Noninvasive detection of lipids in atherosclerotic plaque using ultrasound thermal strain imaging: in vivo animal study

Abstract

Objectives

This study examined the feasibility of in vivo detection of lipids in atherosclerotic plaque (AP) by ultrasound (US) thermal strain imaging (TSI).

Background

Intraplaque lipid content is thought to contribute to plaque stability. Lipid exhibits a distinctive physical characteristic of temperature-dependent US speed compared to water-bearing tissues. As tissue temperature changes, US radiofrequency (RF) echoes shift in time of flight which produces an apparent strain (temporal or thermal strain: TS).

Methods

US heating-imaging pulse sequences and transducers were designed and integrated into commercial US scanners for US-TSI of arterial segments. US-RF data were collected while gradually increasing tissue temperature. Phase-sensitive speckle tracking was applied to reconstruct TS maps co-registered to B-scans. Segments from injured atherosclerotic and uninjured non-atherosclerotic common femoral arteries (CFA) in cholesterol fed New Zealand rabbits, and segments from control normal diet fed rabbits (n=14 total) were scanned in vivo at different time points up to 12 weeks.

Results

Lipid-rich atherosclerotic lesions exhibited distinct positive TS (+0.19±0.08%) compared with that in non-atherosclerotic (−0.10±0.13%) and control (−0.09±0.09%) segments (p<0.001). US-TSI enabled serial monitoring of lipids during atherosclerosis development. The co-registered set of morphological and compositional information of US-TSI showed good agreement with histology.

Conclusions

US-TSI successfully detected and longitudinally monitored lipid progression in atherosclerotic CFA. US-TSI of relatively superficial arteries may be a modality that could be integrated into a commercial US system for noninvasive lipid detection in AP.

Introduction

Atherosclerosis is characterized by vessel wall inflammation and thickening, where lipids, cells, and scar tissue deposit (1,2). These compositional changes in the vascular wall lead to the formation of atherosclerotic plaque (AP) (3). APs can rupture and cause major cardiovascular events such as acute coronary syndromes or ischemic stroke (4). Post-mortem correlational studies linking morphological and compositional features of plaque to acute cardiovascular events have motivated efforts to develop imaging methods for interrogating plaques in vivo; some of these imaging approaches are in clinical use, undergoing clinical trials, or still under pre-clinical development (5). To the extent that lipids in a plaque may be prognostically important (6,7), we sought to develop a method for detecting lipid in AP.

Ultrasound (US) thermal strain imaging (TSI) can identify lipid-bearing tissue (LBT) surrounded by water-bearing tissue (WBT) (8,9). US-TSI is based on the observation that the speed of sound is temperature-dependent in LBT vs. WBT in opposite directions (10). As sound speed increases with temperature rise in WBT, radiofrequency (RF) US echoes arrive back sooner to the US transducer (10). Conversely, sound speed decreases with increasing temperature in LBT, such that echoes from LBT return later. Because in US imaging, time of flight translates into distance, these phenomena make LBT appears further away during imaging, which can be estimated as positive thermal strain (TS), whereas WBT appears closer, which can be estimated as negative TS (9). These apparent temporal or thermal strains have no relation to the mechanical strains generated by tissue deformation in elastography.

Based on these considerations, we hypothesized that lipids in AP can be detected by US-TSI. Using a rabbit atherosclerotic model, we present the first in vivo study demonstrating the feasibility of detecting lipids in AP noninvasively using US-TSI.

Methods

Animals

Fourteen male New Zealand white rabbits (3.5–4 kg) were studied (4 controls, 10 accelerated atherosclerosis (11)) under the approval of the Institutional Animal Care and Use Committee of the University of Pittsburgh. The atherosclerosis rabbit group was fed an atherogenic diet (peanut oil 6%, cholesterol 1%) for 5 weeks. One week after commencing the diet, a balloon catheter (2F Fogarty, Edwards Life Sciences LLC, CA) was introduced into the common femoral artery (CFA) to induce injury. Injured right CFAs served as “atherosclerotic” vessels (n=10), while uninjured contralateral CFAs in the same atherosclerosis group were used as “non-atherosclerotic” vessels (n=10). In four rabbits without balloon injury that were fed a normal diet, the right CFA served as a negative control for atherosclerosis (“normal diet control”).

