AI-Based MRI Radiomics for Preoperative Detection of Tumor Deposits in Rectal Cancer: A Bivariate Diagnostic Accuracy Meta-analysis

AI-based MRI radiomics for preoperative tumor deposit detection in rectal cancer shows pooled sensitivity 0.765 and specificity 0.790, with AUC 0.845 across 9 studies (2589 patients).

• Why it matters: Tumor deposits in rectal cancer are linked to worse prognosis and influence risk stratification and treatment planning, but MRI radiomics results vary across studies.
• What they did: A Cochrane- and PRISMA-DTA–adherent systematic review and bivariate diagnostic accuracy meta-analysis pooled MRI radiomics studies in adult rectal cancer patients using histopathology as the reference standard.
• The result: Validation datasets retained similar performance (AUC 0.806; sensitivity 0.751; specificity 0.751), and combining radiomics with clinical/radiological variables increased sensitivity (0.842 vs 0.657, p<0.001) and AUC (0.827 vs 0.752).

AI summary of the abstract · may contain errors · not medical advice · read the paper.