Perioperative Pembrolizumab/Chemo Regimen Yields Superior OS in Resectable Gastric Cancer

Kohei Shitara, MD

Perioperative pembrolizumab (Keytruda) plus chemotherapy followed by pembrolizumab monotherapy significantly improved overall survival (OS) outcomes and maintained superior pathologic complete response (pCR) rates vs neoadjuvant chemotherapy alone in patients with locally advanced, resectable gastric and gastroesophageal junction (GEJ) adenocarcinoma, according to results from the final OS analysis of the phase 3 KEYNOTE-585 trial (NCT03221426) presented at the2024 ESMO Gastrointestinal Cancers Congress.¹

At a median follow-up of 59.9 months (range, 39.0-75.8), patients treated with pembrolizumab plus chemotherapy in the main cohort (n = 402) achieved an OS rate of 45% and a median OS of 71.8 months (95% CI, 52.5-not reached [NR]). Conversely, those treated with the placebo regimen in this cohort (n = 402) experienced an OS rate of 51% and median OS of 55.7 months (95% CI, 41.7-NR; HR, 0.86 [95% CI, 0.71-1.06]). Notably, the 36- and 60-month OS rates in the pembrolizumab group were 65% and 54%, respectively; the corresponding rates with placebo were 59% and 48%.

In the intention-to-treat (ITT) population, which included the populations of patients treated with the main and FLOT (fluorouracil, leucovorin, oxaliplatin, and docetaxel) regimens, the OS rate in the pembrolizumab group (n =502) was 43%, and the median OS was NR (95% CI, 59.2-NR). In the placebo group (n = 505), the OS rate was 50%, and the median OS was 55.7 months (95% CI, 42.8-NR; HR, 0.86 [95% CI, 0.71-1.03]). The 36- and 60-month OS rates in the pembrolizumab group were 66% and 55%, respectively; the corresponding rates with placebo were 59% and 49%.

OS outcomes trended in favor of pembrolizumab across most patient subgroups in both the main cohort and the ITT population, except for patients 65 years of age or older, along with those with stage II or IVa disease, GEJ tumors, and mixed or unknown histologic subtypes.

“Efficacy outcomes from the final analysis were consistent with those of prior reports,” presenting author Kohei Shitara, MD, director of the Department of Gastrointestinal Oncology at the National Cancer Center Hospital East in Kashiwa, Japan, detailed in a presentation of the data.

The updated pCR rate in the main cohort was 13.4% (95% CI, 10.3%-17.2%) with pembrolizumab and 2.0% (95% CI, 0.9%-3.9%) with the placebo regimen, for an overall difference of 11.4% (95% CI, 8.0%-15.3%). In the ITT population, these rates were 14.2% (95% CI, 11.3%-17.6%) vs 2.8% (95% CI, 1.6%-4.7%), respectively, for an overall difference of 11.4% (95% CI, 8.1%-15.0%).

Event-free survival (EFS) rates in the main cohort were 48% vs 57% with pembrolizumab vs placebo, respectively. The median EFS was 44.4 months (95% CI, 33.0-69.8) with pembrolizumab vs 25.7 months (95% CI, 20.8-36.5) with placebo. In the ITT population, the EFS rates in these respective arms were 47% and 56%, and the median EFS was 47.0 months (95% CI, 36.2-NR) vs 26.9 months (95% CI 22.1-34.7).

The addition of PD-1/PD-L1 inhibitors to chemotherapy has induced survival benefits in first-line gastric cancer and shown promising antitumor activity in the perioperative setting. However, the overall benefit of combining neoadjuvant and adjuvant chemotherapy with immune checkpoint inhibition in patients with locally advanced, resectable gastric/GEJ adenocarcinoma is still unknown, Shitara and colleagues explained.

Previously reported data from the interim analysis of KEYNOTE-585 showed a pCR rate of 12.9% (95% CI, 9.8%-16.6%) with perioperative pembrolizumab plus chemotherapy vs 2.0% (95% CI, 0.9%-3.9%) with chemotherapy alone (P < .00001). The median EFS was 44.4 months (95% CI, 33.0-NR) vs 25.3 months (95% CI, 20.6-33.9; HR, 0.81 [95% CI, 0.67-0.99; P = 0.0198]); however, this improvement did not reach the threshold for statistical significance. Accordingly, OS was not formally tested at this time.²

The final OS analysis of KEYNOTE-585 was conducted 12 months after the third interim analysis and after 361 OS events had been recorded in the main cohort. The data cutoff was February 16, 2024.¹

“No alpha was allocated to OS at final analysis, as EFS was not significant at the time of the third interim analysis,” Shitara noted.

