Devlin G. Morrison, Amanda Chan, Doug Hill, Eric C. Parent, Edmond H. M. Lou

February 2015, Volume 24, Issue 2, pp 306 - 312 Original Article Read Full Article 10.1007/s00586-014-3684-1

First Online: 21 November 2014


To investigate the accuracy and reliability of the Cobb angle, the spinous process angle (SPA), and apical vertebral rotation (AVR) for measuring adolescent idiopathic scoliosis (AIS), and to evaluate the correlations between these measurements.


A retrospective study of two sets of standing posteroanterior radiographs of patients with AIS was performed. The first set was 59 consecutive patients with AIS with Cobb angles <45° and the second set was 25 patients with Cobb angles >45°. The Cobb angle, SPA and AVR of each curve was measured twice by three observers with varying measurement experience. The mean absolute difference, standard deviation, and intra- and inter-rater reliability coefficients for each measurement were examined. The Pearson correlation coefficients between any two parameters were reported. The association of the Cobb angle with the SPA and AVR was examined using a multiple regression model.


The average intra- and inter-observer reliabilities (ICC [2, 1]) of the Cobb angle, SPA, and AVR were 0.99, 0.95, 0.92 and 0.98, 0.88, 0.83, respectively. The correlation coefficients (r) between Cobb angle and SPA, Cobb angle and AVR, and SPA and AVR were 0.93, 0.68, and 0.60, respectively. Using multiple regression, the association between the Cobb angle and SPA combined with AVR was R 2 = 0.90. The resulting regression model was: $${\text{Cobb angle}} = 4.1 + 0.88 \times {\text{SPA}} + 0.44 \times {\text{AVR}}$$ Cobb angle = 4.1 + 0.88 × SPA + 0.44 × AVR .


The SPA has high correlation with the Cobb angle. Including the AVR as an additional factor in multiple regression improves the prediction of the Cobb angle.

Read Full Article