Transcript: Johanna C. Bendell, MD: Let’s get to the meat: immunotherapy in gastric cancer. Where do you use it? Do you use it in the first line, second line, third line? In combination? PD-L1 [programmed death-ligand 1]—positive is the FDA approval. In Japan, you can use it for anybody. What’s the deal here? And tell us a little about the pembrolizumab data. Kohei, you have done so much work in this field. Why don’t you kick us off by telling us a little about the use of immunotherapy?

Kohei Shitara, MD: Yeah, the first study is ATTRACTION-2, which compared nivolumab and placebo in chemotherapy-refractory patients, in third- or later-line patients. In the Asian population, it clearly showed a survival benefit. The hazard ratio was 0.64, and this resulted in the Japanese approval of nivolumab. Currently, in Japanese guidelines, nivolumab is recommended as third-line or later-line treatment.

We have a nivolumab, irinotecan, and in the future TAS-102 [trifluridine, tipiracil] as third-line treatment options. Actually, a biomarker analysis from this ATTRACTION-2 study did not show any difference in efficacy between PD-L1—positive and the PD-L1–negative patients in tumor tissues. But the tumor sample is available only from 39 patients. Very small, limited information.

In contrast, pembrolizumab is approved by the FDA for patients with a CPS [combined positive score] of 1 or more based on KEYNOTE-059, which suggests a relatively higher response rate and a better PFS [progression-free survival] in the CPS of 1 population compared with a CPS of less than 1 or PD-L1—negative population.

In the third-line setting, there is some difference. Maybe in the third line we need to clarify which patients should receive the checkpoint inhibitor first, compared with irinotecan and with TAS-102 [trifluridine, tipiracil]?

In the second line, we previously compared pembrolizumab with paclitaxel. This study did not show a significant improvement in overall survival with pembrolizumab versus paclitaxel in the CPS of 1 patient population. Very disappointing. But it still showed that cross of the survival curve. Initially, paclitaxel did better. But the long tail suggested the benefit in pembrolizumab. A subgroup analysis suggested better outcomes in patients with a good performance status, high PD-L1 score, and a CPS of 10 or higher, and importantly, for MSI [microsatellite instability]—high tumors. This subgroup analysis may support the use of pembrolizumab in MSI-high tumors even in the second-line setting. This was also described in the Japanese guidelines as well as the ESMO [European Society for Medical Oncology] guidelines published in Annals of Oncology last year.

After that, as you know, many trials are ongoing. At this ASCO [American Society of Clinical Oncology Annual Meeting], now we have a first-line agent, KEYNOTE-062, which compared pembrolizumab plus chemotherapy, and pembrolizumab monotherapy, and chemotherapy in patients with PD-L1—positive tumors.

Very disappointingly, chemotherapy plus pembrolizumab did not show a statistical improvement in overall survival and PFS. This is very disappointing because many trials already showed the benefit with the chemotherapy combination in other types of tumors. Another point is a noninferiority of pembrolizumab monotherapy. There may be some negative comments because the upper limit of the hazard ratio is set at 1.2. The actual hazard ratio of upper limits showed 1.18, and some investigators say this is not acceptable.

These are the pros and cons. This hazard ratio is based on 90.2% confidence interval, because this trial is very complicated with many primary endpoints. If you look at a 95% confidence interval, the upper limit is 1.1. I think this hazard ratio itself is acceptable as a noninferiority. And clear safety—better safety results that support noninferiority. But the most important point is a closer survival curve. Still, we need to pick up after patients who achieved a good outcome.

A CPS of 10 showed a very good outcome. There was a hazard ratio of 0.69, and an improvement of median survival by 6 months. This cannot be achieved with any targeted therapy from before. I still think a CPS of 10, and maybe in the future we can report MSI. This is also very important.

Johanna C. Bendell, MD: You reported the KEYNOTE-181 study, which also looked at patients with a CPS of 10 in the second-line setting. Tell us a little about that.

Kohei Shitara, MD: Yeah, this is actually slightly different from the previous trials. This is mainly the squamous type. The subgroup analysis, primary analysis, has 3 groups—the ITT [intent-to-treat] group, the squamous group, and a CPS of 10 subgroup. The CPS of 10 subgroup met the primary endpoint to show an improvement in overall survival. This also supported a CPS of 10. These are important predictors for better outcomes with pembrolizumab.

Transcript Edited for Clarity