Innovative Machine Learning Solutions for Automated Kidney Tumor Detection in CT Imaging Through Comparative Analysis

The most common urinary system cancer is kidney cancer, the 7th leading cause of cancers death in men and women worldwide. The early and prompt diagnosis of renal cancer is very important for the survival rate in patients. In this work we provide a study concerning whether machine learning techniques—specifically Random Forest (RF) and Neural Networks (NN)—could be helpful for the computer-aided diagnosis of kidney tumors in CT images. Both models were trained and tested on a dataset from Kagglewhich contain 10000 CT scans, the set is balanced between normal (5000 image) & tumor type of tumors case images. The results indicate that the NN model have higher accuracies than those of RF, 99% versus 98%. Moreover, in all evaluation measures like precision score and recall values F1-score etc, this NN model improved so that it could be treated as a more accurate tool for diagnosis. The automated method minimises the time taken to diagnose and lightens radiologists' load, aiding a timely discovery of cancer at an early stage. The research indicates that adding machine learning tools to clinical practice would allow us to improve the speed and accuracy of diagnosing kidney tumors, with potential benefits for patient care and survival. In the future, bettering these models and studying on larger samples with more diversity are hoped to be achieved.