2018 Annual Meeting

Wednesday, November 28, 2018

Friday, November 30, 2018

Moderator: Giske Ursin

---

Welcome and Introductions
Kathy Helzlsouer and Giske Ursin

Dr. Kathy Helzlsouer welcomed participants and introduced the session moderator Dr. Giske Ursin, who is the current 2018 Chair of the Cohort Consortium Steering Committee. Dr. Helzlsouer also introduced 2019 Chair-Elect Dr. Celine Vachon and 2020 Chair-Elect Dr. Lynne Wilkens, as well as new Steering Committee members. She thanked current Steering Committee members and the NCI organizers of this meeting and related strategic planning activities.

Dr. Helzlsouer noted that a highlight of the meeting would be a followup to the strategic planning activities conducted during the 2017 meeting and the subsequent surveys. Strategic planning at this year’s meeting will focus on ways for the Cohort Consortium to advance the science in the future.

Current National Institutes of Health (NIH) Requests for Information (RFIs) that are relevant to the Consortium include the following: (1) the Proposed Provisions for the Draft Data Management and Sharing Policy and (2) the Draft Strategic Plan for NIH Nutrition Research. Dr. Helzlsouer encouraged participants to review these RFIs and provide feedback.

---

Steering Committee Activities
Giske Ursin

Dr. Giske Ursin discussed activities of the Steering Committee. She highlighted a previously published paper, The National Cancer Institute Cohort Consortium: An International Pooling Collaboration of 58 Cohorts from 20 Countries, which was co-authored by a few members of the Steering Committee, under the leadership of former Steering Committee Chair Dr. Anthony Swerdlow.

The Steering Committee worked with NCI staff to implement changes to the Cohort Consortium newsletter. Dr. Ursin asked participants to submit suggestions for newsletter content to one of the Steering Committee members or to either Nonye Harvey or Camille Pottinger at the NCI.

The Steering Committee recently updated the Cohort Consortium bylaws. Dr. Ursin expects to distribute the updated bylaws to voting members of the Cohort Consortium sometime in January 2019.

The Steering Committee reviews WG project proposals for consistency and appropriateness and provides suggestions on improving the proposed projects. Once the review is complete, the Steering Committee disseminates the proposals to Cohort Consortium members so that they can decide whether they wish to participate in the new project. In 2018, the Steering Committee received one updated report from one WG. Five new WG/Projects started in 2018 with a focus on glioma and bladder, breast, and stomach cancers.

The Steering Committee is evaluating the publications generated by the WG/Projects. To facilitate this task, the bylaws have been changed to require recognition of the Cohort Consortium in all publications that use Consortium data. The Cohort Consortium can be recognized in the study title or as an acknowledgement. Recommended acknowledgement wording will be included in the updated bylaws.

Dr. Ursin encouraged participants to provide feedback on this meeting, particularly the 14 Project/WG sessions that were held either immediately before or immediately after the main plenary sessions.

Dr. Ursin recognized the Steering Committee members. She announced that Steering Committee member Dr. Elisabete Weiderpass will be leading the International Agency for Research on Cancer in 2019 and as such will resign from the Steering Committee.

---

Opening Remarks and NIH Updates
Giske Ursin

Dr. Ursin introduced Dr. Robert Croyle, Director of the Division of Cancer Control and Population Sciences (DCCPS) at NCI. She noted that Dr. Croyle was given the Distinguished Achievement Award by the American Society of Preventive Oncology.

She also introduced Dr. Stephen Chanock, Director of NCI’s Division of Cancer Epidemiology and Genetics (DCEG). Dr. Chanock is the leading expert in the discovery and characterization of cancer susceptibility regions in the human genome. He has received several awards for scientific contributions to the understanding of common inherited genetic variance associated with cancer risk and outcomes.

Robert Croyle, Director of the Division of Cancer Control and Population Sciences at NCI

Dr. Croyle also thanked the Steering Committee members and NCI staff who organized the meeting. Dr. Croyle is Interim Director of NCI’s Center for Global Health. He announced that the Center is seeking a new director, as is the NCI’s Office of Cancer Survivorship.

Dr. Croyle noted that the U.S. life expectancy had declined further, a finding that has particular relevance to the activities of the Cohort Consortium. He also mentioned a special section on rural cancer control in a recent issue of Cancer Epidemiology, Biomarkers, and Prevention. Rural cancer control has been a special area of focus for DCCPS over the past few years. A DCCPS research priority is understanding rural/urban health disparities, particularly those experienced by special populations such American Indians/Alaska Natives and immigrant populations.

DCCPS also is supporting studies of the risks of marijuana exposure and methods for measuring those risks. Dr. Gary Ellison is the DCCPS coordinator for research on the effects of marijuana.

The U.S. Food and Drug Administration tobacco regulatory funds transferred to NCI have supported 109 grants. Dr. Croyle noted that the Altria Group is investing heavily in JUUL Labs, the e-cigarette company that now dominates approximately 70 percent of the U.S. market.

The NCI’s U.S. National Cancer Advisory Board (NCAB) is reviewing the cohort portfolio. Two members of the Cohort Consortium, Drs. Julie Palmer and Les Robison, are co-chairing the NCAB Working Group on Strategic Approaches and Opportunities in Population Science, Epidemiology, and Disparities. Dr. Deborah Winn is the Executive Secretary of this WG. This WG will focus on strategic planning for etiological and survivor cohort studies.

The NCI Cancer Moonshot has funded a large number of special initiatives, resulting in more funding opportunities in the past 2 years than in the history of the Institute. Many Moonshot-funded projects focus on population science and implementation science. NCI recently released a new Request for Applications (RFA) for Centers of Excellence in Implementation Science. An implementation science project on prostate cancer in African American men will be presented at the next NCAB meeting.