The randomized, double-blind, placebo-controlled trial study enrolled patients at least 18 years of age with treatment-naive localized gastric or GEJ adenocarcinoma, as defined by a T3 or higher primary lesion or the presence of any positive lymph node; a plan to proceed to surgery following neoadjuvant chemotherapy; an ECOG performance status (PS) of 0 or 1, and provision of a tumor sample for PD-L1 testing.

Patients were stratified according to geographic region (Asia vs non-Asia), tumor staging (stage II vs III vs IVa), and chemotherapy backbone.

A total of 1254 patients were screened between October 9, 2017, and January 25, 2021. Those enrolled onto the study were assigned to either the main study cohort or the FLOT cohort. Patients in each cohort were then randomly assigned 1:1 to the experimental arm or the placebo arm. Patients in the experimental arm received 200 mg of intravenous (IV) pembrolizumab on day 1 of each 3-week cycle, plus a chemotherapy regimen comprising either cisplatin plus capecitabine (Xeloda; XP), cisplatin plus 5-fluorouracil (FP), or FLOT. Patients randomly assigned to the placebo arm received 1 of these chemotherapy regimens. After undergoing surgery, patients in either arm received up to 3 additional cycles of the same regimen. This was followed by either pembrolizumab monotherapy or placebo administered once every 3 weeks for up to 11 additional cycles.

The trial’s primary end points included pCR rate per blinded independent central review, EFS per investigator’s assessment, and OS in the main cohort, as well as safety in the FLOT cohort. Key secondary end points included safety in the main cohort, safety, OS, and EFS in the overall study population.

The efficacy-evaluable population was defined as all patients randomly assigned to the main cohort (n = 804) or ITT population (n = 1007); the safety-evaluable population included all patients who underwent random assignment and received at least 1 dose of study treatment (n = 799; n = 1001).

Most patients completed all planned cycles of neoadjuvant treatment in the main cohort (94% with pembrolizumab; 91% with placebo) and FLOT cohort (95%; 94%).

Within the main cohort, 6% vs 9% of patients in the pembrolizumab vs placebo arms, respectively, discontinued all study therapy permanently; 13% vs 15% of patients in these respective arms discontinued treatment prior to resection. Reasons for discontinuation included adverse effects (AEs; 3%; 4%), disease progression (5%; 4%), patient/physician withdrawal, (2%; 3%) and a surgically unresectable tumor (4%; 4%). Five and 11 patients did not undergo surgery in the pembrolizumab and placebo arms, respectively. On-study surgery was completed by 85% and 83% of patients in these respective arms, with 11% of patients discontinuing treatment after surgery in each group. One or more doses of adjuvant treatment were received by 75% of patients in both arms, with 48% of patients in both arms completing adjuvant treatment and 27% of patients in both arms discontinuing therapy in this setting.

In the FLOT cohort, 5% and 6% of patients in the pembrolizumab vs placebo arms, respectively, discontinued all study therapy permanently; 9% vs 12% of patients in these respective arms discontinued treatment prior to resection. Reasons for discontinuation included AEs (3%; 3%), disease progression (2%; 2%), patient/physician withdrawal, (1%; 5%) and a surgically unresectable tumor (3%; 2%). Three patients did not undergo surgery in the pembrolizumab arm vs 1 patient in the placebo arm. On-study surgery was completed by 88% and 87% of patients in these respective arms, with 13% and 15% of patients discontinuing treatment after surgery. One or more doses of adjuvant treatment were administered to 78% of patients in the pembrolizumab arm vs 74% of patients in the placebo arm; 45% vs 54% of patients completed adjuvant treatment and 32% vs 19% of patients discontinued therapy in this setting.

Most patients achieved complete resection with pembrolizumab or placebo in both the main cohort (80%; 75%) and the FLOT cohort (79%; 79%). Within the ITT population, the median age was 64 years (range, 22-90) for patients in the pembrolizumab arm and 63 years (range, 25-84) in the placebo group. Most patients were male (73%; 71%), White (54%; 54%), from Asia (38%; 38%), and had an ECOG PS of 0 (73%; 72%). Regarding PD-L1 status, most patients in these respective arms had a combined positive score (CPS) of 1 or greater (73%; 76%), followed by a CPS of less than 10 (66%; 66%), 10 or greater (25%; 28%), and less than 1 (18%; 18%); PD-L1 status was not evaluable or missing for the remaining patients.

