ClinNet: Evidential Ordinal Regression with Bilateral Asymmetry and Prototype Memory for Knee Osteoarthritis Grading

TL;DR

This work addresses KOA radiographic grading, which is hindered by subtle inter-grade differences and annotation uncertainty, by reframing grading as an evidential ordinal regression problem. ClinNet combines a Bilateral Asymmetry Encoder, a Diagnostic Memory Bank, and an Evidential Ordinal Head based on the Normal–Inverse-Gamma ($NIG$) distribution to output a continuous severity estimate and principled uncertainty. It achieves state-of-the-art performance (e.g., Quadratic Weighted Kappa = $0.892$ and Accuracy = $0.768$) while providing uncertainty signals that flag out-of-distribution and misdiagnosed cases, supporting safe clinical deployment. The framework demonstrates clinically meaningful attention shifts toward the medial compartment with disease progression and enables uncertainty-driven selective prediction to balance automation with expert review, offering a practical path toward triage-enabled radiographic KOA assessment.

Abstract

Knee osteoarthritis (KOA) grading based on radiographic images is a critical yet challenging task due to subtle inter-grade differences, annotation uncertainty, and the inherently ordinal nature of disease progression. Conventional deep learning approaches typically formulate this problem as deterministic multi-class classification, ignoring both the continuous progression of degeneration and the uncertainty in expert annotations. In this work, we propose ClinNet, a novel trustworthy framework that addresses KOA grading as an evidential ordinal regression problem. The proposed method integrates three key components: (1) a Bilateral Asymmetry Encoder (BAE) that explicitly models medial-lateral structural discrepancies; (2) a Diagnostic Memory Bank that maintains class-wise prototypes to stabilize feature representations; and (3) an Evidential Ordinal Head based on the Normal-Inverse-Gamma (NIG) distribution to jointly estimate continuous KL grades and epistemic uncertainty. Extensive experiments demonstrate that ClinNet achieves a Quadratic Weighted Kappa of 0.892 and Accuracy of 0.768, statistically outperforming state-of-the-art baselines (p < 0.001). Crucially, we demonstrate that the model's uncertainty estimates successfully flag out-of-distribution samples and potential misdiagnoses, paving the way for safe clinical deployment.

Figures (7)

• Figure 1: Overview of the ClinNet architecture. The model takes paired knee images, extracts features via the BAE, aligns them using the Memory Bank, and outputs evidential parameters $\gamma,\nu,\alpha,\beta$.
• Figure 2: The Evidential Learning process. Unlike Softmax, the evidential (EDL) head predicts a distribution over possible severity scores, producing $\gamma,\nu,\alpha,\beta$ and enabling uncertainty estimation through evidence strength.sensoy2018edlamini2020deepedlregression
• Figure 3: Radar Chart Comparison (Normalized). ClinNet demonstrates superior performance across Accuracy, Kappa, F1, and Recall.
• Figure 4: Detailed diagnostic analysis including OA detection performance and ordinal consistency.hanley1982aucdelong1988rocdavis2006prroccao2020coral
• Figure 5: Clinical utility and cost-effectiveness analysis.vickers2006decisionvickers2016netbenefitgeifman2017selective