Ex Vivo Evaluation of IVUS-VH Imaging and the Role of Plaque Structure on Peripheral Artery Disease

Abstract

Peripheral artery disease (PAD) results from the buildup of atherosclerotic plaque in the arterial wall, can progress to severe ischemia and lead to tissue necrosis and limb amputation. We evaluated a means of assessing PAD mechanics ex vivo using ten human peripheral arteries with PAD. Pressure-inflation testing was performed at six physiological pressure intervals ranging from 10–200 mmHg. These vessels were imaged with IVUS-VH to determine plaque composition and change in vessel structure with pressure. Statistical analysis was performed to determine which plaque structures and distributions of these structures had the greatest influence on wall deformation. We found that fibrous plaque, necrotic core, and calcification had a statistically significant effect on all variables (p<0.05). The presence of large concentrations of fibrous plaque was linked to reduced vessel compliance and ellipticity, which could lead to stent fractures and restenosis. For the plaque distribution we found that clustered necrotic core increased overall compliance while clustered calcification decreased overall compliance. The effect of plaque distribution on vessel wall deformation must be considered equally important to plaque concentration.

Introduction

Peripheral artery disease (PAD) is a condition commonly characterized by the buildup of atherosclerotic plaque in arteries of the peripheral system. The associated risk factors for PAD are similar to those for coronary heart disease and include diet, age, smoking and genetic factors [1]. Plaque severity and vulnerability can be assessed using intravascular ultrasound (IVUS), where an ultrasound probe is mounted on a catheter and the catheter is fed to the region of interest through an opening in the vessel. The probe then gives a cross-sectional ultrasound image of the vessel allowing the plaque size, type, and distribution to be identified. Plaque components can be further distinguished by spectral analysis of the back scattering, known as virtual histology (VH) [2,3].

While plaque components can be identified and distinguished using this method, there is limited information on the effect of these components on wall deformation and compliance which can have a large impact on disease complications and interventions such as angioplasty and stent performance [4–6]. This knowledge has clinical implications, particularly when performance of mechanical interventions such as balloon angioplasty and stenting will likely depend on disrupted mechanical behavior of the artery with plaque [7–9]. The knowledge of which plaque types have the greatest effect on vessel expansion may provide additional clinical insight to assist in the selection of treatment options for a specific patient. Currently, computed tomography angiography is utilized to calculate the approximate sizing of the vessel lumen to help with the selection of appropriate balloon catheters and stents. Observational analysis of coronary revascularization trends from the Medicare and Medicaid Services database between 2001 and 2009 demonstrated low utilization of IVUS (<50%) in percutaneous intervention [10]. Additionally, multiple large meta-analyses illustrated a reduction in major adverse events (MACE) with the routine use of intracoronary imaging for procedural guidance during percutaneous intervention [11,12]. However, plaque characterization and the resulting deeper underlying mechanical response are not available, or are not widely adopted, and can impact clinical decision making. Additionally, the presence of calcification pockets can greatly alter stress concentrations on the stent struts, as these regions will not deform like the softer plaque components [8,13].

In this study, we demonstrate a means of evaluating the mechanical performance of atherosclerotic peripheral arteries ex vivo using IVUS-VH imaging and pressure inflation testing. IVUS-VH is utilized to determine plaque composition and its influence on changes of wall shape and size at increasing internal pressure (as evidenced by cross-sectional area, major axis diameter, and aspect ratio). We anticipate that this knowledge will provide a patient- and location-specific indicator for artery stiffness, which will be useful as a strategic clinical tool for individualized interventional therapies.

