Reliability of surgeon dependent agreement of classification and treatment planning in adolescent idiopathic scoliosis

Scoliosis is the most common deformity of the spine. Idiopathic scoliosis constitutes about 80% of structural scoliosis, and the vast majority of these deformities with unknown aetiology consist of adolescent idiopathic scoliosis. Treatment options for scoliosis includes; follow up, non-surgical treatment and surgical treatments. Preoperative planning of scoliosis surgery, which requires a series of difficult decisions, is a complicated process. Although many publications related to the application of idiopathic scoliosis are published in the literature, very little of them are related to the selection of the site of specific fusion surgery or how this decision is made. In this study, 24 female and 3 male, 27 patients who were operated for adolescent idiopathic scoliosis in Dokuz Eylul University Faculty of Medicine, Department of Orthopaedics and Traumatology between the years 1994-2014, and had preoperative AP / Lat, left bending, right bending, AP after traction and postoperative X-rays with the appropriate standards, were included in the study. 2 professors, 2 associate professors and a specialist doctor all of whom are interested in spinal surgery in Dokuz Eylul University Faculty of Medicine, Department of Orthopaedics and Traumatology evaluated preoperative X-rays of included patients. Investigators were asked to assess upper end vertebra, lower end vertebra, apex, Cobb angle in preoperative PA x-rays; Cobb angle in preoperative traction, right and left bending xrays for upper thoracic segment; upper end vertebra, lower end vertebra, apex, Cobb angle in preoperative PA X-rays; Cobb angle in preoperative traction, right and left bending x-rays for thoracic segment; upper end vertebra, lower end vertebra, apex, Cobb angle in preoperative PA x-rays; Cobb angle in preoperative traction, right and left bending x-rays for thoracolumbar/lumbar segment; upper instrumented vertebra and lower instrumented vertebra, measurement of T2-T5, T5-T12, T11-L1, T12-S1 sagittal Cobb angles; King Moe and Lenke classifications; preoperative choice of preferred implant, implantation methods. Agreement and consistency of measurements, classifications and preoperative plans between investigators was established and differences were determined. Measurements, classifications and preferred segment for fusion were compared between investigators and the instrumentation actually applied to the patient; and surgical planning was investigated to evaluate compliance. Mean thoracic Cobb angle was found 44,62° (±3,1°) with a good level of interobserver agreement (ICC=0,764). Mean lumbar Cobb angle was found 41,22° (±3,35°) with a good level of interobserver agreement (ICC=0,775). In determination of fusion segments, mean preferred upper instrumented vertebra was T4-T5, with a good level of interobserver agreement (ICC=0,714). In determination of fusion segments, mean preferred lower instrumented vertebra was L2-L3, with a good level of interobserver agreement (ICC=0,717). When these preferred upper and lower instrumented vertebrae were compared with actual postoperative x-rays; higher levels for both upper and lower instrumented vertebrae were preferred. Assessment of King Moe classification between investigators were found at excellent level of reliability. (ICC=0,806). Total agreement for Lenke classification between investigators was 18.5%, rate of agreement above 50% was 66,6%, , rate of agreement below 50% was 11%, and fully mismatched for 1 case (3,7%). When all cases are considered, interobserver agreement was good. (ICC=0,754). Choice of preferred implant and implantation methods were found to be intermediate. (ICC=0,533). Mean planned segment for fusion was found 10.58 (±2.8), the number of theoretically used implants was 21.16 (±5,7), and interobserver agreement was found to be intermediate (ICC=0,533) Choice of preferred implant and implantation methods varies between surgeons and patient satisfaction should be closely considered to understand how these differences reflect on patients.