[Validation of a tool to measure pelvic and spinal parameters of sagittal balance]

Parameters determining sagittal balance are essential for optimal analysis and treatment of many spinal disorders. The purpose of this work was to validate a software designed to measure the principal parameters involved in sagittal balance of the spine.

Six parameters (lumbar lordosis, thoracic kyphosis, sagittal tilt at T9, pelvic index, pelvic tilt, slope of the sacrum) were measured on lateral views of the spine from 100 healthy volunteers free of any spinal disease. Two measurement techniques were used: manual measurement and automatic computerized measurement with this software. We hypothesized that manual measures could be accepted as reference values because they are currently the most widely used and because the computerized measurements were obtained using digitalized images that may have modified interpretation. The software was therefore validated by comparing the angles measured manually with the computer output. Inter- and intraobserver coefficients of variation were calculated for the two measurement techniques. One operator performed both series of measurements (manual and computerized). Two other operators preformed two series of independent measures using one of the measurement techniques. Finally, two new operators performed a complete series of measurements using both measurement techniques. Comparisons were performed with the t test for paired variables with calculation of the coefficients of correlation. Intraclass coefficients of correlation were determined for inter- and intra-observer variability.

There was an excellent correlation between the manual measurements and the computerized measurements with intra-class coefficients of correlation varying from 0.82 to 0.96. Inter- and intra-observer variabilities were comparable for the two measurement techniques used to determine thoracic kyphosis, lumbar lordosis, pelvic index, pelvic tilt, and slope of the sacrum. Inter- and intra-observer variability was lower when the sagittal tilt was measured with the software specially designed.

This comparison between two techniques for measuring pelvic and spinal parameters of sagittal balance of the spine demonstrated a good correlation between manual and the computerized measurements obtained with the software to be evaluated. Computer-assisted measurements not only provided a saving in time but also minimized inter- and intra-observer variability for the estimation of certain parameters.