Stephen Chanock, Director of the Division of Cancer Epidemiology and Genetics at NCI

Dr. Chanock also thanked Steering Committee members. He began by emphasizing the need for a firm commitment to evidence-based science and dissemination of evidence in the face of misinformation via social media and other channels. At both the national and international levels, the defense of scientific method and evidence is critical; thus, Dr. Chanock encouraged participants to support the NCI and the International Agency for Research on Cancer (IARC) in this endeavor. Dr. Chanock added that NCI Director Dr. Norman Sharpless is strongly supportive of epidemiology and population science.

Dr. Chanock noted that the NCI’s Intramural Program is working closely with the extramural community. He described a few Cancer Moonshot activities, including a human papillomavirus (HPV) vaccine trial in Costa Rica to test single-dose efficacy and international studies of HPV screening methods, to control cervical cancer. For screening studies, artificial intelligence and cellphone technology offer important new tools.

The DCEG portfolio includes studies of classical and descriptive epidemiology and methodology in the areas of microbiomics, metabolomics, and genomics. A new focus area is the sequencing of pathogens, such as HPV and H. pylori, to better understand the host response to pathogens and somatic alterations that result. Dr. Chanock mentioned the poster presentation by Dr. Montserrat Garcia-Closas on an international study to understand the genetic architecture of breast cancer and encouraged participants to become involved in this study. He also described Dr. Teri Landi’s Sherlock Lung etiological study on lung adenocarcinoma in nonsmokers around the world, which will involve high-volume sequencing and mutational signatures.

DCEG supports a new cohort and various methodologic activities, which are part of the Cohort Consortium, in areas such as physical activity, metabolomics, and microbiomics.

---

National Institute of Environmental Health Sciences (NIEHS) Exposure Analysis Resources: Children’s Health Exposure Analysis Resource (CHEAR) and Human Health Exposure Analysis Resource (HHEAR)
Claudia Thompson

Dr. Claudia Thompson, Chief of the Population Health Branch at the National Institute of Environmental Health Sciences (NIEHS), discussed the Children’s Health Exposure Analysis Resource (CHEAR) and Human Health Exposure Analysis Resource (HHEAR). Dr. Thompson asked participants to consider the role of environmental exposures in outcomes studied by Cohort Consortium members. Environmental exposures should be measured and included in Cohort Consortium analyses whenever possible.

The CHEAR’s focus on environmental exposures and their effects on children’s health and development is transitioning to environmental exposures and their effects on the health of humans of all ages (HHEAR). The CHEAR funding ends on August 31, 2019, but NIEHS has received approval to issue five RFAs for the continuation of the resource as HHEAR. CHEAR and HHEAR resources are available to the NIH research community free of charge. The purpose of creating these resources was to provide infrastructure for comprehensive and reliable exposure analysis of biological samples that would allow investigators to explore the breadth of the exposome. CHEAR and HHEAR involve a network of laboratories across the United States supporting targeted and untargeted analyses of biomarkers of exposure in biological samples. Targeted and untargeted studies of exposure biomarkers will allow cross-validation to better characterize features of exposure and will improve the understanding of biological response mechanisms (e.g., epigenetics, microbiome, inflammatory markers, etc.). Investigators with minimal or extensive environmental exposure expertise are encouraged to use these resources.

Currently, 39 CHEAR projects are funded. Some of these projects are developing panels of known features that will allow investigators to identify specific chemicals in the blood. CHEAR provides technical support to help investigators determine what resources and methods to use and how. CHEAR includes approximately 54,000 samples to date.

The HHEAR laboratory network will support targeted and untargeted analyses, as well as a new component, environmental monitoring. The environmental monitoring component currently is in the developmental stages; it will test new methods to track environmental exposures (e.g., wearable devices). HHEAR analyses will examine the impact of exposures on health across the life course by providing a data/biospecimen resource with quality control on par with national laboratories. HHEAR will help investigators design the best study for the question they want to ask so that samples are not wasted or used inefficiently. NIEHS is encouraging other NIH Institutes and Centers (ICs) to contribute to HHEAR. The NCI has agreed to contribute so that NCI-funded investigators will have access to HHEAR.

---

Discussion

Participants noted that the technology available to measure exposure biomarkers is well developed but within-person variation presents a problem. Dr. Thompson indicated that CHEAR and HHEAR staff have considered the challenges of within-person variation and have analyzed many studies that have performed longitudinal data collection with numerous samples collected from the same individuals over time. Staff also have considered the short half-life of certain chemicals. Data also are collected from individuals to determine what exposures they might have experienced in their lives. The laboratory hubs with in CHEAR have extensive experience in biospecimen quality control and provide guidance to potential clients (i.e., users of CHEAR) to work towards improving the quality of samples.

Moderator: Susan Gapstur

Dr. Gapstur began by recognizing the Cohort Consortium as a model of team science, which is critical to the success in the field of epidemiology. The evidence generated by the Cohort Consortium allows expert panels to make judgments on the causality of disease.

Dr. Gapstur discussed the reasons for strategic planning within the Cohort Consortium. A strategic plan is needed to set priorities, focus energy and resources, and strengthen operations. The strategic plan helps ensure that the Cohort Consortium identifies the most important and current cancer-relevant public health gaps in scientific knowledge and responds to those gaps through multiple prospective cohort studies. Therefore, the strategic plan must evolve, and strategic planning activities should focus on what needs to change in the plan.

Strategic planning started with the World Café at the 2017 Cohort Consortium meeting. The World Café discussions focused on methods for enhancing operations, best practices for operationalizing scientific activities, and new scientific directions. The results of the World Café activity are available on the Cohort Consortium portal. Ms. Harvey also can send the results to those who have not seen them.