Materials and Methods

Pressure Inflation Testing

After receiving Mayo Clinic Institutional Review Board approval (IRB 14–009261), ten human cadaveric femoral arteries were collected from patients who had PAD and were stored at −70°C until testing. Demographics and characteristics of the patients from which the arteries were excised are summarized in Table 1, along with the number of VH frames collected for each artery. Before testing the samples, they were thawed in a refrigerator at 4°C overnight. Excess fatty tissue surrounding the artery was carefully removed using surgical scissors, and any vessel branches were closed by suturing. The fatty tissue was removed to allow the plaque components to be distinguished more clearly and to allow bifurcations to be identified so that they could be sealed. As fatty tissue has low stiffness it would be expected that the changes to the mechanical response will be negligible. A short piece of plastic tubing with an outer diameter slightly smaller than the vessel lumen diameter was inserted into each end of the vessel and secured with cable ties. Custom fixtures were attached at the top and bottom of the in vitro chamber to enable vessel pressurization, and the fixtures were connected to the plastic tubing using clamps. The samples were mounted such that the artery was held straight with the minimum axial tension applied. The IVUS catheter was inserted into the vessel through a side port that minimized water outflow, and then the chamber was sealed and filled with water. Pressure was measured above and below the clamps and controlled using custom software (Harvard Apparatus, Holliston, MA) [14]. The full configuration can be seen in Fig. 1). IVUS images were captured with an Eagle Eye Platinum catheter (Philips, Andover, MA, USA), and acquisition was performed on a Volcano IVUS s5 machine (Volcano Corporation, San Diego, CA, USA). Automatic pullback was performed at 0.5 mm/s at the baseline pressure of 10 mmHg, which trial experiments showed fully perfused the artery, and then pressures of 60, 80, 120, 160 and 200 mmHg were sequentially applied and the IVUS imaging procedure was repeated for each pressure. VH analysis was performed at the baseline pressure, while at the subsequent pressures only IVUS data was acquired. VH images were acquired at 5 frames/mm using an ECG simulator (HE Instruments, Lake Worth, FL, USA) to replicate patient cardiac electrical activity, due to VH being ECG gated. VH results identified arterial wall (grey) and four plaque components; calcification (white), fibrous plaque (dark green), fatty plaque (light green), and necrotic core (red) (Fig 2).

Table 1.

Patient information and number of VH frames collected from each artery.

Case Number	Gender	Age	BMI	Height (m)	Race	No. VH Frames
1	M	58	24.12	1.82	White	219
2	F	73	21.08	1.65	White	322
3	F	63	23.56	1.67	White	321
4	F	58	42.58	1.55	Hispanic	230
5	F	49	24.12	1.68	White	271
6	F	96	29.10	1.60	White	401
7	M	84	30.92	1.73	White	295
8	M	87	20.79	1.70	White	299
9	F	79	32.66	1.84	White	328
10	M	66	18.12	1.60	White	107

Patient gender, age, BMI, Height in meters, race, and the number of VH frames collected from each artery are shown. Age listed is the patient’s age at the time of artery explant.

Fig 1. Images detailing the experimental assembly.

Plastic tubing is inserted into each end of the vessel and held with cable ties (a). This tubing in turn is inserted into the fixtures by clamps (b) that are attached into the bioreactor (c), which is then filled with water. The schematic (d) shows the full assembly.

Fig 2. Representative VH slices from case 3 (left) and case 8 (right).

Figure demonstrates the arterial wall (grey) and four different plaque components; calcification (white), fibrous plaque (dark green), fatty plaque (light green), and necrotic core (red).

Segmentation and Registration

Semi-automatic segmentation of the artery lumen was performed using 3D Slicer [15]. Prior to segmentation, the IVUS DICOM images were processed using a custom Python script (Python Software Foundation), implementing SimpleITK functions [16,17]: (1) a vector image was converted to a scalar image using the first image index, (2) the image was resampled to every 5^th slice, and finally (3) a 5 × 5 × 5 pixel median filter was applied. Using the segment editor module in 3D Slicer, the high intensity pixels (representing primarily the artery wall) were segmented using the “threshold” effect. Next, the lumen was labeled as a second segment using the “level tracing” effect in 15 to 20 slices along the length of the artery. The IVUS probe artifact was labeled with the lumen label using the “scissor effect” through the entire length of the artery. The “grow from seeds” effect was then implemented to perform the automatic segmentation. Manual adjustments to the segmentation were implemented as needed. A raw model (no smoothing or decimation) was created from the segmentation and exported as a triangulated surface in .stl format for geometric analysis.

