Histological validation of simultaneous non-contrast angiography and intraplaque hemorrhage imaging (SNAP) for characterizing carotid intraplaque hemorrhage

Objectives This study sought to validate the performance of simultaneous non-contrast angiography and intraplaque hemorrhage (SNAP) imaging in characterizing carotid IPH by histology. Methods Thirty-five patients with carotid atherosclerotic disease (symptomatic 50-70% stenosis or > 70% stenosis) scheduled for carotid endarterectomy underwent 3.0-T carotid MR imaging by acquiring SNAP and magnetization-prepared rapid acquisition gradient-echo (MP-RAGE) sequences. Presence and area of IPH were separately evaluated on SNAP and MP-RAGE images. Presence and area of IPH were also assessed on histology. Agreement between SNAP/MP-RAGE and histology was determined in identify and quantify IPH using Cohen kappa, Spearman correlation, and Bland-Altman analyses. Results Of all 35 patients (mean age: 63.1 +/- 8.8 years; 27 males), 128 slices with successful registration were eligible for analysis. The accuracy, sensitivity, specificity, and positive and negative predictive values were 86.7%, 85%, 89.6%, 93.2%, and 78.2% for SNAP, and 76.6%, 75%, 79.2%, 85.7%, and 65.5% for MP-RAGE in identification of IPH, respectively. In identification of IPH, the kappa value between SNAP and histology and between MP-RAGE and histology was 0.725 and 0.520, respectively. The correlation between SNAP and histology (r= 0.805,p< 0.001) was stronger than that between MP-RAGE and histology (r= 0.637,p< 0.001) in measuring IPH area. Bland-Altman analysis showed that, in measuring IPH area, the bias of SNAP (1.4 mm(2), 95% CI: - 0.016 to 2.883) was smaller than that of MP-RAGE (1.7 mm(2), 95% CI: - 0.039 to 3.430) compared with histology. Conclusions This validation study by histology demonstrates that SNAP sequence better identifies and quantifies carotid intraplaque hemorrhage compared with traditional MP-RAGE sequence.