 This study demonstrates the potential of deep learning networks to accurately predict mortality risk in non-small cell lung cancer patients. By analyzing standard of care CT scans, the researchers were able to identify prognostic signatures which could be used to stratify patients into low and high mortality risk groups. These signatures were found to be more accurate than traditional methods such as age, gender, and tumor node metastasis staging. Furthermore, the researchers were able to identify specific features of the CT scan which contributed to the accuracy of the model. This suggests that deep learning networks can be used to improve patient stratification and potentially lead to improved outcomes. This article was authored by Ahmed Hosni, Chintan Parmer, Tbopee Kohler, and others.