Skip to Main Content
An official website of the United States government
Epidemiology and Genomics Research Program

2016 Annual Meeting

The annual NCI Cohort Consortium meeting, sponsored by EGRP and the Division of Cancer Epidemiology and Genetics (DCEG), was held on November 1-2, 2016, at the Natcher Conference Center on the NIH campus in Bethesda, MD. Project/Working Group meetings were also held during this time.


Day 1: Tuesday, November 1, 2016

Refer to Working Group Schedule
Time Event
1:00 p.m. - 1:40 p.m.

Session I: Opening Plenary Session

Dr. Susan Gapstur, American Cancer Society, Inc.

1:00 p.m. – 1:10 p.m.

Welcome and Introductions
Dr. Kathy Helzlsouer, National Cancer Institute
Dr. Susan Gapstur, American Cancer Society, Inc.

1:10 p.m. – 1:20 p.m.

DCEG Cohort Activities
Dr. Montserrat Garcia-Closas, National Cancer Institute

1:20 p.m. – 1:30 p.m. Open Discussion
1:40 p.m. – 2:30 p.m.

Session II: Metabolomics Updates

Dr. Wei Zheng, Vanderbilt University
Dr. Mattias Johansson, International Agency for Research on Cancer

1:40 p.m. – 1:55 p.m.

An Atlas Describing Integrated Genetic, Epigenetic, and Metabolic Human Individuality
Dr. Christian Gieger, Munich University, Germany

1:55 p.m. – 2:10 p.m.

Metabolomics: Applications to Cancer Epidemiology
Dr. Steve Moore, National Cancer Institute

2:10 p.m. – 2:30 p.m. Open Discussion
2:30 p.m. – 3:30 p.m

Session III: Poster Session
Presenters are asked to be by their posters during this time.

3:30 p.m. – 3:45 p.m. Break
3:45 p.m. – 4:45 p.m.

Session IV: Micro RNAs

Dr. Giske Ursin, Cancer Registry of Norway
Dr. Celine Vachon, Mayo Clinic College of Medicine

3:45 p.m. – 4:05 p.m.

Prospects and Challenges of Using Archived Serum for miRNA Biomarker Discovery
Dr. Trine Rounge, Cancer Registry of Norway

4:05 p.m. - 4:25 p.m.

Using Circulating Nanovesicles to Detect, Monitor, and Treat Cancer
Dr. Philip Bernard, University of Utah School of Medicine

4:25 p.m. – 4:45 p.m. Open Discussion
4:45 p.m. – 5:45 p.m.

Session V: Precision Prevention

Dr. Anthony Swerdlow, University of London Institute of Cancer Research
Dr. Deborah Winn, National Cancer Institute

4:45 p.m. – 5:05 p.m.

Incremental Precision in Primary and Secondary Prevention
Dr. Robert Hoover, National Cancer Institute

5:05 p.m. – 5:25 p.m.

Precision Prevention: Let's Avoid Exacerbating Cancer Disparities
Dr. Graham Colditz, Washington University School of Medicine

5:25 p.m. – 5:45 p.m. Open Discussion
5:45 p.m. Adjourn

Day 2: Wednesday, November 2, 2016

Time Event
8:30 a.m. - 12:00 p.m. Cohort Consortium Main Meeting
8:30 a.m. – 9:30 a.m.

Session VI: Use of E-Technology in Cohort Studies

Dr. Leslie Bernstein, City of Hope
Dr. Kathy Helzlsouer, National Cancer Institute

8:30 a.m. – 8:50 a.m.

Building Research Technology Platforms for Participants with a Consumer Engagement Experience in Mind
Dr. James McClain, National Cancer Institute

8:50 a.m. – 9:10 a.m.

Non-Invasive Personal Wristband Samplers Applications
Dr. Kim Anderson, Oregon State University

9:10 a.m. - 9:30 a.m. Open Discussion
9:30 a.m. - 10:00 a.m. Break
10:00 a.m. – 11:40 a.m.

Session VII: Cohort Consortium Subcommittee Updates

Dr. Meir Stampfer, Harvard School of Public Health

Poster Awards
Dr. Anthony Swerdlow, University of London Institute of Cancer Research

10:00 a.m. – 10:30 a.m.

MTA Subcommittee Update
Dr. Joanne Elena, National Cancer Institute

NIH Data Sharing Policy Update
Dr. Elizabeth Gillanders, National Cancer Institute

10:30 a.m. – 10:50 a.m. Open Discussion
10:50 a.m. – 11:20 a.m.

Consortium Working Groups Subcommittee Update
Dr. Celine Vachon, Mayo Clinic College of Medicine

Working Group Survey Results
Ms. Camille Pottinger, National Cancer Institute

11:20 a.m. – 11:40 a.m. Open Discussion
11:40 a.m. – 12:00 p.m.

Wrap Up
Dr. Susan Gapstur, American Cancer Society, Inc.

12:00 p.m. – 1:00 p.m. Lunch
1:00 p.m. – 5:00 p.m.

