Recommendations for Molecular Testing in Melanoma

Transcript:

Hussein A. Tawbi, MD, PhD: BRAF mutations occur in half of all melanoma patients. Again, they may actually carry prognostic value. But most importantly, I consider them to be the most predictive test that we have in melanoma because they predict for responses to targeted therapy with BRAF and MEK inhibition. Generally speaking, since we think about them in the setting of targeted therapy, we started testing in metastatic disease. That’s where the combination of dabrafenib and trametinib, vemurafenib, cobimetinib, and now encorafenib and binimetinib have all proved benefit. However, I think the earlier you can find out, the better off your patient will be. You will always know whether you have that option at your disposal. This became very true after the adjuvant data from COMBI-AD was published, and now we have an approved adjuvant therapy for that setting. So now stage III patients need to be tested.

We generally test all patients who have a positive lymph node, and we’re starting to think about whether we should even test earlier-stage patients. There are now considerations for trials for stage II patients. And so, I guess my answer to this would be: Test as early as possible and as soon as possible, so you know where your patient fits in the treatment paradigm.

Biomarkers of response in melanoma are not well established. Again, I think the BRAF mutation is probably the best biomarker we have because it consistently predicts a response to BRAF inhibitors. PD-L1 [programmed death-ligand 1] testing can kind of help us understand the response of the patient, but I don’t honestly think it predicts the response as well as we would like it to. So we don’t reflexively test all our patients for PD-L1. We do test it typically on clinical trials and, most of the time, more retrospectively. Again, in melanoma, PD-L1 is not very useful as a marker of what you should do for your patient.

Similarly, TMB [tumor mutation burden] has been reported to be associated with response with a higher TMB. As you would expect, you have a higher mutation burden, so you would expect most patients to respond. I’m not sure it’s necessarily an independent predictor, though. Again, you would find that some of the TMB-high patients will be PD-L1—positive. They also may have a high interferon gamma signature, so all these subsets end up overlapping. Again, TMB itself is not independently predictive, so we don’t use it for clinical management at this point.

Transcript Edited for Clarity