The strategic planning activities in 2017 led to a draft of the NCI Cohort Consortium Strategic Initiatives for 2018–2021 document. The goals in this document are communication, career development, and research facilitation. The Steering Committee has updated the bylaws based on the Strategic Initiatives document.

Dr. Gapstur noted that the NCI wants to add a session on the Cohort Consortium at the next American Association for Cancer Research conference. NCI also is launching new activities on career development. The Steering Committee is looking for new investigators to lead and co-lead Projects/WGs and is forming an associate council of new investigators. New investigators interested participating in the associate council should contact Dr. Vachon.

Data sharing will be another focus of the new strategic plan, including biospecimen data. The Steering Committee is seeking opportunities to enhance discussion on these topics. Surveys of Cohort Consortium members on scientific directions will be conducted.

---

Strategic Planning Update
Susan Gapstur

The World Café session at the 2017 Cohort Consortium meeting generated 20 topic areas. The brainstorming session at this 2018 Cohort Consortium meeting focused on strategies and activities for the next 5 years. An online survey was conducted in April 2018 and sent to all Cohort Consortium stakeholders for whom the Steering Committee had contact information. Members of 35 cohorts responded with a 26 percent response rate. Overarching themes generated by the survey included the need for adequate statistical power for analyses and the need for improved understanding of mechanisms, etiology, and outcomes. Other key survey findings included:

• Circulating biomarker data from multiple cohorts is important because they allow investigators to refine and validate new types of exposure assessments.
• The Cohort Consortium can provide data on rare cancers for analyses with sufficient statistical power.
• Another important advantage of the Cohort Consortium is the molecular pathoepidemiology that includes timely collection of tumor tissue and integration of tissue biomarker data with other data.
• Data linkages across cohorts have been challenging.
• Coalitions need to be formed to resolve data-linkage problems.
• More survivorship studies are needed.

---

Charge and Instructions for the Brainstorming Session
Nonye Harvey

The purpose of the brainstorming session is to identify research gaps and opportunities for the following five topic areas:

• Circulating biomarkers.
• Linkages, including those to electronic medical records and other passive data sources.
• Molecular pathoepidemiology (timely collection of tumors and integration of tissue biomarkers with other data sources).
• Rare cancers.
• Survivorship research.

Participants were asked to identify research directions for each topic area (which could be combined, if necessary). Specifically, they were asked to consider:

• Why should the topic area be a Cohort Consortium effort?
• What is the potential public health impact?
• How do we advance research on this topic (other than funding)?

Reports of the brainstorming sessions were not delivered at the meeting. Ms. Harvey agreed to compile the information collected during these sessions from flip charts and notes, summarize that information, and share it with the attendees.

Moderator: Roger Milne

---

Virtual Tissue Repository Pilot Studies (VTR) Pilot Studies: Lessons Learned and Future Directions
Alison Van Dyke

The future SEER (Surveillance, Epidemiology, and End Results Program)-linked Virtual Tissue Biorepository (VTR) program will establish an infrastructure for researchers to access tissue and associated data at the population level through the SEER registries. The VTR is perfect for studying rare cancers and rare outcomes.

SEER registries will serve as honest brokers for providing researchers with deidentified, but linked, clinical history data and formalin-fixed, paraffin-embedded tissue. Seven SEER registries are participating in VTR pilot program, six of which are collecting medical history data and from pathology laboratories in their catchment areas for two studies of unusual cancer outcomes, one in pancreatic cancer and the other in breast cancer. These pilot studies are important for determining feasibility of scaling the VTR and identifying best practices.

In the future, VTR users will be able to search clinical and demographic information as well as deidentified ePath reports to select what cancer cases on whom they would like to request data and tissue collection.

In a laboratory inventory conducted through the VTR pilot program, approximately 75 percent of laboratories reported that they destroy tissue blocks after reaching the minimum time required to retain those blocks, as established by the College of American Pathologists. The vast majority of pathology laboratories (69%) indicated that they would be willing to share tissue for research. These findings suggest a need to expand residual tissue repositories (RTRs) and have them function in parallel with the VTR.

More than 50 percent of laboratories surveyed are not aware of the consent requirements for releasing tissue for research, thus suggesting the need for education and collaboration with these laboratories.

Ethical and regulatory issues are important concerns when using tissue samples. The revised Common Rule (Final Rule) states that research on deidentified specimens and data is not considered human subjects research.

The NCI VTR team is developing processes for project review and approval, investigator search interface, and the tracking of approved projects. These processes will be implemented in phases beginning with infrastructure and training of SEER staff and promotion of the VTR and RTR, especially with pathology laboratory directors.

---

Discussion

Roswell Park Comprehensive Cancer Center is collecting tissue blocks for a case-control study that will use data from SEER registries. Registries will obtain tissue samples from hospital laboratories and access medical records to obtain additional data. Registries will maintain links to patient identifiers for research purposes. The study will attempt to link biospecimen and registry data to questionnaire data. The infrastructure from this study could be used for VTR studies.

Cancer Moonshot initiatives developed a hashed and encrypted application that allows linkage to external data sources without risking personally identifiable information (PII).

---

The Los Angeles Residual Tissue Repository (RTR): A 36-Year Population-Based Tumor Block Archive
Wendy Cozen

The SEER RTR was implemented at the Iowa, Hawaii, and Los Angeles cancer registries.

The Los Angeles registry selects only some discarded tissues and has its own pathologist review them.

The tissues collected at the Los Angeles RTR overrepresent minority populations, with a high representation of Korean Americans compared with other registries. The registry has information on five major cancer sites by race/ethnicity. The registry also collects childhood cancer tissues.

The Los Angeles RTR has tissue samples for many rare cancers, especially Kaposi’s sarcoma because of the high population of people living with HIV/AIDS in the area.

