Jonathan E. Rosenberg, MD, and Thomas Powles, MD, PhD, on Urothelial Carcinoma: Expert Commentary on Two Key Abstracts

Jonathan: Great to talk to you today, Tom. Tom: It's nice to see you too. Jonathan: Here at ASCO. I'd like to ask you about your opinions on CheckMate-901, the lymph node-only analysis that- Tom: Beautiful. Jonathan: We've heard, it seems like many of those patients do extraordinarily well with gemcitabine, cisplatin, and nivolumab. Can you tell me what you think about that data? Tom: Yeah, so we know 901 already the positive trial. It's got a PFS advantage, it's got an OS advantage, it has a ratio 0.78. There's always been a discussion about how much of that's coming from the maintenance period and how much is coming from the concurrent period. We've had negative trials before in this setting, but the previous trials have been all-comers. This has been just the gemcitabine and cisplatin population and there may be something special about gemcitabine and cisplatin. And the reason why that's the case, which is the other side of the argument, is the response rate and the CR rate was a little bit higher. And so what the group have done, and I'm involved in that from a distance, is that we've looked at the lymph node-positive patients and characterized the CR patients. So who are the winners, essentially? And what we are seeing is they're predominantly those patients with lymph node-only disease and they do exceptionally well. CR rate, they're above 60%, which is really, really attractive. Also, we're now seeing really nice PFS and OS curves. So there's a tail on those. You expect that. Now, don't get me wrong, the control arm are performing well as well. And actually when you look at, there's a very nice, I think it's a nice swim lane plot there, and you can see that swim lane plot. There are patients who stopped the nivolumab at two years and remain disease-free. Those who are relapsed to about half of the CRs end up relapsing. So this isn't like melanoma and kidney cancer where CR is cured. That's not, I think, what we're seeing in both arms incidentally. So it's also, but there are a lot more CR patients in the lymph node-only, or in the nivolumab arm. So overall, I would summarize it and say we can see now that this group do really well. We don't know from EV-Pembro about how they do in that sub-room. One would imagine they might, who knows, they might do even better. And we need to wait and see. I don't think this is a reason to give this combination in that instead of EV-Pembro as it currently stands, but it does allow us to learn more about that disease. It also in the background, what suggests that those patients that don't have that characteristic, some patients with lymph node disease for example, might be struggling a little bit. So I think there's a lot to this. I really like the analysis and it underlines the importance of proper staging at baseline. Jonathan: For EV-302, we saw data presented looking at quality of life of patients treated on either EV-Pembro or gemcitabine and cisplatin. We saw that many of the patients have maintained quality of life on both arms. What are your thoughts about the differences that we're seeing in this study? Tom: I think EV-302, EV-Pembro versus GemCis GemCarbo chemotherapy, 50% reduction in the risk of death, 50% improvement in progression of pre-survival, response rates of 70 versus 40%. Very positive standard of care in the guidelines, and I think it is the standard care for unselected patients. The quality of life data is always challenging. I'm not sure we've got perfect tools for quality of life analysis. They certainly weren't designed to assess EV-Pembro, but the adverse event profile associated with that compared with chemotherapy, it was a slightly novel technique in that we kept going with the assessment. We didn't stop at progression. So it's a complete assessment, but the compliance drops off in the control arm, chemotherapy arm quite a lot. Once they stop chemotherapy, many patients stop filling out the questionnaires. I think we need to be careful of that because there's quite a lot of censoring the arm. So we need to look at this data as practice-informing, but not practice- changing, I don't think. The overall data suggests similarities between the two arms. There was an initial primary analysis, a secondary endpoint, time-to-pain progression in unselected patients that didn't show a big difference between the two. But when you focused on those patients that had pain at baseline, it looked like EV-Pembro was outperforming GemCis GemCarbo. You might expect that if you've got pain at baseline and your response rate's higher, it's going to control it a bit better. And so I see that. And when you move to the global health scale in the EORTC Quality of Life questionnaire, you can see overall more similarities and differences in the two arms. You can show actually, if you focus on those patients with symptoms at baseline, you can see there's an improvement, which actually gets into the clinically meaningful boundary with EV-Pembro, the chemotherapy arm just misses that for what it's worth. But again, I'm not prepared at all. It's not clearly much better. What do I take from this overall, I think is that it reinforces the fact that EV-Pembro is a very active regime. I think both regimes have adverse events. I think EV-Pembro skin toxicity at the beginning, cumulative and peripheral neuropathy, chemotherapy, diarrhea, vomiting, particularly that difficult, neutropenic sepsis. Neither regime is like abiraterone and prostate cancer. And so you wouldn't expect with these relatively blunt tools to see massive differences between the two. Jonathan: Very subtle differences. Right. Tom: And yeah, we didn't see subtle differences. And so I would say it sort of reinforces what we know already without changing too much. Jonathan: Does this affect their regulatory landscape in any way outside the US in terms of quality of life testing? Tom: No, I don't think so. Jonathan: For the study in particular? Tom: The cost. So there's an organization in the UK called NICE and they look cost per quality. And they look at quality of life and whether there's a massive detrimental effect on quality of life with the drugs. And there isn't a massive detrimental effect. In fact, if anything, those patients with symptoms, their symptoms improve. And that's slightly more marked with EV-Pembro. So I don't think it's going to have a dramatic effect. But what I would say, Jonathan, is we did avelumab versus placebo or best supportive care and we didn't show quality of life differences between those two. Now avelumab is a well tolerated agent, but 10% of patients get toxicity or maybe more than that. And so if these tools were super brilliant, you'd expect those patients having two weekly drug associated with adverse events you would expect to see. That's not a subtle difference. So you'd expect to see a difference and there's no visible difference. And so I'm a bit worried that if we're using the wrong tool, it's not a surprise that we're not showing differences. It may be what's better, it may be not be better, who knows? But so I'm as worried about the tools as I am that we're using. I'd like to see us using step counts, wearable devices, sleeping patterns, real-time, filling out questionnaires, asking patients specific time points questions. I'm not super keen on this. I'd like this field to move on. Jonathan: I think the next generation of PROs are going to be very important to understand all this. Thanks so much, Tom, great to talk to you. Tom: Thank you Jonathan.