Application of artificial intelligence in cancer rehabilitation: A scoping review

Background Cancer rehabilitation faces challenges, including resource limitations, workforce shortages, and a lack of personalized care. Artificial intelligence (AI) offers promising solutions with the potential to personalize and scale rehabilitation services. However, realizing this potential requires a systematic understanding of current applications and knowledge gaps, which remains lacking. Objective This scoping review maps AI applications in cancer rehabilitation, identifies relevant technologies and clinical scenarios, and highlights existing knowledge gaps. Methods Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews guidelines, eight databases (CNKI, Wanfang, VIP, Chinese Biomedical Literature Database, PubMed, Web of Science, Embase, and the Cochrane Library) were searched from January 2011 to March 2026. Studies applying AI in cancer rehabilitation were included. Two reviewers independently extracted data and assessed study quality using the Quality Assessment with Diverse Studies (QuADS) tool. Results Of 3,032 records identified, 49 studies were included. Breast cancer was the most frequently studied cancer type (27%), followed by studies involving multiple cancer types (35%), lung cancer (12%), and head and neck cancer (6%). Machine learning was the predominant AI approach (41%), followed by conversational AI (23%) and neural networks (12%). Six application scenarios were identified: postoperative rehabilitation, adverse drug reaction management, psychological rehabilitation, personalized nutrition and exercise, survival prediction, and remote follow-up. Remote monitoring systems consistently reduced symptom burden and improved quality of life. AI-powered platforms improved functional recovery, whereas conversational AI was associated with reduced distress, anxiety, and depression. Ensemble machine learning models outperformed traditional staging systems in predicting recurrence and survival. Conclusions AI shows considerable promise in cancer rehabilitation, particularly for symptom monitoring, functional recovery, and survival prediction. However, gaps remain, including a lack of large-scale randomized controlled trials, limited long-term follow-up, and insufficient attention to psychological rehabilitation. Future research should prioritize rigorous trials, standardized outcome measures, and equitable access to AI-enabled rehabilitation services.