Artificial Intelligence in Radiotherapy for Rectal Cancer

Radiotherapy is a key component of neoadjuvant and radical treatment for rectal cancer, yet it faces challenges such as inefficient target delineation, significant individual variability in treatment response, and difficulties in toxicity prediction. Artificial intelligence (AI), particularly deep learning models, has significantly enhanced the precision and efficiency of radiotherapy by enabling high‐accuracy automatic organ‐at‐risk contouring (Dice similarity coefficient > 0.85), intelligent plan optimization (time efficiency improved by 40%–60%), and the construction of multimodal dose‐toxicity prediction models (AUC 0.82–0.93). This review systematically summarizes recent advances in AI applications for rectal cancer radiotherapy, focusing on convolutional neural network‐based auto‐segmentation, generative adversarial network‐assisted dose prediction, and toxicity risk stratification models integrating radiomic and genomic features. It aims to provide a theoretical basis and clinical practical guidance for AI‐enhanced precision radiotherapy in rectal cancer.