MULTIMODAL DEEP LEARNING FOR PULMONARY DISEASE DIAGNOSIS USING CHEST X-RAYS AND CLINICAL REPORTS

Jonathan Jonathan; Benediktus Lysander Tristan Setiawan

Authors

Jonathan Jonathan Universitas Buddhi Dharma Author
Benediktus Lysander Tristan Setiawan Binus University Author

Keywords:

Chest X-Rays, Clinical Reports, Deep Learning, Multimodal Diagnosis, Pulmonary Diseases

Abstract

ABSTRACT
Pulmonary diseases continue to pose significant global health challenges, necessitating improved diagnostic methods. Chest X-rays (CXRs) are widely utilized for pulmonary assessment but are limited by interpretive complexity and variability. Recent advances in artificial intelligence, particularly deep learning, have enhanced medical imaging analysis. This systematic review examines the application of multimodal deep learning approaches that combine CXRs with clinical reports to improve diagnostic accuracy. The study synthesizes findings from 30 peer-reviewed articles published between 2020 and 2025, analyzing model architectures, fusion strategies, and performance outcomes. Results demonstrate that multimodal systems significantly outperform
unimodal approaches, achieving higher diagnostic accuracy and enhanced interpretability. Challenges such as data scarcity, model generalization, and clinical integration persist, emphasizing
the need for standardized datasets, explainable AI, and cross-institutional validations. The review highlights the transformative potential of multimodal deep learning in pulmonary diagnostics and offers recommendations for future research and clinical implementation.