Predicting EGFR mutation status in lung adenocarcinoma on computed tomography image using deep learning

What is it about?

We proposed a deep learning model to predict epidermal growth factor receptor (EGFR) mutation status in lung adenocarcinoma by computed tomography images. This model is non-invasive and easy to use which requires only the raw CT image without further human-assist. After training the model in 14926 CT images, the deep learning model achieved encouraging predictive performance (AUC=0.81), which is better than using hand-crafted CT features or clinical characteristics.

Featured Image