Advancements and Challenges of Deep Learning in Diagnostic Radiology: A Systematic Literature Review

Abstract

The rapid integration of Deep Learning (DL) in medical imaging is revolutionizing radiology and addressing critical challenges in diagnostic accuracy and healthcare delivery. In Indonesia and other developing countries, the shortage of radiologists and uneven distribution of healthcare services underline the urgency of exploring DL applications as potential solutions. This study aims to systematically review recent trends, effectiveness, and challenges of DL in diagnostic radiology, as well as to provide insights into its potential adaptation in the Indonesian healthcare system. Using a systematic literature review of peer-reviewed articles (2020–2025) from PubMed, IEEE Xplore, ScienceDirect, and Google Scholar, we identified and synthesized evidence on DL applications across multiple imaging modalities, including CT, MRI, X-ray, and ultrasound. Results show that DL achieves radiologist-level accuracy in tasks such as disease detection, segmentation, and automated report generation, while also improving workflow efficiency and clinical decision-making. However, challenges remain in terms of data availability, model interpretability, ethical issues, and clinical integration. This study provides recommendations for advancing DL adoption in radiology, emphasizing the need for standardized validation, clinician training, and context-specific implementation strategies in Indonesia. The findings highlight both the global and local significance of DL in enhancing healthcare access and equity.