Detection of concentric left ventricular wall hypertrophy by contrast-enhanced non-electrocardiogram-gated chest computed tomography.

BACKGROUND: No established measure of concentric left ventricular wall hypertrophy (cLVH) on routine computed tomography (CT) of the adult chest currently exists. The objective of this study was to identify and test linear measures for the detection of cLVH using transthoracic echocardiography (TTE) as the reference standard.

METHODS: Contrast-enhanced non-electrocardiogram-gated chest CTs acquired within two weeks of TTE were retrospectively evaluated. Two radiologists independently made trans-axial measurements in the proximal half of the left ventricle at its approximate widest internal diameter: maximum septal thickness (sept), maximum lateral wall thickness (lat), and inner (Id ) and outer (Od ) wall-wall diameters at the level of greatest combined myocardial thickness. The sum of sept and lat, hereafter Thmax , and modified cross-sectional area (Amod  = Od 2 - Id 2 ) were calculated. Sept, lat, Thmax , and Amod were evaluated by receiver operating characteristic (ROC) curves using TTE as the reference standard. Thresholds were optimized for specificity and applied to a validation cohort. Inter-rater agreement was assessed by a simple unweighted Kappa statistic (κ).

RESULTS: Sept and Amod were selected based on areas under the ROC curves of 0.75 and 0.71, respectively, using 100 CTs. Thresholds of 1.6 cm and 30 cm2 , respectively, showed similar specificities of 98% with sensitivities of 27% and 30%, respectively. Applied to a validation cohort of 100 CTs, sept had higher combined positive predictive value (75%), inter-rater agreement (κ = 0.58), specificity (91%), and sensitivity (24%).

CONCLUSION: Linear measures demonstrate utility in the diagnosis of cLVH on routine contrast-enhanced chest CT.