UNDERSTANDING THE PAIN IN KNEE OSTEOARTHRITIS: INSIGHTS FROM THE SEMI-QUANTITATIVE MRI OSTEOARTHRITIS KNEE SCORE ANALYSIS

This study aimed to analyze and predict various types of knee pain in knee osteoarthritis (KOA) patients based on Semi-Quantitative MRI Osteoarthritis Knee Score (SQ-MOAKS) features. The goal was to identify structural abnormalities in the knee joint associated with different types of pain, thereby aiding in the development of personalized treatment plans. Subjects were selected from the Osteoarthritis Initiative (OAI) database, including 297 patients who underwent both SQ-MOAKS assessment on knee MRI images and WOMAC knee pain assessment. Univariate and multivariate logistic regression analyses were used for feature selection, followed by the construction of multi-risk factor models and corresponding nomograms for each pain type. The predictive values of the models were evaluated using ROC curves and related metrics (AUC, accuracy, sensitivity, specificity), and the DeLong test was used to assess the significance of differences between various ROCs and AUCs. Significant associations were found between specific SQ-MOAKS features and different types of knee pain. Bone marrow lesions (BML) and cartilage morphology (CM) features were crucial in predicting the walking pain and rest pain, while osteophytes (OS) were more associated with nocturnal pain and pain during stair climbing. The models demonstrated high predictive accuracy, particularly for walking pain and rest pain. This study demonstrated that SQ-MOAKS features could effectively predict different types of knee pain, revealing the structural basis of knee pain in KOA patients. These results prompted the use of SQ-MOAKS in developing personalized treatment strategies, enhancing clinical decision-making, and improving patient quality of life.