Overall survival was meaningfully improved with treatment of pembrolizumab in first and later lines for patients with advanced gastric/gastroesophageal junction cancer with a high number of PD-L1–expressing cells in the tumor, lymphocytes, and macrophages compared with chemotherapy.
Zev A. Wainberg, MD
Overall survival (OS) was meaningfully improved with treatment of pembrolizumab (Keytruda) in first and later lines for patients with advanced gastric/gastroesophageal junction (G/GEJ0) cancer with a high number of PD-L1expressing cells in the tumor, lymphocytes, and macrophages compared with chemotherapy.1
A collective analysis of the use of pembrolizumab as first, second, or third or greater line of therapy in patients with advanced G/GEJ cancers across three clinical trials revealed enhanced OS compared with chemotherapy in patients with a combined positive score (CPS) ≥10, said Zev A. Wainberg, MD, at the 2020 GI Cancers Symposium.
“CPS ≥10 is a recognized cutoff, at least for gastric cancer. A CPS ≥10 cutoff represents approximately 30% of patients with gastric cancer,” said Wainberg, co-director of the GI Oncology Program, University of California, Los Angeles. “I think the benefit of pembrolizumab is pretty much in the MSI (microsatellite instability)-high and in the CPS ≥10 populations. I think the best bang for your buck is in the CPS ≥10 patients.”
When used as initial therapy in patients with CPS ≥10 who were enrolled in the KEYNOTE-062 study (n = 182), pembrolizumab monotherapy was associated with a median OS of 17.4 months compared with 10.8 months in the chemotherapy arm (cisplatin plus 5-fluorouracil [5-FU] or capecitabine).
“There’s a higher response rate with chemotherapy in the frontline, but in survival analyses, which is the most important endpoint, there is a much better median OS in the patients with CPS ≥10 with pembrolizumab,” he said.
The objective response rate (ORR) in KEYNOTE-062 in the CPS ≥10 subcohort was 25.0% in the pembrolizumab arm (complete response [CR], 7.6%; partial response [PR], 17.4%) compared with 37.8% in the chemotherapy arm (CR, 4.4%; PR, 33.3%), but the duration of response was longer in the pembrolizumab arm at 19.3 months versus 6.8 months in the chemotherapy arm. The ORR in the overall study population was 14.8% in the pembrolizumab group versus 37.2% in the chemotherapy group, but the duration of response was longer with pembrolizumab (13.7 vs. 6.8 months, respectively).2
When used as second line in patients with CPS ≥10 who were enrolled in KEYNOTE-061 (n = 108), the median OS was 10.4 months in the pembrolizumab arm compared with 8.0 months in the chemotherapy arm (paclitaxel).
When used as third line or later in the KEYNOTE-059 study, pembrolizumab was associated with a median OS of 7.9 months in the subset with CPS ≥10 (n = 46). Because KEYNOTE-059 was a single-arm study,3there was no comparator; however, in KEYNOTE-061 and KEYNOTE-062, pembrolizumab prolonged OS in the CPS ≥10 subset compared with chemotherapy.
The median OS in the pembrolizumab arms in the entire study cohorts was 5.8 months in KEYNOTE-059, 9.1 months in KEYNOTE-061 (vs 8.3 months in paclitaxel arm), and 10.6 months in KEYNOTE-062 (vs 11.1 months with chemotherapy).
In the chemotherapy arm in KEYNOTE-061, patients received paclitaxel at 80 mg/m2on days 1, 8, and 15 of 4-week cycles. In KEYNOTE-062 the chemotherapy regimen comprised at cisplatin, 80 mg/m2every 3 weeks, plus 5-FU at 800 mg/m2/day for 5 days every 3 weeks, or capecitabine at 1000 mg/m2twice daily on days 1 to 14 every 3 weeks (with cisplatin capped at 6 cycles in accordance with guidelines in the country in which patients were enrolled). Pembrolizumab was administered at 200 mg every 3 weeks for up to 35 weeks in all 3 trials.
The 12-month OS rate was 32.6% and the 24-month OS rate was 15.2% with pembrolizumab in patients with CPS ≥10 in KEYNOTE-059.
In KEYNOTE-061, in patients with CPS ≥10, the 12-month OS rates were 45.3% in the pembrolizumab arm and 23.6% in the paclitaxel arm, and the 18-month OS rates were 35.7% and 18.8%, respectively. In KEYNOTE-062, in the subpopulation with CPS ≥10, 12-month OS rates were 56.5% and 46.7% in the pembrolizumab and paclitaxel arms, respectively, and 24-month OS rates were 39.0% and 22.2%, respectively.
The median progression-free survival (PFS) was 2.1 months with pembrolizumab in the CPS ≥10 subset in KEYNOTE-059, with 12- and 24-month PFS rates of 8.7% each. Data from the other 2 studies showed that the median PFS was not longer with pembrolizumab compared with chemotherapy in the first- and second-line settings in the CPS ≥10 subset. In the second-line setting in KEYNOTE-061, the median PFS was 2.7 months and 3.4 months in the pembrolizumab and chemotherapy, arms, respectively, and in the first-line setting in KEYNOTE-062, the median PFS was 2.9 months in the pembrolizumab arm versus 6.1 months in the chemotherapy arm.
The ORR in KEYNOTE-059 with pembrolizumab in the CPS ≥10 subset was 17.4%, with a CR rate of 2.2% and a PR rate of 15.2%. The duration of response in this trial with pembrolizumab was 20.9 months. In patients with PD-L1positive tumors who were enrolled in KEYNOTE-059, the ORR was 15.5% and the duration of response was 16.3 months.3
The ORR in KEYNOTE-061 in the CPS ≥10 subcohort was 24.5% in the pembrolizumab arm (CR, 9.4%; PR, 15.1%) compared with 9.1% in the chemotherapy arm (CR, 1.8%; PR, 7.3%), with a duration of response that was not reached in the pembrolizumab group versus 6.9 months in the chemotherapy group. In the overall KEYNOTE-061 population, the ORR was 16.6% with pembrolizumab and 14% with chemotherapy, with a longer duration of response with pembrolizumab (18.0 vs 5.2 months).4
“Keep in mind that this is a retrospective posthoc analysis of CPS ≥10, but it is, at least so far, the most comprehensive efficacy analysis of CPS ≥10 in the gastric cancer population out there,” said Wainberg.