Manish A. Shah, MD: A lot of data have been reported, as Kohei just reviewed. I’ll highlight some of the data. I’ll start at the third line and move up. In the third-line setting for PD-L1 [programmed death-ligand 1]—positive tumors, pembrolizumab is in fact approved for gastric cancer. And the approval is gastric and GE [gastroesophageal] junction.

In the second-line setting, based on your data, which was unfortunately a negative study, pembrolizumab is not approved. So we don’t use immunotherapy in the second-line setting outside mismatch repair deficiency. In the United States, if you’re mismatch-repair deficient, or MSI [microsatellite instability] high, that’s when you should use immunotherapy in the second-line setting. That’s not unique to gastric cancer but all solid tumors. So that’s the second-line use of immunotherapy.

In the first line, which becomes a bit complicated based on the study, I think the first question is, should we combine immunotherapy with chemotherapy? And the answer is a resounding no. There’s no evidence that says that’s effective. So the standard of care for most patients remains chemotherapy, a doublet combination, as we’ve discussed.

The study did have a unique feature for PD-L1—positive patients, where it was technically nonequivalent based on their study design. There were about 500 patients, and it compared PD-L1–positive tumors who received pembrolizumab versus chemotherapy, and there appears to be a nonequivalence. I don’t know if that’s going to be practice changing because of the points that Kohei made about the acceptance of detriment of effect. I think we’ll have to wait and see. I personally probably wouldn’t use pembrolizumab alone because, remember, most gastric cancer patients, at least in the West, are diagnosed with advanced disease. They’re symptomatic. If we don’t actually achieve a response, we may not be able to continue treatment. In fact, in this study only about 50% of patients in either arm received second-line therapy. So it’s an important point for me to mention that I think we need to consider the patient as we’re making these decisions.

Johanna C. Bendell, MD: David, what’s your take?

David H. Ilson, MD, PhD: I think Manish summarized it very, very clearly. But I think to emphasize, all these studies show late benefit for immunotherapy. The curves don’t separate until late, and many of these patients fail immunotherapy early. I think at least the indication in the United States will remain MSI-high, second-line, and chemotherapy-refractory PD-L1—positive gastric cancer—that’s where we can consider the use of pembrolizumab.

I have a bit of an issue with this noninferiority claim of pembrolizumab monotherapy. If you look at the progression-free survival, 70% to 80% of patients get a few doses of pembrolizumab, fail it, and go on and get chemotherapy. Arguably, you’re comparing chemotherapy with chemotherapy. Almost all the pembrolizumab patients are getting chemotherapy very early on.

To argue a survival benefit for up-front pembrolizumab, the problem is that it wasn’t a crossover. All the patients who got pembrolizumab up front got pembrolizumab during the course of their illness. If you look at the chemotherapy-alone patients, 5% to 10% got pembrolizumab. So you really can’t make a survival comparison. I think, to Manish’s point, these are sick patients. You have a 25% response rate at best, so you’re permitting a 70% to 80% fail rate in a sick patient. I don’t think this is going to be practice changing. It’s going to be interesting to see what the regulatory take is on this. I think up-front pembrolizumab is going to be for a highly selected, asymptomatic patient with a high CPS [combined positive score]. Maybe someone who has had prior gastrectomy, who’s had prior adjuvant chemotherapy. But I think to me, this trial does not change the current standard of care, which is late-line use of the therapy. We’ll have to see what other ongoing up-front studies of chemotherapy combined with checkpoint inhibitor therapy yield.

Kohei Shitara, MD: I completely agree, because I was deeply involved in that trial. Maybe there’s some bias, but I completely agree. Even if pembrolizumab is available, I may carefully select the patients based on MSI-high, CPS score and the patient’s symptom and general status. Very select patients can receive pembrolizumab monotherapy...

Johanna C. Bendell, MD: So aggregating everything, maybe a very select patient with all the, “I’m going to respond to immunotherapy markers” of…

David H. Ilson, MD, PhD: If it gets approved.

Johanna C. Bendell, MD: If it gets approved.

David H. Ilson, MD, PhD: Outside a study, I would not do it.

Johanna C. Bendell, MD: Yeah. So high CPS, MSI, asymptomatic minimal disease to give enough time to respond if it gets approved. In general, don’t do it in first line. In the second line, for MSI-high, would you do if the CPS is greater than 10?

Manish A. Shah, MD: It’s indicated for anybody who’s MSI high, independent of CPS.

Johanna C. Bendell, MD: But if you had a CPS that was greater than 10, would you ever consider using pembrolizumab in the second line?

David H. Ilson, MD, PhD: We have a negative randomized trial showing it’s not better.

Kohei Shitara, MD: Actually, we cannot use it even in Japan.

Johanna C. Bendell, MD: OK. Also, the wrong comparator, right? We should have been using paclitaxel plus ramucirumab. And then in the third line in the US for PD-L1—positive patients.

Manish A. Shah, MD: Right.

Johanna C. Bendell, MD: And in Japan?

Kohei Shitara, MD: Results, patient enrichment by biomarkers. But we have 3 options in Japan, and maybe this is the same in the US? We have TAS-102 [trifluridine, tipiracil], nivolumab, and irinotecan. So patient selection might be very important. If I have a good PD-L1 score, the high expression, maybe I can start with nivolumab first. But if the patient had a very large tumor and some symptoms, I usually start with irinotecan first. So it depends.

David H. Ilson, MD, PhD: Yeah, and TAS-102, I think, in the PD-L1—negative patients, would now be the standard third-line treatment. For irinotecan, TAS-102. For PD-L1–positive, we have the approval of pembrolizumab to consider.

Transcript Edited for Clarity