Virtual histology images were registered to the 10 mmHg IVUS images by specifying the IVUS location of the first and last VH image. The remaining VH images were assumed to be equally spaced between the first and last IVUS longitudinal coordinates. The end point of the IVUS data collection (first sign of narrowing at the clamp site) was used to register IVUS images from the remaining pressure levels to the 10 mmHg IVUS coordinates. The major and minor axis diameters of the virtual histology images and .stl models created from the IVUS images at each pressure level were determined using a direct least squares ellipse fit to the lumen boundary points at the image cross-section [18].

Data Processing and Statistical Analysis

For this analysis, we evaluated five sets of independent variables: the applied pressures, the plaque composition, the plaque circumferential variation (represented as a coefficient of variation, or COV), the uniformity of the lumen, and the effective wall thickness. Each of these variables (barring the first, which is self-explanatory) is described below:

Plaque composition:

The proportion of each plaque component in each slice, measured by finding the ratio of the number of pixels of the respective plaque color to the total number of non-black (i.e., non-background) pixels.

Plaque circumferential variation:

This was found first by drawing a line between the centroid of the lumen and the edge of the image and calculating the proportion of non-black pixels along this line. The line was then rotated about the centroid at regular angular intervals, similar to a clock hand, and the pixel proportion counted at each interval. The variation was then found by taking the standard deviation normalized by the mean (the coefficient of variation, COV). Clustered plaques gave a higher COV, while evenly distributed plaques gave a low COV.

Uniformity of the lumen:

First an ellipse is fitted to the lumen, and then an exclusive OR (XOR) operation is performed between the binary image of the lumen and the overlaid binary image of the fitted ellipse. The ratio is then found between the area from the XOR operation and the lumen area.

Effective wall thickness:

This was calculated from 2(A_outer/P_outer − A_inner/P_inner), where A_outer and A_inner are the areas within the outer and inner walls of the artery, respectively, and P_outer and P_inner are the perimeters of the outer and inner walls, respectively.

For the following analysis, the plaque composition and plaque circumferential variation each have five parameters (arterial wall and four plaque components), and the uniformity of the lumen and the effective wall thickness are both single variables. Thus with the six applied pressures there were sixteen independent variables for each slice (S. Tables 1–3). The dependent variables were cross-sectional area, major axis diameter, and ellipse aspect ratio (the ratio between major and minor axis diameter) for each slice.

The effect of pressure on the change of the vessel’s cross-sectional area, major axis diameter, and ellipse aspect ratio was analyzed by performing a one-way ANOVA followed by a multi-comparison test using the statistical toolbox within MATLAB 2016a (The Mathworks Inc, Natick, MA, USA). In order to assess the association between vessel diameter and vessel composition, three linear mixed-effects models were fit to the data, one for each dependent variable. The vessel composition manifested in the model as percentage variables for four plaque types and one for arterial wall. The arterial variable (grey in the VH images) was excluded from the model. As such, all interpretations are in reference to a corresponding increase or decrease in the percentage of arterial wall (grey region in the VH images). The correlation of repeated measurements within subjects was accounted for using random effects. Also, the nested effect that the location on the vessel within the subject was modeled assuming an order 1 autoregressive covariance structure. This structure was used to capture the assumed higher correlation of measurements made nearer to each other within each subject. Plaque circumferential variation, uniformity of the lumen, and effective wall thickness were scaled by their respective standard deviations prior to regression. The fixed effects with accompanying 95% confidence intervals are included in the results. The Wald test was used to test the significance of fixed effects. For the purposes of this paper, a 2-tailed p-value of less than .05 was considered significant. This statistical analysis was performed using R (version 3.4.2).

