Harvard hopes database will speed cancer cures
Since World War II, many cancer patients who have had surgery at a Harvard-affiliated teaching hospitals have left a small piece of their tumor to science.
These clumps of human cells have been frozen in liquid nitrogen or preserved in paraffin blocks the size of small Post-it notes -- and they now fill giant freezers and floor-to-ceiling shelves in hospital basements and off-site warehouses.
The value of this tissue trove has soared in recent years with the successful cataloging of humans genes. Researchers need to study hundreds of specimens to find genetic mutations, proteins, and other molecules linked to cancer, in hopes of developing new medicines and tests to diagnose cancer early and help customize treatment for individual patients.
But there are so many samples in so many different places -- each of the five major Harvard hospitals has its own central ''tissue bank" and individual departments maintain their own stores -- that researchers often don't know where to find the specimens they need. Even when they do, it's cumbersome to negotiate each hospital's and department's rules for using them. And some researchers have been stymied by colleagues' reluctance to share specimens that they have painstakingly collected for their own studies, particularly when the cancers are rare and the samples small -- meaning there's enough to use in only a few experiments.
''You spend hours on end trying to locate the specimens," said Dr. Massimo Loda, a pathologist at Dana-Farber Cancer Institute and Brigham and Women's Hospital. And then ''you can't just walk into another institution and say 'I want 500 cases of your prostate cancer.' "
Now the Harvard hospitals are establishing a ''virtual specimen locator" to encourage sharing among researchers and speed up discovery of life-saving cancer screening tests and treatments -- a concept they hope to expand to the University of California, Los Angeles, the University of Pittsburgh, and the University of Indiana School of Medicine.
The Harvard hospitals this summer took the first step, loading descriptions of about 1 million tissue samples from the central hospital tissue banks onto a shared database. Samples have been saved since World War II, but only those kept since the mid-1980s are recorded in electronic databases and used for the virtual specimen locator.
Harvard researchers can search the database, which contains information about each patient's diagnosis and date of surgery and specifics about the tumor, though not the patient's name or other identifying details. They can then request tissue under a new set of standardized rules. Directors of the tissue banks and boards at each hospital that oversee research must sign off before researchers can pick up the samples.
Harvard eventually hopes to expand the database to include the department tissue banks and the out-of-state hospitals so that researchers at all the institutions potentially would have access to millions of cancer samples. The National Institutes of Health has supported the effort with a $7 million grant and is deciding whether to give another $7 million. The project faces a number of ethical and legal hurdles because the hospitals must develop ways to protect patients' identities, as information about them and their tumors rockets around the country via the Internet.
Loda was one of the first researchers to use the virtual specimen locator. He is looking for biomarkers in prostate cancer, molecules that predict how aggressive a particular cancer will be. He needs to look at cancers from thousands of patients to end up with a subset with a particular biomarker. He began by searching the database for prostate cancer specimens since 1990, and the computer coughed up 6,284 samples. After getting approval from the hospitals' institutional review boards, he was able to search the pathologists' reports on the tumors, which are in the database, to identify tumors from patients between 50 and 70 years old who had a common but unpredictable type of prostate cancer, giving him about 500 cancer samples. He's requested those from Brigham and Women's and Massachusetts General hospitals and Beth Israel Deaconess Medical Center.
He's never tried to gather such a large number of tissue samples before. ''The best possible outcome is that we would identify patients who could benefit from a new drug," Loda said.
But allowing researchers to easily locate tissue is one matter; convincing those in charge of the individual tissue banks to actually give it away is another.
Researchers who run the department banks have invested hundreds of thousands of dollars into not only collecting the tissue, but in obtaining consent from patients to collect detailed medical information and contact them years later about whether their disease progressed or was cured. This information is crucial because it allows researchers to connect a patient's eventual outcome to specific mutations. The central hospital tissue banks don't contain as much medical information, so convincing doctors who oversee these individual banks to participate in the virtual specimen locator is key.
Dr. Chin-Lee Wu, who heads the prostate cancer bank at Mass. General, has collected several hundred cancers during the last four years. His group is leaning toward allowing only half of their patient specimens -- the ones of which they have extra -- onto the database. Prostate cancers are small, producing little extra tissue for research, and they're slow-growing, so patients need to be followed for years before doctors know if a treatment worked.
''I don't want to receive 10 requests per day and then I have to explain to them that I can't accommodate everybody," Wu said. ''The tissue we have so far will be very, very valuable in 10 years."
