Quantifying Radiographic Knee Osteoarthritis Severity using Deep Convolutional Neural Networks

TL;DR

This paper introduces a deep learning approach using CNNs to automatically quantify knee osteoarthritis severity from radiographs, improving accuracy over previous methods by framing the task as a regression problem.

Contribution

It demonstrates that pre-trained CNNs fine-tuned on knee OA images significantly outperform shallow classifiers and advocates for using regression metrics for evaluation.

Findings

Deep CNNs outperform shallow classifiers in KL grade prediction

Regression approach with mean squared error improves accuracy

Results show significant improvement over previous state-of-the-art

Abstract

This paper proposes a new approach to automatically quantify the severity of knee osteoarthritis (OA) from radiographs using deep convolutional neural networks (CNN). Clinically, knee OA severity is assessed using Kellgren \& Lawrence (KL) grades, a five point scale. Previous work on automatically predicting KL grades from radiograph images were based on training shallow classifiers using a variety of hand engineered features. We demonstrate that classification accuracy can be significantly improved using deep convolutional neural network models pre-trained on ImageNet and fine-tuned on knee OA images. Furthermore, we argue that it is more appropriate to assess the accuracy of automatic knee OA severity predictions using a continuous distance-based evaluation metric like mean squared error than it is to use classification accuracy. This leads to the formulation of the prediction of KL grades as a regression problem and further improves accuracy. Results on a dataset of X-ray images and KL grades from the Osteoarthritis Initiative (OAI) show a sizable improvement over the current state-of-the-art.