AI based advances in diagnosis of chronic obstructive pulmonary disease: A systematic review

Chronic Obstructive Pulmonary Disease (COPD) is one of the major global health problems, and early detection plays a great role in improving outcomes for patients. Traditional methods of diagnosis are generally based on subjective interpretation, thus delaying diagnosis in many cases. Artificial Intelligence presents a disruptive opportunity, making the detection and classification of COPD possible with a variety of data types. This paper reviews the use of AI-based approaches in COPD diagnosis by using three primary types of datasets: text data, such as clinical notes and electronic health records; audio data, including lung sounds, cough signals, and so on; and image data from chest X-rays and CT scans. Discussing the use of deep learning techniques, specifically CNNs, in analyzing images, we identify how these networks can successfully classify COPD cases along with the level of severity. The potential of AI models in COPD diagnostics is very promising, though there are areas of challenges like data standardization, model generalizability, and interpretability. This review emphasizes the AI potential for COPD diagnostics revolution and outlines future research directions: integration of multi-modal data and advancements in model transparency to support clinical adoption.