Mismatch repair deficiency (MMR-d) and microsatellite instability (MSI) are prognostic and predictive biomarkers in oncology. Current testing for MMR/MSI relies on immunohistochemistry (IHC) for MMR proteins and molecular assays for MSI detection. This combined diagnostic strategy, however, lacks tumor specificity and does not account for gene variants. This study provides an in-depth analysis of MMR mutations frequency, spectrum, and distribution in solid tumors. Data from 23,893 patients across 11 tumor types, using 66 publicly available studies, were analyzed. MMR-mutated (MMR-m) status was defined by alterations in MLH1, PMS2, MSH2, and/or MSH6; MSI was assessed by MSIsensor. Cases with indeterminate labelling were excluded. Survival was analyzed using the Kaplan-Meier method. Among 19,353 tumors, 949 MMR variants were identified, comprising 432 pathogenic and 517 variants of unknown significance (VUS), as defined by OncoKB. MSH6 mutations were the most frequent (n = 279, 29.4 %), followed by MSH2 (n = 198, 20.9 %), MLH1 (n = 187, 19.7 %), and PMS2 (n = 161, 16.9 %). MMR-m cases were more frequent in endometrial (EC, 20.5 %), colorectal (CRC, 8.2 %), bladder (BLCA, 8.7 %), and gastroesophageal cancers (GEC, 5.4 %). Pathogenic mutations were more common than non-pathogenic in EC, CRC, and GEC (p < 0.001, p = 0.01, p = 0.32, respectively). MMR-m status was not associated with MSI in 247 (48.9 %) cases, including 67 (13.2 %) with pathogenic mutations. The highest concordance between MMR-m and MSI was observed in CRC (65.7 %), EC (91.2 %), and GEC (69.6 %), while the lowest in pancreatic (0.2 %) and lung cancers (0.1 %). MMR-m GECs showed improved overall survival compared to MMR-wt (p = 0.009), a relationship not observed in other tumor types. This study demonstrates that the MMR spectrum is extremely hetoerogeneous in solid tumors, highliting the need for comprehensive and tumor-specific testing strategies.
Read full abstract