Histopathologic Characterization of Chronic Total Occlusions by Directional Atherectomy: The HIPACT Study

Background. Chronic total occlusions (CTOs) and long lesions have been associated with higher reocclusion rates in femoropopliteal arteries and increased need for revascularization. While several studies have analyzed atherectomy samples, no study to date has correlated the tissue characteristics of CTOs with clinical outcomes. This pilot study assessed lesions in order to predict clinical outcomes based on lesion characteristics. Methods. Patients presenting with femoropopliteal (FP) CTO lesions, including in-stent restenosis and de novo lesions, were enrolled in a prospective, observational study. With patient consent, CTOs were crossed using a crossing catheter guided by optical coherence tomography (OCT) imaging. Atherectomy was performed with a directional atherectomy device and tissue samples were collected and subjected to histopathological analysis for the presence of adventitial tissue and thrombus, and the amounts of hypercellular cells, fibrocellular material, fibrous tissue, and lipid-rich tissue in the excised tissue were measured. The compiled data were correlated with clinical outcomes, as recorded at each patient’s clinical follow-up visit. Results. All CTO lesions (n = 19) were successfully crossed, with no dissections or perforations. Adventitial tissue was found in excised tissue from all 19 lesions (up to 57% of total lesion area), and thrombus was found in 15 lesions (up to 34% of total lesion area). The amount of hypercellular cells, fibrocellular material, fibrous tissue, and lipid-rich tissue present in the excised tissue did not correlate with the incidence of target-lesion revascularization (TLR). At 6-month follow-up exam, 79% of the treated lesions had TLR. Risk of TLR was assessed based on weighted risk of each variable; the results determined that occurrence of TLR was associated with elevated levels of adventitia and thrombus in the lesions and with lesions >15 cm in length. There was a significant correlation (P<.05) between TLR and the lesion characterization as set forth in the present study. Conclusions. Pairing the histological analysis, including content of adventitia and thrombus, with lesion length and binary clinical outcomes enabled the predictive incidence of TLR in this pilot study. Further work needs to be conducted to validate these findings in larger studies.

J INVASIVE CARDIOL 2021;33(6):E443-E450. Epub 2021 May 17.

Key words: critical limb ischemia, femoropopliteal arteries, peripheral arterial disease, popliteal arteries

Critical limb ischemia (CLI) represents advanced peripheral arterial disease (PAD), with a quarter million patients in the developed world requiring amputation within 6 months of diagnosis, and up to half resulting in death after amputation within 5 years.^1-3 Most lesions causing CLI are chronic total occlusions (CTOs), which can be difficult to treat and often lead to low revascularization rates.^4-6 CTOs of the femoropopliteal (FP) arteries are a subset of PAD patients most amenable to percutaneous revascularization, but nonetheless remain challenging both with respect to initial recanalization procedures and long-term patency following the procedure.^7-10 Several factors determine the suitability and potential success of percutaneous revascularization, including lesion length, vessel calcification, duration of occlusion, and CTO anatomy. However, even with knowledge of these factors, long-term patency cannot yet be predicted reliably.

Imaging of occluded vessels has previously been performed using extravascular techniques, including fluoroscopy, duplex ultrasound imaging, computed tomography, and magnetic resonance imaging to determine lesion characteristics. However, these techniques do not provide information regarding plaque morphology or composition. Advancements in imaging with the introduction of intravascular ultrasound (IVUS) devices allow further characterization of plaque, albeit limited in some cases by its moderate level of resolution. More recently, optical coherence tomography (OCT), which has very high levels of resolution, has gained traction to assess lesions in peripheral vessels.^11,12 Although useful for in vivo arterial assessment of luminal dimensions, plaque distribution and morphology, presence of aneurysmal disease, and stent malapposition, OCT imaging is used primarily for plaque definition of lesions and for guiding devices in their crossing of stenotic lesions. In addition, OCT has been utilized to predict long-term patency of vessels following treatment based on visual cues of lesion morphology.^13,14 The incorporation of OCT imaging into treatment modalities, such as atherectomy, to assist in directing plaque resection in PAD is a recent innovation to assist procedural strategies. In addition, assessing histopathologic plaque characteristics may aid in further delineating lesion characteristics associated with high restenosis rates, where data on plaque characteristics remain scant.

This pilot phase of the HIPACT (Histopathologic Characterization of Chronic Total Occlusions Using Directional Atherectomy) study was conducted to assess the value of incorporating histopathologic specimen analysis in reviewing lesion morphology and predicting restenosis rates by correlating lesion characteristics with clinical outcomes.

The pilot phase of the HIPACT study was a prospective assessment of histologic samples obtained from patients who underwent atherectomy of obstructive lesions in FP arteries. Patients were eligible for the study if they had chronic PAD (Rutherford-Becker categories 2–5) and occlusions within a FP segment (Figure 1). Patients were ineligible if they were taking oral anticoagulants, had a history of contrast allergy, or had undergone attempted treatment or atherectomy of the target CTO previously. Patients returned to the clinic within 6 months after the procedure for symptom assessment, ankle-brachial index evaluation, and documentation of any repeat procedures since treatment.

