Introduction
Bioresorbable vascular scaffolds (BRS) are a novel approach to the interventional treatment of coronary artery disease (CAD), providing short-term vascular scaffolding combined with drug-delivery capability. They may offer potential advantages compared to metallic drug-eluting stents (e.g. adaptive remodeling, restoration of vasomotion and late luminal enlargement). The so called 4th revolution in coronary artery disease revascularization steered extensive scientific research in BRS developments [
1‐
3].
It has been shown that the designs and materials of BRS platforms—either metallic or polymeric—influence the resorption process [
3‐
5]. Considering the variety of possible platforms, it is necessary to establish tools capable of monitoring the degradation process and its correlated mechanical characteristics.
Intravascular ultrasound-derived parameters have shown to be useful to assess the BRS resorption of metallic and polymeric scaffolds in humans [
6‐
8]. One of the most studied intravascular ultrasound (IVUS) techniques to evaluate the resorption process is called differential echogenicity [
8,
9]. This method consists in an automated and quantitative three-dimensional analysis of coronary tissue components scored for echogenicity using as reference the mean level of the adventitia brightness [
9] where scaffold struts appear as bright hyperechogenic structures. In clinical studies, a continuous decrease of echogenicity over time has been shown in regions treated with BRS, being putatively correlated to BRS degradation [
7,
8]. However, in serial human assessments, changes in the adventitia and plaque-media compartment of the treated regions during the follow-up period could possibly affect these interpretations [
10‐
14].
The objectives of the current study were: (1) to describe a novel method of echogenicity for tissue analysis; (2) to evaluate its reproducibility; and (3) to assess its aptitude to assess the BRS degradation process through a direct correlation with the molecular weight (Mw) in a preclinical model using a drug-eluting poly-l-lactide-acid (PLLA) bioresorbable scaffold (Absorb BVS, Abbott Vascular, Santa Clara, California).
Discussion
In the present study, using IVUS grey scale derived parameters we attempted to assess the degradation process of the Absorb poly-L-lactide bioresorbable everolimus-eluting scaffold at multiple time points in a porcine model. The major findings of this study can be summarized as follows: (1) hyperechogenic and upperechogenic thresholds had strong and positive correlations with the scaffold molecular weight assessment; (2) the combination of hyper and upperechogenicity could be used as a surrogate for the chromatographic assessment of scaffold molecular weight and (3) echogenicity demonstrated good inter- and intra-observer reproducibility (Supplementary Material).
The present manuscript describes a new software designed to assess the differential echogenicity and, for the first time, ascertained the correlation between IVUS grey scale intensities and quantitative assessment of Mw by GPC. The first novelty is that it was not necessary to use ECG gating and therefore, it is not needed a dedicated IVUS console or post-processing correction. The robustness of this method and the aforementioned good reproducibility demonstrate, for the first time, good correlation of echogenicity with the degradation of the scaffold without being mandatory correction for motion artifacts [
30].
Image resolution can be defined as the capability of making distinguishable the individual parts of an object. Therefore, the use of 40 MHz IVUS catheter in the present study has potential to be more precise to detect scaffold degradation than the previous methodology with the 20 MHz ultrasound [
7,
31]. Ultrasound at a center frequency of 10 MHz has demonstrated to detect decline in the acoustic impedance of PLA when molecular weight varied from 60 to 24 kDa, but further decrease in molecular weight to 15 kDa did not result in discernible change [
32]. In the present study, working with the higher resolution of the 40 MHz IVUS catheter, we were able to detect acoustic differences in 150 m thick samples degrading from ~100 to <4 kDa.
The use of ultrasound to monitor the degradation process of polymers has been initially proposed with a wave pulse-echo method in an in vitro essay [
31]. Wu succeeded to monitor by ultrasound the degradation process of three biodegradable polymers: poly(glycolic acid) (PGA), poly(L-lactic acid) (PLLA) and 50:50 poly(D, L-lactide-co-glycolide) (PDLLG) [
33]. Another IVUS based approach to detect the resorption process in human is virtual histology [
6]. The spectral analysis of the raw backscattered ultrasound misrepresents polymeric struts as dense calcium (DC) and necrotic core (NC). As these parameters are shown to decrease over time after implantation, they have been correlated putatively with resorption [
6,
16,
34,
35].
