During the coming decade, a generation of large-scale cancer prevention trials, early detection trials, and prospective observational studies will be coming to their conclusion. When funding for participant follow-up ends, the extensive biorepositories these trials have carefully collected will be at risk. Currently, there is no mechanism to provide funding to maintain a trial-associated biorepository once participant follow-up has ended. It is timely now, even urgent, to assure the availability of funding for the physical safekeeping and quality control of these large biorepositories and for personnel to maintain the associated databases. To address this funding shortfall for infrastructure support and importantly to assure the continued optimal use of these samples, we propose the establishment of a grant-supported, web-based national biorepository and scientific review board. This commentary discusses the importance of establishing a system for continued maintenance and assured accessibility of these biorepositories to investigators globally.
Advances in molecular and genetic epidemiology have enhanced the possibility of identifying individuals at genetic risk for cancer. Likewise, advances in proteomics and DNA/RNA technology hold the promise of developing biomarkers for the earlier detection of cancers and the identification of tumor genetic profiles, which will predict prognosis and select optimal treatment. Although these potentials are as yet largely unrealized, definitive confirmatory studies will require large populations to investigate the multiple variables and their interactions. One of the designs for these studies is the retrospective case-control analysis nested within a prospective population-based trial. This design avoids the biases of differential collection of specimens by health status and is feasible when serial biological samples are collected on the entire trial population at the time of enrollment and periodically during follow-up. Participants developing disease during the trial and matched controls who remain healthy can then be selected for a more efficient analysis. While planning, organizing, conducting, and completing large long-term prospective population trials is time-consuming and expensive, many have already been completed. Our cardiovascular colleagues have been conducting these types of observational and prevention trials for decades. Starting in the early 1980s and based on these earlier models, cancer researchers planned and initiated several large cancer prevention trials, observational studies, and early detection/screening trials. Part of the infrastructure and, in fact, one of the specific aims of many of these trials was the serial collection of biological samples that included serum, plasma, whole blood, buffy coats, and buccal swabs. In many cases, tumor samples were also collected and stored.
Table 1 describes biorepositories from some recent cancer prevention trials. Collectively, among just these repositories, there are >400,000 well-annotated samples from men and women who have been closely followed for the major cancer end points. Similar repositories exist for many early detection trials and observational studies. Although the populations recruited, the intervention tested, and the type and timing of samples collected varied from study to study, all were meticulous in sample, demographic, and end point collection. Most of these samples have been carefully stored at −80°C or in liquid nitrogen at the parent institution, commercial vendors, or the National Cancer Institute repository at Frederick, MD.
Biorepositories of recent cancer prevention trials
Trial and year enrollment began . | No. participants (sex) . | Age (y) . | End point(s) . | Intervention . | Specimens collected . | Frequency of collection . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
ATBC, 1982 (3, 4) | 29,133 M | 50-69 | Lung cancer, mortality | Beta-carotene, 20 mg and/or Vitamin E, 50 mg/d | Serum, whole blood | Baseline and periodic | ||||||
PHS, 1982 (5) | 22,071 M | 40-84 | Total cancer, CVD, mortality | Aspirin, 325 mg and/or beta-carotene, 50 mg QOD | Serum, plasma, buffy coat | Baseline | ||||||
CARET, 1985 (6, 7) | 12,025 M, 6,289 F | 45-74 | Lung cancer, mortality | Beta-carotene, 30 mg and Vitamin A, 25,000 IU/d | Serum, plasma, whole blood | Baseline and periodic | ||||||
WHS, 1992 (8, 9) | 39,876 F | ≥45 | Total cancer, CVD, mortality | Vitamin E, 600 IU and/or aspirin, 100 mg/d | Serum, plasma, DNA | Baseline | ||||||
PCPT, 1993 (10) | 18,882 M | ≥55 | Prostate cancer | Finasteride, 5 mg/d | Serum, tissue blocks | End of trial (7 y) | ||||||
WHI, 1993 (11) | ||||||||||||
Dietary modification | 48,836 F | 50-79 | Breast cancer, total cancers, CVD, fractures, mortality | Low-fat diet | Serum, plasma | Baseline | ||||||
Postmenopausal hormone therapy | 27,347 F | Premarin, 0.625 mg; progesterone, 2.5 mg/d | Serum, plasma | Baseline | ||||||||
