Patient-specific 17-segment myocardial modeling on a bull's eye map

The purpose of this study was to develop and validate cardiac computed tomog-raphy (CT) quantitative analysis software with a patient-specific, 17-segment myocardial model that uses electrocardiogram (ECG)-gated cardiac CT images to differentiate between normal controls and severe aortic stenosis (AS) patients. ECG-gated cardiac CT images from 35 normal controls and 144 AS patients were semiautomatically segmented to create a patient-specific, 17-segment myocardial model. Two experts then manually determined the anterior and posterior interven-tricular grooves to be boundaries between the 1st and 2nd segments and between the 3rd and 4th segments, respectively, to correct the model. Each segment was automatically identified as follows. The outer angle of two boundaries was divided to differentiate the 1st, 4th, 5th, and 6th segments in the basal plane, whereas the inner angle divided the 2nd and 3rd segments. The segments of the midplane were similarly divided. Segmental area distributions were quantitatively evaluated on the bull's-eye map on the basis of the morphological boundaries by measuring the area of each segment. Segmental areas of severe AS patients and normal controls were significantly different (t-test, all p-values < 0.011) in the proposed model because the septal regions of the severe AS patients were smaller than those of normal controls and the difference was enough to divide the two groups. The capabilities of the 2D segmental areas (p < 0.011) may be equivalent to those of 3D segmental analysis (all p-values < 0.001) for differentiating the two groups (t-test, all p-values < 0.001). The proposed method is superior to the conventional 17-segment in relation to reflection of patient-specific morphological variation and allows to obtain a more precise mapping between segments and the AHA recommended nomenclature. It can be used to differentiate severer AS patients and normal controls and also helps to understand the left ventricular morphology at a glance.