Introduction: The study aims to build and validate a nomogram for estimating the probability of patients developing type A aortic dissection at a diameter less than 55 mm.
Methods: A primary cohort of 896 patients diagnosed with acute type A aortic dissection by computed tomography angiography (CTA) were used for model development, with data collected between January 2005 and March 2012. The subjects were assigned to two groups based on ascending aorta diameter (group A<55 mm, Group B ≥ 60 mm). Univariate and multivariate logistic regression analyses were employed for the development of the prediction model. Demographic factors, as well as clinical and imaging characteristics were taken into account. The resulting nomogram was evaluated for performance traits, e.g. calibration, discrimination and clinical usefulness. After internal validation, the nomogram was further assessed in a different cohort containing 385 consecutive subjects examined between January 2013 and December 2015.
Results: The individualized prediction nomogram included 9 predictors derived from univariate and multivariable analyses, including gender, age, weight, hypertension, liver cyst, renal cyst, bicuspid aortic valve, and bovine arch. Those predictors were double confirmed with Lasso regression. Internal validation showed good discrimination of the model with area under the curve (AUC) of 0.854 and good calibration (Hosmer-Lemeshow test, P = 0.876). Application of the nomogram in the validation cohort still revealed good discrimination (AUC = 0.802) and good calibration (Hosmer-Lemeshow test, P = 0.398). Decision curve analysis demonstrated that the prediction nomogram was clinically useful.
Conclusions: The current work presents a prediction nomogram incorporating demographical data as well as clinical and imaging characteristics that could help identify patients who might develop type A aortic dissection at a diameter less than 55 mm with convenience.
Keywords: Aortic dissection; Diameter; Nomogram; Preventive surgery timing.
Copyright © 2018 IJS Publishing Group Ltd. Published by Elsevier Ltd. All rights reserved.