VARIABILITY IN TRANSVERSE VOLUMETRIC ASYMMETRY MEASUREMENTS IN ISIS2

Abstract

Background: Scoliotic deformity has been traditionally measured by Cobb angle using radiography. This parameter gives a measure of the lateral curve in the spine in the coronal plane. However, patients are often more concerned about their rib humps or other volumetric asymmetries in the surface of their backs. There is often little relation between Cobb angle and the magnitude of the asymmetry. A method of quantifying volumetric deformity, especially if it requires no radiation, would therefore be useful for spinal surgeons and patients alike.

Methods: The three dimensional shape of the back is measured using structured light and digital photography with ISIS2, a non-commercial surface topography system. Markers are placed on bony landmarks so that the surface can be related to body axes. A zero plane is defined through the sacrum and the vertebra prominens, parallel to the line between the markers on the dimples of Venus. A curve is fitted through the markers on the spinous processes on the measured surface and is used as the line of symmetry. The difference in the areas between the surface and the zero plane on each side of the symmetry line is then calculated for each horizontal (transverse) section. The left and right volumetric asymmetry parameters are then calculated by summing the area differences on each side and normalising for back length. These parameters range from zero for a perfect straight back with no transverse asymmetry to over 70 for extreme transverse asymmetry. The variability in these parameters was investigated using pairs of photographs of 59 patients. Two photographs were taken with the patient walking around the room between them. Left and right volumetric asymmetry was then calculated for each measurement and Bland-Altman analysis was carried out.

Results: The mean difference between pairs of measurements was −0.10, the standard deviation was 2.03 and the 95% tolerance limits covering 95% of the population were −4.8 to 4.6 for left volumetric asymmetry; the mean difference was 0.46, standard deviation was 3.13 and the 95% tolerance limits covering 95% of the population were −6.8 to 7.7 for right volumetric asymmetry. There was no evidence of bias from the Bland-Altman plots.

Conclusions: The variability in the volumetric asymmetry was low in comparison to the levels found for subjectively classified ‘moderate’ deformity. Change in degree of volumetric deformity can be monitored by ISIS2 volumetric asymmetry.