Reliability of Radiographic Parameters for Scoliosis Evaluation

Introduction: Evaluation of spinal deformity includes both a clinical and radiographic analysis. Generally, objective radiographic evaluation has been limited to the Cobb measurements. Additional measurements have been described, however, their reliability/reproducibility has not been proven. As new objective radiographic scoring systems are developed, an assessment of their reliability is necessary. This study determines the intra- and inter-observer reliability of these measurements.

Materials and Methods: After approval by our institutional review board, 30 complete sets of full-length radiographs (pre-op PA, lateral and side benders; post-op PA and lateral) were evaluated on 2 separate occasions by 3 fellowship trained spinal deformity surgeons. Pre-op coronal measurements (Cobb, AVT, C7-CSVL, T1 tilt, LIV, disk angle below LIV, Nash-Moe grade, LIVCSVL relationship, triradiate cartilage, left and right supine side bending Cobb), and sagittal measurements (T2-5, T5-12, T2-12, T10-L2, T12-S1, C7 plumbline), in addition to the same post-op measurements (minus supine side benders) were recorded. A total of 5,940 discrete data points were generated. Analysis of variance for repeated measures were performed to estimate intra- and inter-observer reliability using intraclass correlation coefficients (ICC) where a rho of > .8 is considered good, .8-.7 is fair, and <.7 is poor.

Results: All measurements were considered to have good reliability (inter-observer agreement) except the sagittal T2-5 (poor), which also has poor intra-observer reliability, the post-op T2-12 (fair), the preop T12-S1 (poor), and the post-op disk below LIV (poor), which also had only fair intra-observer reliability.

Conclusions: These radiographic parameters have good reliability (inter-observer agreement) except for sagittal T2-5, which can often be difficult to visualize. As more attention is made to the position of the disk below the LIV, this measurement reliability should also improve. As such, these measurements can be used as appropriate process measures for comparison of various treatment techniques. Long-term follow-up of cohorts is necessary to determine if the specific measures are predictive of outcome.

· If noted, the author indicates something of value received. The codes are identified as: a-research or institutional support; b-miscellaneous funding; c-royalties; d-stock options.