WASHINGTON—An unusual clinical trial involving four different drugs offered promise that guiding treatment based on the molecular traits of a tumor can improve survival from lung cancer.
Researchers said the study amounted to "proof of concept" for a new approach to clinical trials that could improve the efficiency of cancer-drug development and eventually shorten the time it takes to get new treatments to market.
The study, called Battle, involved 255 patients with advanced lung cancer, an especially lethal disease in which until recently average survival on chemotherapy was about eight months. New drugs, including so-called targeted medicines such as Tarceva, co-marketed by OSI Pharmaceuticals Inc., and Roche Holding AG, have achieved some improvement in survival, but typically in only a fraction of patients.
Doctors don't know which patients are likely to respond to which treatments, resulting in hit-or-miss use of high-cost medicines.
Researchers at the University of Texas M.D. Anderson Cancer Center mounted the Battle trial to see whether matching tumor characteristics called biomarkers with specific drugs would lead to better results. Lung cancer is the leading cancer killer, accounting for 28% of all cancer deaths in the U.S.
The study is too small to call for immediate new treatment strategies. But researchers believe it sets a path for bringing personalized medicine into the care of lung-cancer patients, just as it is already playing a critical role in breast, colon and other tumors.
"This is a first step to find biomarkers that may help supplant existing toxic therapies and to find the right population for a particular drug," said Edward S. Kim, a cancer researcher at M.D. Anderson and principal investigator of the Battle study. He presented results Sunday at the annual meeting of the American Association of Cancer Research.
Key to the study, Dr. Kim said, was the ability to obtain new tumor biopsies from all patients. The samples were analyzed to find molecular traits driving the patients' cancers. Then patients were randomly assigned to one of four drugs: Tarceva; the kidney and liver cancer drug Nexavar from Onyx Pharmaceuticals Inc. and Bayer AG; Zactima, being studied by AstraZeneca PLC for lung cancer; and a combination of Tarceva and Targretin, a drug for lymphoma-related skin problems from Eisai Inc.
Eight weeks after the start of therapy, doctors used imaging to assess whether disease in each patient was under control. Researchers said absence of progression at eight weeks is considered a reliable predictor of overall survival benefit.
After about 40% of the patients were enrolled, researchers looked at the first imaging results to see which combinations of drug and biomarkers were more likely to result in disease control. Then, in a strategy called adaptive trial design, the randomization of the remaining patients was weighted to steer patients toward therapies they were likely to benefit from based on their biomarkers.
The adaptive design is gaining interest among researchers and drug companies because it could help identify drugs that don't work sooner, and identify biomarkers that would be used to enroll patients in late-stage studies required for market approval.
Currently large clinical trials typically take all comers without evaluating their biomarker status. "The problem is that when you take a drug that has a specific target, but you treat everybody, you dilute the effect" of the drug, said Dr. Kim.
Researchers say that is why many targeted cancer drugs fail in late-stage or Phase III studies.
"This is the future," Tyler Jacks, a cancer researcher at Massachusetts Institute of Technology and president of the AACR, said of the Battle trial. "This is how drugs will be developed and clinical trials organized." Just this month, a study called I-Spy-2, led by Laura Esserman, a cancer doctor at University of California, San Francisco, began enrolling patients in a big breast-cancer trial using adaptive design.
In the Battle study, 46% of patients achieved disease control at eight weeks. The improvement was only slightly higher among patients enrolled by the adaptive technique, but researchers said that likely reflected shortcomings in how biomarkers were defined, among other issues.
"We need better biomarkers and better drugs to improve care," said Roy Herbst, an M.D. Anderson researcher and a leader of the Battle study.
Among other findings: Patients whose biomarker status matched up well with the treatment they got had an average survival of 11.3 months compared with just over 7 among those that didn't match well.
Dr. Kim also said Nexavar showed promise in patients whose tumors had a mutation in a gene called KRAS, a particularly troublesome population to treat.
Nexavar also achieved favorable results in patients without a KRAS mutation. Still, said Anil Potti, a lung cancer expert at Duke University Medical Center who wasn't involved with the study, finding an unexpected potential benefit for Nexavar reflects the potential of an adaptive trial to find new drug candidates that might be missed in more conventional trials. Nexavar previously failed to show a benefit in advanced lung cancer in a large trial in which it was combined with chemotherapy.
Dr. Potti and other researchers said the finding suggested the drug might be reconsidered for a lung-cancer study that included KRAS mutations—but that excluded patients with a mutation in a gene governing what is called the epithelial growth factor receptor. Tarceva, confirming other research, worked well in such patients in Battle, but Nexavar didn't.
Among the medicines, only Tarceva is approved for lung cancer in the U.S. The study was sponsored by the U.S. Department of Defense with support from the National Cancer Institute and some pharmaceutical companies.