The thesis focuses on the application of deep learning models to improve chest X-ray analysis in radiology workflows. This research aimed to bridge the gap between CXR analysis with deep learning and the requirements of healthcare. The thesis identified shortcomings in existing deep learning models, including their reliability, equivalent terminologies, and lack of explainability requirements. The research proposed and evaluated various methods to address these issues, including finding inequivalent terminologies in the reference standard, producing explainable outputs for detecting emphysema, and integrating laboratory parameters with chest X-ray results. The thesis also presented a new approach to extend deep learning models to detect and reject irrelevant cases. The results showed that these methods have significant potential to transform the radiological workflow by improving the accuracy and efficiency of CXR interpretation systems. The public defense will include a presentation of the findings and a discussion of the future directions of this research