Results

Pressure Inflation Testing

From among the ten arteries tested for this study, representative plots comparing percentage change in artery wall structure, cross-sectional area (with respect to diameters at 10 mmHg), and ellipse aspect ratio at 80 mmHg and 120 mmHg (representing diastolic and systolic pressures respectively) are given in Fig 3a (case 3) and Fig 3b (case 8). Case 3 shows a relatively uniform wall structure with little variation along the vessel, while case 8 has a wall structure that varies significantly along the vessel (see also Fig 2). These trends are reflected in the percentage change in vessel cross-sectional area and aspect ratios: the area changes and aspect ratios are relatively constant in case 3, and vary significantly in case 8. For case 3, the percentage change in vessel area is consistently higher for 120 mmHg compared to 80 mmHg, but for case 8 the difference is not as pronounced. However, the aspect ratio does not show a large difference between 80 and 120 mmHg for either case.

Fig 3. Plots detailing arterial wall composition and percentage change in vessel diameter along the length of the artery segments.

Two representative cases are shown for brevity, case 3 (a) and case 8 (b). Pressures shown are 80 mmHg and 120 mmHg to allow comparison between arterial wall composition and wall deformation. The spike in aspect ratio seen in (b) is due to a bifurcation in the vessel, see supplementary figure 2.

Fig 4 shows the percentage changes between the cross sectional area (Fig 4a) and major axis diameter (Fig 4b) at each recorded pressure, compared to the baseline pressure. The ellipse aspect ratio (the ratio between major and minor axis diameters), including the baseline 10 mmHg pressure, is shown in Fig 4c. A summary of the salient features of each is given below:

Fig 4. Box plots detailing the changes in artery slice geometric features when pressurized.

(a) Percentage change in minor axis diameter (compared to 10 mmHg baseline pressure) versus applied internal pressure. (b) Percentage change in major axis diameter (compared to 10 mmHg baseline pressure) versus applied internal pressure. (c) Percentage difference between minor axis and major axis diameters for all applied internal pressures including 10 mmHg baseline. Outliers were omitted for clarity. For (a) and (b), plain asterisks indicate that the respective pressure is significant when compared to all other pressures. For (c), asterisks with numbers adjacent shows the pressures for which the respective pressure is significantly different. Mean and interquartile ranges are shown and whiskers extend to 2.7 standard deviations outside this are outliers. Box plots are calculated for all data lumped across slices and arteries.

Percentage change in cross-sectional area:

There is a consistent increase with pressure (median was 9.012, 10.141, 14.784, 16.601, and 19.878 for each pressure level) and the top whisker extends to 43.61% change at 200 mmHg. While all means and interquartile ranges are positive, the bottom whisker extends below 0% for all cases (−6.129, −6.904, −2.965, −1.627, and −0.696 for each pressure level).

Percentage change in major axis diameter:

As with the change in cross-sectional area, the change in major axis diameter increases with pressure, although the differences are smaller (median was 5.233, 5.841, 8.053, 8.848, and 10.192 for each pressure level). Again, the lower interquartile ranges are positive but the lower whiskers are all approximately −5% regardless of pressure.

Ellipse aspect ratio:

There is little change between each pressure except for 80 mmHg where the median is marginally higher (1.094 compared to 1.086 for 10 mmHg) and 200 mmHg where the mean is slightly lower than the other pressures (1.0718).

Statistical Analysis

Fig 5 shows the results of the statistical analysis for the lumen cross-sectional area, major axis diameter, and ellipse aspect ratio, respectively (further information is available in the S. Tables 1–3). The effect for each plaque component percentage represents variation from the amount that the diameter would be expected to change based on pressure alone with a 1% increase in the predictor (fatty plaque, calcification, necrotic core, and fibrous plaque). The effect for the remaining, scaled, variables represents the amount the expected diameter would change with an increase of one standard deviation, again after adjusting for pressure. Observations for each condition are given below:

Fig 5. Forest plots detailing the statistical analysis of plaque components on arterial geometric features.

Statistical analysis of the effect of plaque constituent proportions (with respect to arterial wall proportion), the lumen uniformity, and effective wall thickness on lumen area, major axis diameter, and aspect ratio as artery is pressurized. For scaled variables, standard deviations prior to scaling are shown in brackets.