All eligible subjects provided consent for the revascularization procedure including atherectomy, but also provided informed consent to participate in the HIPACT study and permit use of their tissue in histopathologic analyses. The study protocol was approved by the institutional review board of Deborah Heart and Lung Center. Subjects were enrolled between August 2012 and April 2014.

The patients were prepared for surgery following standard procedures. The target lesion was identified by angiography and its length was documented. With the lesion identified, the occlusion was traversed with an Ocelot crossing catheter (Avinger) (Figure 2) and then subsequently treated with TurboHawk directional atherectomy (Medtronic). Excised tissue was captured in the nosecone of the atherectomy device and collected at the end of the procedure, fixed in 10% neutral buffered formalin following standard methodology, and sent to the pathology lab.

Upon receipt, the tissue samples were weighed prior to embedding in paraffin. For histologic assessment, the samples were sectioned (4 µm thick), transferred to glass slides, and stained with hematoxylin and eosin, Masson’s trichrome, and elastic van Gieson (EVG) stains. The slides were scanned and converted to digital images using a slide scanner (ScanScope CS System). The EVG-stained slides were used to draw boundaries around the different tissue components. The total area of the excised tissue and the amounts of tissue containing adventitial (full thickness resection with adventitial tissue), medial (presence of media with or without fragments of the internal elastic lamina and plaque, but without external elastic lamina or adventitial tissue), and intimal tissue (presence of atherosclerotic plaque and intima but without medial or adventitial components) were measured as percentages of the total tissue area with ScanScope analysis tools. The samples were also examined for the presence of thrombus, with its area measured, as well as calcium within the tissue.

Histological analysis incorporated identification of the extent of hypercellular fibrocellular, fibrous, and lipid-rich tissues present in the samples. The extent of these tissue components in the samples were ranked as 0 (none present), 1 (minor, occupying <10% of the total area), 2 (moderate, occupying 10%-50% of the total area), or 3 (major, occupying >50% of the total area). Among all variables, risk of TLR was assessed based on weighted risk of each variable, which allowed the correlation of outcomes based on these variables.

Statistical analysis. Continuous data are expressed as mean ± standard deviation and categorical data as absolute numbers or percentages. Measurements of the target and reference sites and intraobserver differences were analyzed with Student’s t-test for paired observations. Differences with P<.05 were considered statistically significant. Statistical analyses were conducted using Microsoft Excel, version 15.27 (Microsoft Corporation) and GraphPad Prism 7 (GraphPad Software).

A total of 19 patients participated in this pilot phase of the HIPACT study; they were predominantly male (67%) with a mean age of 72 ± 11.6 years. The CTO lesions were primarily (67%) in the superficial femoral artery (SFA), with 4 in a popliteal artery, and 1 each in the tibial peroneal trunk and the common iliac artery. All obstructions were crossed successfully by the crossing catheter with no dissections or perforations. Directional atherectomy reduced the obstructions such that blood flow was restored through the area of the obstruction, as determined by angiography. The amount of fluoroscopy time varied with the device used, with less than a minute of time (14 seconds) needed when the OCT-guided catheter was advanced through the obstruction to prepare a path for the atherectomy device (Figure 3).

Lesion characteristics and histologic assessments are provided in Table 1. Lesion lengths ranged from 8-36 cm (mean, 15.7 ± 7.3 cm) and resected tissue weighed 27.9-310 mg (mean, 104.5 ± 76 mg) with a mean total tissue area of 65.62 ± 41.3 mm². Ten of the lesions were de novo and 9 were in-stent restenosis (ISR) in origin. The de novo lesions were shorter (mean length, 15.4 ± 6.5 cm vs 16.4 ± 9.4 cm; P=.10) and bulkier (mean tissue weight, 105.9 ± 73.2 mg vs 102.2 ± 68.3 mg; P=.90) in comparison with the ISR neointimal hyperplastic tissue.

Notably, all samples of tissue from the atherectomy procedure contained adventitial tissue, ranging from 2%-57% of total area of the tissue excised (Figure 4). Between the 2 groups, the neointimal hyperplastic material excised from the ISR obstructions was more intimal (43.2% vs 88%; P<.01), less adventitial (4.7% vs 24.3%; P<.01), and less medial (6.3% vs 32.4%; P<.01) than tissue resected from de novo lesions. There was no significant difference in the mean total tissue area resected between the 2 groups (65.8 ± 43.2 mm² for de novo lesions vs 65.3 ± 57.1 mm² for ISR material).

Histologic characterization determined that the majority of tissue from both the de novo lesions and ISR neointimal hyperplastic tissue was neither hypercellular nor fibrocellular in nature; however, by a slight margin, the ISR tissue was ranked to have more fibrous material than present in the de novo lesions (Table 1). Neither the ISR nor the de novo tissue samples contained significant lipid pools, suggesting that the lesions were less likely to be related to plaque rupture. However, 12.2% of the de novo lesions and 17.9% of the ISR material consisted of thrombus (P=.04).

