Predictive Biomarker Testing for Immunotherapy in Melanoma

Antoni Ribas, MD, PhD: For certain cancers where we know it can only work in a specific situation that has a specific marker, having a companion diagnostic makes a lot of sense. You want to select those patients who have the marker and then give the targeted agent. Immunotherapies act like targeted agents, but they’re using the immune system to attack the cancer, and the immune system cells work by having a specific receptor, the T-cell receptor. The T-cell receptor will recognize a different ligand on every cancer, so it’s hard to think that we’re going to have one biomarker for all immunotherapy. This is not a companion diagnostic. But something that then reaches the population to be more likely to respond on that would make more sense.

I think the majority of melanoma practitioners are not ordering PD-L1 testing. I don’t see it ordered by my colleagues at UCLA and at other academic centers, or in the people that come from the community. That’s because it does not separate out enough patients who are likely to respond or not. But that doesn’t mean that the negative testing will lead to no response. So, I think implementation in the field of a PD-L1 assay for melanoma has been very low.

I do not use a PD-L1 test in my practice for patients with melanoma. The data, I think, do not support it. And with a negative test, it would not preclude me from thinking that a particular patient may still benefit because some patients with negative test results still benefit.

In the year since the first trials of PD-1 and PD-L1 blocking antibodies, there’s been a lot of interest in looking for the ligand and selecting patients based on that. The first studies used assays that were not that reproducible. Both the antibodies that were trying to detect PD-L1, or the assay process, to process the tissue to get the results were not that well established. Now we have very robust assays for PD-L1, where the data that we get back correctly reflect what’s happening in the tumor. But then it’s not only the specificity and the sensitivity of the assay that is important, it’s the biological interpretation. If there’s not an immune response, there’s not going to be interferon-gamma being produced. And then the tumor may not produce express PD-L1, which is produced as a reaction or a protection from the presence of immune response.

So, not detecting PD-L1 and calling a tumor negative may not really mean a negative tumor. If you unleash an immune response, like, for example, with the combination of ipilimumab and nivolumab, it brings a lot of T cells into tumors by releasing CTLA4, which is upstream, and then releasing PD-1, which is downstream. You get a lot of T cells in the tumor. Now the baseline assessment has changed. If there were no T cells, maybe the tumor could not protect itself by expressing PD-L1. If you bring them in, then the tumor may become positive. But it would have been read as negative before.