A multi-center team, led by Vanderbilt-Ingram Cancer Center investigators, has discovered a "signature" of proteins in the blood that predicts which non-small-cell lung cancer patients will live longer when they are treated with certain targeted cancer therapies.
The findings, published June 6 in the Journal of the National Cancer Institute, could one day help physicians decide which lung cancer patients to treat with drugs known collectively as EGF receptor tyrosine kinase inhibitors, a step forward in the era of personalized medicine.
"There's a real clinical need to identify which patients will benefit from targeted therapies," said David Carbone, M.D., Ph.D., Harold L. Moses Professor of Cancer Research at Vanderbilt-Ingram and the senior author of the study. "If our findings are confirmed, we will be able to use a simple and inexpensive blood test to select the most beneficial therapy for each patient."
Targeted cancer therapies are the newest generation of anti-cancer drugs. In contrast to traditional chemotherapy, targeted therapies affect proteins and signaling pathways that are selectively activated in certain malignant cells and not in normal cells. But this selectivity makes these therapies effective only for the subset of patients whose tumors are "driven" by the targeted pathways.
The drawbacks of treating every patient with a targeted therapy include the expense of these drugs, the delay -- for those who do not respond -- in initiation of effective therapy, and the possibility that some patients will be harmed by the targeted therapy.
In the case of the EGF receptor tyrosine kinase inhibitors (TKIs) gefitinib (Iressa) and erlotinib (Tarceva), studies have demonstrated a survival benefit for 30 to 40 percent of lung cancer patients, but there has been no method for identifying these patients prior to treatment, Carbone said.
Investigators at Vanderbilt-Ingram, the University of Colorado in Aurora, Colo., and Biodesix Inc. in Steamboat Springs, Colo., with worldwide collaborators providing patient samples, set out to determine if a protein profile in the peripheral blood could predict clinical benefit -- measured in terms of patient survival -- to EGF receptor TKIs.
Using mass spectrometry, the researchers analyzed pre-treatment blood samples from 139 patients who had been treated with gefitinib (three patient cohorts in Italy and Japan), identified a pattern of eight proteins that was correlated with survival, and developed a prediction algorithm.
They then tested the algorithm in two additional groups of patients -- 67 gefitinib-treated patients in Italy and 96 erlotinib-treated patients in a U.S. Eastern Cooperative Oncology Group protocol. At the time of the mass spectrometry analysis and classification, the researchers were "blind" to the survival status of the patients.
"We classified the patients as being either likely to benefit ("good") or not likely to benefit ("poor") from the TKIs," Carbone said.
The method was highly successful in predicting a survival benefit. In the gefitinib-treated group, patients classified as "good" had a median survival of 207 days whereas those classified as "poor" had a median survival of 92 days. In the erlotinib-treated group, median survivals for "good" and "poor" groups were 306 and 107 days, respectively. The fact that the signature, which was developed from gefitinib-treated patients, also accurately predicted a survival benefit for erlotinib-treated patients buoys Carbone's confidence in the algorithm, he said, since these two drugs share a common mechanism of action.
The method was not prognostic -- it did not predict a survival benefit in three different control groups of patients treated with either chemotherapy or surgery alone.
The investigators are currently working with the Eastern Cooperative Oncology Group to develop a prospective national phase III trial that will test the prediction method's clinical benefit in lung cancer patients who are just beginning treatment for advanced disease.
"This is a convincing set of data from multiple institutions and multiple cohorts of patients, and we're excited to test the prediction algorithm in a prospective way," Carbone said. "If it holds up, we will be able to separate patients into two groups -- one that would benefit more from chemotherapy and the other from targeted TKIs -- and treat them accordingly. The overall survival of the whole group of patients would be better by virtue of this biomarker-based test."
The research was supported by the National Cancer Institute's Specialized Programs of Research Excellence (SPOREs) in Lung Cancer and cancer center core grants to Vanderbilt-Ingram and the University of Colorado.
Cite This Page: