A clinical decision support system for interventional urinary stone management planning

Urinary stone disease is a very common disease of the urinary tract where excessive mineral content in the kidneys form stones and obstruct the urinary tract, due to which affected patients experience severe pain and uneasiness. Identification of the appropriate interventional stone management method is usually done based on the size and location of the stone. Improper planning of the interventional method can lead to multiple revisits and unnecessary radiation exposure and thus an intelligent clinical decision system to precisely provide a treatment plan is required. In this paper, an intelligent system that recommends an appropriate interventional stone management system is proposed. Stone management data of 600 patients containing information about the stone size, location and the treatment provided to them was used to train machine learning models. The training and testing performance of different machine learning models with the dataset has been compared. Results showed that decision trees and support vector machines showed better results in predicting the right treatment planning method while given necessary inputs (stone size and location). This system can be useful in clinical setups in assisting urologists in planning treatment for urinary stone disease.