Genomic Test Improves Forecast of Recurrence in NSCLC

DURHAM, N.C., Aug. 9 -- Using a form of genomic profiling for early non-small-cell lung cancer (NSCLC), researchers here believe they can cherry pick those patients who will benefit most from adjuvant chemotherapy.

In laboratory models, the so-called Lung Metagene Predictor was up to 90% accurate in picking which patients would have a recurrence of their cancer after surgery, according to Anil Potti, M.D., of the Duke University Medical Center.

"We now have a tool that can be used to move these high-risk patients from the 'no chemotherapy' group into the aggressive treatment group," Dr. Potti said in a statement.

The test, based on gene microarray technology, was significantly better than clinical prognostic factors at predicting which patients in several clinical trials would go on to suffer a recurrence, Dr. Potti and colleagues reported in the Aug. 10 issue of the New England Journal of Medicine.

In fact, it was so successful in the lab that the researchers have started a multicenter randomized clinical trial, which will begin enrolling patients within six months, according to David Harpole, M.D., also at Duke, who will be principal investigator of the trial.

"If we can use the test to increase patient survival by even 5%, we would save 10,000 lives a year," Dr. Harpole said. "In reality, I think we can do much better than that."

The test is one of the first instances of so-called "personalized medicine" that can actually be used to change a clinical decision, said Joseph Nevins, Ph.D., a molecular geneticist at Duke and senior author of the study.

"Instead of placing all patients with small tumors in the same early-stage category, as physicians currently would do, we can now assess their risk based on the tumor's genomic profile," Dr. Nevins said in a statement.

The researchers noted that adjuvant chemotherapy is recommended for patients with stages IB, II, or IIIA NSCLC, but not those with stage IA. Nonetheless about 25% of stage 1A patients have a recurrence, suggesting that the staging system is imprecise.

Using tissue samples from a cohort of 89 NSCLC patients at Duke, the researcher analyzed the genetic make-up of the tumors and zeroed in on a collection of genetic profiles that appeared to be associated with recurrence.

In the Duke cohort, the final metagene test was 93% accurate in predicting recurrence, compared with 64% accuracy using a model based on standard prognostic factors, including age, sex, tumor diameter, stage of disease, histologic subtype, and smoking history.

To confirm the results, Dr. Potti and colleagues analyzed tissue samples from 25 patients in the American College of Surgeons Oncology Group Z0030 study and from 84 patients in the Cancer and Leukemia Group B (CALGB) 9761 trial. They found:

• In the American College of Surgeons Oncology Group patients, the metagene test was 72% accurate.
• In the Cancer and Leukemia Group B patients, the test was 79% accurate.

The researchers also calculated a C statistic for both cohorts, where a result of 0.5 means there is no information and a result of 1.0 means the test perfectly predicts an outcome -- in this case, recurrence.

For the American College of Surgeons Oncology Group patients, the C statistic for the standard prognostic model using clinical data was 0.67; adding the genomic data increased the value to 0.84. For the Cancer and Leukemia Group B patients, the clinical model had a C statistic of 0.73; adding the metagene data increased it to 0.87.

"Clearly, the genomic data transformed a limited clinical-based prognosis to one with substantial capacity to identify patients who were likely to have disease recurrence," the researchers concluded.

They also focused on a subgroup of patients, the 68 patients from the Duke, American College of Surgeons Oncology Group, and Cancer and Leukemia Group B cohorts who were classified clinically as having stage IA disease.

Kaplan-Meier survival curves were generated for the group as a whole, as well as for the subgroups predicted to be at high or low risk for recurrence by the lung metagene model. Although the survival rate for the group was approximately 70% at four years, the survival rate for those predicted to be at high risk was less than 10%, thus identifying the subgroup of patients with stage IA NSCLC at risk for recurrence.

Drs. Nevins and other authors reported holding equity in Expression Analysis, a DNA microarray service provider established by Duke University. Drs. Nevins and other authors reported having served on the advisory board of Expression Analysis. Dr. Harpole reported having served on the advisory board of Genentech (OSI Pharmaceuticals).