Methods
Database and patient population . Institutional electronic medical record system was queried from years 2013-2016 to identify patients with age ≥ 50 and \(<\)85 years who received CT chest imaging for either in inpatient, outpatient, or emergency department encounters. We set the lower age limit of 50 years because thoracic aortic aneurysm below this age is rare unless associated with connective tissue disease, and the upper age limit was set as the benefit of detection leading to intervention decreases at old age. After excluding multiple scans obtained on the same patient during the study period, there were 21,336 scans obtained on unique patients. In patients with multiple scans, the scan with the earliest date was used for the analysis. The Institutional Review Board approved this study. All CT scan reports were screened for any comments on the aortic arch and great vessels. Those reports were further reviewed for the description and subcategorization of arch anomalies. Also, all CT scans were screened for words and permutations related to thoracic aortic aneurysm (TAA) (aneurysm, dilation, dilatation, ectasia, enlargement) and reports containing such words were further reviewed for the site and the size of the aneurysms. Aneurysms of the abdominal aorta were excluded. The site of the thoracic aortic aneurysm was the site of the maximum diameter of the aorta. Subjects with missing clinical or demographic data were excluded.
Validation sample . Due to the concern that the aorta was not accurately assessed on scans done without contrast and routine scans done for non-thoracic pathology, a random sample of 200 scans was generated and reviewed by two radiology trainee and a board certified cardiothoracic radiologist with 10 years of experience in reading aorta studies. All the 200 scans were non-contrast. The prevalence and the type of the anomaly was documented, if found.
Patient characteristics . Indicated age, height, and weight values were recorded at the time of the CT scan. Race was categorized into Caucasian, African American, Asian, and other. Body surface area was calculated using Mosteller equation.11 Smoking history was dichotomized based on those with ≥ 5 pack-year smoking history indicating positive smoking history. Comorbidities (aortic valve disease, hypertension, diabetes, dyslipidemia, congestive heart failure, cocaine use, chronic kidney disease, myocardial infarction, chronic obstructive pulmonary disease) were chosen as commonly evaluated cardiovascular comorbidities and were defined using ICD-10 codes. To define aortic dilatation and aneurysm; 4 cm cut-off value was used for the root, ascending and the aortic arch, and 3 cm was used for the descending thoracic aorta.
Statistical analysis . Differences in the patient characteristics were compared with two-tailed t-test for normally distributed continuous variables and Fisher’s exact test for categorical variables. Multivariable logistic regression model was fitted to identify independent risk factors of TAA. Age, sex, race, BSA, history of hypertension, diabetes, dyslipidemia, smoking history, COPD, arch anomalies and aortic valve diseases were included in the model. P value of <0.05 and 95% CI were used to define statistically significant difference. Analysis was conducted using Microsoft excel 2019 and Prism 8.2 (GraphPad Software, San Diego, CA) for simple analysis and SAS 9.4 (SAS Institute Inc, Cary, NC) for modelling.