Predicting and Visualizing STK11 Mutation in Lung Adenocarcinoma Histopathology Slides Using Deep Learning

Studies have shown that STK11 mutation plays a critical role in affecting the lung adenocarcinoma (LUAD) tumor immune environment. By training an Inception-Resnet-v2 deep convolutional neural network model, we were able to classify STK11-mutated and wild-type LUAD histopathology images with promising accuracy (per slide AUROC = 0.795). Dimensional reduction of activation maps before output layer test set revealed fewer cells accumulated around cancer STK11-mutation cases. Our study demonstrated model can automatically identify mutations based on slides confirmed cell density was main feature used by distinguish