M2Gen, the for-profit subsidiary of Moffitt Cancer Center, has launched an analysis service that uses a database of clinical and molecular information from consenting patients to select participants for clinical trials.
The system is targeted primarily at pharmaceutical companies, Timothy Yeatman, Moffitt's executive vice president and founding chief scientific officer, told BioInform this week.
The system uses software called TransMed that Moffitt developed in partnership with Oracle and Deloitte. The software searches the center's database of clinical data and gene expression profile information to locate patients that meet the required criteria for pharmaceutical drug trials, Daniel Sullivan, M2Gen's chief medical officer, explained.
The database contains clinical information on 85,000 patients and has gene expression profile data on about 18,000 of these individuals, Sullivan said.
The data is collected from consenting patients at Moffitt and 17 other sites that are part of a network of hospitals and cancer clinics that make up the Total Cancer Care consortium — a Moffitt-led effort to provide individualized, evidence-based treatment decisions for cancer based on the large-scale integration of information technology, scientific discovery, and health outcomes.
Sullivan said that M2Gen has so far signed contracts with several "large US pharma companies" for services based on the system.
Some of the services that M2Gen is providing to these firms include identifying and validating biomarkers for specific diseases ahead of the trial, as well as selecting patient cohorts and signing them up for trials. Sullivan said that the cost of the service is based on the specific project needs.
Although Sullivan could not disclose details about most customers for confidentiality reasons, he did say that Merck — which provided some of the funds that were used to launch M2Gen in 2006 — is in "the very early stages" of using the system.
A Better Approach to Cohort Selection
Moffitt's system selects cohorts based on molecular profiles; disease/diagnosis; symptoms; and demographic information and family history.
In the case of a drug company with a cancer drug candidate that targets a particular receptor, "we would look to the database for patients that meet the ... criteria for their trial and also have over-expression of that target based on gene expression profiling," Yeatman explained.
The TransMed software sits on top of the database and uses a series of filters to select patients based on the trial's inclusion/exclusion criteria, he explained.
For example, in a trial for breast cancer treatments, the system would select patients from the database that have the disease and then it might filter the short list of candidates to find patients with invasive ductal carcinoma, and then further filter that list to find patients that are estrogen receptor-positive and who have undergone microarray analysis and whose results show an over-expressed signature, Yeatman said.
The system is currently being used to select patients for cancer clinical trials but it could also be used to select participants for trials in other diseases such as cardiology, rheumatology, and neurodegenerative disease, he said.
Ultimately, having an existing pool of patients speeds up the drug development process since it shortens the time that would be spent on recruiting and screening patients who meet the trial criteria, Moffitt said.
Moreover, using a combination of molecular genetics and clinical data makes it possible to develop "molecular fingerprints for each patient's tumor" that can later be used to identify similar subpopulations of patients, Yeatman said.
Since all that information is concentrated in a single database, it's possible to find participants for clinical trials faster than it would be to select patients prospectively or retrospectively, he noted.
In these cases, participants are screened "as they walk through the door" to see if they meet the trial requirements; or patients might be screened for a "standard" set of target genes that are believed to respond positively to a particular treatment, he explained.
For example, the ongoing I-SPY2 breast cancer trial, which is using tumor biomarkers to assign patients to particular treatment regimens, accepts women with tumors of a certain size and then screens them to find out more about the cancer and to discuss their eligibility to participate in the treatment phase of trial.
"It's a very long and laborious task" to select the final participants using a prospective approach, Yeatman said.
Sullivan said that Moffitt doesn't see much competition for M2Gen's services. He noted that while contract research organizations like Covance and reference labs like Laboratory Corporation of America offer CLIA testing for biomarkers that could help with choosing trial participants, there aren't any firms that offer a "complete package" covering preclinical testing activities though to actual cohort selection
Furthermore, the addition of the TransMed software tool is "pretty novel," Sullivan said. "I don’t know anybody else who has that."
Have topics you'd like to see covered in BioInform? Contact the editor at uthomas [at] genomeweb [.] com.