Milind Javle, MD: As you mentioned, there was a benefit seen with durvalumab, perhaps more in the Asian population. However, there was also a benefit seen in the Western population, although it was somewhat lower. I have to give you my personal anecdote. As you mentioned, the hazard ratio in gallbladder cancer was much lower, and the benefit appeared to be lower. Also, a smaller number of patients are included with gallbladder cancer. I just finished a discussion on the same topic with my fellow in the clinic, and the next patient I co-shared with a colleague of mine ... She had 2 patients with gallbladder cancer, and they had a very remarkable response with arterial disease. I went back and found that 1 had an ETM mutation, and I think the other had another CHEK1 or CHEK2 mutation—some DNA-repair mutation family. Do we need to learn a little more about these patients, including their molecular profile and their PD-L1 status? I’d like to get your opinion, Ruth—as well as Mark’s—about what’s known about the underlying immune profile of these tumors in TOPAZ-1 and what, if anything, is known about the underlying molecular profile.

Aiwu Ruth He, MD: I’ll start. For the next analysis, we’re trying to roll out data, looking at the TOPAZ-1, and explore ration of the TOPAZ-1 data more with the subsequent meetings. Next, the data set is for a biomarker analysis. This is difficult, but there are many patients for whom we’re able to collect baseline tissue. The first role of analysis is looking at MSI [microsatellite instability] status, PD-L1 staining, and the tumor mutational burden. That will be the first round. The second role is to profile those samples and the relationship between the typical heart tumor vs the tumor with other genetic alterations, such as FGFR2 fusion and IDH mutation. That’s in the works.

On the same line, a lot of researchers are looking at this. There’s a project sponsored by NCI [National Cancer Institute] under the Moonshot project, led by Tim Greten and ... and we collaborate on those projects. It’s very interesting. With the initial data rolling out, we should see more by looking at HCC [hepatocellular carcinoma] and intrahepatic cholangial as a spectrum. For sure, a subset of patients have heart tumors, and we’re trying to understand. I believe there’s subset of patients who benefit from immunotherapy combination. You have a global study that’s heterogeneous, and we haven’t found the way to define those patients.

Milind Javle, MD: Mark, the paper presented some data on PD-L1 and TAP [tumor area positivity] score. What’s your interpretation of these data?

Mark Yarchoan, MD: The main takeaway is that we don’t have a biomarker coming out of TOPAZ-1. The benefit of durvalumab was small but relatively consistent across the different groups. There appeared to be some benefit in PD-L1–positive and –negative patients, with slight enrichment for benefit in PD-L1–positive. But it was not enough that anybody would use that to make a treatment decision. I’d love to see the detailed molecular analysis. We’ve published an analysis of about 400 biliary tract cancers looking at DNA and RNA. We did some cluster analysis and found that patients with FGFR2-driven cholangiocarcinomas appear to have a particularly quiet tumor-immune microenvironment without a lot of immune infiltration. It would be interesting to see if there’s less benefit in that group of patients with TOPAZ-1, for example. Those data have been slow to come out, in part because it’s positive across all patients and there’s not much interest in splitting. But we need to move in that direction.

Milind Javle, MD: Thank you. Ruth, you pointed out very well the need for profiling across different populations as you’re doing. I want to refer you to a paper by a colleague of ours ... who published in Hepatology and performed a similar cluster, as Mark mentioned. He pointed out that about 15% of patients had an immune hot cluster in cholangial ... population. Not coincidentally, that’s the number that you see in response with immune checkpoint inhibitors. Suddenly more direction is required, and Dr Yarchoan’s colleague Dr [Nilofer] Azad and some of us are doing an international study profiling across different countries.

Now Flavio, I was interested in the TOPAZ-1 study that some of the localized tumors, the earlier tumors, the hazard ratio was seemed to be even more substantial there than in the advanced. And same was for the current tumors after surgery. Do you see these results as having some implication from the surgical point of view?

Flavio G. Rocha, MD, FACS, FSSO: As a surgeon, I’m always very keen to see what these new regimens and targets are because I want to move them up in the treatment spectrum. I want to see these agents used in the neoadjuvant setting precisely to downstage tumors to resection, or even in resectable cases to see if we can minimize recurrence rates. This is the Achilles’ heel. We do this big operation, and then 2 years later half of them recur. The more we can get this information and these agents out, I just wish we could slow the delay and basically get these agents out getting studied in the neoadjuvant setting.

Mark Yarchoan, MD: I want to make a quick comment out of turn. I clearly think the TOPAZ-1 study is an important landmark for cholangiocarcinoma, but the response rate to gemcitabine-cisplatin-durvalumab in the study was unfortunately only about 26% if I remember correctly. Someone can correct me if I’m off by a couple of numbers. But we do have single-arm phase 2 data for gemcitabine-cisplatin-Abraxane, and the response rate is well over 40%. At this point, the locally advanced setting is in which I wouldn’t use gemcitabine-cisplatin-durvalumab. I’d spring for gemcitabine-cisplatin-Abraxane even though it’s only phase 2 data. I’m not sure if others agree with me, but it’s an important point.

Flavio G. Rocha, MD, FACS, FSSO: I’ll jump in, Mark. I was going to mention this later, but as you may know, we’ve taken the gemcitabine-cisplatin-Abraxane data based on the phase 2 study. With Milind and others, we’ve put together a 30-patient cohort for neoadjuvant therapy. We had 30 patients, and 22 of them went to surgery for intrahepatic cholangiocarcinoma; we’re waiting for long-term data. This was presented at ASCO GI [American Society of Clinical Oncology Gastrointestinal Cancers Symposium] earlier this year. But we agree. Having that combination, obviously we’re waiting for the phase 3 advanced trial to report. We’re excited. We want to see what you think has good efficacy, safety profiles, and so forth, and we want to move them up.

Milind Javle, MD: Thank you for that comment, Mark.

Transcript edited for clarity.