Previously, echogenicity has been used to assess paired serial acoustic properties of coronary plaques in BRS-treated segments in the clinical setting [
7,
8,
35]. It has been shown that these segments had an increase in hyperechogenic tissue after implantation which decreased over time [
7,
8,
35]. The aforementioned methodology succeeds to document the progressive decrease of high intensity grey level tissues in both metallic and polymeric BRS [
8,
35].
However, until now, the link between echogenicity and the degradation process has been hypothetically assumed. The pending question was whether temporal plaque changes could interfere with the multistage degradation of the polymer and confound the echogenicity analysis. It has been shown that coronary atherosclerosis is a dynamic phenomenon and numerous factors can influence the atherosclerotic changes as detected by IVUS-derived parameters. For instance, statin treatment may reduce the percentage lipid volume index over time [
13] and may increase the calcified plaque component [
11]. Additionally, there is a significant decrease in NC (16 %) and DC (30 %) content in coronary plaque located behind the struts of the everolimus-eluting bioresorbable vascular scaffold [
36]. All the above-mentioned confounding factors might influence the acoustic properties in the lumen-vessel compartment and hinder the clinical relevance of echogenicity for BRS degradation assessment.
As we have used a porcine non-atherosclerotic model, we did not have the confounding presence of coronary artery disease, thus enabling the evaluation of Poly-L-Lactide’s echogenic characteristics over time. Hyperechogenic, upperechogenenic and hyper + upperechogenic tissues had strong and positive correlations between echogenicity and the degradation process. Echogenicity is determined by the difference in acoustic impedance between two mediums, which is proportional to density and acoustic velocity. The acoustic velocity is proportional to the square root of the stiffness (bulk and shear moduli). Many factors impact the stiffness of PLA, including molecular weight, polydispersity, crystallinity, orientation of crystalline microstructure, and other environmental conditions [
37]. As a result, one would expect to change the impedance of PLA as it degrades and molecular weight to have a generalized relationship to this decline.
Qualitatively, the correlation was however not perfectly linear. For instance, at 1-month the combination hyper + upper tended (without statistical significance; Table 6, supplementary material), in average, to be lower than at 3-months whereas the molecular weight had a continuous decrease in the same period. From the ultrasonic point of view, the significantly higher neointimal hyperplasia (Fig.
2) at 1-month might have affected the ultrasound penetration and therefore the echogenicity interpretation. Additionally, the scaffold-vessel hyper + upperechogenicity at 30-months was numerically comparable to that at 18-month. However, the degradation process may be influenced by individual biological factors and it has to be emphasized that these assessments were not serial. However, we showed a consistent individual positive correlation between the molecular weight and echogenicity (Figs.
4 and
5).
Limitations
Arteries used for molecular weight assessment could not be evaluated histologically. Therefore, changes in the observed echogenicity (both lumen-scaffold and scaffold-vessel) could not be related to the histologic changes over time [
19,
38]. As this study has been performed in a non-atherosclerotic model, it should be acknowledged that the rate of degradation has not been confirmed in atherosclerotic coronary arteries. However, as the degradation of PLLA is a hydrolytically driven and not enzymatically driven process, it is expected that the rates would be largely equivalent.
We could not test the reproducibility of the echogenicity IVUS findings in the two different consoles. However as we worked at the same ultrasound frequency (40 MHz) and the tissue classifications were normalized by the individual adventitia grey scale intensity we could show a robust correlation between scaffold degradation and high echogenic parameters. It has been shown that the comparison with the adventitia allows for normalization with respect to transducer variability, gain settings and across populations [
21].
The changes in vessel, lumen, scaffold and neointima volumes over time are in line with the serial IVUS findings in the pre-clinical model and clinical setting showing progressive increase in vessel, lumen area and scaffold area [
16,
18,
39,
40]. However, in the porcine model the somatic growth can influence our findings [
39]. As we do not have the IVUS at baseline we could not normalize these geometrical changes for the increase in the reference vessel size. Nevertheless, this information has been described in the literature and are beyond the main scope of the current manuscript.
Acknowledgments
The authors acknowledge Dave Pinson and Katherine Fu (Abbott Vascular) for their technical contributions in GPC analysis. The authors also acknowledge Michael Frie and the staff of American Preclinical Services (Minneapolis, MN) for their care and attention in successful completion of the in life phase of this study. The present study was sponsored by Abbott Vascular, Santa Clara, California.