DEEP LEARNING-BASED PREDICTION OF KNEE OSTEOARTHRITIS PROGRESSION USING 3D MRI: AN ANN APPROACH
DOI:
https://doi.org/10.62643/ijerst.2024.v20.i2.pp1329-1335Keywords:
Knee Osteoarthritis (OA), Artificial Neural Network (ANN), Cartilage Damage Index (CDI), Joint Space Narrowing (JSN), Support Vector Machine (SVM)Abstract
Knee osteoarthritis (OA) is a progressive degenerative disease that disproportionately affects older adults, leading to significant morbidity, disability, and socioeconomic burden. Despite its high prevalence, the pathology of OA remains poorly understood, and no effective disease-modifying interventions currently exist. Medical imaging, particularly magnetic resonance imaging (MRI), plays a crucial role in diagnosing OA and assessing disease progression by measuring cartilage changes. However, manual segmentation of knee cartilage from 3D MRI scans is labor-intensive, timeconsuming, and requires specialized expertise, making it impractical for large-scale clinical applications. To address these challenges, this study leverages deep learning techniques to predict OA progression using 3D MRI data. We computed the Cartilage Damage Index (CDI) from 36 key locations in the tibiofemoral cartilage compartment and applied principal component analysis (PCA) for feature extraction. A supervised learning model, support vector machine (SVM), and an artificial neural network (ANN) were implemented to predict OA progression, as measured by changes in Kellgren and Lawrence (KL) grade, medial joint space narrowing (JSM), and lateral joint space narrowing (JSL). By integrating artificial intelligence with MRI-based biomarkers, this approach aims to enhance the accuracy, efficiency, and reproducibility of OA progression prediction, offering a promising tool for early intervention and treatment evaluation
Downloads
Published
Issue
Section
License

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.













