Automatic Detection and Grading of Knee Joint Space Narrowing

Automatic Detection and Grading of Knee Joint Space Narrowing

Introduction : Osteoarthritis (OA) is the most prevalent form of arthritis commonly affecting the knees, spine, hips, and hands. Elderly females, people with obesity, and those involved in heavy physical occupational activity are at the highest risk of developing osteoarthritis. Since damage caused by the onset of osteoarthritis is irreversible, doctors are currently focusing on symptom control in the early stages of OA.


Our knee joint space narrowing (JSN) detection model has been designed to address this problem, by being able to detect and assess the severity of knee JSN caused by knee OA, hence improving the accuracy and efficiency of radiologists.



Result : 

● The AI model proved its clinical utility by being tested on an external dataset.

● The model achieved an overall AUC of 0.80 [0.78, 0.82] in the external test set.



Conclusion : Since different grades of JSN require different interventions, it is important to identify the grade of JSN afflicting the knee of a patient. By automatically assigning KL grades to knee radiographs, our AI models ensure an early and accurate diagnosis that ensures the correct treatment strategy. Our model can assist healthcare practitioners in structured

Share this

Request A Whitepaper

Related Posts