The Los Angeles RTR has variables from its SEER program on demographics, residence, tumor type, cancer site, vital status, and treatment.

A pancreatic tissue microarray was created with RTR tissue.

An investigator at the Los Angeles registry also used the RTR to investigate HPV-positive tumors and found the incidence of these tumors to be increasing. HPV-positive oropharyngeal cancer likely will surpass cervical cancer as the most common HPV-associated cancer in the near future.

Dr. Cozen is examining Hodgkin’s lymphoma and the tumor microenvironment in non-Whites by using the RTR. Another investigator is using the Los Angeles RTR to study triple-negative breast cancer and social class.

---

Discussion

Dr. Cozen recommended that investigators interested in using the Los Angeles RTR resource develop a list of patients with cancer for whom they would like to obtain tissue samples. Consent or IRB exemption also should be obtained for these patients before requesting RTR tissue blocks. The RTR will use the list to locate a tissue block for the indicated patients if a tissue block is available.

Moderator: Julie Palmer

Three Cohort Consortium Working Groups that oversee long-running studies delivered presentations on their research activities.

---

Prostate Cancer Cohort Consortium (PC3)
Michael Cook and Lorelei Mucci

The motivation for PC3 stems from the clinical and molecular heterogeneity in prostate cancer as well as the importance of cross-validation of epidemiology findings across different populations

First meeting was in 2017. In the Cohort Consortium, >125,000 cases of prostate cancer have been diagnosed, including 12,500 fatal outcomes. Very powerful for etiologic and survivorship research.

Blood samples are available from 51,000 patients in the PC3, and tissue samples are available from more than 10,700 patients.

The PC3 WG is focusing on the following objectives: (1) defining clinical subtypes, with a plan to develop a white paper, reviewing definitions, and attempting to develop consensus on prostate cancer subtypes; (2) determining risk factors beyond the five nonmodifiable ones (the PC3 is developing a strategy for an evidence review of lifestyle factors followed by a consensus building activity); and (3) engaging in collaborative activities. The WG is planning additional face-to-face meetings at conferences other than the annual Cohort Consortium meeting.

---

Discussion

In response to the participants’ questions about the planned evidence review, Dr. Cook explained that the review systematically will examine known risk factors and other consensus-building efforts regarding modifiable prostate cancer risk factors. Participants encouraged the PC3 WG to perform a systematic, published review.

---

Diabetes and Cancer Initiative
Marc Gunter

The Diabetes and Cancer Initiative has not generated results at this point. Research questions being examined include:

• Is Type II diabetes associated with cancer independent of shared risk factors?
• Is there an association between the two diseases after considering surveillance/ascertainment bias? Many people are diagnosed with diabetes and cancer concurrently.
• Is diabetes related to cancer prognosis and survival once adiposity and treatment are taken into consideration?
• What is the biological mechanism for the link between diabetes and cancer? Direct effects of hyperinsulinemia, inflammation, and alterations in the sex hormone axis will be examined.

Investigators are working on metadata acquisition and data harmonization. They are developing a cohort metadata repository with a set of harmonized variables that can be used in any project (Cohort Metadata Repository, CMR). Anyone can use the CMR to save time on data harmonization for Cohort Consortium studies. Harmonized variables in CMR include demographics, anthropometry, smoking, alcohol, health history, and reproductive history. The Project is now establishing data transfer agreements (DTAs) and first results are expected by November 2019.

---

Discussion

Participants asked how incident diabetes cases are ascertained. The ascertainment method varies by the cohort. Each center must complete a questionnaire asking how it documents diabetes. Incident case data are validated using medical records.

Participants asked whether the initiative identified prediabetes and Type I diabetes. Most cohorts did not collect data on Type I diabetes diagnoses. Age at diabetes diagnosis could serve as a proxy for cases of Type I diabetes. The lack of biomarker data has made identification of people with prediabetes difficult. Participants noted that cases of prediabetes are important for studying windows of susceptibility. The value of collecting data on these cases would depend on the demand for this type of data but is being examined.

Participants added that health care data could be useful. Many health organizations, such as Kaiser, have diabetes registries. These registries could be used to identify people with prediabetes.

---

Calcium Intake and Lung Cancer Pooling Project
Xiao Ou Shu

The Lung Cancer Pooling Project involves 1.7 million of individuals and approximately 25,000 lung cancer cases, including 21, 850 cases who are never smokers.

The Pooling Project found that in general dietary calcium intake was not associated with lung cancer risk. However, a positive association was found for high calcium intake and lung cancer among heavy drinker, and for small cell carcinoma. Low pre-diagnosis calcium intake, however, was associated with high mortality.

Fat intake (especially saturated fat) was associated with an increased risk of lung cancer. Polyunsaturated fat, however, was inversely associated with lung cancer risk. These associations were consistently seen among smokers and never smokers, and more prominent for small cell and squamous cell lung cancer. A 5% energy substitution of saturated fat with polyunsaturated fat was associated with 16-17% lower risk of small cell and squamous cell lung cancer

Body mass index (BMI) was inversely associated with lung cancer risk regardless of smoking status, with the exception of small-cell lung cancer in which BMI is positively associated with the risk. Waist to hip circumference ratio (indicate of central obesity), on the other hand, was positively associated with lung cancer risk.

Yogurt and higher dietary fiber intake were inversely associated with lung cancer risk, especially among current and former smokers.

Coffee or tea drinking was both associated with increased lung cancer risk. The associations did not vary by smoking status, caffeine content of coffee and type of tea.