High microsatellite instability (MSI-H) was observed in 9% and 8% of patients in the pembrolizumab vs placebo arms, respectively; non–MSI-H disease occurred in 78% and 76% of patients in these respective groups. Tumors were located in either the stomach (75%; 76%) or the GEJ (25%; 23%). Patients had diffuse (40%; 44%), intestinal (49%; 44%), or mixed/unknown (11%; 12%) histological subtype, and stage II (21%; 22%), III (75%; 74%), or IVa (4%; 4%) disease. In both arms, 80% of patients received the XP/FP chemotherapy backbone.

Regarding safety, Shitara reported that, “The observed safety profile was consistent with prior reports for pembrolizumab plus chemotherapy in metastatic disease.”

Across all phases, 99% of patients in both the pembrolizumab and placebo groups experienced an AE of any grade. Treatment-related AEs (TRAEs) were reported in 95% and 96% of patients in these respective groups, and 64% and 63% of AEs were grades 3/4. Grade 5 AEs were observed in 4 patients treated with the pembrolizumab regimen and 2 patients treated with the placebo regimen. AEs leading to discontinuation of any study drug occurred in 26% and 20% of patients in the pembrolizumab and placebo arms, respectively. Surgery-related AEs occurred in 23% and 20% of these patients, respectively.

In order of prevalence, common TRAEs of any grade occurring in at least 10% of patients included nausea, decreased neutrophil counts, decreased appetite, neutropenia, anemia, vomiting, fatigue, white blood cell decreases, palmar-plantar erythrodysesthesia, stomatitis, asthenia, constipation, decreased platelet counts, dysgeusia, and increased blood creatine.

Furthermore, 29% and 10% of patients treated with the pembrolizumab vs placebo regimens, respectively, experienced any-grade immune-mediated AEs and infusion reactions, 11% and 3% of which were grade 3/4, respectively, in these 2 arms. Seven patients in the pembrolizumab arm discontinued treatment with any drug due to these AEs compared with 4 patients in the placebo group. No grade 5 immune-mediated AEs and infusion reactions were reported, and 1 patient in the pembrolizumab group experienced a surgery-related AE of this type.

Any-grade immune-mediated AEs and infusion reactions included hypothyroidism, colitis, pneumonitis, hyperthyroidism, adrenal insufficiency, severe skin reactions, hepatitis, thyroiditis, pancreatitis, hypophysitis, sarcoidosis, type 1 diabetes mellitus, Guillain-Barré Syndrome, uveitis, and vasculitis.

“Additional studies are needed to elucidate the effectiveness of immune checkpoint inhibitors in this setting,” Shitara concluded.

Disclosures: Dr Shitara reports serving in consultancy/advisory roles for Eli Lilly, Bristol-Myers Squibb, Takeda Pharmaceuticals, Pfizer, Ono Pharmaceutical, Novartis, AbbVie, Daiichi Sankyo, Taiho Pharmaceutical, GlaxoSmithKline, Amgen, Boehringer Ingelheim, MSD, Astellas, Guardant Health Japan, and Janssen; receiving honoraria from Bristol-Myers Squibb, Takeda Pharmaceuticals, and Janssen; and receiving institutional grant/research funding from Astellas Pharma, Ono Pharmaceutical, Daiichi Sankyo, Taiho Pharmaceutical, Chugai Pharmaceutical, MSD, Amgen, Eisai, and Medi Science.

References

1. Shitara K, Rha SY, Wyrwicz LS, et al. Final analysis of the phase III KEYNOTE-585 study of pembrolizumab plus chemotherapy vs chemotherapy as perioperative therapy in locally-advanced gastric and gastroesophageal junction cancer. Ann Oncol. 2024;35(suppl 1):S212-S214. doi:10.1016/annonc/annonc1497
2. Shitara K, Rha SY, Wyrwicz LS, et al. Neoadjuvant and adjuvant pembrolizumab plus chemotherapy in locally advanced gastric or gastro-oesophageal cancer (KEYNOTE-585): an interim analysis of the multicentre, double-blind, randomised phase 3 study. Lancet Oncol. 2024;25(2):212-224. doi:10.1016/S1470-2045(23)00541-7