Cohort Consortium Working Group Meetings
(See working group schedule below)

5:00 p.m. Adjourn

Cohort Consortium Project / Working Group Meeting Agenda

Tuesday, November 1, 2016

  F1/F2 A G1/G2 E1/E2 C1/C2
9:00 a.m. Physical Activity Working Group
(Steven Moore)
    Vitamin D Pooling Project of Breast and Colorectal Cancer
(Closed to WG members only; Stephanie S-Warner)
Second Cancers / Survivorship Working Group
(Joanne Elena)
9:30 a.m.   NCI Meet the Expert Session
(Joanne Elena)
10:00 a.m.      
10:30 a.m.        
11:00 a.m. Markers of HPV Infection and Risk of Head and Neck Cancer
(Aimee Kreimer)
Steering Committee Working Lunch
(11:30 a.m. - 12:45 p.m.)
Diabetes and Cancer in the Cohort Consortium (Marc Gunter) Biomarkers and Breast Cancer Risk Prediction in Young Women
(Closed to WG members only; Anne ZJacquotte)
11:30 a.m.  
12:00 p.m. Lunch

Wednesday, November 2, 2016

  G1/G2 B A E1/E2 H
12:00 p.m.     African American Working Group
(Closed to WG members only; Julie Palmer)
12:30 p.m.        
1:00 p.m. Premenopausal Breast Cancer Collaboration Group
(Closed to WG members only; Hazel Nichols)
    Diet and Cancer Pooling Project
(Stephanie S-Warner)
Prostate Tissue Biomarkers WG
(Lorelei Mucci, Eric Jacob)
1:30 p.m.    
2:00 p.m.      
2:30 p.m.   PanScan
(Rachel S-Solomon)
Ovarian Cancer Cohort Consortium
(Shelley Tworoger)
  Lymphoid Malignancies
(Brenda Birmann)
3:00 p.m.    
3:30 p.m.       General Breast Cancer Working Group
(Mia Gaudet)
4:00 p.m.        

Meeting Summary

The 2016 Annual Meeting of the NCI Cohort Consortium was held on the NIH campus in Bethesda, MD on November 1-2. This summary reflects the portions of the meeting involving all participants. The other portions of the meeting were dedicated to multiple, simultaneous working group meetings.

Session I: Introduction

Moderator: Susan Gapstur

Dr. Kathy Helzlsouer thanked Drs. Leslie Bernstein and Wei Zhang for their service and welcomed the incoming chair, Dr. Anthony Swerdlow. In 2017, the Cohort Consortium will continue to improve its strategic planning and engage in discussions about how the committee can advance, including a review of the bylaws to determine whether they remain effective.

Dr. Susan Gapstur thanked the Steering Committee, noting that three subcommittees had been formed to promote more active discussion about the annual meeting, the working groups, and material transfer applications (MTAs). Several new cohorts have been approved by the Steering Committee, including the Polish Cohort Study and the Generation Scotland: Scottish Family Health Study. New working groups for 2016 also are studying such topics as coffee drinking and mortality, inflammation and breast cancer risk, and oral cancer risk prediction models. A new procedure for proposals has helped to streamline and advance papers and projects.

Six working groups presented on Steering Committee calls this year; these presentations are used to gather feedback, inform future planning, and act on data sharing and MTA issues. Once the data from the cohorts have been harmonized, the next question is to determine how to keep high-impact research moving forward.

As of November 2016, 59 cohorts and 43 working groups exist. Of the active working groups, 23 are open to new members; Dr. Gapstur noted that many would display posters at this meeting, and she invited attendees to engage with the presenters. Several groups have accomplished their goals, including the Breast and Prostate Cancer Cohort Consortium (BPC3) and the Tumor Tissue Working Group, which helped to launch efforts now in use in other groups. Dr. Gapstur noted that ideas for next year's meeting and opinions about this year's format change would be collected in surveys after this meeting.

Dr. Robert Croyle welcomed the attendees on behalf of NCI Acting Director Dr. Douglas Lowy. He noted that the Cohort Consortium continues to be one of the NIH's most significant efforts in terms of teamwork and international collaboration. He noted that the Precision Medicine Initiative (PMI) has two aspects, Cohort and Oncology; the latter was launched only recently. The next step for PMI Oncology is to turn attention to the Vice President's Blue Ribbon Panel, which has several recommendations already underway and is encouraging new collaborations with governmental branches that may have a role in cancer studies that has not been fully leveraged.

The starting point for Blue Ribbon Panel recommendations is treatment- and development-focused. However, plans in development cannot be actuated until the NIH budget is approved. Dr. Croyle noted that he has heard support for an increase in the NIH budget from Republicans and Democrats. All plans in progress have been made scalable in case the budget is approved at a lower or higher level than anticipated. Dr. Croyle's team also has been working toward the transition of administrations by creating briefing books and assisting the transition teams.

In recent years, a greater degree of attention has been given to health disparities. The PMI Cohort has been renamed "All of Us," and the first wave of initiatives has been funded. A query will be sent to attendees for ideas of other areas that could be included. Dr. Croyle thanked all attendees and emphasized that "international collaboration is not the problem—it's the solution."

DCEG Cohort Activities
Monserrat Garcia-Closas

Dr. Monserrat Garcia-Closas updated the attendees on two new initiatives. The Prostate, Lung, Colorectal and Ovarian Cancer (PLCO) Screening Trial is scanning the entire PLCO population to facilitate research on clinical outcomes. The participants in the intervention arm of the cohort receive regular screenings for these cancers; those in the control arm continue usual care. There are 113,000 subjects from both groups who provided consent for DNA scanning; 33,000 already have been scanned, and genome-wide association studies (GWAS) have been done. Researchers are planning to scan the rest of the cohort using the Infinium Global Array, which allows scanning at a low cost. The genotyping data will be made available to the scientific community through the Database of Genotypes and Phenotypes, and information on phenotypes and outcomes already is accessible through the Cancer Data Access System. The genome-wide scan enables evaluation of the effects of genetic risk stratification, validation of risk prediction models, study of cancer susceptibility alleles, and discovery of genetic determinants of risk factors and biomarkers. DNA extractions will commence before the end of the year, and the genotyping is estimated to occur on a 15 month timeline.

