The new SUMPOT to predict postoperative complications using an Artificial Neural Network

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Surgical Risk Assessment

Background:

• Accurate preoperative risk assessment is crucial for optimizing hospital resources and reducing patient morbidity and mortality.
• Identifying high-risk surgical patients preoperatively can significantly improve outcomes.

Purpose of the Study:

• To implement an automated surgical risk calculator using Artificial Neural Network (ANN) technology.
• To identify patients at high risk for postoperative complications (PoCs).

Main Methods:

• Developed SUMPOT, an ANN-based risk calculator, using established risk factors.
• Trained and tested the model on a cohort of 560 surgical patients.
• Validated the ANN's predictive accuracy on a separate testing set.

Main Results:

• The ANN model achieved an average classification accuracy of 90% in the testing set.
• Successfully identified patients likely to experience postoperative complications with high accuracy (90.2% control, 88.9% PoC groups).
• Out of 560 patients, 77 (13.7%) experienced PoCs.

Conclusions:

• The ANN-based SUMPOT demonstrates strong performance in predicting postoperative complications.
• This tool has potential value for perioperative management of surgical patients.
• Further clinical validation is recommended for routine practice integration.