US-TSI

Imaging and heating were performed using a single US transducer at 6 MHz (axial resolution ≈ 200 μm) attached to a clinical US scanner (SonixTOUCH, Ultrasonix Medical Corp., Canada). Rabbit ECG signal served as a trigger to synchronize US-TSI frame acquisition to end-systole (or end-diastole) to eliminate the mechanical strain from cardiac pulsation. An US-TSI sequence (12) was adapted as follows: The sequence began with imaging for 5 ms upon receiving the trigger, followed by heating for 192 ms, and then a short pause until the next trigger.

In a subset of 8 animals, high-resolution US-TSI was investigated using an ECG synchronized high-frequency US machine (Vevo2100, VisualSonics Inc., Canada). A custom US heating array transducer was attached to the imaging transducer (axial resolution ≈ 75 μm) to efficiently increase tissue temperature (13). An US center frequency of 21 MHz was used for imaging, while 3.55 MHz was used to operate the heating array.

US-RF data were processed offline using Matlab 7.12.0 (MathWorks Inc., MA). A temperature increase of 1.1 ± 0.1 °C in 5 s was measured in vivo near the CFA using a temperature sensor. In this study, a number of US frames (heating time) corresponding to approximately 1.5°C temperature rise was used.

Signal Processing

US-RF frames were acquired before and after temperature rise. Due to temperature-induced sound speed change, US-RF echoes arrive sooner or later based on tissue composition, which appears as negative or positive time shift, respectively. These shifts were tracked using a phase-sensitive speckle tracking technique (14). TS was then estimated as the derivative of time shifts along the US propagation direction (9). TS maps for arterial segments, color coded such that red and blue indicated the positive and negative strain, respectively, were superimposed on B-mode images.

Histology

Post-mortem, CFAs were perfusion-fixed, excised, embedded in OCT compound, and frozen at −80°C. Vessel cross-sections were stained with general hematoxylin and eosin for nuclei staining, and oil red O for lipid staining. Histology sections were identified relative to the distance from CFA bifurcation (BF) to enable comparison with US at anatomically concordant sites.

To quantify lipid progression relative to an atherosclerotic segment, a region of interest, from lumen boundary to adventitial layer, was manually segmented. Then, the area of red stained lipids (oil red O) was divided by the total segmented area to estimate percentage lipid. Similar procedures were followed to quantify percentage lipid in approximately matched segments within US-TSI for comparison. Histological quantitative measurements were performed using Image J 1.46r (National Institutes of Health, Bethesda, MD).

Statistical analysis

Data analyses were performed using the Statistics Toolbox of Matlab 7.12.0. All values are expressed as the mean ± SD. TS assessments in the atherosclerotic, non-atherosclerotic, and control vessel groups were compared at terminal days using Student’s t test. Linear regression analysis was performed to compare US and histology measurements. A p-value <0.05 was considered significant (two-tailed).

Experimental Protocol

Serial US-TSI, co-registered to Duplex US, was performed at week 0 (day of injury), and 4, 6, 8, 10 or 12 weeks post-injury. US-TSI was performed for multiple views of the CFA from the CFA-BF and up to ~12 mm proximally. Regions of interest in the CFA were within the heating beam and at depths ranging 9–13mm from the skin surface. Rabbits were euthanized at week 4 (n=1), 6 (n=1), 8 (n=2), 10 (n=1) and 12 (n=5) post-injury.