The other new initiative is a prospective effort on cancer research through the Division of Cancer Epidemiology and Genetics (DCEG) Cohort. Key design features include repeated exposure assessments, collection of biospecimens, and use of electronic health records (EHRs). Building effective infrastructure will be a key component of this effort, as will the collection of data that have elements in common with existing cohorts to enable future data sharing. The cohort will operate within integrated healthcare systems, which provide whole care of patients; plans in areas such as infrastructure and specimen collection will build on these collaborative efforts, but specific systems have not yet been determined. The project is still in the design phase, but Dr. Garcia-Closas estimated that the startup and pilot phases each would require 1 year to complete; the project should be ready for full-scale operation by 2020.

The baseline biospecimens would be blood, urine, and buccal specimens; other potential biospecimens and collection protocols will be determined and finalized once the participants have been confirmed. The PMI Cohort questionnaire will be made public when it clears institutional review board (IRB) review, as it is intended to be an open resource with open activities.

Session II: Metabolomics

Moderators: Wei Zhang and Mattias Johansson

An Atlas Describing Integrated Genetic, Epigenetic, and Metabolic Human Individuality
Christian Gieger

Dr. Christian Gieger was involved in a study titled Collaborative Health Research in the Region of Augsburg, known by its German acronym, KORA. KORA was a cohort study involving 18,000 participants whose biospecimen data were banked for later analysis. The goal of the study was to integrate omics data, particularly metabolomic, genomic, and epigenomic data. Because the metabolome represents all metabolites in a biological cell or tissue, the metabolome describes the end product of an organism's genetic makeup and all exposure influences. The study design included the KORA F4 and TwinsUK cohorts. Researchers used semiquantitative, nontargeted metabolic profiling of serum samples to detect 529 metabolites, both known and unknown, covering 60 biological pathways. They found significant genomic associations at 145 metabolic loci.

The N-acetyltransferase 8 (NAT8) locus was one such association. N-acetylation is an important detoxification mechanism in the kidney, and the researchers found a highly significant association of variation at the NAT8 locus with the metabolite N-acetylornithine. Based on this association, they investigated whether N-acetylornithine concentrations were linked to kidney function and discovered a clear association with estimated glomerular filtration rate. Such associations point to pathways that warrant further exploration to better understand disease etiology.

Another association was found between the SLC16A9 locus and the metabolite carnitine. The hypothesis that SLC16A9 might code for a carnitine transporter was validated experimentally. Metabolomics potentially could be used to determine the function of orphan genes and the effects of genetic variants. Dr. Gieger shared the network of gene-gene and gene-metabolite associations that was constructed from KORA F4 and TwinsUK data, creating an atlas of genetic influences on human blood metabolites.

Similar to investigating links between metabolomics and genomics, the researchers examined the influence of epigenetic changes on the metabolome. One epigenetic change is DNA methylation, which occurs at CpG sites, DNA regions where cytosine and guanine nucleotides are separated by a single phosphate. The analytical method used created quantitative, genome-wide DNA methylation profiles that covered approximately 450,000 CpG sites. The researchers conducted an epigenome-wide association study with metabolite concentrations, then corrected for genetic variance near or in CpG sites that might be a confounding factor behind some of the associations. They found few highly significant CpG-metabolite associations after removing confounding genetic effects and concluded that causality might be difficult to establish.

Dr. Gieger summarized the conclusions that could be drawn from the compilation of an atlas of genetic, epigenetic, and metabolic variation among individuals. Such an atlas, together with the genetic and epigenetic associations it reveals and the gene functions of those associations, can pave the way for improved understanding of complex diseases, and metabolomics reveals new pathways for drug targets for such diseases. He predicted that comparisons across omics in cohorts will help to better phenotype patients and identify novel risk markers and pathways.

Metabolomics: Applications to Cancer Epidemiology
Steve Moore

Dr. Steve Moore presented his studies on the relationship of body mass index (BMI) to postmenopausal breast cancer risk. Using the Metabolon, Inc. platform, samples from postmenopausal women in the PLCO trial were processed to reveal the abundance of 617 metabolites. BMI was associated with 67 metabolites that met the study's threshold of significance, and six metabolites were associated with breast cancer risk, all of which were higher in women with higher BMIs. Twenty-three metabolites were associated with estrogen receptor positive (ER+) breast cancer; none were associated with estrogen receptor negative (ER-) breast cancer, although the small sample size could have affected that result.

The researchers used a forward modeling process to select 67 BMI-related metabolites by determining the top hit and incorporating it in the model, then repeating the modeling for the remaining 66 metabolites until a threshold for false discovery was established. Four metabolites were associated with ER+ breast cancer risk; the top hit was a precursor of estriol, an estrogen metabolite, but the other three were novel. When the sum of the metabolite levels, the metabolite index, was plotted, the odds ratio was higher than that seen in most metabolite studies. Overweight and obese subjects showed no increase in risk unless they also had a high metabolite index value, which suggests the potential ability to distinguish between low-risk and high-risk obesity. These results are what would be expected if the index was mediating associations. Each metabolite in the index was therefore characterized as a marker, with three of the marker metabolites related to low enzyme activity in the degradation of amino acids. This could be evidence that proliferating tumors need amino acids that can be scavenged, but much work remains.

