Prediction of standing radiographic lumbar lordosis by using supine MRI

Abstract

Objective: To study the correlation between lumbar lordosis (LL) measured by standing radiographs and supine magnetic resonance imaging (MRI), and the possibility of predicting standing radiographic LL based on supine MRI-LL and severity of lumbar disc degeneration (LDD). Materials and Methods: LL was measured from L1 to superior endplate of S1 on standing radiographs and supine MRI in 65 southern Chinese subjects aged 39 to 77 years. LDD score was calculated as sum of Pfirrmann gradings of the five discs levels. Relationship between X-ray and MRI-measurements were analysed by a paired sample t test and a scatter plot. Pearson’s correlation coefficient was calculated to study the effect of MRI lordosis and LDD on the correlation. A prediction equation was also developed. Results: Mean LL in standing radiographs (XR-LL) and supine MRI (MRI-LL) were 53.0° (± 8°) and 43.9° (± 8°) respectively. Intraindividual difference ([°] standing XR-LL minus [°] supine MRI-LL) ranges from -0.9° to +21°, with a significant mean of +9.1° (± 5°) (p<0.05). A significant positive correlation between the two LL measurements was found (R=0.81, p<0.05). Regression equation with use of LDD score and MRI-LL as predictors of radiographic lordosis was y=24.18 - 0.562D + 0.815x, where y was XR-LL, D was the LDD score, and x was MRI-LL. R-square of this equation was 71.0%. Conclusion: A good clinical correlation between LL measured in standing radiographs and supine MRI was found. Combined use of LDD score and LL yielded a moderately good prediction model for standing radiographic LL.