The promise and challenges of deep learning models for automated histopathologic classification and mutation prediction in lung cancer

The histologic subclassification of tumors as either squamous cell carcinoma or adenocarcinoma is used by convention to guide systemic chemotherapy for patients diagnosed with non-small cell lung cancer (NSCLC) (1,2). The current gold standard for histopathologic classification of tissue specimens is visual microscopic inspection of tissue specimens by pathologists.