The Role of Deep Learning in Diagnostic Imaging of Spondyloarthropathies: A Systematic Review

Abstract

Aim: Diagnostic imaging is an integral part of identifying spondyloarthropathies (SpA), yet the interpretation of these images can be challenging. This review evaluated the use of deep learning models to enhance the diagnostic accuracy of SpA imaging. Methods: Following PRISMA guidelines, we systematically searched major databases up to February 2024, focusing on studies that applied deep learning to SpA imaging. Performance metrics, model types, and diagnostic tasks were extracted and analyzed. Study quality was assessed using QUADAS-2. Results: We included 22 studies demonstrating that deep learning aids in diagnosing and classifying SpA, differentiating arthritis forms, and estimating disease progression and structural changes. These models, particularly those using advanced U-Net architectures, consistently outperformed traditional diagnostic methods, showing a notable increase in diagnostic accuracy. Conclusion: Deep learning models are excellent for augmenting the accuracy of SpA imaging diagnostics. Despite their potential, challenges in overcoming retrospective study biases and integrating these models into clinical practice remain. Future directions should aim to validate these models in real-world clinical settings.