Results

US-TSI of atherosclerotic and non-atherosclerotic arteries

Figure 1 compares US-TSI of atherosclerotic and contralateral non-atherosclerotic CFAs in a cholesterol-fed rabbit 12 weeks post-injury. The dashed lines in B-scans approximately mark the heated area. US-TSI (Fig. 1B) for the atherosclerotic vessel (Fig. 1A) shows positive TS at sites, where histology (Fig. 1C) indicated lipids. In the non-atherosclerotic vessel (Fig. 1D), US-TSI (Fig. 1E) did not show positive TS near the lumen. Histology (Fig. 1F) showed normal vessel.

Figure 1. US-TSI of Atherosclerotic Versus Non-atherosclerotic Vessels.

(A) B-mode of atherosclerotic CFA 12 weeks post-injury. US-TSI (B) for (A) identified lipids (positive TS) in the vessel wall. (C) Histology of atherosclerotic vessel (A) (left) with rectangular inset at 5x magnification (right). (D) B-mode of the contralateral uninjured CFA of the same rabbit. US-TSI (E) for (D) shows no positive TS near the inner vessel layer. (E) Histology of non-atherosclerotic vessel (D) (left) with rectangular inset at 5x magnification (right). Histology cross-sections are ~5 mm proximal to BF.

Monitoring lipid progression using US-TSI

The B-mode of the atherosclerotic vessel (Fig. 2A) shows a noticeable luminal stenosis (arrows) and correspondingly reduced US Doppler signal (Fig. 2B) at 10 weeks post-injury. Histology (Fig. 2C) confirmed the presence of lipid-laden AP within the segment. Serial US-TSI was performed for this CFA at 6, 8, and 10 weeks post-injury. Positive TSs were observed in the segment at week 6 (Fig. 2D), and progressed in intensity and spatial extent in weeks 8 and 10 (Figs. 2E–F).

Figure 2. Longitudinal Monitoring of Lipids by US-TSI.

B-mode (A) of atherosclerotic CFA 10 weeks post-injury shows a luminal stenosis (arrows) and reduced US Doppler (B) signal (arrows). Oil red O histology (C) confirms lipid-rich AP (arrows). (D–F) US-TSI for the same CFA 6, 8, and 10 weeks post-injury, respectively.

Quantitative assessment of lipid-rich AP using US-TSI

TS in lipid-rich atherosclerotic segments (n=10) was significantly higher (p<0.001) than that measured in both uninjured non-atherosclerotic (n=10) and control segments (n=4) (Fig. 3A). Atherosclerotic segments exhibited positive strains of +0.19±0.08%, whereas non-atherosclerotic and control segments exhibited mostly negative strains of −0.10±0.13% and −0.09±0.09%, respectively. No significant difference in TS was observed between non-atherosclerotic and control segments (p=0.748).

Figure 3. Quantitative Analysis of TS and Correlation with Histology.

TS in lipid-rich atherosclerotic segments significantly differed (^*p<0.001) from that measured in both non-atherosclerotic and control segments (A). No significant difference in TS was observed between non-atherosclerotic and control segments (^†p=0.748). US-TSI and histology percentage lipid measurements exhibited close correlation (B).

A close correlation was found between histological and US-TSI measurements of percentage lipid (n=10) (Fig. 3B).

Discussion

Using a rabbit atherosclerotic model, this study is the first in vivo demonstration that differences in TS between atherosclerotic and normal segments can be spatially mapped and those areas with positive TS co-localize with histologically proven lipid-laden AP. These preliminary data have major implications for the in vivo detection of lipid-rich plaques.

