Validating a New Computed Tomography Atlas for Grading Ankle Osteoarthritis

• Published in: The Journal of foot and ankle surgery 2015-03, Vol.54 (2), p.207-213
• Main Authors: Cohen, Michael M., DPM, FACFAS, Vela, Nathan D., DPM, MPH, Levine, Jason E., DPM, Barnoy, Eran A., MS
• Format: Article
• Language: English
• Subjects: Adult; ankle; Ankle Joint; Atlases as Topic; computed tomography; Female; grading scale; Humans; Male; Observer Variation; Orthopedics; osteoarthritis; Osteoarthritis - classification; Osteoarthritis - diagnostic imaging; Reproducibility of Results; Severity of Illness Index; Tomography, X-Ray Computed
• ISSN: 1067-2516; 1542-2224
• DOI: 10.1053/j.jfas.2014.06.017

Abstract As the most common joint disease, osteoarthritis (OA) poses a significant source of pain and disability. It can be defined by classic radiographic findings, particular symptoms, or a combination of the 2. Although specific grading scales have been developed to evaluate OA in various joints, such as the shoulder, hip, and knee, no definitive classification system is available for grading OA in the ankle. The purpose of the present study was to create and validate a standardized atlas for grading (or staging) ankle osteoarthritis using computed tomography (CT) and “hallmark” findings noted on coronal, sagittal, and axial views extrapolated from the Kellgren-Lawrence radiographic scale. The CT scans of 226 patients at the Miami Veterans Affairs Medical Center were reviewed. An atlas was derived from a retrospective review of 30 remaining CT scans taken from July 2008 to November 2011. After this review, 3 orthogonal static CT images, obtained from 11 remaining patients, were chosen to represent the various stages on the OA scale and were used to test the validity of the atlas developed by 2 of us (M.M.C. and N.D.V.). A multispecialty panel of 9 examiners, excluding ourselves, independently rated the 11 CT scan subjects. The differences among examiners and specialties were calculated, including an intra-examiner agreement for 2 separate readings spaced 9 months apart. Although the small number of subspecialty examiners made the intraspecialty comparisons difficult to validate, the findings nevertheless indicated excellent agreement among all specialty groups, with good intra-investigational (intraclass correlation coefficient 0.962 and 1) inter-investigational (intraclass correlation coefficient 0.851) values. These results appeared to validate the CT ankle OA atlas, which we believe will be a valuable clinical and research tool, one that will likely be more beneficial than less relevant generalized OA grading scales in use today.