Vessel Lumen Area:

All plaque component proportions except fatty plaque (−0.0965; p=0.6317) showed a significant negative effect on the increase in vessel lumen area when pressurized. Fibrous plaque proportion had the greatest effect magnitude (−0.9605; p<0.0001) followed by necrotic core (−0.5230; p<0.0001) then calcification (−0.4930; p<0.0001). Necrotic core circumferential COV demonstrated a significant positive effect (3.1299; p<0.0001) while calcification circumferential COV demonstrated a significant negative effect (−1.3619; p<0.0001) and fatty plaque circumferential COV also showed a smaller significant negative effect (−0.3050; p=0.0342). Fibrous plaque circumferential COV showed a small non-significant effect (0.0072; p=0.9600). Finally, the effective wall thickness demonstrated a significant large negative effect (−3.6415; p<0.0001), and there was no significant effect on vessel uniformity (−0.22678; p=0.0872).

Major Axis Diameter:

For the plaque component proportions, fibrous plaque had the greatest negative effect on increasing diameter with pressure (−0.0227; p<0.0001). Necrotic core had the next strongest effect (−0.0081; p =0.329), followed closely by calcification although it was not significant (−0.0070; p = 0.0585). Fatty plaque had a small non-significant effect (−0.0029; p=0.6946). Necrotic core circumferential COV had the only positive effect (0.1247; p<0.0001). Calcification circumferential COV had the largest-magnitude negative effect with the highest magnitude (−0.0272; p<0.0001). Both fibrous and fatty plaque circumferential COV showed small non-significant effects (−0.0013; p=0.8068 and −0.0035; p=0.5136 respectively). As with area, effective wall thickness had a significant negative effect on the major axis diameter (−0.0618; p<0.0001), but there was no significant effect for vessel uniformity (0.006987; p=0.1573).

Ellipse Aspect Ratio:

All plaque component proportions except fatty plaque (3.19e–4; p=0.8053) showed a significant positive effect on lumen aspect ratio when pressurized. Necrotic core proportion had the strongest effect (0.0035; p<0.0001) followed by calcification (0.0033; p<0.0001) then fibrous plaque (0.0032; p<0.0001). Fatty plaque, necrotic core and calcification circumferential COV had significant positive effects (0.0021; p=0.0252, 0.0094; p<0.0001, and 0.0055; p<0.0001 respectively), while fibrous plaque circumferential COV had a small non-significant effect (2.22e–4; p=0.8105). Lastly, the effective wall thickness had a large positive effect (0.0155; p<0.0001), and vessel uniformity had a significant positive effect (0.005929; p<0.0001).

Discussion

In this study, we developed a test protocol to evaluate the mechanical properties of peripheral arteries with atherosclerosis ex vivo. This was then used to assess the effect of each plaque constituent on wall deformation under the application of internal pressure. Previous studies have developed alternate means of evaluating mechanical properties of atherosclerotic arteries and the plaque constituents. Chai et al. performed micro indentation on atherosclerotic carotid artery transverse slices while imaging from beneath with an inverted confocal microscope. This allowed them to evaluate local anisotropic mechanical properties of the plaque and they found no significant differences in fiber stiffness or dispersion at the different indentation sites. Though, this only probed the axial direction while the major loading in vivo is in the circumferential direction [19]. Recently, Sanders et al. performed a ring-inflation experiment on transverse slices of carotid arteries where the slice is expanded under internal pressure while held between to glass slides and imaged by a high speed camera [20]. They found that the setup produced comparable results to pressure-inflation experiments and uniaxial tension tests. However, this approach has not been evaluated for heterogeneous materials such as atherosclerotic arteries. For uniaxial testing multiple studies have used varying approaches with resulting varying mechanical behavior for each plaque type. Carotid plaques have been tested the most extensively with varying sample size and test parameters [21–24]. However, the major drawback of this approach is the inability to distinguish the plaque constituents during the test and commonly the plaques are categorized prior to the test and segregated by this. The proposed approach allows characterization of the plaque constituents during the test and characterization of the entire vessel. The utility of this approach was demonstrated with an investigation of the wall compliance for each plaque constituent and other easily measurable variables.

