Classification of medical images with convolutional networks

Abstract

This study explores using Convolutional Neural Networks (CNN) to predict microsatellite instability (MSI) and stability (MSS) from histology images in gastrointestinal cancer. A deep learning model was developed with Keras and TensorFlow in R, applying advanced techniques to histology images. The results show that deep CNN architectures effectively predict MSI and MSS, providing clinicians with a reliable tool to identify the microsatellite stability of tumor tissues.