Purpose: Quantitative ultrasound parameters enable noninvasive visualization of carotid plaque vulnerability. We evaluated two methods for computing attenuation coefficient (AC) in-vivo and their relationships to surgical scores and histopathological assessments of plaque.
Methods: Ultrasound images and radiofrequency (RF) data (n=62 frames) were acquired from 29 individuals scheduled for carotid endarterectomy. Reference phantom data were acquired after clinical scanning. Surgical side internal carotid artery plaques were segmented at end diastole. AC (dB/cm-MHz) was computed using optimum spectral shift estimator (OPSSE) and reference phantom method (RPM). Surgical scores for cholesterol, calcium, thrombus, and ulceration (1-4), and histopathology assessments of representative H&E sections for percent cholesterol/fibrinoid, percent calcium, percent hemorrhage, and for hemosiderin and inflammation (0-3) were performed. Wilcoxon signed rank test evaluated differences between AC computation methods. Kendallâ€™s tau assessed relationships between median AC, surgical scores, and histopathology assessments. Kruskal-Wallis H test evaluated median ACâ€™s ability to distinguish between surgical scores and histopathology assessments.
Results: Participants were median(interquartile range[IQR])=74(10) years-old, 65.5 percent female. Median AC differed between OPSSE (0.51[0.19]) and RPM (0.69[0.60]), p<0.01. Percent of pixels containing negative AC estimates ranged 0-25 percent with OPSSE and 0-100 percent with RPM. OPSSE-computed median AC correlated (correlation tau[p-value]) with calcification scores (-0.31[0.002]), percent cholesterol/fibrinoid (0.27[0.003]), percent hemorrhage (-0.31[0.002]), and hemosiderin (-0.21[0.035]). RPM-computed median AC correlated with hemosiderin (-0.23[0.027]). OPSSE-computed median AC can distinguish between calcium scores (Î§2[df=3]=12.25, p=0.007). OPSSE- and RPM-computed median AC can distinguish between hemosiderin (Î§2[df=3]=10.48, p=0.005; Î§2[df=3]=10.28, p=0.006).
Conclusion: OPSSE- and RPM-computed median AC differ statistically. OPSSE outperforms RPM relative to negative AC estimates and relationship to surgical scores and histopathology assessments. OPSSE-computed median AC correlates with percent hemorrhage, percent cholesterol/fibrinoid, and calcification and hemosiderin scores, while RPM-computed median AC correlates with hemosiderin scores. OPSSE-computed median AC can distinguish between calcification and hemosiderin scores, providing noninvasive insight into carotid plaque vulnerability.
Funding Support, Disclosures, and Conflict of Interest: This work was supported by NIH grants F31HL141008 and R01NS064034. Steffel, Salamat, Wilbrand, Dempsey: no disclosures. Mitchell: Davies Publishing Inc, textbook author; Elsevier, Wolters-Kluwer, author textbook chapters, royalties; contracted research grants from W.L. Gore and Associates to UW-Madison. Varghese: research agreement with Siemens Medical Solutions for ultrasound research interface use.