Microsatellite Instability and Mismatch Repair Subclonality in Human Cancers: Biologic Basis, Diagnostic Pitfalls, and Therapeutic Implications with a Focus on Colorectal Cancer.

Microsatellite instability-high (MSI-H) and deficient mismatch repair (dMMR) define a molecular subtype of colorectal cancer (CRC) and predict benefit from immune checkpoint inhibition. Clinically, MSI/MMR assessment may yield discordant results across assays, anatomic sites, or time points. A challenging scenario occurs when tissue analysis identifies microsatellite-stable (MSS) or mismatch repair-proficient (pMMR) CRC, whereas circulating tumor DNA (ctDNA) analysis indicates MSI-H. This review evaluates evidence for MSI/MMR heterogeneity and subclonality in CRC. We reviewed literature on spatial, temporal and subclonal MSI/MMR heterogeneity across human cancers, focusing on CRC. Explanations for tissue-plasma and tissue-tissue discordance, including assay limitations, sampling bias, lesions misattribution, biological evolution, and treatment-related selection, were assessed. Although discordance more commonly reflects assay limitations, sampling bias, or profiling of different lesions, increasing evidence supports genuine biological heterogeneity. Distinct tumor regions may show retained MMR protein expression in one area and regional loss with MSI in another. Noncanonical MMR defects, epigenetic heterogeneity, post-treatment evolution, adaptive mutator-state, and immune selection may also generate dynamic or subclonal instability. Therapeutic relevance may depend not simply on MSI detection, but on whether the unstable clone is sufficiently dominant to generate broadly shared neoantigens across the disease burden. MSI/MMR discordance requires careful interpretation and should not automatically be considered as true biological heterogeneity. Nevertheless, genuine subclonality occurs in CRC and other cancers and may affect responsiveness to immune checkpoint inhibition. Integrated tissue, plasma, spatial and longitudinal analyses may improve treatment decisions and biomarker development.