Stratifying Colorectal Cancer stages through CT scan images using Convolutional Neural Networks

Abstract

Artificial intelligence and deep learning have propelled cancer treatment, achieving a 25% increase in treatment accuracy. By analyzing vast datasets, AI has identified 30% more nuanced patterns, revolutionizing tailored therapies and patient outcomes. The proposed research investigates the feasibility of utilizing Convolutional Neural Networks (CNNs) to determine colorectal cancer staging using CT scan images. We have used VGG16 as the base model by fine- tuning the hyperparameters and the layers to accomplish the desired outcomes. The focus lies in demonstrating CNN’s effectiveness in automating the staging process, potentially providing a reliable and efficient tool for precise cancer diagnosis and treatment planning. In further work the result with tumor stage with patient’s other parameters are integrated to assess the risk level of cancer. The model results in 94.6% accuracy with minimal error rate.