New Publication(Continual Learning in Medical Imaging) in JKSR

We are pleased to announce that our team has published a new review paper in the Journal of the Korean Society of Radiology (JKSR), Special Issue: Informatics and AI.

This review examines continual learning as a practical framework for maintaining artificial intelligence models in medical imaging. AI models for medical imaging are typically trained once and deployed unchanged, yet scanners, protocols, and patient populations continue to evolve, and performance degrades accordingly. Continual learning offers a way to incorporate newly acquired data while limiting the loss of previously learned diagnostic knowledge, making it particularly suited to clinical environments in which historical images cannot be freely retained or reused. The paper outlines the continual learning scenarios and methodological families most relevant to radiology, examines when continual updating is preferable to periodic retraining or simple recalibration, and summarizes applications in classification, segmentation, detection, and multi-joint radiographic grading. It further addresses the clinical constraints that determine feasibility—privacy and regulatory limits on data retention, delayed and noisy labels, class imbalance, and evaluation requirements beyond average accuracy—and argues that continual learning should be implemented as a governed and auditable maintenance process rather than as a single algorithmic choice.

"Continual Learning in Medical Image Analysis: Appropriate Use and Lifecycle Governance" (2026). Journal of the Korean Society of Radiology.

Fig. 1. Lifecycle-aware continual learning for medical imaging AI.