Dr. Moore pointed out that he did not mention insulin or estradiol; these were not studied because the researchers had not known what their findings would be prior to beginning the study. He also noted that results obtained using particular platforms can change over time and that larger sample sizes and replication studies are needed.


An attendee asked Dr. Gieger if it would be more correct to refer to GWAS as "metabolome-wide association studies." He concurred, adding that his team already has begun to look at other epigenetic techniques. Another participant asked Dr. Gieger about the maturation of sequences in his laboratory; he responded that his team currently does not sequence but uses more of a screening approach.

A participant asked about the reliability of genetic markers for cancer. Dr. Moore has looked broadly at this and estimated that an intraclass correlation coefficient (ICC) would demonstrate a reasonable degree of reliability for biomarkers. In response to a question about the platform used, Dr. Moore explained that they have compared the Metabolon platform to that of the Broad Institute, which is more customized. The typical overlap is 120 metabolites, with an ICC of 0.8, but there are many caveats to consider.

A participant asked about factors that affect reproducibility. Dr. Moore explained that fasting has a large effect on 5 percent of metabolites and a smaller effect on another 10 to 15 percent, so even relatively minor differences in sample collection can affect the results. An attendee commented on a more recent comparison of the Metabolon, Inc. platform to the Broad Institute platform that showed 400 overlapping metabolites and a high ICC, noting that as peaks are identified, old datasets can be reviewed to extract new information.

A participant asked about the justification for studying BMI; Dr. Moore replied that it was his area of interest and the subject of his dissertation. He also noted that studying one marker results in a set of tractable associations, commenting that metabolomics is "the Wild West" and focusing on one marker can help scientists better assess the landscape. Dr. Mattias Johansson asked about the level of confidence in the causal involvement of these markers. Dr. Moore replied that, though he is optimistic about causality, it is always challenging to confirm; picking a specific exposure such as BMI has an advantage in developing a causal story. Dr. Moore and Dr. Gieger emphasized the need for further studies and improved randomization approaches.

Session III: Poster Session

The attendees were provided time to view the posters, including those of the active working groups, and interact with the presenters.

Session IV: Micro RNAs

Moderator: Giske Ursin

Prospects and Challenges of Using Archived Serum for miRNA Biomarker Discovery
Trine Rounge

Dr. Trine Rounge explained that microRNA (miRNA) is the best-characterized small noncoding RNA (sncRNA). In cancer, miRNAs often can target more than one messenger RNA (mRNA), and multiple miRNAs can target the same transcript; this complex system is well-known to be associated with hallmarks of cancer.

sncRNA in archived serum has a long follow-up time, large cohorts are available, and its use is cost- and time-efficient, but tests using the Janus Cohort showed that it has low stability and is subject to degradation during freeze-thaw cycles and errors in the sampling process. However, miRNA in serum can be stable for more than 10 years and is unaffected by freeze-thaw cycles. When measuring sncRNA in serum, hemolysis testing can help researchers rule out confounders.

The methodology has long been established for analyzing archived samples of sncRNA. The challenge is to ensure that the sequencing represents the underlying abundance of input RNA molecules. Dr. Rounge's team performed tests using manual phase separation followed by automatic extractions.

Researchers also must decide which molecules to analyze; a wider sample of sncRNA will have a smaller fraction of miRNA. To address this, Dr. Rounge's team used a process of repeated sequencing to attain the desired amount of miRNA, achieving a promising success rate with 40-year-old samples. Dr. Rounge noted that many promising tools exist, but the lack of good databases and standard normalizations across datasets are challenges that remain.

Dr. Rounge reported that her team recently completed its first study of serum sncRNAs as early detection biomarkers of cancer, in which a large group of prediagnostic samples was analyzed. Samples from the Janus Cohort were collected up to 10 years before diagnosis; these data were supplemented with data gathered from cancer registries, health surveys, and screening registries. Some patients had serial samples, which can show the change in miRNA expression over time. The data shows the expected peak for miRNAs, but there also are varied peaks indicating the presence of other sncRNAs that have not been characterized. Dr. Rounge pointed out promising studies are underway using archived serum samples for miRNAs as biomarkers for lung cancer in smokers.

Using Circulating Nanovesicles to Detect, Monitor, and Treat Cancer
Philip Bernard

Dr. Philip Bernard provided an overview of the rapidly evolving fields of noncoding RNAs and extracellular vesicles. Many miRNAs are expressed in normal tissue and few in tumor tissue, which is the opposite of the desired expression of a good biomarker. Exosomes, which are released by most cells, including tumor cells, are part of the endosomal sorting process, generated from the multivesicular endosomal system; microvesicles, by contrast, originate in the plasma membrane. Dr. Bernard described the discrepancy seen between the size of exosomes on scans and the hydrodynamic size, but this is caused by proteins used for docking and messaging; after treatment with a proteinase, the exosomes are smaller. Dr. Bernard showed types of endosomal proteins and methods of exosome isolation, noting that many publishers require confirmation of endosomal origin with a Western blot.

Exosomes can be isolated with nanotracking analysis, and high variability is shown across individuals in the concentration of exosomal vesicles. As an example, mi-R21 is highly expressed in cancers and seen early in breast cancers. Its expression in exosomes after surgery is variable, so researchers must determine when to take a blood sample to show that it is free of disease. Another application of exosomal vesicles to cancer can be seen in the way tumors progress predictably to certain other sites in the body; this can be explained by the proteins attached to the exosomes from the tumors.

