A Novel Approach Based on Hypergraph Convolutional Neural Networks for Cartilage Shape Description and Longitudinal Prediction of Knee Osteoarthritis Progression
Knee osteoarthritis (<i>KOA</i>) is a highly prevalent muscoloskeletal joint disorder affecting a significant portion of the population worldwide. Accurate predictions of <i>KOA</i> progression can assist clinicians in drawing preventive strategies for patients. In this paper...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-04-01
|
| Series: | Machine Learning and Knowledge Extraction |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2504-4990/7/2/40 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Knee osteoarthritis (<i>KOA</i>) is a highly prevalent muscoloskeletal joint disorder affecting a significant portion of the population worldwide. Accurate predictions of <i>KOA</i> progression can assist clinicians in drawing preventive strategies for patients. In this paper, we present an integrated approach based on hypergraph convolutional networks (<i>HGCNs</i>) for longitudinal predictions of <i>KOA</i> grades and progressions from MRI images. We propose two novel models, namely, the <i>C_Shape.Net</i> and the predictor network. The <i>C_Shape.Net</i> operates on a hypergraph of volumetric nodes, especially designed to represent the surface and volumetric features of the cartilage. It encompasses deep <i>HGCN</i> convolutions, graph pooling, and readout operations in a hierarchy of layers, providing, at the output, expressive 3D shape descriptors of the cartilage volume. The predictor is a spatio-temporal <i>HGCN</i> network (<i>ST_HGCN</i>), following the sequence-to-sequence learning scheme. Concretely, it transforms sequences of knee representations at the historical stage into sequences of <i>KOA</i> predictions at the prediction stage. The predictor includes spatial <i>HGCN</i> convolutions, attention-based temporal fusion of feature embeddings at multiple layers, and a transformer module that generates longitudinal predictions at follow-up times. We present comprehensive experiments on the Osteoarthritis Initiative (<i>OAI</i>) cohort to evaluate the performance of our methodology for various tasks, including node classification, longitudinal <i>KL</i> grading, and progression. The basic finding of the experiments is that the larger the depth of the historical stage, the higher the accuracy of the obtained predictions in all tasks. For the maximum historic depth of four years, our method yielded an average balanced accuracy (<i>BA</i>) of 85.94% in <i>KOA</i> grading, and accuracies of 91.89% (+1), 88.11% (+2), 84.35% (+3), and 79.41% (+4) for the four consecutive follow-up visits. Under the same setting, we also achieved an average value of Area Under Curve (<i>AUC</i>) of 0.94 for the prediction of progression incidence, and follow-up <i>AUC</i> values of 0.81 (+1), 0.77 (+2), 0.73 (+3), and 0.68 (+4), respectively. |
|---|---|
| ISSN: | 2504-4990 |