Click here to


Are you sure ?

Yes, do it No, cancel

Integrated Backscatter Computation for Quantitative Ultrasound Evaluation of Carotid Atherosclerotic Plaque In-Vivo

C Steffel*, S Wilbrand, S Salamat, R Dempsey, C Mitchell, T Varghese, University of Wisconsin-Madison School of Medicine & Public Health, Madison, WI


(Wednesday, 7/17/2019) 1:45 PM - 3:45 PM

Room: 303

Purpose: Quantitative ultrasound parameters may provide visualization of heterogeneous carotid plaque tissue composition by quantifying plaque echogenicity. We compute attenuation-corrected integrated backscatter (IB [dB]) in-vivo using an optimum spectral shift estimator (OPSSE) and calculate IB parameters. We hypothesize that IB parameters will relate to percent cholesterol/fibrinoid and percent calcium in plaque as assessed by histopathology.

Methods: Radiofrequency (RF) and clinical ultrasound data of longitudinal internal carotid artery plaque were collected from 29 individuals (n=62 frames) scheduled to undergo carotid endarterectomy (CEA). RF data were collected at a center frequency of 11.43MHz. Reference phantom RF data were acquired at identical imaging settings. Phantom attenuation coefficient was 0.5 dB/cm-MHz. A research sonographer segmented plaque at end-diastole using MITK software. IB was computed at end-diastole using a reference phantom technique. Attenuation coefficient was computed using an optimum spectral shift estimator (OPSSE) method. Negative attenuation coefficient estimates were floored before performing IB attenuation correction via an additive term. Then, minimum and maximum IB were extracted from plaque using MITK segmentation as a mask. Representative hematoxylin and eosin sections of CEA plaque specimens were assessed for percent cholesterol/fibrinoid, percent calcium, and percent hemorrhage, and for degree of hemosiderin and inflammation (0-3). Histopathological markers were compared against minimum and maximum IB in-vivo using Kendall’s tau correlation.

Results: Participants were median(interquartile range[IQR])=74(10) years-old, 65.5 percent female. Minimum IB correlated with percent cholesterol/fibrinoid, correlation tau(p-value)=-0.216(0.017), and maximum IB with percent calcium (0.301[0.001]). Percent of pixels containing nonphysical IB estimates ranged median(IQR)=11.10(38.86).

Conclusion: Attenuation-corrected IB was computed using OPSSE method and a reference phantom. Minimum and maximum IB relate to histopathological markers of plaque vulnerability. Clinical implementation will require regularization or filtering, as a significant percentage of pixels yielded nonphysical IB estimates. Our results suggest IB parameters may provide insight into carotid plaque tissue composition.

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.


Ultrasonics, Quantitative Imaging, Stroke


IM- Ultrasound : Quantitative imaging/analysis

Contact Email