Translating these studies to clinical practice is limited by the lack of consistent method, consensus on the best biofluid, or agreed-on controls for normalization. Dr. Bernard issued a call for methods, asking for demonstration of a collection of intact exosomes, high reproducibility and yield of oncosomes, and sufficient yield of nucleic acids. He noted that healthy subjects have less circulating RNA, so some of the questions he requested may be difficult to compare to healthy subjects.


Dr. Rounge explained that it is too early to determine whether miRNAs will be good biomarkers, but they can be good additions to established methods and personalized risk assessments. In response to a question about the comparison of circulating miRNA to exosomes, Dr. Rounge explained that archived serum samples cannot address details of this because of the bias of processing and storage. An attendee asked about the differences between samples in plasma and serum; Dr. Rounge responded that each cohort has its own personal challenges and recommended a thorough pilot study on plasma samples. Dr. Bernard added that remaining consistent is the most important consideration; his study used serum because the banks they had were of serum samples and the biomarkers they wanted were in serum.

Dr. Giske Ursin asked about ways to better understand the normal variation that can complicate studies. Dr. Bernard recommended better characterization and larger cohorts, noting that variation strongly affects niRNA because of the small scale. Dr. Rounge added that studying large sets of healthy controls can help researchers learn about normal variation. A participant asked whether there is tension between needing large cohorts and the concept of precision medicine, and Dr. Bernard acknowledged that this is contradictory, noting that the studies need to be done regardless, but the resolution may depend on how personalized science truly is able to become. In response to a question about when diagnosis researchers could expect to find miRNA biomarkers, Dr. Rounge explained that answering this question is one of the goals of her study; they may discover that the 10-year limit currently set on their samples is not enough. Dr. Rounge and Dr. Bernard agreed that the main goal is to identify the biomarker before the cancer becomes physically detectable.

Session V: Precision Prevention

Moderators: Anthony Swerdlow and Deborah Winn

Incremental Precision in Primary and Secondary Prevention
Robert Hoover

Dr. Robert Hoover described opportunities in cancer prevention, including identification of high-risk groups for interventions and the recent speed of discovery that has led to opportunities for more targeted prevention development. Precision prevention is the use of data to reduce the incidence of cancer in specific groups of individuals; it is a relatively new field, but some critics already have suggested that this effort has failed. Dr. Hoover offered three examples demonstrating the usefulness of precision prevention.

Cervical cancer was the subject of one of the earliest attempts at prevention with the invention of the Pap smear in the 1950s. In the early 1980s, scientists did not know the etiology but knew the risk factor was number of sexual partners. By the end of the 1980s, the cause had been determined to be the human papilloma virus (HPV); since then, new methods of prevention and screening have been developed. Each refinement of the testing method removes more low-risk women from the cycle of testing and eliminates the costs associated with those tests. There are numerous types of HPV with genomes that can be sequenced, which reveals sublineages with different levels of risk. Dr. Hoover predicted that future tests will screen for specific genes in all potentially harmful sublineages, resulting in very precise identification.

Dr. Hoover outlined the success of the HPV vaccine, noting that it is good for developed countries but less helpful in countries where it is challenging to follow through with multiple doses given at specific intervals. He showed that rates of success are similar for those treated with three or two doses; the success rate is lower after a single dose but still better than the rate with zero doses or natural infection.

Dr. Hoover commented on the state of precision prevention as it relates to genetics. Many genes have been found that relate to cancer, though most have a very small risk. Many of the genes that have been identified do not yet have a clear effect. Dr. Hoover theorized that more avenues to prevention will be found when scientists have a better understanding of the function of these genes. He explained a study from the BPC3 that showed a narrow spread of risk when only modifiable risk factors were graphed and a much wider range when genetic methods were included. Even without specific knowledge of these genes' actions, including them in the risk model can better identify extremely low-risk and high-risk individuals.

Dr. Hoover's final example involved recommending patients for a national lung cancer screening trial. Patients were screened using low-dose computerized tomography (CT) scans or chest radiography, and the incidence of and death from lung cancer were shown. The data suggested that risk-based selection may be more effective and efficient than current guidelines. Dr. Hoover commented that this experiment brought modern biostatistics into the identification of which patients to screen. If those at the highest risk can be identified for screening, lives can be saved.

Dr. Hoover theorized that pessimism about the prospects of precision prevention could result from the expectation of a faster timeline for discovery; most of this progress has occurred incrementally. He predicted that in the next 40 to 50 years there would be a very low rate of cervical cancer because of the success of the HPV vaccine and improved precision prevention. Dr. Hoover emphasized that epidemiologists "need to be in it for the long run."

Precision Prevention: Let's Avoid Exacerbating Cancer Disparities
Graham Colditz

Dr. Graham Colditz considered how researchers can work from the beginning of a study to engage stakeholders and avoid exacerbating disparities. One example study suggested that health disparities may result from genomic misdiagnosis and a dearth of diverse control populations. He echoed Dr. Hoover's definition of precision prevention and explained the idea of "geospatial epidemiology," a community-engaged delivery of prevention. Interventions should be able to be accessed by the whole population across economic and cultural differences.

