Despite a growing armamentarium of molecular techniques available to drug developers and clinical researchers, accurate prognosis of ovarian cancer still depends largely on clinical characteristics, along with a few genetic and biological factors. But it appears that molecular determinations may be finally appearing on the scene.
The Ovarian Cancer Prognostic Profile has so far been able to differentiate the outcomes of late-stage ovarian cancer cases using the expression profile of 115 genes with better predictive capability than methods currently in use. The OCPP divides patients into “favorable” and “unfavorable” prognosis groups, a strategy that has the potential to save drug developers time and money by allowing them to stratify trial subjects, according to OCPP officials. It may also help clinicians by enabling them to use a version of the test to specifically tailor treatment strategies.
“Ovarian cancer is a pretty aggressive disease, so right now we use our most aggressive therapy for pretty much every patient,” said Andrew Berchuck, a professor of Gynecologic Oncology at the Duke University Medical Center. “I think the idea is that if we knew patients are destined to do poorly with our present aggressive therapy — you watch those patients a little more carefully and move them more quickly to experimental therapies, biological therapies,” he said.
Alternatively, if a clinician knows ahead of time that certain therapies are not going to be successful, he or she has the option of avoiding their toxicity, said Berchuck.
Using the approximately 12,000 genes available on Affymetrix U95A2 Gene Chips, Dimitrios Spentzos, who developed the OCPP with colleagues at the Beth Israel Deaconess Medical Center in Boston, used statistical methods to narrow the field of relevant data to the 115 genes. After identifying the important genes in 34 patients, the OCPP was able to separate 34 additional late-stage ovarian cancer patients into two groups with “dramatically different likelihood of being alive several years down the road,” Spentzos told Pharmacogenomics Reporter. “It was a pretty robust performance,” he added.
The median survival for the “favorable” and “unfavorable” prognosis groups in the study were 47 months and 30 months, respectively. The results of the study appear in the Dec. 1 Journal of Clinical Oncology.
Because ovarian cancer is rarer than many other forms of cancer, a dataset of 68 ovarian cancer patients, “even though relatively small in absolute terms, is probably one of the biggest you’re going to have in this discovery phase,” Spentzos said. Subsequent prospective validation studies will probably feature “several fold” more subjects, he said.
“If I have a new drug, which is in development, who am I going to test it on?” Spentzos. By testing new drugs on patients with the poorest prognoses, “first of all, you are giving this group of patients another chance; second, you are likely to find which therapies or drugs can add” most effectively to therapies already in use, said Spentzos.
Armed with information about some of the 115 genes revealed by the OCPP, cancer researchers may be able to fashion therapies directed at known mechanisms, said Spentzos. The OCPP will probably not become a common fixture of stratification studies until the profile of associated genes is refined to the point where PCR and other easily standardized assays can be used, he said. “We want to make it into a real clinical tool,” although that process will take at least five years, said Spentzos.
Such a tool could allow clinicians to perform “minimal therapy” with patients expected to do well, avoiding “overly aggressive” treatment, whereas clinicians might prescribe additional therapies for patients diagnosed with less favorable outcomes, said Spentzos.
No diagnostic or drug companies have shown interest in the research yet, Spentzos said.