# Automated Joint Space Width Assessment in Patients Treated for Juvenile Osteochondritis Dissecans of the Distal Femur: A Cross-Sectional Study and Systematic Review of the Literature

## TL;DR

This study uses AI to assess joint space width in patients treated for a knee condition, finding no significant differences between treatment types and showing AI's potential for monitoring cartilage health.

## Contribution

The study introduces AI-based joint space width assessment as a feasible and standardized tool for monitoring JOCD treatment outcomes.

## Key findings

- AI-based JSW measurements showed no significant differences between affected and contralateral knees.
- Clinical scores were high and comparable across treatment groups.
- A moderate positive correlation was found between JSW and Lysholm scores.

## Abstract

Background/Objectives: Juvenile osteochondritis dissecans (JOCD) of the knee is commonly treated using conservative or joint-preserving surgical techniques. While clinical outcomes are generally favorable, the risk of early cartilage degeneration remains unclear. Joint space width (JSW) on weight-bearing radiographs serves as an indirect marker of cartilage health. Artificial intelligence (AI)-based JSW assessment may enable sensitive and reproducible detection of early degenerative changes. Methods: This cross-sectional feasibility study included 21 skeletally immature patients treated for JOCD of the distal femur between 2002 and 2017. Treatment modalities comprised conservative management, retrograde drilling, and fragment refixation. Fully automated JSW measurements were performed on standardized anteroposterior knee radiographs using a validated AI-based software IB Lab KOALA™, Version 2.4. JSW of the affected compartment was compared with the contralateral knee and between treatment groups. Clinical outcomes were assessed using the Lysholm Knee Scoring Scale and the International Knee Documentation Committee (IKDC) score. Additionally, a systematic review of the literature on post-treatment degenerative changes following OCD therapy was conducted according to PRISMA guidelines. Results: Compared with manually reviewing images, the software IB Lab KOALA™, Version 2.4 as easy to implement. AI-based analysis revealed no significant differences in JSW between the affected and contralateral knees, nor between treatment modalities. Average JSW exceeded 6 mm in all groups after a median follow-up of 64 (min. 27, max. 177) months. Clinical scores were high and comparable across treatments. A moderate positive correlation was observed between the JSW and Lysholm score, while increasing age and longer follow-up were associated with a reduced JSW. The systematic review identified ten relevant studies, reporting generally favorable long-term clinical outcomes with a low but present risk of osteoarthritis progression. Conclusions: Our AI-based analysis showed no differences in JSW between conservative and joint-preserving surgical treatments of JOCD in the follow-up. This technology can provide a valuable tool for standardized and sensitive radiographic monitoring in young patients.