---

Discussion

Participants had asked whether secondhand smoke exposure might be associated with tea and coffee drinking in nonsmokers and raised a concern on confounding from passive smoking. Dr. Shu replied that passive smoking information was collected in the Shanghai Women’s Health Study. Additional adjustment for passive smoking status in that study did not change the tea drinking and lung cancer association. She also pointed out that a recent IRAC study reported that secondhand smoke is not a strong risk factor for lung cancer.

Participants recommended that Dr. Shu stratify by sex in the nonsmoker group.

Responding to a question regarding the involvement of other investigators in the pooling projects, Dr. Shu mentioned that a Canadian investigate team is leading a project evaluating the association of physical activity with lung cancer risk in which Vanderbilt team helped out by performing the data analyses. Vanderbilt team has consulted the Pooling Project of Prospective Studies of Diet and Cancer at Harvard University during the data harmonization.

Moderators: Robert Hoover and Celine Vachon

---

Asia Cohort Consortium
Wei Zheng

The Asia Cohort Consortium (ACC) has harmonized demographic, anthropometry, lifestyle, diet, medical history, and family history data collected from ~20 cohorts in several Asian countries. Thirty-one cohorts now participate in this Consortium, and a few of these cohorts are supported by the NCI. Data from the ACC have been used in ~40 active research projects. To date, nearly 20 papers from the ACC have been published or submitted for publication. The ACC conducts two Consortium-wide meetings per year. One meeting is held in Asia, and a WG meeting is held in the United States, often coinciding with the annual AACR meeting. Dr. Zheng encouraged interested participants to visit ACC website (https://www.asiacohort.org) or contact him if they are interested in learning more about the ACC and/or contributing to ongoing ACC projects.

---

Discussion

In response to a question from one of the participants, Dr. Zheng indicated that DNA samples are not available for many ACC studies because many of the participating cohorts were initiated many years ago. However, some of these cohorts did collect biological samples and have contributed to several molecular and genetic epidemiologic studies, such as the Asia Breast Cancer Consortium and Asia Colorectal Cancer Consortium led by Dr. Zheng.

---

Summary of Report from the International Cohorts Summit
Anthony Swerdlow

The International Cohorts Summit (ICS) is intended to enable leaders of large longitudinal cohort studies worldwide to discuss and share best practices, data-sharing standards, common challenges, and the potential for larger collaborative sequencing. ICS was initiated by the Heads of the International Research Organizations group and is co-chaired by Drs. Francis Collins (NIH) and Jeremy Farrar (Wellcome Trust).

The criteria for invitation, were, in general:

• ≥100K participants
• Not selected based on a specific disease
• Biospecimens available
• Potential for longitudinal follow-up

The cohorts currently include approximately 25 million people; their aggregate target is more than 36 million. Nine cohorts from the NCI Cohort Consortium participate in the ICS.

---

Discussion

The Global Genomic Medicine Collaborative, which organized the ICS, is holding a meeting in South Africa on low-income countries, concurrent with the NCI Cohort Consortium meeting. ICS is aspirational, with the goal of expanding research across different countries. One objective of ICS might be capacity building. A member of the Cohort Consortium has joined one of the teams developing work plans for ICS and commented that the ICS’s interest seems to be primarily on diseases other than cancer, and there has be interest in pharmacogenomics. NCI Cohort Consortium members who wish to participate and meet the ICS criteria could contact the ICS secretariat (teji@g2mc.org).

---

UK Biobank: A Visionary Health Resource for the 21st Century
Naomi Allen

The UK Biobank is a highly validated data set established to help scientists better understand diseases, including cancer, in individuals of middle and old ages (ages 40–69 years). The Biobank includes data on lifestyle and environmental factors, personal and family medical history, cognitive function and hearing tests, physical measurements, and biological samples (blood, urine, saliva). This open-access resource is available to investigators around the world, with a streamlined, nonpreferential process for accessing the data. Users are expected to return derived variables to the Biobank for use by others. Most applications for access are approved. The UK Biobank currently has 36 different biomarkers that the creators thought would be most valuable to scientists (e.g., lipids, sex hormones, insulin-like growth factor 1, vitamin D, indicators of cardiovascular disease, cancer, and liver and kidney diseases). These are available for all 500,000 participants.

Investigators currently are using the Biobank to conduct cohort-wide assays on metabolomics and telomere length. Large pharmaceutical companies are using the UK Biobank and have developed a consortium to perform exome sequencing for all 500,000 participants (data will be publicly available for the first 50,000 in March 2019 and for the remainder at the end of 2020). A genome-wide array was performed for the full cohort, resulting in genome-wide array data for hundreds of traits. These data are available (with significant imputation) and searchable. A vanguard study to perform whole-genome sequencing for 50,000 participants is underway, and cohort-wide sequencing in the UK Biobank is planned.

The UK Biobank is also undertaking a large, population-based, multimodal imaging study. Investigators plan to have imaging data for 100,000 patients by the end of 2023. Imaging data will include scans from MRI (brain, heart and full body), carotid ultrasound, and dual-energy X-ray absorptiometry. Imaging-derived phenotypes are being produced by the imaging community using automated analysis pipelines. A repeat imaging assessment is planned for a subset, with a focus on cognitive decline (as measured from the brain scans), although these data can be used to answer a wide range of research questions.

The UK Biobank has health data from the National Health System on death (date and cause), cancer (date at diagnosis, cancer type, morphology and histology – and will soon have data on disease aggressiveness, tumor size, treatment, and prognostic indicators), hospital admissions (date of admission, diagnoses and procedures), and primary care data (diagnoses, symptoms, referrals, prescriptions, and laboratory results), which is pending. The primary care providers use different coding systems, so harmonizing primary care data will be challenging.

