One of the National Community Cancer Centers Program sites setting up its own biorepository is Christiana Hospital in Newark, Del., where Mary Iacocca, MD, is chief pathologist and medical director of the tissue procurement section. Dr. Iacocca headed up the original team that established tissue procurement at Christiana in 2003; her section is part of the Helen F. Graham Cancer Center of the Christiana Care Health System, which serves patients in Delaware, Maryland, Pennsylvania, and New Jersey.
Nicholas J. Petrelli, MD, the Bank of America endowed medical director of the Helen F. Graham Cancer Center, came to Christiana in 2001. A surgical oncologist with experience in tissue procurement, Dr. Petrelli says he sized up the natural resources: excellent pathology work on campus, a variety of pathology, affiliations with outstanding research institutions, and excellent physicians participating in the cancer program. “If you put those components together,” Dr. Petrelli says, “quality, clinicians, scientists, and a variety of pathology, plus a state that used to be No. 1 in cancer incidence and mortality, it made sense to me that there were components for a translational cancer research program.”
The staff supported setting up an independent biorepository and found strong backing within the administration. “The two biggest obstacles to starting a biospecimen procurement facility,” Dr. Iacocca says, “are money and manpower. There has to be a very strong coordinated effort on the part of the hospital administration and the cancer center.”
Dr. Iacocca and her team attended a tissue procurement training workshop in 2003 sponsored by the University of Alabama Comprehensive Cancer Center at Birmingham and the Cooperative Human Tissue Network. William Grizzle, MD, PhD, a professor of pathology and former president of the International Society for Biological and Environmental Repositories (ISBER), who coordinates the workshops with Katherine Sexton, assistant director of the UAB tissue repository, says the repository has been a part of the comprehensive cancer center since 1981. (ISBER is a division of the American Society for Investigative Pathology.)
“In 1987, we became one of the original sites of the Cooperative Human Tissue Network,” he says, which is an independent, NCI-funded group of six organizations experienced in providing human tissue for research. The NCI started the network to increase access to human cancer tissue for basic and applied science in academia and industry.
“The Cooperative Human Tissue Network is different from tissue banks in that it’s a prospective tissue selection effort in which we actually get requests from investigators and try to meet their exact requests,” Dr. Grizzle says.
“One of our goals is to provide as much tissue and information as we can to investigators,” he adds, so local institutions can set up their own biorepositories. “Then we’d supplement what is available from the local repositories. That’s our goal in life,” he says.
Dr. Grizzle stresses that different researchers have different needs. “Some investigators seek what tissues they can obtain when they can be obtained; I call that the catch-as-catch-can approach,” he says. These types of samples are not usually tied to quality assurance or standard operating procedures. Tumor banks would be the next category; there investigators have the benefit of a standard operating procedure and a large number of tissues related to a known disease, as well as information about clinical outcomes when available.
In prospective tissue procurement, Dr. Grizzle says, “investigators can say exactly how they want tissue collected—fresh, frozen, fixed, paraffin blocks —and we try to meet their needs. But the investigator has to wait until we collect specimens and it’s early in the disease process, so they have to wait for clinical outcome.”
The workshop in Birmingham ran several days, Dr. Iacocca says, and she and her team returned home with much to read. Duties were parceled to work groups. IT staff built a database to capture the clinical and pathologic information that researchers would need. Others worked out critical details—access to liquid nitrogen, a –80°C freezer, a mechanism to keep the database under locked password.
The medical staff at the Graham Cancer Center recognize they are laying a foundation. “The key was not saying we had a timetable and a specific goal of collecting a specific number of specimens,” Dr. Petrelli says. “There was no pressure, nothing like, ‘We have to collect 100 colorectal or thoracic.’ There will be [some pressure] if we participate in the cancer genome atlas project, but we have such a strong infrastructure now that I am not concerned.”
The Christiana biorepository started with lung cancer specimens. “Our cardiothoracic surgeons identify patients with tumors that are at least two centimeters and patients are consented prior to the day of surgery,” Dr. Iacocca says. “We send to the operating room a collection bucket on ice. As soon as the lung cancer is harvested from the patient, it is placed on ice and sent immediately to the pathology laboratory, where a pathologist evaluates it.” If there is sufficient tissue to collect, they divide approximately a cubic centimeter into five aliquots under sterile conditions: four snap-frozen in liquid nitrogen and one processed by routine methods for quality control. The snap-frozen pieces are stored at –80° indefinitely in a freezer that is alarmed with a backup electrical supply. Dr. Iacocca confirms the diagnosis with the surgical pathology report and provides information about the tumor, and the anonymous data are entered into the databank.
Consents proved to be among the biggest challenges. “There are two aspects,” Dr. Iacocca says. “One is educating the patient about what it is they’re doing when they volunteer to donate tissue. The second is feeling confident that the consent is truly informed. That’s the trickiest part of the consent.” At Christiana Care, oncology research nurses ensure that patients know their tissue will be anonymous, that there would be no possibility of personal benefit from the research. “Patients are okay with that,” she adds. “They just want future people to benefit.”
Physicians at Christiana are engaged in quality assurance activity with the NCI, formulating a plan to evaluate their ability to comply with best practices. “Some of the [best practices] recommendations are very specific and not easily attainable,” Dr. Iacocca says. “One example would be for us to have knowledge of the exact anesthesia time for a patient undergoing surgery. While that’s best practice, we would have to work very hard for pathologists to document it. So we were asked to identify officially what would need to be done to come into alignment with these recommendations.”