Calcium and vitamin D | 36,282 F | D3, 200 IU; Ca, 500 mg/d | Serum, plasma, buffy coat | Baseline | ||||||||
Observational study | 93,676 F | Observation only | Serum, plasma, buffy coat | Baseline and year 3 | ||||||||
BCPT, 1997 (12, 13) | 13,388 F | ≥35 | Breast cancer, mortality | Tamoxifen, 20 mg/d | Plasma, buffy coat | Baseline | ||||||
STAR, 1999 (14, 15) | 19,000 F | ≥35 | Breast cancer | Tamoxifen, 20 mg vs Raloxifene, 60 mg/d | Serum, buffy coat, tissue blocks | Baseline and periodic | ||||||
SELECT, 2001 (16) | 35,534 M | ≥55 White M; ≥50 Black M | Prostate cancer, total cancer, mortality | l-Selenomethionine, 200 μg. and/or Vitamin E, 400 mg/d | Serum, whole blood, tissue blocks | Baseline and periodic |
Trial and year enrollment began . | No. participants (sex) . | Age (y) . | End point(s) . | Intervention . | Specimens collected . | Frequency of collection . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
ATBC, 1982 (3, 4) | 29,133 M | 50-69 | Lung cancer, mortality | Beta-carotene, 20 mg and/or Vitamin E, 50 mg/d | Serum, whole blood | Baseline and periodic | ||||||
PHS, 1982 (5) | 22,071 M | 40-84 | Total cancer, CVD, mortality | Aspirin, 325 mg and/or beta-carotene, 50 mg QOD | Serum, plasma, buffy coat | Baseline | ||||||
CARET, 1985 (6, 7) | 12,025 M, 6,289 F | 45-74 | Lung cancer, mortality | Beta-carotene, 30 mg and Vitamin A, 25,000 IU/d | Serum, plasma, whole blood | Baseline and periodic | ||||||
WHS, 1992 (8, 9) | 39,876 F | ≥45 | Total cancer, CVD, mortality | Vitamin E, 600 IU and/or aspirin, 100 mg/d | Serum, plasma, DNA | Baseline | ||||||
PCPT, 1993 (10) | 18,882 M | ≥55 | Prostate cancer | Finasteride, 5 mg/d | Serum, tissue blocks | End of trial (7 y) | ||||||
WHI, 1993 (11) | ||||||||||||
Dietary modification | 48,836 F | 50-79 | Breast cancer, total cancers, CVD, fractures, mortality | Low-fat diet | Serum, plasma | Baseline | ||||||
Postmenopausal hormone therapy | 27,347 F | Premarin, 0.625 mg; progesterone, 2.5 mg/d | Serum, plasma | Baseline | ||||||||
Calcium and vitamin D | 36,282 F | D3, 200 IU; Ca, 500 mg/d | Serum, plasma, buffy coat | Baseline | ||||||||
Observational study | 93,676 F | Observation only | Serum, plasma, buffy coat | Baseline and year 3 | ||||||||
BCPT, 1997 (12, 13) | 13,388 F | ≥35 | Breast cancer, mortality | Tamoxifen, 20 mg/d | Plasma, buffy coat | Baseline | ||||||
STAR, 1999 (14, 15) | 19,000 F | ≥35 | Breast cancer | Tamoxifen, 20 mg vs Raloxifene, 60 mg/d | Serum, buffy coat, tissue blocks | Baseline and periodic | ||||||
SELECT, 2001 (16) | 35,534 M | ≥55 White M; ≥50 Black M | Prostate cancer, total cancer, mortality | l-Selenomethionine, 200 μg. and/or Vitamin E, 400 mg/d | Serum, whole blood, tissue blocks | Baseline and periodic |
Abbreviations: ATBC, Alpha-Tocopherol, Beta-Carotene; PHS, Physician's Health Study; CARET, B-Carotene and Retinol Efficacy Trial; WHS, Women's Health Study; PCPT, Prostate Cancer Prevention Trial; WHI, Women's Health Initiative; BCPT, Breast Cancer Prevention Trial; STAR, Study of Tamoxifen and Raloxifene; SELECT, Selenium and Vitamin E Cancer Prevention Trial; M, male; F, female; CVD, cardiovascular disease.
The costs for the collection, maintenance, and use of these biorepositories and the all-important associated databases have historically been part of the parent trial. However, as these prospective population trials mature, reach their end point(s), and end participant follow-up, funding for the parent trial inevitably ends. In many cases, this leaves the biorepository in funding limbo. Currently, the National Cancer Institute has an initiative to address the important and complex issues of biorepositories (1). However, there is currently no granting mechanism available to provide ongoing support of existing biorepositories once the parent trial has ended. Although some institutions maintain and fund these biorepositories internally and others receive support from RO1 grants using their specimens, many lack stable sources of funding. These large biorepositories and the associated valuable clinical database have been carefully collected from dedicated study participants. The costs in terms of both federal research dollars and investigator effort spent in participant accrual, follow-up, sample collection and storage, and data collection are large. On the other hand, the cost for maintaining and assuring the continuity and availability of these biorepositories is small. For example, more than one million specimens are maintained and monitored at the B-Carotene and Retinol Efficacy Trial Coordinating Center for about US$105,000 direct costs per year. These biorepositories are an irreplaceable national resource. Collectively, they can provide the sample sizes needed for anticipated retrospective cancer risk factor, screening, early detection, and prognostic analyses. We have a scientific and moral obligation to provide funding to ensure that existing samples from these trials as well as many smaller trials continue to be properly stored and the associated databases maintained.