---

Discussion

UK Biobank exome sequencing for the first 50,000 participants is funded by Regerenon and Glaxo SmithKline (GSK); the remaining 450,000 participant samples are funded by a wider consortium of pharmaceutical/tech companies. Metabolomic assays are funded by Nightingale. These companies have invested in exome sequencing and metabolomic research to accelerate drug discovery. Participants of the meeting asked about the criteria for applications to use UK Biobank resources. Applicants must register with the UK Biobank as part of a bona fide research organization. For data-only application, the research must be defined in scope and be of public health interest. The criteria for access to biospecimens is much more stringent and ideally the assay should cover thee full cohort study, be well-established, and be of high scientific value to answer questions of broad scientific interest. Dr. Allen added that some duplication of research would be expected. The Biobank also presents the opportunity for open peer review, as researchers are obliged to return their results back UK Biobank to be shared by others.

Moderator: Lynne Wilkens

---

Introduction of the Keynote Speaker
Robert Croyle

Dr. Croyle provided some history of the NCI Cohort Consortium, which initially was formed to respond to a new NCI strategic priority, gene-environment interaction research, and the need to develop infrastructure and methods for this research. Dr. Croyle then introduced NCI Director Dr. Ned Sharpless.

---

Keynote Address
Ned Sharpless

Dr. Sharpless began by drawing attention to the recently released Annual Report to the Nation on the Status of Cancer, which presents data showing that overall cancer rates have declined from 1999 to 2015. A few cancers have not followed this pattern (e.g., liver, pancreatic, endometrial).

NCI appropriations have been increasing since fiscal year (FY) 2015. The 21st Century Cures Act funding has added to these increases since 2017 as a result of funding from the Cancer Moonshot initiative. Dr. Sharpless noted that FY 2019 probably will have the greatest number of new funding announcements under the Cancer Moonshot. Cancer Moonshot funds do not need to be spent in the fiscal year for which they are appropriated. Moonshot funding will end in 2023.

Dr. Sharpless highlighted the strengths of the NCI, which include:

• Basic science that drives new approaches and technologies.
• Big data to accelerate cancer research.
• Workforce development. Congress has asked the NCI to focus on workforce development, especially in fields with a shortage of specialists, such as data scientists. The first R01 funding for new investigators should be a workforce development priority.
• Innovative clinical trials that employ new designs, administrative practices, and analytic approaches. The NCI can play an important role in improving the efficiency of clinical trials, reducing redundancy, and determining the roles of industry and government.

A challenge in pursuing big data efforts is the tension between those who prioritize the protection of privacy and those who prioritize broad availability of data and transparency. Scientists need to better explain to the public the importance of large, aggregated data sets. The importance of big data might be best communicated by its importance in determining the best care for patients, because conclusions that lead to new treatments cannot be based on small samples of patients. The NCI Big Data infrastructure investments include:

• Cancer Cloud resources.
• Data Commons Framework services.
• New reporting tools for better insight into active clinical trials.
• Department of Energy collaborations.
• Prototype Cancer Research Data Commons nodes for imaging and proteomics.
• Collaborations that enable the integration of electronic medical record and insurance claims data with cancer registry data.

The Data Commons are important for integrating genomic, proteomic, imaging, clinical trial, immuno-oncology, and imaging data by applying a common data model and standards that facilitate interoperability. Dr. Sharpless highlighted the Genomic Data Commons (GDC), which includes data on a large number of patients and has large storage capability and several thousand users daily. GDC has been used in many innovative projects. GDC generates custom cohorts for genomic and clinical comparison, but clinical annotation could be improved.

Cancer Cloud resources began as Cloud Pilots for placing The Cancer Genome Atlas (TCGA) and other widely used data sets in the cloud to increase access. The NIH has negotiated with cloud service providers for low-cost cloud computing for NIH-supported investigators. At some point, cloud computing and analysis will be readily available to NIH grantees at a low cost.

Dr. Sharpless discussed NCI’s SEER cancer registry program, which covers approximately 34 percent of the U.S. population through 16 population-based registries. A current challenge for SEER registries is integrating information from the large number of pathology reports received every year. Dr. Sharpless pointed out that few other diseases have registries and that the SEER program could serve as a model for other important disease registries in the future. Important current efforts of SEER registries include the testing of linkages to pharmacy data at the Georgia registry, studies of genetic testing trends at the California and Georgia registries, and analyses of cancer disparities across the program.

Dr. Sharpless mentioned Confluence, a new research resource that combines the data of new breast cancer genetic studies and existing genome-wide association study consortia data. The program provides new cases and controls to support studies.

All of Us is an important NIH-wide initiative to collect data for personalized medicine research. Data from this initiative will be used in future studies. An important limitation is that participants are self-selected; therefore, investigators will need to consider how to conduct quality cancer research using data from the All of Us cohort.

Dr. Sharpless concluded with some challenges facing cohort research at the NCI, including:

• Balancing the research portfolio of new and old cohorts and determine what types of new cohorts will advance cancer research.
• Expanding childhood cancer and survivorship research, as required by the Survivorship, Treatment, Access and Research Act.
• Determining approaches for facilitating interactions with other initiatives, such as All of Us.
• Hiring new directors for Center for Biomedical Informatics and Information Technology and the Division of Cancer Prevention.

---

Discussion

Participants asked about the possibility of an observational cohort in the Data Commons. Dr. Sharpless indicated that this type of cohort is a possibility but that the Cohort Consortium would need to examine the type of large observational cohorts needed by the research community. He added that all Cohort Consortium data should be put in a format that is accessible and usable by the larger research community. A tiered system could be used to make sensitive data available. NCI leadership is discussing this option. The Database of Genotypes and Phenotypes (dbGaP) is an example of a security measure, but it is difficult to use. A different portal might be implemented in the future.

