Machine learning in gastrointestinal surgery.

Area of Science:

• Surgical Informatics
• Artificial Intelligence in Medicine
• Data Science in Healthcare

Background:

• Machine learning (ML) enables computers to learn from data without predefined equations, widely used for analyzing "big data".
• Gastrointestinal surgery generates diverse data (clinical, imaging, pathology) essential for risk stratification, surgical safety, and patient prognosis.
• The "big data" era accelerates the need for ML algorithms to process and interpret this complex information.

Purpose of the Study:

• To review the applications of machine learning across preoperative, intraoperative, and postoperative phases in surgical practice.
• To highlight the potential of ML in prediction and risk stratification within surgery.
• To identify challenges and essential requirements for integrating ML into routine surgical care.

Main Methods:

• Review of current literature on machine learning applications in gastrointestinal surgery.
• Categorization of ML applications based on surgical phases (preoperative, intraoperative, postoperative).
• Discussion of ML subfields: supervised learning, unsupervised learning, and reinforcement learning.

Main Results:

• ML demonstrates promise in predicting surgical risk and patient outcomes.
• Applications span various surgical phases, leveraging diverse data types.
• Key challenges include the availability and integration of ML models into clinical workflows.

Conclusions:

• Machine learning is a fundamental technology for processing complex surgical data beyond human cognitive limits.
• Effective integration requires robust information systems capable of managing "big data" and interfacing with electronic health records.
• Artificial intelligence, particularly ML, is poised to fundamentally transform daily surgical practice.