Dr. Colditz presented an example of the way colon cancer rates had changed during several decades in women and men, including African American men. In asking whether a disparity was created in colon cancer rates, he showed that white mortality decreases after screening. He noted that differentiation of health systems could affect the rates; racial or ethnic differences are reduced in an integrated health system. In another study, larger differences in survival were seen between black and white patients in cancers more amenable to intervention. He noted that colon cancer deaths, which were spread across the Midwest and Northeast in the 1970s and 1980s, now are largely concentrated in hotspots, indicating that implementation of screening has created disparities. Dr. Colditz also used the HPV vaccine as an example, noting that it has not been delivered equally to all Americans at all income levels and to all races. He emphasized that action is needed to avoid worsening this disparity.

Other factors that could affect disparities include low literacy levels, low interest in participating in research, and variability in awareness of cancer and heritability. The challenge is how to address issues of disparity at the cultural, individual, and system levels from the beginning of the process so that the results of advances in precision prevention can reach all Americans at the appropriate time. Dr. Colditz commented on the disappointing ranking of the United States in rates of cancer and heart disease compared to other countries. His suggested that potential goals for the future of this field are to avoid inducing disparities, create an effective implementation platform, and recruit diverse collaborators.


Dr. Gapstur asked whether disparities exist in women's attitudes when told of their genetic risk. Dr. Colditz explained that studies of genetic testing and smoking cessation did not show action, but in the breast clinic there is a variation in understanding based on education and knowledge of family. Dr. Gapstur wondered whether there is value in asking cohort participants about risk perception and whether there is a research agenda in the notion of risk communication to which cohorts can contribute. Dr. Hoover noted that knowledge is always good but that cohorts are not a representative sample of the population because they tend to comprise people who are interested. Dr. Colditz added that strategies to translate science from cohorts to the whole population is an important consideration in relation to disparities because men, and especially minority men, participate less often in research. Dr. Ursin asked what cohorts can do to recruit the whole population; scientists have long known that the most vulnerable populations are the most difficult to reach. Dr. Hoover noted that the broadest cohort recruited from a health care center in a rural Southern black community rather than recruiting volunteers. Participants added that partnerships with churches and community organizations can be fruitful ways to recruit more diverse populations, though followup remains a challenge. Dr. Hoover advocated for more health care to be made part of standard packages, such as government vaccine recommendations.

A participant asked Drs. Hoover and Colditz to comment on the emerging role of tumor biomarkers and epidemiological research in precision prevention, noting that factors affecting survival can vary across molecular subtypes. Dr. Hoover responded that enthusiasm exists for reclassifying tumors based on genetics; he continues to be optimistic that malignancy could depend on exposure. An attendee asked whether risk-based prevention could be increasing the risk of other diseases as an unanticipated consequence of focusing more on the risk for cancer than other diseases. Dr. Hoover responded that knowledge of risks is helpful in trying to convince patients to be tested; he advocated for increased integrated healthcare. Dr. Colditz commented on the worrying trend in this country that mothers know best, for example in the antivaccination movement, noting that population-wide public health strategies must be retained concurrently with the growth in precision prevention.

In response to a comment about populations that are more motivated to understand their risk factors, Dr. Hoover noted that it becomes more important to account for risk of disease when the potential solutions are risky. Determining the optimal risk ratio is a risk/benefit question that will be different for each case.

Session VI: Use of E-Technology in Cohort Studies

Moderators: Leslie Bernstein and Kathy Helzlsouer

Building Research Technology Platforms for Participants with a Consumer Engagement Experience in Mind
James McClain

Dr. James McClain spoke about All of Us, the new name of the PMI, pointing out that the original name had little meaning to those outside the federal system; the new name represents diversity and inclusion. The initial awards were funded 4 months prior to this meeting, and the initial IRB protocol was reviewed days before the meeting; the first formal submission to the IRB is anticipated around December 1. This is a relatively quick timeline for a startup initiative across a large and diverse consortium.

The initiative aims to recruit more than 1 million volunteers, beginning with adults because issues of consent and vulnerable populations still are under review. Each site has its own focus, but the initiative is on track to mirror the projected diversity of the 2040 census. Diversity is being considered not only in terms of people but also health status—meaning that there is no disease-specific cohort—and geography as a means to study environmental exposure.

All of Us intends to make volunteers partners in the initiative, involving them in the types of data collected, analyses conducted, research performed, and the approaches by which data are returned. Various styles of communication will be supported to fit participants' preferences, and each participant will own his or her data. Dr. McClain commented that it is both contentious and groundbreaking that no enrollment centers have specialized data access. The initiative currently is collecting data from established models, such as EHRs and basic physical tests. The methods of data collection will expand as the study and technology evolves; for example, personal health technology, which currently is modest, is anticipated to grow into a rich resource that can provide different forms of data and empower subjects.

This initiative offers numerous scientific opportunities, such as the development of risk estimates and targeted therapies, the identification of biomarkers and causes of individual variation, and the ability to empower participants and launch targeted therapy trials. The initiative will expand and develop with each new release; the intent is to support long-term innovation and advanced platform development. Dr. McClain noted that the project is just getting started, and there are many opportunities to contribute; he added that he envisions future funding centered on data access and expanding the initiative's methods.

The initiative is organized into four quadrants. The backend data building is organized into the Data and Research Support Center at Vanderbilt University, all biobanks are hosted at the Mayo Clinic, and health care provider organizations participate across the country. Dr. McClain represents the fourth quadrant, the Participant Technology Center (PTC), which is solely focused on the user experience. Two methods of engagement exist: direct volunteers and recruitment from health care provider organizations. The PTC is tasked with building a network that can reach anyone in the United States, and it utilizes partnerships with a number of organizations for outreach, platform development, and biospecimen collection.

