The Cutting Edge: A Systematic Review of Artificial Intelligence and Machine Learning in Predicting Esophagectomy Outcomes

Background: Esophagectomy carries high risks and complications. Traditional outcome prediction models often fail to account for the multifactorial nature of postoperative outcomes. This systematic review evaluates the use of artificial intelligence and machine learning (AI/ML) models in predicting outcomes after esophagectomy. Methods: A systematic review of PubMed and Embase was conducted to identify studies published from January 2014 to June 2024. Inclusion criteria were observational studies or clinical trials using AI/ML models to predict esophagectomy outcomes. Studies were assessed for quality, and data were extracted on study characteristics, patient demographics, AI/ML models used, and predictive outcomes. Results: Of 187 studies identified, 19 met the inclusion criteria. These studies, conducted across various countries, employed models such as random forest, neural networks, and gradient boosting machines. The most common outcomes predicted were postoperative complications, long-term survival, and early recurrence and readmission. Random forest models frequently demonstrated strong predictive performance, with area under the curve values often exceeding 0.8. The studies demonstrated the potential of AI/ML in enhancing prediction accuracy but highlighted significant variability in models and methodologies. Conclusions: AI/ML models show promise in enhancing the predictive accuracy of esophagectomy outcomes. However, the variability in methodologies underscores the need for standardization in model development and validation. Future research should focus on larger, multi-institutional studies to improve the generalizability and clinical applicability of these models.