Comprehensive Summary
Radical cystectomy (RC) is a major surgery involving the removal of the entire bladder. In order to create a machine learning model for prediction of overall RC survival rates and evaluate contributory inflammatory markers, a retrospective analysis using an institutional cystectomy database was used. Three data sets were created and analyzed using ML models with information on patient demographic, clinical, and pathological data. ML results indicated a maximum F1 score of 0.78, and preoperative albumin was found to be the strongest predictive factor for inflammatory markers and survival.
Outcomes and Implications
The use of machine learning models for survival analysis using inflammatory markers provides the opportunity for more efficient, streamlined patient care. This can improve consistency within perioperative care decisions and predictions, allowing for greater standardization among health care providers and medical institutions.