ASSESSMENT OF CAROTID INTRAPLAQUE HEMORRAGE BY PHOTOACUSTICS IMAGING (PAI) IN PATIENTS UNDERGOING TROMBOENDOARTERIECTOMY: FIRST IN-VIVO HUMAN VALIDATION STUDY

Abstract

Objective: The aim of the study is the validation of a portable multimodal photoacoustic imaging (PAI) system, for the identification of intraplaque haemorrhage and compare with MRI and histology (gold standard). Design and method: 25 patients with carotid stenosis>70% and clinical indication to tromboendoarterectomy were recruited. Angio-MRI for intraplaque hemorrhage assessment (Cube sequence) was performed. PAI clips (5 seconds, Frame rate 1000/sec) were acquired. Each clip was scored for the presence of PAI signal by means of an integrated scoring system (semiquantitative, from 0 to 12). Semiquantitative grading scales were used to assess plaque histological features of hemorrage and vulnerability. Results: 18 patients had no missing MRI, PAI and histology data and were included in this analysis. Mean age was 73 ± 8 years, 60% men, 80% Caucasians, 92% hypertensives, 60% with a previous stroke. Only 3 plaques out of 21 showed no signs of intraplaque hemorrhage, 4 showed small hemorrage, while 14 (67%) showed large hemorrhages. PAI score (best cut-off >4) correctly classified 14 out of 18 patients (Sensitivity 73.3%, specificity 100%, AUC 0.867). MRI performance was substantially similar (Sensitivity 60%, specificity 100%, AUC 0.800, 12 patients correctly classified), with a non-significant difference in AUC compared to PAI (p = 0.420). Conclusions: In this first in-vivo human study, PAI is able to identify histological intraplaque hemorrhage with an excellent specificity and acceptable sensitivity, equivalent to MRI. The very high specificity, with a low number of false positives, make PAI a good candidate for evaluation of plaques prior to surgery to i.e. reinforce the decision to perform surgery.