Abstract
OBJECTIVE: To compare the degree of kyphosis among patients with old thoracolumbar fracture kyphosis (OTFK) in various positions and to assess kyphosis flexibility. METHODS: A total of 32 patients with OTFK who met the inclusion criteria were retrospectively included between February 2017 and August 2022. The cohort consisted of 4 males and 28 females with a mean age of 66.47 years (range, 55-88 years). All patients underwent preoperative standing full-length spine x-ray, prone full-length spine CT scout view (FLS-CT), and supine MRI. Among them, 29 patients had single-segment fractures and 3 had double-segment fractures. The local kyphosis Cobb angle (LKCA) was measured on all imaging modalities. The LKCA measured on standing x-ray and FLS-CT were recorded as LKCA(X) and LKCA(FLSCT), respectively. On MRI, LKCA was measured on three sagittal slices (left parasagittal, midsagittal, and right parasagittal), recorded as LKCA(LMR), LKCA(MMR), and LKCA(RMR), respectively. Kyphosis flexibility (KF) was calculated based on these measurements. Pairwise comparisons were performed using the Wilcoxon signed-rank test with Bonferroni correction after an overall Friedman test. Equivalence analysis between prone FLS-CT and supine MRI was performed using a prespecified margin of ±5°. Interobserver reliability was assessed using the intraclass correlation coefficient (ICC). RESULTS: The mean standing LKCA was 39.58 ± 9.00°. The LKCA measured on prone FLS-CT was 29.61 ± 6.96°. On supine MRI, the LKCA values were 28.34 ± 6.37° (LKCA(LMR)), 27.64 ± 6.18° (LKCA(MMR)), and 28.97 ± 5.92° (LKCA(RMR)). The mean LKCA of the three MRI planes was 28.32 ± 5.91°. The corresponding KF values were 24.45% ± 10.86% for prone FLS-CT, 27.36% ± 11.08% for the left parasagittal slice, 29.16% ± 10.89% for the midsagittal slice, 25.52% ± 11.20% for the right parasagittal slice, and 27.35% ± 10.16% for the mean of the three MRI planes. LKCA was significantly lower in the prone and supine positions than in the standing position (all adjusted p < 0.001). No significant differences were found between prone FLS-CT and any supine MRI measurement (all adjusted p > 0.05). In equivalence analysis, all 95% confidence intervals of the paired mean differences between prone FLS-CT and supine MRI measurements were entirely within the prespecified equivalence margin of ±5°. Interobserver reliability was excellent across all imaging modalities, with ICC values ranging from 0.985 to 0.992. CONCLUSION: Kyphosis severity was significantly reduced in the preoperative recumbent position in patients with OTFK. Prone FLS-CT and supine MRI provided clinically comparable estimates of positional kyphosis correction, suggesting that both modalities may be useful for preoperative assessment of kyphosis flexibility in OTFK.