Atherosclerotic segments exhibited distinct positive TS as would be expected in LBT, whereas control and non-atherosclerotic segments possessed negative TS typical of WBT. Similar findings were observed in vitro using excised tissue (9). We measured a relatively large SD of TS in atherosclerotic segments, probably for several reasons. First, because we combined measurements at different terminal days (4–12 weeks), variations in lipid content increased the range of positive TS. Second, atherosclerotic segments contain WBT such as smooth muscle cells, connective tissue, and fibrosis, which have negative temperature dependences of sound speed. This would result in heterogeneous TS maps, as the varied tissue components cannot be completely excluded during image segmentation. TS in control and non-atherosclerotic segments also had large SD, as some areas exhibited positive TS, which could be due to normal adventitial fat (15). Three common factors that may vary TS in all groups are the variation in the heating beam, US attenuation in muscle/vascular tissue of ~1 dB/(MHz.cm) (16), and potential residual mechanical deformation due to physiological motion. Notwithstanding these inherent limitations, the average TS clearly differentiated lipids from WBT in AP with high significance (p < 0.001). Terminal study points were varied to compare US-TSI and histology at different stages of atherosclerosis progression. Although some atherosclerotic vessels did not show a significant stenosis by duplex US, US-TSI was able to detect non flow-limiting AP fatty lesions. Furthermore, US-TSI serially tracked lipid accumulation during AP progression, exhibiting a continuous change in the area and/or value of positive TS in AP over time (Figs. 2D–F). The continuous remodeling in the shape of AP at different stages (17) may suggest corresponding changes in lipid distribution. Although a slight misalignment in imaging planes between each week is possible, the overall observations were consistent with the interpretation of AP progression.

This study has several limitations. Relatively few animals were studied in this proof of concept study, precluding complete statistical analysis and quantitative measurements of the sensitivity and specificity of US-TSI. However, our quantitative analysis, albeit limited, showed good correlation between US and histology (Fig. 3B). US and histology measurements did not match in their absolute magnitudes, which could be due to the limited resolution of US compared to high-magnification histology images, and/or the potential post-mortem changes in vessels during histology processing (18). Although this rabbit model exhibited some features of AP similar to human disease, the pathophysiology and extreme nature of the lesions may not be fully representative of those causing human carotid plaque or acute coronary syndromes (11). Nonetheless, these lipid-rich arterial lesions (19) enabled us to demonstrate the feasibility of noninvasive lipid detection in AP.

US-TSI relies on inducing a gentle temperature rise in tissue, so temperature control is important for safe in vivo use. At this small temperature (<10°C), a linear relationship between sound speed and temperature can be assumed (20). Accordingly, the temperature rise during US-TSI can be estimated using the linear relationship between heating time and TS (12). According to the American Institute of Ultrasound in Medicine (AIUM) (21), “there have been no significant, adverse biological effects observed due to temperature increases less than or equal to 2°C above normal, for exposure durations up to 50 hours”. For future clinical implementation, we chose an US-TSI temperature increase of ~1.5°C in less than 10 s, which generated sufficient signal/noise without conferring significant adverse biological effects. The peak negative pressure of US heating pulse was estimated experimentally. The clinical system exhibited a pressure amplitude of −1.27 MPa, while it was −1.40 MPa using the custom heating array. We did not observe adverse biological effects due to US-TSI. Also, to the best of our knowledge, there have been no previous reports to suggest adverse effects of the temperature rises used in this study on plaque stability. Our limited observations thus far and the acoustic parameters used would predict safety of our approach, although important next steps at this juncture would be to perform serial studies on the heating effect on plaque biology and other tissues along with rigorous measurements for temperature and time-average power. However, a comprehensive systematic investigation of bioeffects was outside the scope of the current proof of concept study.

Conclusion

US-TSI was able to detect and longitudinally monitor lipids in AP. Lipids in atherosclerotic arterial segments exhibited distinct positive TS, whereas control and non-atherosclerotic segments, comprising mostly WBT, showed predominantly negative TS. Lipid-rich lesions identified by US-TSI co-localized and were quantitatively concordant with histology. These preliminary findings demonstrate the potential utility of US-TSI for noninvasive lipid detection in AP in vivo. This novel strategy is worthy of additional study to determine its sensitivity and specificity in detecting lipid in AP and the potential clinical utility of these findings.

Abbreviations and Acronyms

AP

Atherosclerotic Plaque

BF

Bifurcation

CFA

Common Femoral Artery

LBT

Lipid-Bearing Tissue

RF

Radiofrequency

TS

Temporal (or Thermal) Strain

TSI

Temporal (or Thermal) Strain Imaging

US

Ultrasound

WBT

Water-Bearing Tissue