RELIABILITY OF A NOVEL KOCHER CLASSIFICATION FOR OSTEOCHONDRITIS DISSECANS

Background: Current radiographic classification systems for osteochondritis dissecans (OCD) of the knee utilize different characteristics of lesions to rate the stability and severity of disease. Classification systems such as Hefti and Nelson are widely used as the current standards for radiographic imaging and arthroscopic evaluation is the gold standard to assess stability. The purpose of this study was to compare the reliability of this new method with that of older, more established methods, as the first step in establishing its validity and clinical utility. Methods: One hundred twenty-four standardized knee MRIs of established patients with knee OCD were pre-selected to capture the spectrum of lesion types both regarding progression and location of the lesion. Each of the MRIs were classified independently by 2 readers into the Kocher, Hefti, and Nelson classification systems and a random sample was re-reviewed by one rater 6 weeks after initial review. The inter-rater and intra-rater agreement was evaluated by estimating Krippendorff’s alpha. Results: 108 knees were classified by the Kocher, Hefti, and Nelson classification systems, as 16 of the studies had an absence of the appropriate imaging sequences necessary. There were no differences in agreement across classification systems. Krippedorf’s alpha for interrater agreement was 0.51 (CI 0.33-0.66) for the Hefti classification, 0.50 (0.34-0.64) for the Nelson classification, and 0.49 (0.32-0.65) for the Kocher classification. The intrarater agreement was 0.88 (0.75-0.97) for the Hefti classification, 0.94 (0.86-0.99) for the Nelson classification and 0.98 (0.94-1.00) for the Kocher classification system. Conclusions: The novel Kocher classification for knee OCD had almost perfect intrarater agreement and moderate interrater agreement, consistent with well-established classification systems. This new classification system would be simpler with only three categories, whereas the Hefti (five) and Nelson (four) sysems had more. This simpler classification system could be widely applicable because the results could more accurately drive clinical treatment decision making for clinicians.