Abstract
Emphysema detection has traditionally relied on clinical assessments and imaging modalities, such as computed tomography (CT) scans. Recently, artificial intelligence (AI)-based imaging modalities have been explored for their potential to enhance detection accuracy and efficiency. I conducted a systematic review to assess the efficacy of AI-based imaging modalities in detecting emphysema. Multiple databases were searched using tailored MeSH phrases, keywords, and Boolean operators. Each study was evaluated for quality using the QUADAS-AI tool and the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach to assess the certainty of the evidence. Ten studies were included, that demonstrated varied efficacy of AI in the detection and quantification of emphysema. AI tools were generally found to correlate well with clinical pulmonary function tests and were able to accurately quantify emphysema from imaging studies. Some studies reported that AI was comparable or superior to radiologists in detecting emphysema, particularly in the context of lung cancer screenings. However, effectiveness varied with different AI approaches and settings. AI also showed potential in differentiating severity levels of chronic obstructive pulmonary disease (COPD) and outperforming traditional densitometry methods in texture analysis. AI-based imaging modalities show promise in enhancing the detection and management of emphysema. They offer potential not only as supplementary tools to traditional methods but also, in some cases, as superior alternatives. The integration of AI into clinical practice could improve diagnostic accuracy and patient outcomes, although variability in AI performance across different contexts should be considered.

Key words: AI imaging, emphysema detection, systematic review, QUADAS-AI, GRADE methodology, pulmonary function tests, COPD, texture analysis