A large, recently concluded clinical trial of Tarceva demonstrated improved survival for all patient groups, so will the drug's development partners Genentech, OSI Pharmaceuticals, and Roche Pharmaceuticals remain happy with the drug's second-line status and never couple it with a diagnostic that could make it a first-line contender?
In November 2004, the US Food and Drug Administration approved Tarceva, also known by its scientific name erlotinib, for patients with advanced non-small cell lung cancer who had received at least one regimen of chemotherapy.
But a recent study describing the broad survival of 731 patients receiving the epidermal growth factor receptor inhibitor, along with a companion study that strongly implies — but doesn't prove — that molecular diagnostics might be capable of identifying patients whose survival would benefit from Tarceva. Genentech, OSI, and Roche now face an interesting commercial choice: Should they pursue a first-line indication for Tarceva, a path that might require the companies to conduct a new trial to prove its superior efficacy in a specific subset of patients, perhaps as identified by a diagnostic; or should the firms leave it as a second-line treatment that is broadly applicable to non-small cell lung cancer patients?
"Gene-copy number and protein expression still appear to be in the running, but our results were not definitive."
"Overall, there could be a treatment effect which could be better than chemotherapy" should researchers develop a diagnostic or another means to identify responders, said Chandra Belani, head of the lung-cancer program at the University of Pittsburgh School of Medicine. Belani was cautious in his assessment, however, noting that any first-line application "needs to be validated," and pointing out that the pair of studies only "add to the complexity, with less clarity at this point in time for proper patient selection for treatment." Belani and colleague Suresh Ramalingam wrote a summary of EGFR-targeting agents for CancerConsultants.
"Gene-copy number and protein expression still appear to be in the running, but our results were not definitive," said Frances Shepherd, a medical oncologist at Princess Margaret Hospital and professor of medicine at the University of Toronto, who co-wrote the study, and is the principal investigator of the study of overall survival, both of which appear in the July 14 issue of the New England Journal of Medicine. Ming-Sound Tsao, also of Princess Margaret, was the principal investigator of the study of molecular markers of Tarceva response and survival. That study's sample size of 325 and 125 tumors was too small to draw significant conclusions about the use of protein expression or gene-copy number, respectively, as possible indicators of efficacy, Shepherd said. In addition, the study found no link between the presence of EGFR tyrosine-kinase domain mutations and survival or tumor response.
Diagnostics and the Number of Prescriptions
The three owners of Tarceva are "working … on determining the best options for evaluating Tarceva in the first-line setting," said Colleen Wilson, Genentech associate director of product communications, in an e-mail exchange with Pharmacogenomics Reporter. However, "at this point we don't have a final protocol for a first-line study," she said.
Furthermore, Genentech is cautious about developing diagnostics that would identify which patients benefit most from Tarceva. Wilson noted that the Shepherd et al. study shows that the drug produces a survival improvement in most second-line patient subsets, does not test its use in first-line treatment, and that "we don't believe patients should be denied access to Tarceva based on the EGFR biomarker data generated to date."
Howard Liang, who covers Genentech for A.G. Edwards & Sons, noted that the company has previously pointed to studies showing the drug's broad efficacy for many patient groups, in the face of evidence that it may be more efficacious for patients whose tumors carry EGFR tyrosine-kinase domain mutations. "I think they're trying pretty hard not to narrow the patient population," he said.
According to Liang, Genentech and its partners may ultimately have to weigh a broad market for a second-line therapy against a narrow market for the drug as both a first-line treatment and a second-line treatment. It's possible that establishing biomarkers of response or survival, which may be necessary if the companies are to pursue its approval as a first-line drug, could limit Tarceva's market to a particular subset of patients. "It is a little bit of a dangerous thing for Genentech," said Liang. "If you do identify some sort of test that you can use to narrow down patient population, I presume [doctors] would use it in second-line [therapies] as well, so you can potentially reduce the market size of the drug. So it is a double-edged sword," he said.
Shepherd said biomarkers for survival would likely end up being the same for both first- and second-line therapies. Fred Hirsch, a University of Colorado Cancer Center researcher who has studied EGFR-related drugs extensively, said the existence of a diagnostic for first-line treatment using Tarceva may not weigh on doctors' decisions to prescribe it in second-line treatment without a diagnostic.
But whether a diagnostic would be required to bring Tarceva into first-line treatment is not yet known, and neither is the degree to which it might limit prescriptions for either line of therapy. "It's too early to say if there will be a diagnostic for Tarceva and what it might look like," said Wilson via e-mail. In general, "there are more patients receiving treatment for first-line NSCLC (approx. 69,000) than [for] second- or third-line NSCLC (approx. 41,000)," she said.
In Tsao's study, approximately half of tumors tested positive for either protein expression or high gene-copy number. In the total clinical trial population of 731, 325 tumors were tested for EGFR expression and 125 were successfully tested for high gene-copy number in Tsao's study. About 57 percent of tumors, or 184 samples, were EGFR-positive: 50 percent of adenocarcinoma samples and 63 percent of samples of tumors of other types. Forty-five percent of tumors tested successfully with FISH showed high gene-copy numbers: 48 percent of adenocarcinoma samples and 41 percent of samples of other types.
Study Findings and Prevalence of NSCLC
Shepherd and colleagues found that patients who received Tarceva as second- or third-line therapy survived an average of 6.7 months, compared to 4.7 months for patients receiving placebo. Thirty-one percent of patients receiving Tarceva survived at least one year, compared to 22 percent of patients receiving placebo.
In Tsao's study, patients treated with Tarceva lived significantly longer than placebo patients when their tumors exhibited a high EGFR copy number or EGFR protein expression, but this correlation was not statistically significant. "Here's the problem," said Shepherd. "When you compare [high gene-copy number to low copy number], there is a 1-in-10 chance that these differences were by chance alone — we need to get it to less than five percent," she said. The statistics for patient tumors showing protein expression, versus those having no expression, exhibited the same problem, she said.
The researchers used Spectrum Green and Spectrum Orange FISH, a product of Abbott Diagnostics subsidiary Vysis, to test tumor gene-copy number. Dako Cytomation's EGFR IHC kit was used to test for protein expression.
In the United States, about 60 percent of lung cancer patients die within one year of diagnosis, and between 70 and 80 percent die within two years, according to the US National Cancer Institute.
About 80 percent of lung cancer cases involve non-small cell lung cancer. An estimated 163,510 deaths from lung cancer will occur in the United States during 2005, according to Cancer Facts and Figures, 2005, an American Cancer Society publication. An estimated 350,679 Americans currently have the disease, and an estimated 172,570 new cases of lung cancer will be diagnosed in 2005, according to the NCI and Cancer Facts and Figures, 2005, respectively.
An estimated 1 million people worldwide die from lung cancer annually, and as the most common diagnosed cancer, more than 3 million people live with the disease, mostly in developed countries, according to the World Health Report, 2004, a publication of the World Health Organization.
— Chris Womack ([email protected])