The All of Us initiative aims to ensure that participants look forward to their next engagement with the program; participants must understand that the initiative values them and their contributions. This goal is supported by offering users the ability to compare their data to the rest of the cohort and using flexible tools that can help shape and record the participant experience. All of Us uses many innovative processes, but many components are built on existing best practices. For example, much work has been done in the field of informed consent, and All of Us is building on these foundations. Other measures that are likely to be fairly advanced are the integration of online portals, system security, and testing of that security.

The core app is a "wall"-style experience with a conversational flow allowing participants to filter and visualize their results as soon as their data are submitted. The "dashboard"-style experience also is in use, which will support links to other sections and allow participants to explore potential insights. Another section of the app is the "Donation Hub," an engagement feature emphasizing the value of participants' donations that researchers hope will redefine the typical one-way conversational flow of research. The Donation Hub will allow participants to connect to publications as they emerge and also offer content related to the survey that traditionally would have been packaged in newsletters. Several features will allow participants to engage with the science of the survey at various levels—as a brief video, a narrative, or links to publications. The apps will include behavioral analytics technologies to help researchers learn about and adjust user-engagement methods.

Dr. McClain commented that the initiative has a large amount of quality protocols and modest engagement plans, and they require more effort on material development. The initiative aims to be "ultra-transparent," easy to understand, and engaged with the participants so they look forward to using it.


Participants asked about the cost of the initiative and whether it would consume all available population science resources. Dr. McClain responded that the resources it will consume will largely be in terms of recruiting interested scientists in the field, and the work of the initiative hopefully will support other studies and lead to partnerships. He noted that though the cost over time will be reasonably high to maintain, the startup award of $207 million over 5 years is relatively low for the extent of the initiative and the number of charges that have been given to the PTC. Dr. McClain confirmed that participants will be encouraged to suggest research questions and identify gaps.

A participant expressed concern at the amount of money given to this project given frequent budget restrictions for projects that meet a greater number of expected standards. Dr. McClain noted that because he was asked to present on technology innovation, he did not have authority to comment on priorities or research goals; however, from a scientific perspective, the initiative is optimized to fill in gaps and address new questions rather than duplicating research questions successfully answered by existing studies.

Participants described challenges experienced with other cohorts utilizing online registry and immediate feedback, such as difficulty accommodating highly engaged participants on limited computer systems. Dr. McClain commented that he suspected All of Us is better funded for this purpose, having been resourced with the intent of delivering that large engagement. Dr. McClain concurred with an attendee's theory that the kinds of data gathered will shift over time and verified that the majority of study participants wanted data comparison abilities. Participants asked about the regulatory requirements for engagement features, and Dr. McClain suggested ways that engagement could evolve as necessary within the bounds of the strategy initially sent to the IRB. In response to a question about whether feedback has affected recruitment or engagement, Dr. McClain explained that the digitally driven cohort is less likely to follow the traditional cohort model of poor retention, and the initiative will take any action it can to drive engagement.

A participant noted that existing cohorts have been trying similar initiatives with less funding and asked whether cohorts could add their data to the All of Us initiative and in turn share resources developed by the initiative. Dr. McClain responded that the initiative plans to release the platform as open-source when it is ready. A participant with the California Teachers' Study noted that when participants in that cohort were asked what additional features they would like, one-half requested the ability to see their own data and compare them to the rest of the cohort, so the plans for All of Us to include these features are in line with what this cohort has experienced.

Session VII: Cohort Consortium Subcommittee Updates

Moderator: Meir Stampfer

MTA Subcommittee Update
Joanne Elena

Dr. Joanne Elena explained that the mission of the MTA Subcommittee this year was to make the data transfer process more efficient. The first step is to access previously successful agreements. A template approved by NCI was distributed to other institutions, but at each it was edited until it no longer matched the original. The subcommittee then crowdsourced three examples of previous successful agreements, which were anonymized using dummy organizations. These examples are going through the NCI review process and will be ready in early 2017, at which time they will be posted on the Cohort Consortium website. Another option would be to use the Uniform Biological Material Transfer Agreement (UBMTA), the preferred format of NCI specialists.


A participant asked whether this process will be extended for biospecimens, as it would be helpful to obtain examples of successful agreements and problems experienced. Dr. Elena responded that the subcommittee had asked for feedback on problems and tips but had received none; she noted that the UBMTA is a first step and encouraged participants to submit feedback. In response to a question about a trial period for the agreement; Dr. Elena explained that the terms and scope were defined during NCI development, and satisfying templates have been developed, but there usually are continuations and amendments. Participants discussed the necessity of planning agreements to accommodate any evolution of the study and the possibility of creating a database to help track the development of this agreement.

NIH Data Sharing Policy Update
Elizabeth Gillanders

Dr. Elizabeth Gillanders commented on initiatives from the highest levels of government mandating that federal agencies increase access to federally funded data. The Vice President's Cancer Moonshot is focused on developing revolutionary approaches for sharing cancer data and enhancing existing data sharing. The NIH Genomic Data Sharing Policy affects all NIH-funded research that generates large-scale genomic data; it also may apply to smaller scale but high-priority programs.

Many data sharing projects exist in the field of epidemiology, but the process still is in development. NCI held a workshop in 2014 to discuss broadening data sharing and ways to maximize the public benefit from federally funded studies. Dr. Gillanders noted that NCI has heard and appreciates the tension between the benefits and burdens of data sharing, and discussions of improvements are ongoing. She emphasized that every NIH-funded cohort should have a webpage that outlines data sharing opportunities for that cohort, present a transparent process for data-sharing requests and track these requests, and utilize an updated informed consent process.


