Real-world effectiveness and safety of pembrolizumab in mismatch repair-deficient (dMMR) gastrointestinal non-colorectal tumors.

Abstract

733 Background: The KEYNOTE-158 trial demonstrated durable clinical efficacy with pembrolizumab in long-term follow-up of previously treated unresectable or metastatic MSI-H/dMMR non-colorectal cancer. Our study, the GAIN study, aims to evaluate effectiveness and safety of pembrolizumab in dMMR Gastrointestinal Non-Colorectal Tumors in real-world clinical practice. Methods: We conducted a multicenter, retrospective, observational study on patients with Mismatch Repair-Deficient (dMMR) Gastrointestinal Non-Colorectal Tumors treated with pembrolizumab in real-world practice at six university hospitals affiliated with the Galician Research Group on Digestive Tumors (GiTUD) in Northwest Spain. Results: Between January 2018 and April 2023, 35 patients were enrolled. The median age was 66 years (range 33-88), with 54.3% being male. Tumor locations included 23 (65.7%) gastroesophageal adenocarcinomas, 3 (8.6%) pancreatic adenocarcinomas, 2 (5.7%) duodenal adenocarcinomas, 2 (5.7%) intrahepatic cholangiocarcinoma, 2 (5.7%) jejune adenocarcinomas, 2 (5.7%) esophageal squamous cell carcinomas, and 1 (2.8%) tumor of unknown origin. 54.3% of tumors were poorly differentiated. The most frequent alteration was the loss of MLH1-PMS2 expression (71.4%). By tumor location, 100% of gastroesophageal, jejune, and unknown origin tumors showed MLH1-PMS2 loss; while 100% of duodenal, pancreatic, and biliary tumors exhibited MSH2-MSH6 loss. All patients received pembrolizumab, with 57.1% having no prior metastatic disease treatment, 80% having ECOG PS0-1, and 31.4% having hepatic metastases. Among 33 evaluable patients, the overall response rate (ORR) was 81.8% (including 30.3% complete response) and the disease control rate (DCR) was 90.9%. Specifically, in gastroesophageal adenocarcinomas, the ORR was 81%, and the DCR was 90.5%. In the overall population, the median progression-free survival (PFS) was 34.76 months (95% CI 5.33 – 64.19 months), and the median overall survival (OS) was 39.75 months (95% CI 4.33-75.18 months). Treatment was well-tolerated; only 3 patients discontinued treatment due to nephritis, sarcoidosis-like symptoms, or diarrhea. No treatment-related deaths occurred. Conclusions: Our study underscores the real-world effectiveness of pembrolizumab in patients with dMMR Gastrointestinal Non-Colorectal Tumors. The high ORR, DCR, and favorable safety profile support its use in this challenging patient group, providing valuable real-world evidence for clinical decision-making.