PURPOSE: To evaluate the inter-rater reliability of Rajasekaran's kyphosis classification through a multicenter validation study.
OVERVIEW OF LITERATURE: The classification of kyphosis, developed by Rajasekaran, incorporates factors related to curve characteristics, including column deficiency, disc mobility, curve magnitude, and osteotomy requirements. Although the classification offers significant benefits in determining prognosis and management decisions, it has not been subjected to multicenter validation.
METHODS: A total of 30 sets of images, including plain radiographs, computed tomography scans, and magnetic resonance imaging scans, were randomly selected from our hospital patient database. All patients had undergone deformity correction surgery for kyphosis. Twelve spine surgeons from the Asia-Pacific region (six different countries) independently evaluated and classified the deformity types and proposed their surgical recommendations. This information was then compared with standard deformity classification and surgical recommendations.
RESULTS: The kappa coefficients for the classification were as follows: 0.88 for type 1A, 0.78 for type 1B, 0.50 for type 2B, 0.40 for type 3A, 0.63 for type 3B, and 0.86 for type 3C deformities. The overall kappa coefficient for the classification was 0.68. Regarding the repeatability of osteotomy recommendations, kappa values were the highest for Ponte's (Schwab type 2) osteotomy (kappa 0.8). Kappa values for other osteotomy recommendations were 0.52 for pedicle subtraction/disc-bone osteotomy (Schwab type 3/4), 0.42 for vertebral column resection (VCR, type 5), and 0.30 for multilevel VCRs (type 6).
CONCLUSIONS: Excellent accuracy was found for types 1A, 1B, and 3C deformities (ends of spectrum). There was more variation among surgeons in differentiating between one-column (types 2A and 2B) and two-column (types 3A and 3B) deficiencies, as surgeons often failed to recognize the radiological signs of posterior column failure. This failure to identify column deficiencies can potentially alter kyphosis management. There was excellent consensus among surgeons in the recommendation of type 2 osteotomy; however, some variation was observed in their choice for other osteotomies.