Participants noted that for rarer cancers, misleading meta-analysis can be assuaged by a pooled analysis of cohorts and emphasized the value of the work that goes into the pooling project. A participant asked how a new policy about a centralized IRB for multi-site studies would affect cohort studies. Dr. Gillanders replied that a central IRB may still have to gather supporting documentation from individual IRBs. In response to a question about tangible encouragement for data sharing, Dr. Gillanders emphasized that her team has heard researchers' concerns and is making an effort to address the cost-benefit balance associated with data sharing. A participant suggested supporting additional time for intramural scientists. Dr. Elena replied that often only a finite amount of money can be directed toward either infrastructure or research, emphasizing that cohorts should plan for any specialized roles, such as programmers. Dr. Gillanders reiterated that her team has heard the concerns and is working on finding solutions to these complex problems. A participant suggested developing a special request for application (RFA) for large consortial products, and Dr. Elena explained that instruction can be given to study sections, and infrastructure support can be included in an R01 application. She suggested that the best way to increase awareness is to serve on and speak out at study sections.

A participant suggested a series of articles describing how the Consortium has transformed decision making and influenced policies and asked whether an NCI study section could be created. Dr. Elena responded that this is a possibility but it may be a case of "be careful what you ask for" because NCI study sections often are not standing sections and would involve repeated logistical challenges. She informed the audience of the National Heart, Lung, and Blood Institute's Biologic Specimen and Data Repository Information Coordinating Center (BioLINCC), which is predominantly composed of cardiovascular disease datapoints but can be seen as a successful model. Dr. Leslie Caroll, the project manager for BIOLINCC, added that the database only accepts curated data packages, so cohorts must fund that curation at the cohort level. Dr. Gillanders emphasized that many options exist for facilitating sharing, so researchers must consider the method that best supports the science.

Consortium Working Groups Subcommittee Update
Anthony Swerdlow

In the absence of Dr. Celine Vachon, Dr. Swerdlow offered brief comments on the Working Groups Subcommittee. He noted that committees are the essence of organizing working groups and that the goal is to help the main Steering Committee think about working groups. Dr. Swerdlow encouraged attendees to consider whether the programming schedule was successful and submit feedback to assist in the development of future Consortium meetings.

Working Group Survey Results
Camille Pottinger

Ms. Camille Pottinger, a fellow within NCI's Epidemiology and Genomics Research Program, summarized the results of the 2016 NCI Cohort Consortium Working Group/Project Survey. The purpose of the survey was to provide the Cohort Consortium Steering Committee with a better understanding of the activities and experiences of working groups in the Cohort Consortium, as well as improve the support for the different working groups and the Cohort Consortium as a whole in achieving their goals.

A brief survey was conducted to determine the active status of each of the 43 working groups; this was followed by a more comprehensive survey to collect more detailed data on working group activities.

Respondents to the preliminary survey indicated that 88.4 percent of the working groups are active. Groups had become inactive for reasons including limited funding, a change in direction, and achievement of goals. Of the 35 working groups or projects that completed the follow-up survey, 76.5 percent reported that new members are welcome to join their groups. Ms. Pottinger pointed out that 50 percent of working groups open to new members already have active spinoff projects in place. Working groups that were uninterested in recruiting new members gave reasons including being past the data-collection phase, limited funding, and the lack of other studies with common data points.

The comprehensive survey also provided a snapshot of junior investigator involvement in the projects. Junior investigators serve as project leaders in 62.2 percent, primary or corresponding authors in 53 percent, and other roles (e.g., Steering Committee members, research assistants) in 21 percent of the working groups reporting.

Challenges identified by working groups included data limitations, communication and collaboration among members, funding, data transfer, internal working group issues, and Cohort Consortium annual meetings. To address data limitation challenges, suggestions included establishing a centralized harmonization database, developing guidelines to reduce redundancy in data management and increase harmonization across consortia, and maintaining a repository of cohort data and information. Suggested solutions to improve communication and collaboration included promoting and facilitating opportunities to engage in cohort, working group, and project activities; establishing clear guidelines to allow easier and faster approval of subprojects within established consortia; providing additional points of contact to receive timely input; and directing proposals to working groups with similar aims to review.

Funding issues are best addressed by providing separate RFAs for groups that perform consortium work and allowing for budgets that reimburse individual cohorts, as well as creating mechanisms by which investigator effort at multiple institutions can be included in NIH grants to the Cohort Consortium. Suggested solutions for difficulties with data transfer included developing standard templates for data transfer agreements and MTAs, as well as addressing legal issues by developing a policy that requires fees to obtain data and samples. Internal working group issue solutions included agreeing on policies early, creating small grant mechanisms, and publishing a commentary by the Cohort Consortium that would document the complexity and importance of pooling projects.

Regarding the Cohort Consortium annual meeting, working group suggestions included making scheduling changes to reduce meeting overlap and providing better audiovisual technical support.

Ms. Pottinger concluded by noting that less than one-half of working group leaders identified multiple challenges within their working groups or across the Consortium. The next step will be further review of the survey results by the Steering Committee to address identified challenges. For the future, one of the subcommittee's goals is to track publications attributed to the working groups; input is welcome on strategies to accomplish this goal.

Dr. Gapstur encouraged attendees to submit their surveys, noting that past feedback has been very helpful, and a high response rate would be appropriate for this consortium of epidemiologists.

Contact Us

If you have questions about the meeting, contact at the National Cancer Institute.