Recently, Professor Ding Changhai's team at the Clinical Research Center of Zhujiang Hospital, Southern Medical University, proposed a novel Joint Space Radiomic Model (JS-RM) combined with neural networks for predicting the occurrence of knee osteoarthritis. The related findings were published in "Arthritis & Rheumatology."
Knee osteoarthritis is a common degenerative joint disease, particularly prevalent in the elderly population, significantly impacting patients' quality of life and imposing a heavy economic burden on society. Due to the lack of effective disease-modifying drugs, early intervention and sensitive detection methods are crucial.
In addressing this challenge, Professor Ding Changhai's team considered magnetic resonance imaging (MRI) for its high sensitivity in detecting early joint structural changes before X-ray imaging. Based on the Knee Osteoarthritis Initiative cohort, they selected 686 cases of non-radiographic knee osteoarthritis but high-risk knee MRI images, randomly divided into a development cohort and a test cohort (8:2). Image features were extracted from baseline MRI scans and modeled and predicted using neural networks. The JS-RM model integrated radiomic features of femoral cartilage, tibial cartilage, and meniscus for predicting the occurrence of radiographic knee osteoarthritis.
The test results showed that the JS-RM model predicted radiographic knee osteoarthritis with an area under the curve of 0.931, sensitivity of 84.4%, and specificity of 85.6%. With the JS-RM model, the average specificity and sensitivity of 9 hospital physicians in interpreting MRI scans for radiographic knee osteoarthritis increased from 47.4% and 58.6% to 87.4% and 81.2%, respectively.
This study indicates that the radiomic features of femoral cartilage, tibial cartilage, and meniscus have predictive value for the occurrence of radiographic knee osteoarthritis. The JS-RM model can serve as a non-invasive predictive tool, potentially playing a role in personalized clinical decision-making and supporting early intervention for knee osteoarthritis.
Related Paper Information: https://doi.org/10.1002/art.42915
