We designed a general end-to-end DL framework for information extraction from CXR images (X-data) and CT images (CT-data) that can be considered a cross-domain transfer learning model.