Predicting 3D Thoracic Kyphosis Using Traditional 2D Radiographic Measurements in Adolescent Idiopathic Scoliosis

Abstract Study Design Retrospective. Objective To develop and validate a prediction formula to estimate three-dimensional (3D) T5–T12 kyphosis in adolescent idiopathic scoliosis (AIS) from standard two-dimensional (2D) radiographic measurements. Summary of Background Data 2D measurements of thoracic kyphosis in AIS patients overestimate 3D kyphosis; however, there is a lack of widespread availability of 3D imaging technology. Methods Retrospective review was performed for AIS patients with right thoracic curves evaluated with EOS Imaging from January 2010 to June 2014. Standard 2D posteroanterior and lateral radiographic measurements, pelvic incidence, Nash-Moe grade, Perdriolle rotation, and "3D T5–T12" sagittal measures (reconstructed with sterEOS, analyzed with custom MatLab code) were input into a multivariate logistic analysis to create a prediction model for 3D T5–T12 sagittal alignment. An initial cohort of 66 patients (curves 14°–85°) was used to create a predictive model, and a separate cohort of 129 patients (curves 16°–84°) was used to validate the formula. Results 2D thoracic coronal Cobb and 2D T5–T12 kyphosis were the only significant predictors in the model. The prediction formula for estimating 3D T5–T12 sagittal measurement from standard 2D measurements, in degrees, was 18.1 + (0.81*2D T5–T12 sagittal Cobb) − (0.54*2D coronal Cobb), r 2 = 0.84. The average model error between predicted and measured 3D T5–T12 kyphosis was ±7°. The predicted 3D T5–T12 kyphosis (8.6° ± 12.1°) and measured 3D T5–T12 kyphosis (8.5° ± 13.0°) were not significantly different (p = .8). 3D kyphosis was less than standard measures of 2D kyphosis (8.5° ± 13.0° vs. 20.2° ± 12.6°, p Conclusion This simple validated formula to predict 3D T5–T12 sagittal alignment using routine 2D thoracic Cobb and T5–T12 kyphosis for thoracic AIS patients has great potential value in assessing historical data collected prior to the development of 3D imaging methods as well as understanding/planning surgical hypokyphosis correction in patients without access to 3D imaging.