Cohorts are becoming more diverse and are including childhood cancer survivors, but will need to increase the diversity of participants. Congress wants research to be more generalizable to the entire U.S. population. Cohort studies increasingly will need to address health disparities experienced by special populations. An NCAB WG is addressing this concern.

Dr. Sharpless expressed interest in leveraging health care data sets developed by such private-sector organizations as Kaiser. He will be meeting with United Healthcare to discuss data-sharing opportunities. Next-generation sequencing is relevant to insurance companies, so these organizations might serve as another source of support for cancer research. Payers need to be more involved in cancer research, at least by sharing data as pharmaceutical companies do. Kaiser Permanente is leading in conducting real-world studies and making results available.

Participants asked about the implications of NIH’s new Science and Technology Research Infrastructure for Discovery, Experimentation, and Sustainability (STRIDES) initiative, which was launched to provide NIH-supported biomedical researchers access to the most advanced, cost-effective computational infrastructure, tools, and services available. These efforts are in the pilot phase. NIH will need to work with multiple cloud providers, and tiered access will be necessary. Some European data sets might be included in STRIDES.

Participants discussed the challenge of continued followup with funded cohorts. A more efficient infrastructure will be needed to expand opportunities to conduct studies using existing databases, such as VTR, national cancer registries, and the National Death Index (NDI). For example, efforts are underway to improve the speed of access and completeness of the NDI and to reduce the cost of using this database. Participants noted that grants to individual cohorts could be used to help maintain infrastructure. The VTR might be the best example of a resource that supports continuing research using cohort data.

With regard to the VTR and RTR, participants asked what could be done to reduce the legal and other barriers to using tissue blocks. Many non-NIH policies affect the use of tissue blocks, such as the requirement of informed consent after 2015, for example. As a federal agency, NCI cannot lobby for changes in legislation, but Cohort Consortium members can work to educate Congress. Legislators and their staff frequently ask what makes cancer research difficult. The first step might be to determine the source of rules and regulations that limit data access. Dr. Sharpless encouraged participants to inform him about policies that create barriers to research.

Moderator: Debbie Winn

Session VII focused on data protection requirements, fundamental changes in policy, and implications for international research (e.g., new European Union data protection policies).

---

U.S. Common Rule, Data Protection Regulations, and Applications to the All of Us Cohort
Katherine Blizinsky

Dr. Blizinsky discussed the All of Us initiative, which will serve as a rich longitudinal resource of clinical, environmental, and genetic data on a large sample of people in the United States. The initiative will strive to reach a diverse group of participants and involve researchers across all relevant fields of study. Currently, the initiative is overrecruiting from historically underrepresented populations.

Protection and privacy of human subjects is especially important to the All of Us initiative because it attempts to engage populations traditionally suspicious of biomedical research. The initiative has detailed criteria for informed consent and Institutional Review Board (IRB) approval and will adhere to the PMI Privacy and Trust Principles and the PMI Data Security Policy Principles and Framework.

Dr. Blizinsky discussed the Health Insurance Portability and Accountability Act (HIPAA) Privacy Rule, Federal Policy for Protection for Human Subjects, and the Common Rule. She noted that a new and improved Common Rule will go into effect in the near future. The new Common Rule will include new or expanded exemptions, involve a more complicated oversight system, reduce burden for convened IRBs, include new elements of informed consent, and mandate periodic review of the interpretation of “identifiability.” One challenge of the Common Rule is discerning private and identifiable information. Investigators who are uncertain about what data are considered private or identifiable should contact the appropriate NIH staff. The Freedom of Information Act has nine types of exemptions.

Certificates of Confidentiality apply to private, identifiable information collected through research. These Certificates have changed since the implementation of the 21st Century Cures Act. Certificates of Confidentiality now are mandatory.

The Precision Medicine Initiative Privacy and Trust Principles were crafted by a task force commissioned by the White House in 2015. These principles state that data access, use, and sharing should be permitted for authorized purposes only and should employ multiple tiers of access and privacy-preserving methods. These principles prohibit unauthorized reidentification and recontact of participants.

---

Discussion

A participant asked whether her consortium should adopt the All of Us model. Dr. Blizinsky recommended that participants not use this model until it has been proven effective. Other models exist that are similar to the one that used All of Us to democratize research. Dr. Blizinsky added that a centralized resource can improve protection of human subjects, but data protection and access need to be balanced. Currently, access to All of Us requires an eRA Commons identifier, but not every researcher can obtain it. All of Us investigators are seeking different methods to make data available to citizen scientists. Data resources from this initiative will become available next year.

---

European Union General Data Protection Regulation (GDPR)
Magnus Stenbeck

Dr. Stenbeck discussed data privacy protection regulation in the European Union (EU). Previously, EU member states had data legislation that adhered to EU regulations. The GDPR will obviate all of this legislation, and national regulations in the union will need to be modified or repealed. The GDPR takes priority over national EU regulations but leaves room for the addition of national regulations to set standards for privacy.

Sections of the GDPR that will be relevant to Cohort Consortium research include the following:

• What is personal data?
• When is processing legal?
• Special categories of sensitive data.
• What is the legal basis of research?
• Mandatory basic principles for all processing.
• Safeguards.
• Rights of data subjects.
• GDPR roles.
• Keeping a register
• Third-country transfers.

Dr. Stenbeck noted that it is already is more difficult to obtain personal data in Europe than in the United States. Article 6 of the GDPR outlines which legal bases exist for personal data processing. For research performed by a government agency, public university or otherwise based on legislation or a government-initiated commission, the default legal basis for processing personal data is performance of a task in the public interest (GDPR Article 6). Hence, consent is normally not the legal basis, but may be used a safeguard along with other measures such as pseudonymization of the data. Investigators must demonstrate that consent was given freely, which can be problematic with persons who are dependent on the good will of the investigator, such as hospitalized patients.