In parallel with maintaining these repositories, we must assure that specimens will be optimally used by making them available to any investigators with a sound hypothesis-driven project and promising preliminary data. As a step toward accessibility, we propose the establishment of a virtual national biorepository, which would operate with a web-accessible central database where trial investigators can “deposit” or register the samples available in their repository. With a core staff to assist inquirers/users, this virtual repository would annotate samples using common data elements, where possible, and be a resource through which any investigator could learn the availability of specimens and the associated deidentified demographics, risk factors, end points, and survival. This web-accessible repository would allow investigators to determine if a proposed hypothesis can be adequately addressed using the nationally available specimens. Of course, such a web-based repository must have assurances and appropriate institutional review board (IRB) approval in place to protect the study participants, to prevent unauthorized use of trial data, and to guide publication.
The National Cancer Institute is already managing large specimen archives at Frederick, MD; one policy to be considered is consolidation of mature biorepositories at that site. However, because many of the existing repositories are already large, there may not be much economy of scale in amassing them centrally, although, for smaller specimen archives, consolidation may be warranted. In addition, shipping specimens to the central location could result in damage. Finally, as in serious but localized disasters, such as Hurricane Katrina, retaining all the specimens in a single location would put the entire bank at risk from a single catastrophic event. Hence, a virtual repository seems a more realistic and safer option than a repository where samples are moved to a centralized location. It would also allow trial investigators to maintain their samples at their current storage site. A biorepository core grant would provide funds for maintenance and establish storage and quality control policies.
Most of these biorepositories and their associated databases are and will continue to be used productively by the institutional investigators who are most familiar with them. However, there are advantages to opening these repositories to investigators nationally. Most of these large population trials and their lead investigators tend to focus on specific disease end points, such as neoplastic, cardiovascular, neurologic, or pulmonary diseases (areas where they have specific expertise and interest). However, trial participants develop multiple disease end points, many of which lie outside the primary interest of the trial investigators. Investigators with a focus in these end points and not associated with the trial may not know of the availability and applicability of specimens to their fields of research. For example, the B-Carotene and Retinol Efficacy Trial biorepository has serum samples on >1,000 individuals with cardiovascular death end points that have not been accessed. In addition, the expertise of the trial investigators may lie outside the newer areas of proteomics, metabolomics, or DNA/RNA technology. Over the long life of these trials and the repository samples, investigators move, retire, and/or change their research focus. Without a champion to promote and acquire funding, many biorepositories become property of the parent institution. Once trials have published the primary end points and have ended follow-up, they can fade from view, and the very existence of the trial-associated biorepository may be forgotten by the research community. This can lead to specimens remaining unused and sitting in aging freezers. These repositories have a limited life span; samples, even the most meticulously stored and cared for, will age and degrade and with time may no longer be suitable to address specific hypotheses.
To assure specimens are used optimally, we also propose the establishment of a rotating board of trial-related and nationally recognized researchers similar to an National Cancer Institute–sponsored study section. The sole purpose of this board would be to review and prioritize projects requesting samples from the national biorepository. Because this board would include sitting members from many fields of research and ad hoc members with a wide range of expertise, they would be better equipped to review and prioritize the broad spectrum of research proposals than would the gatekeepers of most individual repositories. With the board knowledgeable of the contents of many repositories, they could better identify the most appropriate repository(ies) for a project, something individual repositories are not equipped to do. Grant applications for projects using samples in the biorepository would be required to include a letter of approval from the review board.
Currently, most of the large biorepositories have a formalized structure for collaborations with other investigators and to allow access to the samples. We anticipate that a national review board would not and should not replace or usurp the decision making of the individual repository's review committee but rather would prioritize research proposals submitted to it and steer those proposals to the most appropriate repository. The individual repository would still implement their established mechanisms for reviewing proposals and may or may not agree to provide specimens. However, through this route, the quality of research projects, selected from investigators nationwide, would be enhanced and the repository gatekeepers would feel assured that their specimens are being used optimally. Orphan repositories or those led by many coinvestigators may choose to allow the national review board sole control of access to their samples. Funding agencies may use repositories' shown willingness to release specimens to well-reviewed external research proposals in deciding the level of continuing support. The review board could provide oversight for the funding agencies.
IRB approval and the need to obtain informed consent remain concerns for large collaborative studies where samples are obtained from participants under the jurisdiction of many different IRBs. The need for consistent IRB review has proved especially challenging when genetic analyses are planned. As part of the B-Carotene and Retinol Efficacy Trial biorepository, we negotiated with nine U.S. institutions to establish a precedent setting agreement for central IRB review of proposals (2). This allows any investigator wishing to access B-Carotene and Retinol Efficacy Trial samples obtained from participants enrolled at multiple institutions to submit a proposal to a single IRB. This type of agreement would be valuable for a national biorepository to assure that participants' rights are protected while ensuring that IRB review can be conducted expeditiously.
It is in all investigators' interest and should be regarded as an obligation to trial participants to establish a system that assures that the biological samples so generously donated and carefully collected are maintained, monitored, and used in the most efficient manner possible. We urge the National Cancer Institute to support and collaborate with trial investigators and move forward in the establishment of a web-based national biorepository for existing and future specimen banks. Let's not lose what we have so carefully gathered.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Acknowledgments
We thank Benjamin Bodi for research assistance.