DR DRAKE: I think what’s emerging is that sometimes but not always, PD-L1 expression is a weak predictor of patients who might do better. At the GU ASCO meeting we were just at together, it was interesting. What was shown in bladder cancer, for example, that there was a weak correlation between PD-L1 staining and response.

But the remarkable thing, Neil, is, if you look at the Phase II bladder data from atezolizumab, even in the patients who had absolutely no PD-1 staining, zero, the response rate was 10%. In that setting, that’s really around what you see with chemotherapy. And the part that struck me was about 100 patients with no PD-L1 staining at all. Out of those, Neil, 2 of them had a complete response. Complete response.

DR LOVE: Wow! So this whole discussion of what is it that’s tied into the responses that are seen also is tied into what you see in different tumors? And I think that last meeting we were at together, we showed a slide I’ve shown a bunch of times, which is, which of the following tumors is most sensitive to anti-PD-L1/PD-1 antibodies? And, like, I think 6% of people got the right answer, which is Hodgkin lymphoma. And I guess I’d say, “What does Hodgkin lymphoma, lung cancer and head and neck cancer all have in common?” And you hear things about mutational load. Is there any way you can put this whole puzzle together?

DR DRAKE: So what I would say they do all have in common and you can put it together is, they all elicit a T-cell response, a CD8 T-cell response. This is very clear for Hodgkin’s lymphoma, lung cancer, bladder cancer and kidney cancer and melanoma. So this is the one thing they have in common.

DR LOVE: Do other tumors not elicit a T-cell response?

DR DRAKE: It’s interesting, actually. Even within the field of kidney cancer, not all tumors are infiltrated, actually. Some tumors do, and some tumors don’t, actually.

Bladder cancer

Prostate cancer

Renal cell carcinoma

Checkpoint inhibitors