Radiographic analysis of lumbar lordosis: centroid, Cobb, TRALL, and Harrison posterior tangent methods.

Abstract

Delayed, repeated measures, with three examiners each twice digitizing thirty lateral lumbar radiographs. To determine the reliability and clinical utility of the centroid, Cobb, tangential radiologic assessment of lumbar lordosis (TRALL), and Harrison posterior tangent line-drawing methods for analysis of lumbar lordosis. Cobb's method is commonly used for curvature analysis on lateral lumbar radiographs, whereas the centroid, TRALL, and Harrison posterior tangent methods are not widely used. Thirty lateral lumbar radiographs were digitized twice by each of three examiners. To evaluate reliability of determining global and segmental alignment, all four vertebral body corners of T12-S1 and the superior margin of the femur head were digitized. Angles created were segmental and global centroid, (two-line) Cobb angles, and intersections of posterior tangents. A global TRALL angle was determined. Means, standard deviations, mean absolute differences, interclass and intraclass correlation coefficients (ICC), and confidence intervals were calculated. The interobserver and intraobserver reliabilities of measuring all segmental and global angles were in the high range (ICCs > 0.83). The mean absolute differences of observers' measurements were small (0.6 degrees -2.0 degrees ). Distal segmental (L4-S1) and global angles of lumbar curvature were dependent on the method of measurement. All four radiographic methods had high reliability and low mean absolute differences of observers' measurements. Because it lacks a segmental analysis, the TRALL method is not recommended. The centroid, Cobb, and Harrison posterior tangent methods provide global and segmental angles. However, the centroid segmental method requires three segments and is less useful for a stability analysis.