Predictive biomarkers of response to immune checkpoint inhibitors in mismatch repair-deficient endometrial cancer

Comprehensive Summary

This review examines predictive biomarkers of response to immune checkpoint inhibitors in mismatch repair-deficient (MMRd) endometrial cancer. MMR deficiency causes many mutations and neoantigens to accumulate which causes a tumor microenvironment filled with immune cells susceptible to checkpoint blockage. While these inhibitor therapies have improved outcomes in patients with advanced MMRd endometrial cancer, responses are highly heterogeneous in which half of patients fail to respond or experience early disease progression. This systematic review evaluates established biomarkers and emerging candidates across genomic, epigenomic, transcriptomic, and proteomic domains. It also looks at microenvironmental and tumor-specific mechanisms of resistance, including defects in antigen presentation and interferon signaling. The authors highlight the usage of machine-learning models that incorporate multi-omic and clinical data to develop better therapeutic and patient stratification strategies. Through their review, they found that immune checkpoint inhibitors produce durable responses in some patients with MMRd endometrial cancer, but ~40–50% do not respond, revealing major heterogeneity within this group. Traditional biomarkers like TMB and PD-L1 have limited predictive value, while emerging multi-omic and tumor microenvironment–based signatures show stronger potential to distinguish responders from non-responders.

Outcomes and Implications

This review highlights a major clinical gap: although immune checkpoint inhibitors are standard for MMRd endometrial cancer, a large proportion of patients derive little or no benefit, meaning current treatment selection is imprecise. It underscores that relying on MMRd status alone leads to overtreatment and missed opportunities to direct non-responders toward more effective or combination therapies earlier. In addition, the findings emphasize that current biomarkers (TMB, PD-L1) are inadequate, reinforcing the need for better tools before making individualized treatment decisions. Clinically, it supports a shift toward more personalized, biology-driven care, incorporating tumor microenvironment and multi-omic data rather than single markers. It also suggests that future standard-of-care will likely involve composite biomarker models (potentially ML-based) to guide immunotherapy use. The review also informs practice by encouraging greater use of clinical trials, especially for patients at high risk of immune checkpoint inhibitors resistance. Finally, it has therapeutic implications by supporting development and use of combination strategies to overcome resistance, rather than relying on immunotherapy alone.