A real-world comparison of overall survival (OS) in deficient mismatch repair/microsatellite instability high (dMMR) versus proficient mismatch repair/microsatellite instability low (pMMR) disease in gastroesophageal (GE) cancer patients (pts).

Abstract

400 Background: Immune checkpoint inhibitors (ICIs) have drastically changed the landscape of treatment in GE cancer. The prevalence of dMMR in GE cancer is estimated around 4,7 to 11,3 % depending on reports. The dMMR subpopulation gained the most benefit from immunotherapy with an improved OS compared to pMMR pts in different trials. Whether chemotherapy alone leads to inferior outcomes in this subpopulation remains a question. We aim to characterize and compare long-term survival outcomes in dMMR and pMMR pts. Methods: Consecutive adult GE cancer pts assessed and/or treated at University Health Network (Toronto, Canada) from January 2007 to April 2025 were included. Those with unknown MMR status and/or positive for HER2 overexpression were excluded. dMMR and pMMR pts were matched in 1:2 fashion according to age, sex, ECOG, number of metastatic sites (for metastatic pts), and pathological N (pN) stage (for localized pts). OS was defined as time of diagnosis to time of death/lost to follow-up. Differences in survival curves between dMMR and pMMR cohorts were estimated using the log-rank test. Results: A total of 628 pts was included, including 331 pts with localized disease and 297 with metastatic disease. The prevalence of dMMR pts was 75/628 (11.9%) ((52/331 (15.7%) and 23/297 (7.7%) respectively for localized and metastatic pts). dMMR pts were older compared to pMMR pts in both cohorts 74,1 vs 66,4 and 74.1 vs 61.7 years respectively (p <0,001). In both metastatic and localized cohorts, median OS was longer for dMMR pts compared to pMMR pts (metastatic: 13,5 vs 10,3 months p = 0.9; localized: 72,3 vs 57,8 months p >0.9, respectively), but the differences were not statistically significant. Among pts with metastatic disease, 144 out of 297 (48.4%) received ICIs at any point, including only 34.7% of those in the dMMR subgroup. In the localized cohort, 49 out of 331 pts (14.8%) received ICIs at any point, including only 15% of the dMMR subgroup. Conclusions: This study highlights the notable prevalence of dMMR GE cancer and a higher median age at diagnosis in our subpopulation. We showed a trend toward better OS in the dMMR cohort compared to the pMMR cohort, though not statistically significant which may be due to the small sample size and variations in standards-of-care over time. It can be suggested that pts with dMMR GE cancers do not experience worse outcomes with chemotherapy compared to pMMR pts, as a minority of our dMMR cohort received any ICI in their disease course. Baseline demographics (metastatic cohort).   Full Sample (n=297) MSI (n=23) MSS (n=274) p-value Age at diagnosis       <0.001    Median (Q1,Q3) 62.3 (52.4, 71.5) 74.1 (59.6, 79.5) 61.7 (51.6, 71.0)   Sex       0.18    Female 98 (33) 11 (48) 87 (32)      Male 199 (67) 12 (52) 187 (68)   ECOG       0.028    0-1 239 (80) 14 (61) 225 (82)      2-4 58 (20) 9 (39) 49 (18)