Background: Mismatch repair (MMR) deficiency is the distinguishing molecular feature of a significant portion of endometrial cancers (UCEC), and tumors with MMR deficiency have been identified as candidates for immune checkpoint blockade therapy. We studied MMR deficiency in UCEC using defective mismatch repair associated mutational signatures (MMRd-ams). Methods: WES-derived somatic mutation data of 531 UCEC samples from TCGA Pan-Cancer Atlas were analyzed. COSMIC mutational signatures for each sample were calculated using the R package deconstructSigs. MMRd-ams were correlated with clinical and molecular features for 507 TCGA samples (cBioPortal). Samples were divided into High MMRd-ams (n = 192) and Low MMRd-ams (n = 315) groups by the average of the representative MMRd-ams (0.2396). Fractions of tumor immune infiltrates were derived from CIBERSORT. Results: A significantly higher percentage of patients (47/192, 24.5%) in the High MMRd group had somatic putative driver mutations in at least 1 of the MMR genes (MLH1, MLH3, MSH2, MSH3, MSH6, PMS1, PMS2 and EPCAM), compared with patients in the Low MMRd group (37/315, 11.8%, p = 0.0003). 15% (54/359) of tumors in the non-MSI subtypes expressed significant MMRd-ams and were categorized in the High MMRd-ams group. Patients in the High MMRd-ams group had longer progression-free survival (PFS) (p = 0.0457, log-rank). Analysis of the inferred composition of tumor immune infiltrates revealed that the High MMRd-ams group had significantly higher fraction of CD8+ T cells (p< 0.0001), higher fraction of T follicular helper cells (p< 0.0001) and lower fraction of M2 macrophages (p< 0.001). Tumors in the High MMRd-ams group also displayed higher mRNA expression levels of immune checkpoint genes: PDCD1 (p = 0.0013), and CTLA4 (p = 0.0016). Conclusions: MMRd-ams may be a prognostic and predictive biomarker with significant clinical impact. High MMRd-ams patients prognostically demonstrated longer PFS. Predictively, high MMRd-ams was associated with increased tumor immune infiltrates and elevated expression levels of CTLA4 and PDCD1, known immune checkpoint genes exploitable by immune checkpoint therapies. MMRd-ams importantly characterized a subset of patients that were non-MSI but fit the MMR deficient phenotype by mutational signature. Together these findings open an avenue for recognizing and treating a previously unidentifiable group of patients.