A few weeks ago I described a blood test for predicting survival after getting an oral epidermal growth factor receptor (EGFR) inhibitor that is just becoming commercially available. This is called the Veristrat test and from company called Biodesix. As I noted previously, this technique looks at the protein patterns in a patient's serum and reportedly can reliably predict whether a patient is likely to have a prolonged survival after receiving an EGFR inhibitor. The company gave me the opportunity to send off a few of these tests at no charge, and after a few weeks of trying it in a handful of patients, I'm finding that it seems to do what it says it does, for better or for worse. What they specifically say is that, by separating serum samples into the two thirds that will have a "good" survival and the one third that will have a "poor" survival after getting an EGFR inhibitor, clinicians can use this test to determine which patients shouldn't get an oral EGFR inhibitor like tarceva (erlotinib) or iressa (gefitinib) for advanced NSCLC. The test may well be great for this, but I haven't had the opportunity to send this test off in the last few weeks for this purpose. Instead, what I was hoping we might use this test for is to get a faster answer about whether a patient will do well with an EGFR inhibitor. Even though I and many other lung cancer experts are increasingly becoming converts to the idea that EGFR mutations are quite relevant and override clinical predictors of benefit with EGFR tyrosine kinase inhibitors (TKIs), these mutations aren't always very accessible. Only a minority have enough tumor tissue from their original biopsy to send off for mutation testing, and the test itself takes 3-4 weeks, which is long enough that many patients will not be inclined to sit around and wait for a month before starting treatment.

My hope had been that I could use this test to get an immediate, blood-based answer to the question of whether it would be a bad idea to start a patient on a first line EGFR TKI if they had several clinical factors suggestive of major benefit with an EGFR TKI. Specifically, I now follow two Asian never-smoking young women with an adenocarcinoma, who would be predicted to have an approximately 70% chance of having a tumor with an EGFR mutation, and who actually did very poorly with tarceva as a first line therapy. If the Veristrat test could tell me that they were in the "poor" category, I'd be thrilled to have a blood test that could provide feedback within days that it would likely be helpful to start an EGFR TKI first line. As promised, I received answers on all of the tests I sent within 2 days, after it was sent. The problem is that thus far I've sent it on about 7 patients, and I've only received test results of "good", including from the two women who progressed remarkably quickly on tarceva. I have also sent it on a few patients who are just starting tarceva, so I don't have outcome answers for them yet, and I haven't yet sent it on any patients who have done extremely well on an EGFR TKI. I was disappointed that the test couldn't tell me that the patients who ended up doing poorly on tarceva that an EGFR TKI would be such a poor first line therapy for them. However, in thinking about this further, I realize that the test merely claims to identify which patients will have a longer survival or shorter survival after getting an EGFR TKI. In fact, both of these women have received subsequent chemotherapy and have been doing well, with both showing a convincing modest response (not a 50% shrinkage, but convincing improvement on scans) after being in a tailspin on tarceva. All indicators now suggest that their survival is likely favorable enough to end up on the "good" curves that Veristrat would predict, but it will be despite and not because of the EGFR inhibitor. Specifically, it appears that the Veristrat test is prognostic, telling how a patient is likely to do overall, but not predictive of how someone will do specifically from the EGFR inhibitor. The Veristrat test may still be very helpful in asking the question "is an EGFR TKI worth it" for someone in whom we might be dubious of the benefit -- perhaps a longtime smoker with a squamous tumor (these patients still show a survival benefit overall with an EGFR TKI, despite low response rates compared with other groups) or someone with a marginal performance status. I don't know how many patients would be inclined to not even try an EGFR TKI if the test came out with a result of "poor", but perhaps some would favor focusing on a chemo alternative, clinical trial, or supportive care. But if the test can keep people from wasting time and money on treatment they're likely to better without, I think it could be helpful. Still, I was hoping I'd get a "quick and dirty" proxy for best vs. worst responders, and so far it doesn't look like we can use this test to predict who will do well with an EGFR TKI. In fairness, this isn't what the company has stated the test is for. I've still got to keep testing to see if it can identify the patients who shouldn't even try.