Retrospective observational study.
To assess the accuracy and reliability of standing lateral lumbar radiographs for measurements of spinopelvic parameters, compared with whole-spine EOS® images.
Summary of Background Data.
Lateral lumbar radiographs are commonly used for measurements of spinopelvic parameters. However, variable magnifications by fan-beam x-ray projection at margins may cause measurement errors.
Fifty consecutive patients with standing lateral lumbar radiographs and whole-spine EOS® images were retrospectively reviewed from March to July in 2019. Two orthopedic surgeons (observers) independently measured the spinopelvic parameters including pelvic incidence (PI), pelvic tilt (PT), sacral slope (SS), and lumbar lordosis (LL) on the computers. These spinopelvic parameters of each patient were measured twice on both lateral lumbar radiograph and EOS® image by two observers with the usage of Surgimap software. The measurement difference and agreement of each parameter value between lateral lumbar radiographs and EOS® images were analyzed by paired t test and the interclass correlation coefficient (ICC) respectively. Intra- and inter-observer's reliabilities of both imaging methods were assessed by ICC.
The measurement difference of each spinopelvic parameter on EOS® images and lateral lumbar radiographs was less than 1° on average. Only the measurement difference of PI value revealed statistically significant (P = 0.020) with 0.9° discrepancy (95% confidential interval: 0.2–1.6), which indicated no clinical significance. The ICC values of lateral lumbar radiographs and EOS® images were more than 0.870 (range, 0.872–0.976), thus showed good to excellent measurement agreement between both imaging methods. All the ICC values for evaluating intra- and inter-observer reliability were greater than 0.960 (range, 0.963–0.993), indicating excellent reliability for observer's measurements.
Measurements of spinopelvic parameters (PI, PT, SS, and LL) on standing lateral lumbar radiographs are accurate and reliable, which are comparable to whole-spine EOS® images.
Level of Evidence: 3