Article 5 of the GDPR is important. This Article describes the basic requirements for all handling of personal data.

In comparison with earlier legislation, GDPR means enhanced demands for documentation and organization and higher fines for breaches. In addition, GDPR requires that controllers based in the union follow the GDPR regardless of whether the data processing takes place in the union or somewhere else.

---

Navigating the EU GDPR: Implications for NIH and Early Impacts
Robert Eiss

The EU General Data Protection Regulation (GDPR) took effect in May, creating new obligations on organizations that process the personal data of European Union residents. The regulation is designed to strengthen and harmonize privacy rights across the EU member states, replacing an EU data protection directive that had not kept pace with the rapid development of digital technologies.

Although the European Parliament proposed amendments that would have imposed disproportionate limits on the use of scientific data, these ultimately were discarded due to effective lobbing by the scientific community, and others.

The new regulation now includes exemptions for research. Organizations that process personal data for research purposes may avoid restrictions on secondary analysis and processing sensitive categories of data (e.g., genetics, biometrics, ethnic origin, etc.) with appropriate safeguards, and under specific requirements transfer personal data to third countries for research purposes.

However, there are numerous instances under GDPR where member states have been given the ability to legislate at a national level, including rules governing the use of personal data in research. So there likely will be divergence among EU members states on how the research exemptions and processing of sensitive categories are applied. Many of the practical implications will be better understood as member states adopt the EU legal framework.

Following are some points of divergence between the GDPR and the US framework, and other considerations:

GDPR’s definition of personal data is broader than HIPAA’s “protected health information” and the Common Rule’s “identifiable private information”.

• GDPR definitions for anonymized and “pseudonymized” data set a higher, more difficult standard than the US framework.
• Certain categories of personal data often collected by researchers may require explicit consent under GDPR.
• GDPR creates new processes for personal data protection, which may be costly or present administrative burdens.
• A number of our joint research enterprises with universities and research institutions in the EEA have encountered complications and delays in data transfers under the new regulation. Affected projects range from genomic studies on type 2 diabetes and cancer, to access to epidemiological research databases, to multi-site clinical trials.

There must be a legal base for the transfer of personal data to countries outside of the EEA. This is the most challenging issue that we are confronting due to the options laid out by the GDPR for such transfers. The US has not been granted an ‘adequacy decision,’ demonstrating that appropriate data protections are in place, which would enable transfers to proceed without additional justification or safeguards (Article 45). The US-EU Privacy Shield does not serve that purpose for public agencies such as NIH, since it is focused on commercial transfer.

Several of our EU partners are thus relying on a legal base for transfer laid out in Article 46(2)(c) of the GDPR, which offers standard contractual clauses that have been approved by the EU Commission.

However, NIH is not able to agree to all of these contractual clauses, because (1) as a U.S. government entity, NIH cannot agree to indemnify our collaborators; (2) the U.S. government (likely) cannot agree to auditing of our data systems by a foreign entity; and, (3) NIH (likely) cannot submit to the jurisdiction of foreign courts. Our legal counsel is exploring whether we may have any flexibility with these provisions.

Currently, we are hoping to identify workable alternatives for the transfer of personal data other than contractual clauses, both in the short and long-terms. These include appropriate Article 49 “derogations” or exception and other possible options, as follows:

One mechanism for transfer personal data to countries without an adequacy decision is explicit consent (Art 49), although the conditions for consent are strict. The consent must contain enough information to be a reliable legal base for transfer; specifically, the consent documents must include information about the risk of sending data to a country without an adequacy decision. NIH could rely on explicit consent for any collection of data going forward, but this may still pose problems for data that have already been collected, unless consistent with GDPR. For some projects, perhaps re-consent might be an option, though it may be difficult and burdensome.

A second legal basis under Art 49 is public interest. We likely will need clarification on criteria and procedures if this is a viable option for health research sponsored by public agencies such as NIH.

A ‘last resort’ and more limited provision in the GDPR states that transfer may take place if the transfer is for a compelling legitimate interest, is not repetitive, concerns a limited number of data subjects (which is undefined in the GDPR), and adequate safeguards for the data are in place (Article 49(1)). The controller must provide information to the data subject about the transfer and the ‘compelling legitimate interests’ pursued, though this may already be addressed with the original consent form.

The prospect of a future, perhaps sector specific, adequacy decision of course is the broadest resolution.

The approval by the European Data Protection Board of voluntary codes that enable us to receive data from the EU, provided the provision of appropriate safeguards (Article 46(2)(e).

---

Panel Discussion

The deidentification framework in GDPR deals with commercial data transfers only. Contractual agreements could be displaced by national legislation and, therefore, are inadequate. Contractual clauses need to be negotiable, but courts might decide whether they are binding. Derogation, such as explicit consent, might be a better option.

Participants asked about consent for long-term cohorts. With large numbers of observations, consent might not be feasible.

A law is being proposed to make big international databases available across borders. This law could be incorporated into the GDPR.

Participants noted that 10 projects no longer have access to the Clinical Research Practice Data Link in the United Kingdom because of the difficulty in creating DUAs. This suggests the need to examine possible impediments to data sharing between the United States and the European Union.

Participants suggested that Governance Data Science experts should be involved in developing tools to facilitate research collaborations with EU researchers.

The NCI is attempting to develop approaches for sharing data with international organizations, such as the IARC. Definitions of such terms as anonymization might need to be agreed upon at the highest levels. Participants suggested creating a WG to focus on international definitions.

In response to questions about penalties for violations of the GDPR, Dr. Blizinsky noted that penalties involve fines, probably based on the damage caused. Data Protection Authorities within the European Economic Area member states might take a more permissive approach in the early stages of the GDPR. GDPR applies regardless of the source of the data.