Predicting morbidity and mortality after radical cystectomy using risk calculators: A comprehensive review of the literature

Introduction: Radical cystectomy (RC) with urinary diversion is associated with significant perioperative morbidity and mortality, varying between 30% and 70% and between 0.3% and 10.6%, respectively. Risk calculators have been extensively studied in the general surgery literature to predict 30- and 90-day postoperative morbidity and mortality but have not been widely accepted in the RC literature. Materials and methods: We performed a search of MEDLINE and Embase databases during May 2020 to identify all relevant studies using the following keywords: radical cystectomy, surgical complication predictive model, surgical complication predictive equation, surgical complication predictive nomogram, surgical risk calculator, morbidity, and mortality. We determined the existing surgical predictive nomograms, calculators, and indices and their accuracy in predicting morbidity, mortality, and major complications after RC. Results: National Surgical Quality Improvement Program had poor accuracy at predicting 30-day morbidity at mortality (AUC 0.5–0.6). LACE index showed good discrimination at predicting 90-day mortality (AUC 0.7). The various frailty and sarcopenia indices have shown poor to fair accuracy at predicting (AUC 0.5–0.7). The Isbarn and Aziz nomograms have equivalent accuracy at predicting 90-day mortality (AUC 0.7) but are limited by inclusion of tumor histology and presence of metastatic disease as variables. POSSUM and P-POSSUM have poor ability at predicting morbidity and mortality (AUC 0.5) and are cumbersome calculators. The surgical Apgar score has been able to predict 30-day morbidity and mortality but can only be used in the postoperative setting. Discussion: The currently available surgical risk calculators have either poor accuracy at predicting post-RC morbidity and mortality or are limited by types of variables included. An ideal risk calculator would be comprised of preoperative factors only and have a high accuracy to serve as a tool for preoperative patient counseling prior to surgery. Conclusion: There exists a strong need to develop a comprehensive and accurate preoperative risk calculator that predicts morbidity and mortality after RC.