The Reliability of Assessing Radiographic Healing of Osteochondritis Dissecans of the Knee

• Published in: The American journal of sports medicine Vol. 45; no. 6; p. 1370
• Main Authors: Wall, Eric J, Milewski, Matthew D, Carey, James L, Shea, Kevin G, Ganley, Theodore J, Polousky, John D, Grimm, Nathan L, Eismann, Emily A, Jacobs, Jr, Jake C, Murnaghan, Lucas, Nissen, Carl W, Myer, Gregory D, Weiss, Jennifer, Edmonds, Eric W, Anderson, Allen F, Lyon, Roger M, Heyworth, Benton E, Fabricant, Peter D, Zbojniewicz, Andy
• Format: Journal Article
• Language: English
• Published: United States 01.05.2017
• Subjects: Adolescent; Child; Cohort Studies; Female; Femur - diagnostic imaging; Femur - surgery; Humans; Knee Joint - diagnostic imaging; Knee Joint - surgery; Male; Osteochondritis Dissecans - diagnostic imaging; Osteochondritis Dissecans - surgery; Radiography; Reproducibility of Results; OCD; healing; radiograph; reliability; adolescent; juvenile; osteochondritis dissecans; pediatric
• ISSN: 1552-3365
• DOI: 10.1177/0363546517698933

The reliability of assessing healing on plain radiographs has not been well-established for knee osteochondritis dissecans (OCD). To determine the inter- and intrarater reliability of specific radiographic criteria in judging healing of femoral condyle OCD. Cohort study (Diagnosis); Level of evidence, 3. Ten orthopedic sports surgeons rated the radiographic healing of 30 knee OCD lesions at 2 time points, a minimum of 1 month apart. First, raters compared pretreatment and 2-year follow-up radiographs on "overall healing" and on 5 subfeatures of healing, including OCD boundary, sclerosis, size, shape, and ossification using a continuous slider scale. "Overall healing" was also rated using a 7-tier ordinal scale. Raters then compared the same 30 pretreatment knee radiographs in a stepwise progression to the 2-, 4-, 7-, 12-, and 24-month follow-up radiographs on "overall healing" using a continuous slider scale. Interrater and intrarater reliability were assessed using intraclass correlations (ICC) derived from a 2-way mixed effects analysis of variance for absolute agreement. Overall healing of the OCD lesions from pretreatment to 2-year follow-up radiographs was rated with excellent interrater reliability (ICC = 0.94) and intrarater reliability (ICC = 0.84) when using a continuous scale. The reliability of the 5 subfeatures of healing was also excellent (interrater ICCs of 0.87-0.89; intrarater ICCs of 0.74-0.84). The 7-tier ordinal scale rating of overall healing had lower interrater (ICC = 0.61) and intrarater (ICC = 0.68) reliability. The overall healing of OCD lesions at the 5 time points up to 24 months had interrater ICCs of 0.81-0.88 and intrarater ICCs of 0.65-0.70. Interrater reliability was excellent when judging the overall healing of OCD femoral condyle lesions on radiographs as well as on 5 specific features of healing on 2-year follow-up radiographs. Continuous scale rating of OCD radiographic healing yielded higher reliability than the ordinal scale rating. Raters showed substantial to excellent agreement of OCD overall radiographic healing measured on a continuous scale at 2, 4, 7, 12, and 24 months after starting treatment.