TY - GEN
T1 - Anatomically-Focused Patches for Lightweight and Explainable Knee OA Grading
AU - Chang, Tien En
AU - Lombaert, Herve
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.
PY - 2026
Y1 - 2026
N2 - Current deep learning models for knee osteoarthritis (OA) grading often lack anatomical guidance, limiting their accuracy and explainability. This work proposes a novel framework centered on anatomically-focused patches to overcome these limitations. Our method extracts a set of small image patches from clinically-relevant locations along the joint line, identified by automated landmark detection. These patches are then processed as a bag of instances within an attention-based multiple instance learning (MIL) framework. The MIL model learns to identify and weight the most salient pathological features for an accurate, patient-level diagnosis. Our approach is evaluated on the OAI dataset and achieves state-of-the-art performance with a quadratic weighted Cohen’s Kappa of 0.807. This result outperforms larger baselines such as ResNet-34 while using over 85 times fewer parameters. Furthermore, our attention-weighted visualization method produces sharp, clinically meaningful saliency maps that precisely localize features such as osteophytes and joint space narrowing, in contrast to the diffuse heatmaps of prior work. By combining anatomical guidance with an MIL framework, our work presents a lightweight, accurate and trustworthy solution for automated knee OA grading. The code is available at: https://github.com/tien-endotchang/focused-patch-KOA.
AB - Current deep learning models for knee osteoarthritis (OA) grading often lack anatomical guidance, limiting their accuracy and explainability. This work proposes a novel framework centered on anatomically-focused patches to overcome these limitations. Our method extracts a set of small image patches from clinically-relevant locations along the joint line, identified by automated landmark detection. These patches are then processed as a bag of instances within an attention-based multiple instance learning (MIL) framework. The MIL model learns to identify and weight the most salient pathological features for an accurate, patient-level diagnosis. Our approach is evaluated on the OAI dataset and achieves state-of-the-art performance with a quadratic weighted Cohen’s Kappa of 0.807. This result outperforms larger baselines such as ResNet-34 while using over 85 times fewer parameters. Furthermore, our attention-weighted visualization method produces sharp, clinically meaningful saliency maps that precisely localize features such as osteophytes and joint space narrowing, in contrast to the diffuse heatmaps of prior work. By combining anatomical guidance with an MIL framework, our work presents a lightweight, accurate and trustworthy solution for automated knee OA grading. The code is available at: https://github.com/tien-endotchang/focused-patch-KOA.
KW - Kellgren and Lawrence grading
KW - X-ray
KW - anatomical guidance
KW - knee osteoarthritis
KW - multiple instance learning
UR - https://www.scopus.com/pages/publications/105020023409
U2 - 10.1007/978-3-032-06774-6_6
DO - 10.1007/978-3-032-06774-6_6
M3 - Contribution to conference proceedings
AN - SCOPUS:105020023409
SN - 9783032067739
T3 - Lecture Notes in Computer Science
SP - 74
EP - 86
BT - Shape in Medical Imaging - International Workshop, ShapeMI 2025, Held in Conjunction with MICCAI 2025, Proceedings
A2 - Wachinger, Christian
A2 - Luijten, Gijs
A2 - Egger, Jan
A2 - Elhabian, Shireen
A2 - Gopinath, Karthik
PB - Springer Science and Business Media Deutschland GmbH
T2 - International Workshop on Shape in Medical Imaging, ShapeMI 2025, Held in Conjunction with the 28th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2025
Y2 - 23 September 2025 through 23 September 2025
ER -