Reliability And Reproducibility Of Morphologic Classification For Thoracic Pedicles In Scoliosis

Objectives: To investigate the reliability and inter-observer reproducibilityof scoliosis thoracic pedicle classification system and its clinical applications.Methods: From January1998to January2009,60patients with scoliosisunderwent surgical treatment, preoperative CT scan was performed in allpatients, and the pedicle of vertebral arch axis thin layer reconstruction wasperformed. According to the morphological characteristics of cancellous bone,the thoracic pedicles were divided into normal type, narrow type, isthmicsclerosis, complete sclerosis(including two subtypes of concave sclerosis andstraight sclerosis) and the pedicle absence. Bsaed on the64-slice CTtechnology, multi-layer reconstruction, the1440thoracic pedicle axial CT scanresults from the60cases were classified by three senior spine surgeons whoknew nothing about the research at the same time. A month later, the procedurerepeated. Statistical analysis on multiple categories and multiple observers wasperformed by Stata10.0software and the Kappa statistics to evaluate thecredibility of inter-observer and the repeatability of intra-observer.Results: For the first results of classification among the three senior spinesurgeons: the percentage of result for three spine surgeons was88%,84%and92%respectively, and the Kappa coefficients was0.7647,0.6590and0.8292 respectively; while for the second results from the same spine surgeon: thepercentage of result was95%,97%and96%respectively, and the Kappacoefficient was0.9227,0.9513and0.9117respectively.Conclusions:1、The morphologic classification of thoracic pedicles inscoliosis has moderate reliability；2、The morphologic classification ofthoracic pedicles in scoliosis has excellent reproducibility.