Rajesh C. Dash, MD, director of the Duke University Health System Comprehensive Cancer Center Tissue and Blood Procurement Shared Resource, is creating a system to collect, store, and share frozen and paraffin-embedded tumor samples and corresponding blood components for cancer research. The work at Duke builds on a tissue banking initiative launched in 2000 in collaboration with the for-profit Ardais Corp. in Lexington, Mass.; that activity continues under the direction of Alan Proia, MD, PhD. The shared resource has become part of Duke’s Comprehensive Cancer Center.
To launch the tissue bank in 2000, Dr. Dash explains, it was necessary to decide what kinds of tissue would be saved in the biorepository. “The most desirable specimens were those that had significant pathology associated with them, such as cancer,” Dr. Dash says. “There are some noncancer ones, such as those associated with arthritis, for example, that were also in demand, so a short list of desirable specimens was generated. You have to think about how to bank, the manner in which samples are to be collected, and what proportion of frozen versus formalin embedded will be stored.”
Tissues for gene expression studies are far more useful when frozen immediately after removal from the patient, because freezing limits RNA degradation that occurs at room temperature. Still, while many of the protocols for molecular analysis have been worked out on frozen tissues, researchers have begun to identify better ways to use formalin-fixed, paraffin-embedded samples, so their usefulness will be revisited from time to time.
“The problem with frozen,” Dr. Dash says, “is that you store at minus 85, or in some cases minus 200, and that requires a lot of infrastructure support. You have to have it on an alarm and have service and support available 24/7, even on holidays.” Maintaining an environment for formalin-fixed, paraffin-embedded tissue is far easier. “So there are opportunities for those who have not jumped into biorepositories as of yet; it would be worth examining whether an infrastructure based solely on paraffin-embedded tissue could be investigated for many disease processes. Pathology laboratories, of course, already have an infrastructure to create a formalin-fixed, paraffin-embedded repository because that’s what we’ve always done, so it’s simply a matter of segregating the research from the clinical diagnostic samples.”
Still, Dr. Dash says, the number of analyses suited to formalin-fixed, paraffin-embedded samples will continue to decline because RNA integrity will always be compromised to some degree. “One of the big issues with formalin that big studies don’t convey is that the amount of time a tissue is exposed to formalin will greatly affect its usefulness for molecular analysis,” he says. “In the past, when they used formalin fixation, they never had a need to track the time a tissue spent prior to processing or the time spent at various temperatures.” With the large studies that are contemplated for molecular diagnostics, it will be critical to be able to harmonize data from several institutions so that researchers can draw accurate conclusions without wondering whether collection or storage methods have affected results.
“In many newer protocols, you can use formalin-fixed, paraffin-embedded tissue, but you have to be sure all the tissues have been in paraffin the same amount of time, have undergone the same type of processing, and you’re cognizant of the details associated with that protocol,” Dr. Dash says. “Freeze it and you’re done. You don’t have to worry about it so long as you keep it at the certain temperature to slow down the degradation. With paraffin, if your air conditioning breaks down and it goes to 80° or 90° in the storage facility, that’s going to be a problem. So there are still a lot of issues with paraffin, but if laboratories can establish standards and keep a detailed record of how it is stored, it would still hold great value.”
Standardized collection methods for frozen samples are every bit as critical, however, and minute deviations need to be well documented if biospecimens are to be useful for large-scale, interinstitutional research. Banking 20 samples from a 5x5x5 mm3 tissue, for example, may take 20 minutes, and it may not be clear to the researcher that the difference in the results from tumor in sample No. 1 versus No. 20 is a reflection of the order in which they were submitted and the associated delay prior to freezing or intrinsic properties of the tumor biology being studied.
Dr. Dash recommends that those contemplating a biorepository spend time on the NCI Web site studying the best practices documents that can be downloaded for free. There is a lot to consider—from types of labels that will adhere at –200° to the right IT support to print human-readable bar codes. A sophisticated quality assurance program, Dr. Dash says, is critical. “The important thing is to have a design to maintain the repository and get it all documented and your people trained,” he says.
Data integrity presents separate issues, the first of which is identifying the clinical data elements you will want for each of the specimens. “This may be different for every protocol,” Dr. Dash says. “You have to decide if specimens for research need to be validated by a pathologist to make sure there’s no compromise in clinical care. If the tissue would have otherwise been thrown away, then we don’t require that it be validated by a pathologist.” However, they also have a protocol whereby they take cores of breast tumors, and those cores might contain information that is not contained on the diagnostic cores. So their Institutional Review Board requires them to validate that a pathologist has looked at and compared the diagnostic core with the research core to ensure that clinically significant discrepancies have been brought to the attention of the clinical care personnel. “In that event, the core is moved over to the clinical side and becomes part of the patient’s medical record. It is no longer available for research,” Dr. Dash says.
“One of the primary hurdles, and perhaps the most feared, are the IRB requirements,” he adds, because without a well-designed consent and IRB protocol, investigators in the future could be precluded from carrying out the studies they wish to complete using the biorepository.
Mariann Stephens is a writer in Chicago.