Comprehensive Summary
Perioperative risk stratification is an important process in determining the likelihood a patient is to experience complications before, during, and after surgery. In regards to monitoring acute kidney injuries (AKIs) specifically, current machine learning models have the enormous potential to predict major complications following partial or radical nephrectomy. In a study of almost one thousand patients between January 2017 and March 2023 and the University Medical Center Manheim, it was discovered that from a pool of selected machine learning models, Neural Networks (NN) had the best model fitting for both major complication and acute kidney injury risk stratification. Consequently, this study indicates NN as a valuable tool for future models in predicting and addressing the occurrences of major complications and acute kidney injuries.
Outcomes and Implications
Currently, machine learning models predicting major complications and acute kidney injuries following nephrectomy procedures have not been studied in depth, nor have they been compared to traditional logistic regression models. As such, with continual advancements in technology, this study has revealed the potential of machine learning, specifically NN, to serve as a potential platform for predicting operational risks in urological procedures and surgical interventions for renal tumors. Not only does this offer another tool to be used in the clinical setting, it could also improve patient outcomes and increase the efficiency of procedures treating kidney injuries and diseases.