Impact of Obesity, Osteopenia, and Scoliosis on Interobserver Reliability of Measures of the Spinopelvic Sagittal Radiographic Parameters

Introduction: We investigated the relationships between patient factors, including obesity, osteopenia, and scoliosis, and the reliability of measures of the spinopelvic sagittal parameters using conventional X-radiography (Xp) and slot-scanning Xp devices (EOS) and examined the differences in interobserver measurement reliability between them.Methods: We retrospectively enrolled 55 patients (52.7 +/- 25.3 years, 27 females) with conventional whole-spine Xp and EOS images taken within three months. Patients were classified according to obesity (Body mass index >= 25 kg/m(2)), osteopenia (T score<-1), and scoliosis (Cobb angle>20(degrees)). The associations between patient factors and reliability of radiological parameter measurements were examined with interobserver intraclass correlation coefficient (ICC), defined as poor, <.40; good, 40-.79; and excellent, >=.80.Results: All parameters measured with EOS showed excellent reliability except for L4-S (ICC:.760, 95% CI:.295-.927) in the obesity+ group. All parameters measured with conventional Xp were excellent except for those classified as good: L 4-S (.608,.093-.868) and pelvic incidence (PI) (.512,.078-.832) in the obese+ group; T1 slope (.781,.237-.952), L4-S (.718,.112-.936), sacral slope (SS) (.792,.237-.955), pelvic tilt (PT) (.787,.300-.952), and center of acoustic meatus and femoral head offset (CAM-HA) (.690,.090-.928) in the osteopenia+ group; and lumbar lordosis (LL, L4-S) (.712,.349-.889), SS (.608,.178-.843), and CAM-HA (.781,.480-.917) in the scoliosis+ group.Conclusions: Reliability of EOS measurements was preferable except for L4-S in patients with obesity. The reliability of conventional Xp measurements of pelvic parameters SS, PT, and PI was affected by patient factors, including obesity, osteopenia, and scoliosis. When evaluating lower lumbar and pelvic parameters in patients with these factors, we recommend substituting thoracic parameters, LL (L1-S), sagittal vertical axis (SVA), and T1 pelvic angle (TPA), or combining computed tomography (CT) measurements.