Fibrous plaque had the greatest significant negative effect on both the area and major axis diameter (see Fig 5 and S. Tables 1–2), but its circumferential COV showed a non-significant effect for these same variables. Thus, a high proportion of fibrous plaque is indicative of reduced vessel compliance regardless of how it is distributed, despite its inferior mechanical response compared to calcification [25,26]. It is also likely that the plaque structures that surrounded the fibrous plaque altered both the geometry and the overall mechanical environment. And therefore the presence of fibrous plaque is more indicative of a less compliant vessel wall than stiffer components like calcification, and is consequently of greater concern for stent fracture and restenosis. For aspect ratio, fibrous plaque was also highly significant but had a positive effect, while fibrous plaque circumferential COV again had a non-significant effect. Thus, a high proportion of fibrous plaque, irrespective of the circumferential distribution, may be a predictor of increased vessel ellipticity when pressurized. This may predict early stent failure because, as for the wall thickness, the wall will resist expanding and resist being deformed into a more circular shape, leading to increased stress in the stent. Previous studies have shown that stents deployed into non-uniform plaque distributions show high strut stress compared to uniform circular plaque [7,27]. Therefore, patients with high proportion of fibrous plaque may benefit from repeated balloon angioplasties to establish a more circular vessel before implantation. Additionally stents with higher elastic modulus could be beneficial in these patient populations to account for increased vessel ellipticity.

Similar to fibrous plaque, necrotic core had a significant negative effect on vessel expansion for both area and major axis diameter. As necrotic core has higher compliance than healthy arterial wall, it is counter-intuitive that a high proportion of this component is related to a negative effect on wall expansion [8]. Therefore, it is likely that it is associated with less compliant plaque structures such as calcification. The COV analysis in this work supports this notion, as clustered necrotic core leads to greater overall vessel compliance; thus, isolated pockets of necrotic core may lead to isolated pockets of stiffer associated components, increasing overall wall compliance. Physiologically, it is known that necrotic core is associated with late stage plaques and so it is highly unlikely it is found without other plaque components. Therefore, if necrotic core is evenly spread around the vessel circumference it is also likely that there is greater developed plaque around the vessel resulting in a less compliant wall, agreeing with our findings [6,28,29]. Finally, necrotic core proportion and circumferential COV both had a positive effect on vessel aspect ratio, indicating that an increase in the amount of necrotic core or greater clustering of necrotic core both result in an increase in vessel ellipticity when pressurized. As discussed earlier, this is may be an indicator of increased risk of stent failure.

The most common wall constituent in nearly all cases was calcification (see Fig 3); this is expected as all samples were from elderly patients where calcification in peripheral vessels is common and is highly prevalent in all arteries [30]. Calcification shows a significant effect on the lumen area (although the effect was lower than for fibrous plaque) and aspect ratio, but there was no significant effect on the major diameter. The latter is unexpected, as calcification was present throughout all vessels tested and is by far the stiffest wall component in both tension and compression [31,32]. Additionally, the calcification circumferential COV has a negative effect on wall area and major axis diameter, and thus more clustered calcification resulted in a reduction of artery compliance. The reason for this may be the same as mentioned above for fibrous plaque, that clustered calcifications may be associated with developed plaque structures that result in reduced wall compliance. This is supported by clinical evidence with both calcified and fibrotic plaques associated with late stage atherosclerosis where the wall commonly demonstrates low compliance [33]. Therefore, calcification distribution and proportion are important to consider when choosing the appropriate intervention strategy, as the effects of calcification distribution are not ostensibly clear. Finally, similar to necrotic core, calcification proportion and circumferential COV have a positive effect on the aspect ratio. Therefore, an increased presence of calcification or more clustering of calcification may lead to the vessel becoming more elliptical when pressurized. An interventional application may include selecting a larger stent to account for the elliptical shape.

