Machine learning and deep learning applications in inguinal hernia care show promise for risk prediction and training.

This narrative review examines the current landscape of machine learning (ML) and deep learning (DL) applications within inguinal hernia care. The analysis focuses on diagnostic accuracy, surgical protocols, and patient outcomes, specifically highlighting the potential for prediction of postoperative surgical site infection, surgical site occurrence, intestinal resection in incarcerated inguinal hernia, and postoperative lower extremity venous thromboembolism. Additionally, the review addresses the utility of these technologies in identifying anatomical landmarks, providing real-time feedback, and enhancing surgical training.

Main results indicate that models developed using ML can effectively predict the risks associated with postoperative surgical site infection, surgical site occurrence, intestinal resection in incarcerated inguinal hernia, and postoperative lower extremity venous thromboembolism. Deep learning is described as highly effective for the identification of anatomical landmarks. However, the accuracy and reliability of generative AI were noted to require further validation.

No specific safety data, adverse events, or discontinuations were reported in this narrative review. The primary limitation identified is that the accuracy and reliability of generative AI require further validation. As a narrative review, the study does not provide randomized controlled trial data or specific statistical measures such as absolute numbers or p-values to quantify these findings.

The practice relevance of this evidence lies in offering references for clinical practice and technological innovation. Clinicians should interpret these findings as exploratory. While ML and DL show promise for risk prediction and training, the current evidence is insufficient to mandate their use. Further high-quality studies are needed to confirm efficacy and safety before widespread implementation.