Mitochondria-related chemoradiotherapy resistance genes-based machine learning model associated with immune cell infiltration on the prognosis of esophageal cancer and its value in pan-cancer.

Esophageal cancer, known for its high incidence and low five-year survival rate, poses significant treatment challenges. A key aspect of this challenge is the close link between mitochondria and resistance to chemoradiotherapy (CRT). Currently, there is a scarcity of biomarkers for predicting CRT response and prognosis in esophageal cancer. Our study addresses this gap by developing a prognostic model that incorporates mitochondria-related CRT resistance (MRCRTR) genes, including CTSL, TBL1X, CLN8, MMP1, PDPN, and MRPL37. Survival analysis using Kaplan-Meier curves reveals that patients with high MRCRTR scores have lower survival rates than those with low scores. Utilizing a nomogram, we successfully predict the one-, two-, and three-year overall survival rates for esophageal cancer patients. Cox regression analysis confirms the MRCRTR score as an independent prognostic factor. Furthermore, our single-cell and correlation analyses suggested that MRCRTR genes might influence CRT resistance by modulating the immune microenvironment and impacting angiogenesis. Our pan-cancer analysis also indicates the potential applicability of MRCRTR scores to head and neck squamous cell carcinoma. The validation of these findings, conducted with samples from Xiang-ya Hospital, aligns closely with our bioinformatics results. Our study not only explores the role of MRCRTR genes in predicting the prognosis of esophageal cancer but also enhances the understanding of the interplay between CRT, mitochondria, and patient outcomes.