(Page 2 of 2)
file for FDA approval of larotrectinib either later this year or in early 2018. It’s gearing up for a commercial launch next year. If that were to happen, larotrectinib would join only Merck’s (NYSE: MRK) pembrolizumab as a drug specifically approved for a cancerous mutation, regardless of where a tumor has originated.
There are still key questions for Loxo’s drug to answer. For one, Loxo has yet to report how durable responses to the drug are, though Bilenker says that’s a positive sign—too few people have failed treatment to generate that data, he says, and some responders have been on larotrectinib for over two years. Six patients’ cancers have developed resistance to the drug and progressed, though. Once more granular information is available about relapse rates and durability for specific cancers, it will help doctors see how they should use Loxo’s drug. Pal, for instance, says while there will be a “temptation across disease types” to test for TRK fusions, it’s unclear how Loxo’s drug compares to the myriad other treatments for, say, lung or kidney cancer. “What we really need to know,” he says, is “can patients potentially fare the same or better with standard therapy?”
But the broader question is will patients get the chance to see if they are eligible for Loxo’s drug? That will at least partly depend on adoption of broad cancer genetic tests, like the FoundationOne test sold by Foundation Medicine (NASDAQ: FMI), which can test for hundreds of genetic abnormalities in tumors, among them TRK fusions. These tests have struggled to gain traction for a variety of reasons, from payer pushback to resistance from community oncologists and questions about their utility, and are most commonly used in clinical studies and major cancer centers. “They are considered investigational or exploratory tests, and those are very difficult to get reimbursement for,” says Yuri Nikiforov, the director of the University of Pittsburgh Medical Center’s molecular and genomic pathology division.
Foundation has spent years fighting for insurance coverage, and has applied for FDA approval of the FoundationOne, a decision that is expected this year. A second test from Thermo Fisher Scientific (NYSE: TMO), Oncomine, is also under FDA review, with a decision pending. Approvals for both tests may help, but Bilenker says Loxo is hedging the market anyway. It cut a deal with Roche in March to develop a companion diagnostic specifically for TRK fusions. That test uses old school immunohistochemistry stains, a cheaper and more widely used diagnostic method. “In the intermediate and long-run, we think [broad tumor sequencing] is the modality of the future,” Bilenker says, but “we’re taking a bifurcated approach.”
Loxo’s decision shows the hurdles that these testmakers still face. Pal, for instance, says City of Hope has been using molecular profiling tests in kidney and bladder cancer patients, where they typically aren’t used. Pal used to reserve these tests for patients who had failed a number of treatments, but is being “gradually more motivated” to try them earlier. He’s convinced they can help lead patients to drugs they might not otherwise get, and ultimately better health outcomes—something testmakers like Foundation are fighting to prove. But payers push back. “I have gone to great lengths to justify the utilization of these tests on a case by case basis,” he says.
Nikiforov says UPMC has been finding TRK fusions in the tumors of thyroid cancer patients “every 2 to 3 weeks.” For these patients—often young people who develop brain metastases—insurers may cover the expensive cancer drugs to treat them, but not the sequencing tests. Could Loxo’s data, and Merck’s approval, start to change things?
More validated molecular markers, and approved drugs based on genetic alterations—like a TRK fusion—might mean that oncologists have more reason to test for a wide swath of tumor abnormalities with a single exam. Without these options, broad tumor profiling tests can spit out a lot of data that, in the end, amount to an expensive exam that doesn’t accumulate useful signals, just noise.