Healing Predictors of Stable Juvenile Osteochondritis Dissecans Knee Lesions After 6 and 12 Months of Nonoperative Treatment

Abstract

Background:

Nonoperative treatment of stable juvenile osteochondritis dissecans (JOCD) lesions of the knee fails in up to 50% of cases. Healing predictors are needed to identify potential failures and thus determine treatment options.

Purpose:

A predictive model for healing potential after 6 and 12 months of nonoperative treatment of stable JOCD lesions based on sensitive magnetic resonance imagining (MRI) follow-up measurements was developed.

Study Design:

Cohort study (diagnosis); Level of evidence, 2.

Methods:

A retrospective cross-sectional study was conducted to analyze 62 white patients (76 stable JOCD lesions) who were initially treated by restriction of activity until they were free of pain. The primary end point was healing investigated on MRI with follow-up measurements after 6 and 12 months of nonoperative treatment. Multivariate logistic regression was used to determine the influence of age, sex, JOCD lesion size, clinical symptoms, and the occurrence of cystlike lesions (CLLs) on healing potential. Additionally, optimal prognostic cutoffs were defined to differentiate failures from nonfailures.

Results:

After 6 months of nonoperative treatment, 51 (67%) of 76 stable JOCD lesions showed no progression toward healing or showed signs of instability. Normalized lesion width and area and CLL occurrence differed significantly between failures and nonfailures (P < .05). A multivariate logistic regression best-predictors model that included age, CLL size, and normalized lesion width best predicted healing after 6 months and resulted in an area under the curve (AUC) of 0.779 (P < .001). A cutoff at 48% healing probability, as predicted by a nomogram based on age, normalized lesion width, and CLL size, differentiated failures from nonfailures (sensitivity, 60.0%; specificity, 83.7%). After 12 months, 37 lesions (49%) had progressed toward healing, and the sole observation of CLL size had the highest predictive validity (AUC, 0.766). The optimal cutoff was a healing probability of 61% (lesion size, 1.3 mm; sensitivity, 70.3%; specificity, 74.1%).

Conclusion:

A 6-month period of nonoperative treatment with or without casting might be appropriate if the healing potential is >48%. A 12-month period of nonoperative treatment may be successful if the CLL is <1.3 mm in length as assessed on MRI.