BARCELONA – The immune checkpoint inhibitor pembrolizumab (Keytruda) did not significantly improve overall survival of advanced/metastatic gastric or gastroesophageal junction cancer compared with paclitaxel, the results of the KEYNOTE-061 study show.

Among 395 patients with gastric or gastroesophageal junction (GEJ) cancer who had expression of the programmed death ligand 1 (PD-L1) on 1% or more of their tumor cells, lymphocytes, and macrophages, median overall survival (OS) for patients treated with pembrolizumab was 9.1 months, compared with 8.3 months. This translated into a hazard ratio (HR) for death in the pembrolizumab arm of 0.82, but with the 95% confidence interval crossing 1.00, and a P value (.04205) that did not meet the prespecified threshold for significance (P equal to or less than .0135).

In addition, neither progression-free survival (PFS) nor overall response rate (ORR) were significantly better with pembrolizumab than with paclitaxel, reported Kohei Shitara, MD, of National Cancer Center East in Kashiwa, Japan.

Despite the failure of the trial to reach its primary endpoint, “these results may support further research to identify patients likely to benefit from pembrolizumab monotherapy and ongoing development of pembrolizumab-based combination therapy for gastric cancer,” he said at the European Society of Medical Oncology World Congress on Gastrointestinal Cancer.

Pembrolizumab is approved by the Food and Drug Administration for the treatment of patients with recurrent locally advanced or metastatic gastric or GEJ adenocarcinoma with tumors confirmed to carry PD-L1 for whom two or more prior lines of therapy had failed.

The approval was based on results of the nonrandomized, open label KEYNOTE-059 trial, which enrolled 259 patients with gastric or GEJ adenocarcinoma that progressed on at least two prior systemic treatments for advanced disease. Of the enrollees, 143 patients had tumors with a PD-L1 Combined Positive Score (CPS) of 1 or greater. The primary trial outcome, the objective response rate for these 143 patients, was 13.3% (95% confidence interval; 8.2-20), with a complete response rate of 1.4% and a partial response rate of 11.9%. The duration of response ranged from at least 2.8 months to at least 19.4 months.

The drug is also approved for unresectable of metastatic gastric tumors with high levels of microsatellite instability that progressed on prior therapy.

The KEYNOTE-061 study was designed to test the proposition that pembrolizumab could improve on paclitaxel for treatment of patients with adenocarcinoma of the stomach or GEJ that was metastatic or locally advanced and unresectable, and for which first-line therapy with a platinum agent and fluoropyrimidine had failed.

A total of 592 patients from Europe, Israel, North America, Asia, and Australia were enrolled and randomly assigned to receive either pembrolizumab 200 mg every 3 weeks for 35 cycles, or paclitaxel 80 mg/m² on days 1, 8, and 15 of each 4-week cycle. Each treatment was continued until the maximum number of cycles (for pembrolizumab) or until confirmed disease progression, intolerable toxicity, patient withdrawal, or investigator’s decision.

As noted, OS for patients with a CPS score of 1 or greater, the primary endpoint, did not differ between treatment arms, but there was a numerical tilt that appeared to be in favor of pembrolizumab.

For example, the 12-month OS rates were 39.8% in the pembrolizumab groups vs. 27.1% in the paclitaxel group, and respective 18-month OS rates were 25.7% and 14.8%.

In an analysis of protocol-specified subgroups, there were general trends slightly favoring the checkpoint inhibitor over the taxane, but the only significant difference was among patients with GEJ cancers as the primary tumor location (HR, 0.61, 95% confidence interval, 0.41-0.90).

Pembrolizumab offered a small but significant survival advantage among patients with Eastern Cooperative Oncology Group performance status of 0, with a median OS of 12.3 months vs. 9.3 months (HR, 0.69, 95% CI, 0.49-0.97).

Analyses of PFS and OS by CPS score groups (less than 1, 1-10, or 10 and higher) showed no significant differences, however.

There were no differences in outcomes among patients with high microsatellite instability tumors, but the total number of patients in this subgroup was small, with just 15 in the pembrolizumab arm and 12 in the paclitaxel arm, and there were trends that appeared to favor pembrolizumab.

Treatment-related adverse events were more frequent with paclitaxel (84.1% vs. 52.7% of patients on pembrolizumab), but pembrolizumab was associated with more treatment-related deaths (three vs. one). Two of the deaths in the pembrolizumab arm were immune mediated. Grade 3 or greater adverse events occurred in 14.3% vs. 34.8%, respectively.

In the question-and-response following Dr. Shitara’s presentation, session comoderator David Cunningham, MD, of Royal Marsden NHS Foundation Trust in Sutton, England, asked why the KEYNOTE-061 investigators did not pit pembrolizumab against the combination of paclitaxel and ramucirumab (Cyramza) “since that’s what many people would use in this situation.”

Dr. Shitara replied that ramucirumab was not available or accepted as an option in many countries when the trial was first planned in 2014. He added that paclitaxel and ramucirumab should be the control arm for clinical trials going forward.

SOURCE: Shitara K. et al. ESMO World Congress on Gastrointestinal Cancer 2018. Abstract LBA-005.

Pembrolizumab does not surpass paclitaxel for gastric cancer