Increased wall thickness had a significant effect on all dependent variables, negatively effecting vessel area and major axis diameter, and positively effecting aspect ratio. The first two observations are expected as plaque accumulation will lead to a thicker wall, which would be assumed to be less compliant (and thus in turn may require stents with high radial force to widen the lumen). However, the last observation is less intuitive; it may be that thicker vessels may have non-uniform compliance around the wall leading to increasing ellipticity when pressurized. Thus, if a stent is deployed in a thicker vessel, the vessel will resist both the stent expanding and the stent pushing it into a more circular shape, increasing stress in the stent struts and consequently the likelihood of stent failure in the future. Therefore, clinically, wall thickness may be an indication of necessary repeated balloon angioplasties until the vessel is more uniform before stenting the vessel. By comparison, vessel uniformity, the other vessel geometric feature analyzed, did not have a significant effect on either lumen cross-sectional area or major axis diameter, but did have a significant positive effect on the ellipse aspect ratio as the vessel is pressurized. This implies that protrusions and similar structures do not influence the wall compliance; yet it is likely that a non-uniform lumen will have protrusions that may skew the ellipse aspect ratio and this effect may become more pronounced as the vessel is pressurized.

One limitation of this study was that there were only vessels from 10 cadavers included in the analysis, which limits the generalization of these results to a larger population. Additionally, suture was used when vessel branching was present, although efforts were made to utilize as little suture as possible so that it did not influence the VH images while still ensuring a good seal. Efforts were made to minimize these effects in post processing, and branching occurred in only a small proportion of the vessel wall, so it is unlikely that this had a significant effect on the results obtained. Furthermore, in Fig 4a, the bottom whisker shows negative deformation compared to baseline for all pressures, although the effect decreases with pressure. The vessel branches may explain this as leaking will increase as pressure increases, resulting in reduced vessel expansion.

Due to the low proportions of fatty plaque for all ten vessels, results for this component must be interpreted carefully, as the statistical model had limited data and particularly the COV will be skewed by the low means of this component. A high sample size including arteries with higher proportions of fatty plaque would allow more decisive conclusions to be drawn.

During IVUS segmentation, image artifacts from branching created openings in the artery and manual segmentation was required in these areas to appropriately delineate the lumen of the artery cross section. To decrease the uncertainty in registration errors, future studies may include markers along the artery that are visible within the ultrasound image.

It is very likely that the arteries deform non-uniformly around the artery circumference, and that plaque constituents are a major reason for this. However, there was only a limited means of analyzing the deformation around the artery and correlating that with plaque constituents. In the present study, we attempted this (see Statistical Analysis in the supporting materials and S. Tables 1–3) by selecting the major axis as the reference point and analyzing the difference between the angular location of the major axis diameter and the angular location of the maximum of the circumferential plaque distribution. A more robust approach would be to record the vessel location and morphology in the patient in detail before explant and subsequently register each slice both longitudinally and circumferentially with the anatomy of the patient. Thus, all slices from all arteries would be aligned and reference points would be intuitive. However, this requires substantial endeavor both pre- and post-explant and is beyond the scope of this work.

Conclusions

In this study, we evaluated a test protocol to evaluate the mechanics of ex vivo arteries with PAD. We performed pressure inflation testing on peripheral arteries from patients who had been diagnosed with peripheral artery disease and imaged the vessel using IVUS-VH to identify plaque constituents. Plaque constituents were analyzed with respect to artery deformation to identify which plaque components hinder wall expansion under pressure. We found that fibrous plaque has the greatest negative effect on both lumen area and major axis of the vessel while necrotic core also had a significant negative effect; high proportions decreased vessel compliance, while clustering of the component increased vessel compliance. Calcification had a significant negative effect on wall deformation and clustered calcification also had a negative effect. The latter was unexpected as clustered calcification will result in regions of low compliance. Therefore we theorize that the distribution of plaque plays as large a role as the proportion of plaque on wall deformation and that the presence of components associated with more developed atherosclerotic plaque such as fibrous plaque are of equal, if not more, importance as less compliant components such as calcification.