Risk prediction model for parastomal hernia after colorectal cancer surgery: systematic review and meta-analysis

To systematically assess risk prediction models for Parastomal Hernia after colorectal cancer surgery. We searched databases until October 2025 for studies developing or validating Parastomal Hernia prediction models. Study quality was evaluated using the PROBAST tool. Model performance measures were pooled via meta-analysis, while predictors were summarized qualitatively using narrative synthesis. Seven retrospective studies were included. Four studies contributed to the internal validation pool, yielding a high pooled discrimination (AUC = 0.910). Two studies contributed to the external validation pool, showing good performance (AUC = 0.854). Key predictors consistently identified included stoma diameter, age, sex, BMI, and albumin. All studies were rated at high risk of bias due to retrospective design and methodological limitations. Current prediction models show promising discriminatory performance but are at high risk of bias and have undergone limited external validation. Their clinical applicability remains uncertain. Future research requires rigorous prospective development and broader validation. PROSPERO CRD420251168201.