Artificial intelligence in the surgical treatment of digestive system malignancies: applications and challenges
10.3760/cma.j.cn113855-20250311-00134
- VernacularTitle:人工智能在消化系统肿瘤外科诊疗中的应用与挑战
- Author:
Yuyi ZHANG
1
;
Haoting SUN
1
;
Lunxiu QIN
1
Author Information
1. 复旦大学附属华山医院普外科，上海 200040
- Publication Type:Journal Article
- Keywords:
Artificial intelligence;
Digestive system neoplasms;
Deep learning;
Surgical treatment;
Surgical navigation
- From:
Chinese Journal of General Surgery
2025;40(5):338-346
- CountryChina
- Language:Chinese
-
Abstract:
Artificial intelligence (AI) technology is revolutionizing the precision diagnosis and treatment system in digestive system malignancies. In the preoperative stage, deep learning-driven multimodal data can significantly enhance the accuracy of early lesion detection and organ function assessment, while optimizing the prediction of the efficacy neoadjuvant therapy response with radiomics features. During surgery, the automatic surgical phase alorithm and intelligent pathology testing can effectively coordinate the surgical process; the combination of AI with augmented reality navigation system (AR) and mixed reality technology (MR) can improve the accuracy of intraoperative navigation; enhances intraoperative navigation accuracy; and the autonomous surgical robots can improve their ability to perform key tasks during surgery using sophisticated motion control algorithms. In postoperative management, federated learning promotes secure cross-institutional data sharing and alleviates the problem of data silos; and the development of interpretable models can also provide clearer decision-making basis for complication prediction and prognosis management. However, the current clinical translation process still faces three core challenges: (1) Quality control problems: the dual constraints of insufficient model generalization ability and lack of unified evaluation standards; (2) Methodological challenges: the logical opacity of decision-making and the cognitive barriers of the algorithm's causal reasoning process still affect its credibility; (3) Ethical and legal vacuum: the attribution of medical responsibility is unclear, and the legal system has not yet fully adapted to the development of AI applications. Based on this, the future should focus on: (1) Developing transparent AI systems to achieve a paradigm shift from "black box prediction" to "white box deductions"; (2) Establishing systematic AI evaluation standards and ethical and legal regulatory frameworks; (3) Promoting the construction of specialized biorepositories database and causal graph library for digestive oncology. The aim is to help AI upgrade from an auxiliary tool to an intelligent decision-making partner.
