Chapel Hill
2:00 p.m. - 3:00 p.m. Panel and Audience Discussion
Panel 1 questions:
  • What are strong study designs for the discovery and validation of diagnostic and early detection biomarkers?
  • What has been your experience in utilizing cohorts and/or RCTs for discovery and validation of biomarkers and how does this differ from the ideal?
  • Is the use of existing resources (ongoing RCTs, research cohorts, or practice settings like HMOs) practical for such a study?
  • What evidence is needed before a biomarker should be tested in a cohort study or RCT?

Christine Berg
Independent Research Consultant

Jennifer Cullen
Center for Prostate Disease Research, Department of Defense

Sue Hankinson
University of Massachusetts, Amherst

3:00 p.m. - 3:15 p.m. Break
Session 2: Considerations Regarding Data Quality in Study Design
Moderator: Lisa McShane, Biometric Research Branch (BRB), Division of Cancer Treatment and Diagnosis (DCTD)
3:15 p.m. - 3:45 p.m. Lessons Learned from Biomarker Studies of Ovarian Cancer
Steven J. Skates
Dana-Farber/Harvard Cancer Center
3:45 p.m. - 4:45 p.m. Panel and Audience Discussion
Panel 2 questions:
  • What lessons can be learned from the ovarian cancer marker story? So many groups found "promising" markers initially, but they were not "validated" using specimens from PLCO. What might we prescribe for the field regarding cultivating existing resources for discovery or early validation of biomarkers for diagnosis, if the field were "doing it again," to make the process more efficient and reliable?
  • How might we better leverage existing resources to make the process more efficient and reliable?
  • What should be the study design requirements for the future?

Nicole Urban
Fred Hutchinson Cancer Center

Eleftherios Diamandis
Mt. Sinai Hospital, Toronto

John Baron
Dartmouth Medical School

4:45 p.m. - 5:00 p.m. Wrap up of Day 1
David F. Ransohoff
University of North Carolina at Chapel Hill

View agenda for Thursday, August 29
Thursday, August 29 Topic
7:30 a.m. - 8:00 a.m. Registration
Session 3: Leveraging Cohort Studies and RCTs for Discovery and Validation of Biomarkers for Diagnosis; Part 2 - Practical Strategies and Tactics
Moderator: Steven Skates, Dana-Farber/Harvard Cancer Center
8:00 a.m. - 8:30 a.m. Practical Strategies in Cohorts and RCTs
Garnett Anderson
Fred Hutchinson Cancer Center
8:30 a.m. - 9:30 a.m. Panel and Audience Discussion
Panel 3 questions:
  • What has been your experience in utilizing existing cohorts and/or RCTs for discovery and validation of biomarkers?
  • How do you learn the "true state" (cancer vs. not) of subjects, if it is not a dedicated research study that follows people systematically to outcome?
  • What are the logistical barriers that inhibit performing an ideal study of biomarkers?
  • How can we foster sharing a.m.ong research studies?

William Rom
NYU Langone Medical Center

Steve Shak
Genomic Health, Inc.

Rich Schilsky
American Society of Clinical Oncology (ASCO)

9:30 a.m. - 9:45 a.m. Break
Session 4: Leveraging HMOs and Other Healthcare Delivery Infrastructures for Discovery and Validation of Biomarkers for Diagnosis
Moderator: Sheri Schully, EGRP, DCCPS, NCI
9:45 a.m. - 10:15 a.m. Biomarker Experience from the Cancer Research Network
Lawrence Kushi
Kaiser Permanente Division of Research
10:15 a.m. - 11:15 p.m. Panel and Audience Discussion
Panel 4 questions:
  • How do you get members to participate in research?
  • If it is not a "dedicated" study how do you assess whether people are asymptomatic and that the test was not done because of cancer symptoms?
  • How do you monitor outcomes in participants (i.e., surveys vs. EMRs)?
  • How do you get people to agree to contribute samples and data for research purposes in this setting?
  • How do you minimize problems related to loss of follow-up and what strategies can be implemented to reduce loss of follow-up?
  • What are the challenges for researchers obtaining access to data?

Katrina Goddard
Kaiser Permanente Center for Health Research

Eric Larson
Group Health Research Institute

Paul Doria-Rose
Applied Research Program (ARP), DCCPS, NCI

11:15 a.m. - 12:15 p.m. Lunch
What Lessons Have We Learned and Where Do We Go From Here?
Moderator: David F. Ransohoff, University of North Carolina at Chapel Hill
12:15 p.m. - 2:00 p.m. Audience Discussion
Each participant will be asked to summarize 1 or 2 lessons learned and recommendations for the future.
2:00 p.m. Meeting Adjourns

Return to Top

Workshop Summary

Schully SD, Carrick DM, Mechanic LE, Srivastava S, Anderson GL, Baron JA, Berg CD, Cullen J, Diamandis EP, Doria-Rose VP, Goddard KAB, Hankinson SE, Kushi LH, Larson EB, McShane LM, Schilsky RL, Shak S, Skates SJ, Urban N, Kramer BS, Khoury MJ, Ransohoff DF. Leveraging biospecimen resources for discovery or validation of markers for early cancer detectionExternal Web Site Policy. JNCI. 2015; 107(4) [Epub ahead of print].

Return to Top

Planning Committee

Return to Top

Related Resources

  • Biospecimen Resources for Population Scientists
    Contains links to biospecimen resources that may be of interest to cancer epidemiologists, policies and best practices for biospecimen research, and EGRP contacts for questions about biospecimens related to cancer epidemiology research.
  • NCI Cohort Consortium
    The NCI Cohort Consortium includes investigators from over 40 cohorts involving 4 million people. This extramural-intramural partnership was formed to address the need for large-scale collaborations to conduct a wide range of cancer studies.
  • HMO Cancer Research Network
    The Cancer Research Network (CRN) is an NCI-funded initiative to support and facilitate cancer research based in non-profit integrated health care delivery settings. The CRN welcomes collaborations that result in research projects that improve knowledge about cancer etiology, prevention, early detection, treatment, and prognosis, and that decrease the burden of cancer across the cancer care spectrum.
  • Kaiser Permanente Division of ResearchExternal Web Site Policy
    The Kaiser Permanente Northern California Division of Research has a nearly 50-year history of successful collaboration with other research organizations and acts as part of the HMO Research Network. It uses its collective scientific capabilities to integrate research and practice for the improvement of health and health care among diverse populations.

Return to Top

Key Publications

Return to Top


Return to Top