June 13, 2007 | It’s being called a “biobank on steroids.” The H. Lee Moffitt Cancer Center and Research Institute in Tampa, Florida, is building a gene expression database that will link the molecular profile of cancer tumors with other clinical and historical information about patients in the hope of developing personalized medicines.

The project has received the tactic endorsement of Merck, which has reportedly committed $95 million to help create a cancer tumor research center near the existing Moffitt site. Merck is “further along in its thinking about population-based trial matching” than other large pharmaceutical companies, says Charles Resnick, interim CEO & president of Moffitt Technologies Corporation.

Currently, response rates to experimental cancer drugs are low — as low as 10% in early phase trials — causing many potentially useful drugs to be prematurely retired, says Resnick. Moffitt aims to reverse the traditional clinical trial model by  looking for patients in its gene expression database that are predicted to respond favorably to a therapy based on their profile and what is known about the drug.

Drugs of the future will theoretically target a specific pathway, possibly across two or three tumor types, Resnick says. Patients “may not even be able to get a drug” without diagnostic evidence that it will help them, much like Medicare now limits payment for Herceptin to breast cancer patients who have a specific genetic marker, indicating the drug is likely to be therapeutically useful.

To “operationalize” the Merck-Moffitt collaboration, the cancer center established a for-profit company called M2Gen. Resnick notes that it is attracting $35.5 million in public funds, because it is expected to be the catalyst for a robust biotechnology industry in the Tampa Bay area. The collaborative has also committed to creating 165 mostly high-level jobs.

Resnick says he is already fielding calls from study sponsors, “especially smaller biotechnology companies” that want access to the database. Other IT companies from the Northeast and West Coast want to collaborate with M2Gen to add yet more tumor-related information to the database that can be correlated with clinical data and outcomes.

The endeavor is pricey because of the requisite IT infrastructure, the manpower cost of data collection, and the $1,500 per tumor price tag for molecular profiling, says Resnick. Between 20,000 and 30,000 tumor tissue samples will be collected over the next five years from eight Florida sites, including Moffitt, as well as two out-of-state hospitals. The database currently contains about 4,000 samples and will ramp up to 6,000 per year once all ten sites are actively participating. Simply pulling together the consortium has not been easy “politically,” says Resnick.

The database is expected to have sufficient information for trial matching by 2010. For a fee, any pharmaceutical company will be able to have an experimental drug matched to patients whose cancer has a similar pathway. “If they know we can give them a 60% response rate...there’s nothing they won’t do to get that data,” says Resnick. Trial matching is “the most probable way” M2Gen will make money. “Merck will go away at some point.”

Merck’s investment in the Moffitt initiative buys it exclusive access to the database for purposes of mining for new gene and drug targets, says Resnick. If this leads to the development of a commercially successful drug, Moffitt would reap “some percentage of revenues.”

Moffitt has already put $20 million into a modern data center with a pair of IBM super computers and the development of clinical system interfaces, query algorithms, and standard operating procedures for molecular profiles. Quickly moving tissue samples from a surgery center to pathology and then onto Moffitt for macro-dissection is a “logistical nightmare,” says Resnick.

At the moment, Moffitt is constructing a front-end search engine able to do Boolean searching for patients based on a complex set of criteria, says Resnick. One day, he envisions patients and their physicians being able to view personal medical records in the database from anywhere in the world. They might also go there to get the “survival curve” and best treatment option for a particular cancer type.

The initiative is fully electronic, utilizing a touch screen system of Maitland, FL-based Galvanon to collect data from patients and get their electronic consent via digital signature capture. The absence of paper, Resnick says, is “huge in terms of the ability to go back and query the database.” Paper consent forms are often difficult to find during audits by the U.S. Food & Drug Administration.

Not all scientists are sold on the promise of molecular profiling, Resnick adds, but most believe it will be valuable in “matching the right drugs to the right patients. It either will be or won’t be. This is where Merck, to some extent, is rolling the dice.”

Subscribe to Bio-IT World  magazine.