A review of artificial intelligence-based research on chronic obstructive pulmonary disease

Abstract

In recent years, with the rapid development of artificial intelligence (AI), Chronic Obstructive Pulmonary Disease (COPD), one of the world's three major chronic diseases, has achieved remarkable progress in diagnosis, grading, and prognosis, which is of great significance for promoting the clinical transformation of respiratory diseases. To deeply explore the application of AI in the diagnosis and management of COPD, this paper reviews recent studies based on machine learning and deep learning, covering screening and diagnosis, disease grading and assessment, disease management and monitoring, and treatment. First, the technical basis of COPD-related research is analyzed from five perspectives: traditional research methods, supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning. Then, the commonly used datasets and model evaluation metrics are summarized. Finally, the application scenarios of AI in COPD research are elaborated, focusing on three aspects: early screening and diagnosis, disease monitoring and risk prediction, and disease classification and risk stratification. This paper summarizes the main research hotspots at home and abroad over the past five years with representative models, and analyzes and compares the advantages and limitations of each type of model in specific COPD tasks through comparative experiments. The study also outlines prospects for the future development of this field, aiming to provide theoretical references and insights for subsequent research.

Keywords: Artificial intelligence; Breath sounds; Chronic obstructive pulmonary disease; Clinical decision support; Computer-aided diagnosis; Machine learning.