Is the Cobb technique the most reliable and valid to assess the scoliotic deformity?

Abstract

Introduction Adolescent idiopathic scoliosis (AIS) is a 3D deformity of the spine commonly assessed on frontal X-rays using the Cobb technique. Some authors have found this technique to be poorly representative of the extent of the deformity, suggesting that the Ferguson and Centroid techniques, could better characterize this deformity. The aim was to evaluate the reliability and validity of different measurement techniques of the scoliotic deformity. Material and methods Seventy-three subjects with AIS (55F, age: 14 years [9–18]) underwent full body biplanar X-rays with 3D spinal reconstructions. The subjects were stratified into 4 groups with increasing Cobb values: Gr1 [10–20°] – n = 20, Gr2 [20–40°] – n = 20, Gr3 [40–70°] – n = 20, Gr4 [> 70] – n = 13. 2D frontal X-rays were used to measure the Cobb, Ferguson and Centroid angles digitally using a software, and manually on a film using a goniometer. 3D reconstructions were used to calculate the frontal, sagittal and axial deformity indices (FDI, SDI, ADI: averages of absolute values of upper endplate inclinations or axes of each vertebra in the scoliotic segment, in corresponding plane), which were considered as gold standard for deformity evaluation. Reliability was assessed by measuring each parameter twice manually then digitally, by 3 operators. Intraclass correlation coefficient (ICC) and reproducibility variance (SR) for each parameter were evaluated. As for the validity, stepwise multiple linear regressions (SMLR) were used to investigate which measurement technique is the most representative of the deformity indices in the 3 planes. Results Overall, the SR was higher for the manual technique regardless of the method, and increased with Cobb angle. In Gr1 to 3, the Cobb method showed the lowest SR (2.2°, 2.8°, 3.6° in Gr1 to 3) compared to the Ferguson and Centroid methods (from 2.6° to 4.8° and 3.8° to 5.4°). In Gr4, the Ferguson method showed the lowest SR (5.8° for Ferguson vs. 7.6° for Cobb and 10° for Centroid). The SMLR models found that the only determining factor for all techniques was the FDI. The slope of the linear regressions between the FDI and each parameter was: 1.9 for Ferguson (R2 = 0.94), 2.79 for Centroid (R2 = 0.93) and 2.97 for Cobb (R2 = 0.95). Discussion The Cobb was the most reliable and representative of the 3D deformity and shows that it would better differentiate between spines with different curvatures (highest correlation slope). Although some studies had criticized the Cobb angle for inadequately representing the scoliotic deformity, this study revealed that the Cobb was the most appropriate measurement technique to assess the 3D scoliotic deformity.