Heating the Cold: Overcoming Immunotherapy Resistance in Microsatellite-Stable Colorectal Cancer: A Systematic Review

Colorectal cancer (CRC) has shown significant heterogeneity regarding its response to immunotherapy. Long-lasting, beneficial effects have been observed in mismatch repair-deficient, microsatellite instability-high (dMMR/MSI-H) tumours, while mismatch repair-proficient, microsatellite stable (pMMR/MSS) tumours have remained resistant. Such differences in results have been studied in this review through the "hot" and "cold" tumour concept. It explains how various biological and microenvironmental factors play a role in immune resistance and T-cell priming and infiltration. Key factors include a low neoantigen load and defects in antigen presentation, which reduce the overall immune recognition of tumour cells. The review also studies certain processes such as Wnt/β-catenin and mitogen-activated protein kinase (MAPK) signalling and what input they have in the prevention of effective antitumour immune responses. Conventional treatments like chemotherapy and radiotherapy have been considered alongside more targeted treatments such as the inhibition of vascular endothelial growth factor (VEGF) signalling and the suppression of myeloid-mediated immune evasion, to convert "cold" MSS tumours into immune-responsive lesions. Methods which aim to modify the tumour microenvironment such as metabolic reprogramming and microbiome modulation have also been covered in this review. Artificial intelligence and nanomedicine are new technologies that could provide improved patient stratification and therapeutic precision, although their clinical application in pMMR/MSS CRC remains under investigation. A systematic literature search of PubMed and PubMed Central (PMC) was conducted from 10 June 2026 to 19 August 2026 using predefined eligibility criteria, with study selection reported according to PRISMA 2020. Because of substantial heterogeneity in study design, therapeutic approach and reported outcomes, the included evidence was synthesized narratively rather than by meta-analysis. A total of 158 studies were included.