Dustin G Mark1,2,3, Justin A Davis1, Yun-Yi Hung3, David R Vinson3,4. 1. Department of Emergency Medicine, Kaiser Permanente, Oakland, CA. 2. Department of Critical Care, Kaiser Permanente, Oakland, CA. 3. Division of Research, Kaiser Permanente, Oakland, CA. 4. Department of Emergency Medicine, Kaiser Permanente, Roseville, CA.
Abstract
OBJECTIVE: The objective was to determine if ascending aorta (AscAo) diameters measured by noncontrast computed tomography (CT) allow for meaningful discrimination between patients with and without type A aortic dissection (TAAD), ideally with 100% sensitivity. METHODS: This study was a retrospective analysis of cases of TAAD, as well as controls, undergoing evaluation for TAAD with CT aortography, presenting to 21 emergency departments within an integrated health system between 2007 and 2015. AscAo diameters were determined using axial noncontrast CT images at the level of the right main pulmonary artery by two readers. AscAo diameters were additionally normalized for age, sex, and body surface area (assessed by a Z-score, which is the number of standard deviations between the observed and expected AscAo diameters). Overall model discrimination was assessed using the area under the receiver operating characteristic curve (AUC). Comparative discrimination was assessed using both the change in AUC (∆AUC) and the continuous net reclassification index (NRI). RESULTS: A total of 230 cases of TAAD and 325 controls were included in the study. The median ages for cases and controls were 65 and 62 years, and the median AscAo diameters were 50 and 35 mm, respectively. The raw and normalized AscAo diameters demonstrated similarly excellent discrimination (AUCs of 0.96 vs. 0.97, respectively; ∆AUC = 0.01, p = 0.09) and an NRI of 0.30 (95% confidence interval [CI] = 0.13-0.47), both indicating small incremental improvements in classification with the use of the normalized AscAo measures. A raw AscAo diameter of 34 mm and a normalized Z-score of 1.84 both yielded 100% sensitivity for TAAD, with respective specificities of 35% (95% CI = 29.6%-40.2%) and 67% (95% CI = 61.7%-72.2%). CONCLUSIONS: Nearly all patients with TAAD appear to have enlarged AscAo diameters as measured by noncontrast CT, whereas most patients with suspected but absent TAAD have relatively normal AscAo diameters. Both raw and normalized AscAo measures provided relatively comparable discriminatory value. If validated, these data may be useful in adjudicating risk among patients with suspected TAAD in whom a criterion standard test is unavailable, nondiagnostic, or contraindicated.
OBJECTIVE: The objective was to determine if ascending aorta (AscAo) diameters measured by noncontrast computed tomography (CT) allow for meaningful discrimination between patients with and without type A aortic dissection (TAAD), ideally with 100% sensitivity. METHODS: This study was a retrospective analysis of cases of TAAD, as well as controls, undergoing evaluation for TAAD with CT aortography, presenting to 21 emergency departments within an integrated health system between 2007 and 2015. AscAo diameters were determined using axial noncontrast CT images at the level of the right main pulmonary artery by two readers. AscAo diameters were additionally normalized for age, sex, and body surface area (assessed by a Z-score, which is the number of standard deviations between the observed and expected AscAo diameters). Overall model discrimination was assessed using the area under the receiver operating characteristic curve (AUC). Comparative discrimination was assessed using both the change in AUC (∆AUC) and the continuous net reclassification index (NRI). RESULTS: A total of 230 cases of TAAD and 325 controls were included in the study. The median ages for cases and controls were 65 and 62 years, and the median AscAo diameters were 50 and 35 mm, respectively. The raw and normalized AscAo diameters demonstrated similarly excellent discrimination (AUCs of 0.96 vs. 0.97, respectively; ∆AUC = 0.01, p = 0.09) and an NRI of 0.30 (95% confidence interval [CI] = 0.13-0.47), both indicating small incremental improvements in classification with the use of the normalized AscAo measures. A raw AscAo diameter of 34 mm and a normalized Z-score of 1.84 both yielded 100% sensitivity for TAAD, with respective specificities of 35% (95% CI = 29.6%-40.2%) and 67% (95% CI = 61.7%-72.2%). CONCLUSIONS: Nearly all patients with TAAD appear to have enlarged AscAo diameters as measured by noncontrast CT, whereas most patients with suspected but absent TAAD have relatively normal AscAo diameters. Both raw and normalized AscAo measures provided relatively comparable discriminatory value. If validated, these data may be useful in adjudicating risk among patients with suspected TAAD in whom a criterion standard test is unavailable, nondiagnostic, or contraindicated.