PS1-11: Specimen Collection Within the Cancer Research Network: A Critical Appraisal

Background and Aims: The Genomics Working Group (of the Cancer Research Network (CRN) Pharmacovigilance Project) explored single-site experiences with genomic studies to understand challenges and opportunities for collecting and using biological specimens. Its specific aims were to: describe the availability of existing biospecimens; document successful past specimen collection strategies; identify potential barriers to selected recruitment strategies; and develop standardized study documents for future genomic research. Methods: We distributed a survey on existing biospecimen availability to principal investigators at all CRN Pharmacovigilance sites. Information collected included the number of specimens, storage methods, and type of specimens available. Investigators were also asked to identify past and current studies at their institutions in order to collect information on successful specimen collection strategies. A literature review was conducted both to identify potential barriers to participation and to understand the feasibility of various recruitment procedures. Benefits and limitations of ‘opt-out’ and ‘opt-in’ specimen collection methods were examined. Lastly, we developed study documents, including recruitment letters and informed consent forms, for future saliva and blood specimen collection efforts. Results: Two of the CRN HMOs currently have large-scale banked blood or saliva specimens (>20,000) within existing repositories. One site has two large-scale studies (>20,000 participants each) with banked specimens able to provide DNA for future research. Whole blood or blood components are the most commonly collected specimen types, and multiple CRN sites have access to banked tumor blocks. There were 33 past or current studies that included biospecimen collection. There was a trade-off between intensive data collection requirements and recruitment yield, cost, and participation burden. Conclusions: There is great potential for collection, storage, and use of biospecimen samples for research within the CRN. For future collection efforts, ‘opt-out’ models allow fast sample accrual, but impact data de-identification requirements. ‘Opt-in’ models allow for broad data collection, but are resource and time intensive. Pathology time and support may be a barrier to future use of stored samples.