At 6-month follow-up assessments, TLR had occurred in 79% of the 19 patients, with a TLR rate of 70% in the de novo group and 89% in the ISR group. Risk of TLR was found to be associated with 3 variables: (1) the extent of adventitia present in the resection tissue; (2) the presence of thrombus; and (3) the lesion length (Table 2).

To our knowledge, this is the first study to assess the potential for TLR based on histologic analyses. The utilization of intravascular imaging aided in sufficient targeted plaque removal and, based on size of the histologic samples collected, was fairly efficient in both the de novo and ISR lesion groups. Postprocedural intraluminal imaging has documented that directional atherectomy under the guidance of digital angiographic imaging can result in disruption of the intimal medial layers as well as inappropriate medial and adventitial resection. Such resulting injuries to the vessel wall have been correlated with increased restenosis rates.¹⁵ Although atherectomy under IVUS guidance can mitigate the layered structure disruption and provide larger luminal gains, use of IVUS imaging appears to be insufficient in reducing restenosis rates,¹⁶ possibly because the atherectomy proceeds under angiographic guidance rather than intravascular imaging guidance.

Between the 2 groups in this pilot study, there was a higher restenosis rate in the ISR lesions (89%) than in de novo lesions (70%). This has been noted in previous studies and supports the credibility of these outcomes, which are similar to the patient populations in larger studies.^17,18 As expected, longer lesion length (>15 cm) correlated with the rate of restenosis, which has been proven before as well.¹⁹ Furthermore, the amount of tissue excision was very similar in the 2 groups and slightly, although insignificantly, greater in the de novo group. This may be due to higher intraluminal definition of plaque volume and better targeted atherectomy, which allowed for both groups to be equally identified and excised. Based on previous studies using intravascular imaging guidance, this has likely provided greater intraluminal diameter compared with atherectomy without intravascular imaging.¹⁵

This study was able to identify 2 histologic characteristics, in addition to lesion length, that would aid in risk stratifying TLR in patients with both de novo and ISR lesions. Both of these factors may be related to worsening intravascular inflammation. Adventitial resection is more likely to result in a shorter time between atherectomy and restenosis of the targeted vessel.²⁰ This has been attributed to disruption of the extra elastic lamina, resulting in pronounced inflammation and a more exaggerated rate of collagen deposition and development of restenosis.²¹ Previous studies using IVUS guidance have correlated histopathologic specimen adventitial volume with external elastic lamina (EEL) disruption and risk of restenosis.²² Evidence suggests that image-based identification and treatment with stenting reduces this phenomenon.²³

In this study, the catheter used to cross the CTO has the capability of providing intravascular imaging via OCT on its cannula. This resulted in successful crossing in 100% of CTOs, and it was possible to maintain the cannula of the device within the center of the lumen of the vessel by monitoring the advancement of the catheter with its OCT imaging. Trauma to the vessel wall during CTO crossing has been associated with inflammatory cascades and can serve as a nidus for restenosis. In addition, the atherectomy device used in this study is guided by and monitored with angiography, which makes it a challenge to determine how deeply one is cutting into the layered structures of a peripheral artery during atherectomy. Use of an OCT-guided atherectomy catheter has been noted to result in low TLR rates at 6 months.

This study is the first to correlate intravascular thrombus with further risk of restenosis. To our knowledge, this has not been previously identified in studies using histopathologic or intravascular-imaging based studies.^24,25,26 There are 2 potential reasons for this effect. First, it is possible that this represents an area of endothelial injury that displaces intracellular chemokines with resultant intravascular recruitment of macrophages and enhanced collagen deposition. Previous studies assessing intravascular healing after stenting have alluded to this phenomenon.²⁷ Second, it is possible that intravascular thrombosis represents areas of enhanced angiogenesis, which may independently increase restenosis by virtue of increased neointimal formation and enhanced endothelialization.²⁸ Unfortunately, we did not have entire cross-sectional vessel histopathology to consolidate this hypothesis. Although serial intravascular assessment at the areas of thrombosis would confirm “restenosis points,” this was not logistically feasible.

Study limitations. This study is limited by its small sample size and limited tissue sampling, since this was our first experience with histopathologic assessment, but it was a pilot phase of a study that will build on what was determined with the analyses. To further consolidate our findings, a more detailed assessment in a larger patient population is needed. In addition, with the advancement of OCT imaging capacity being incorporated into directional atherectomy catheters, a more refined and guided intravascular imaging system may be a better option for future studies.²⁹

This proposed TLR risk model utilizing lesion length, extent of adventitial resection, and in situ thrombosis burden is the first to combine conventional risk assessment with intravascular imaging and histopathologic assessment. We believe that such lesion characterization with intravascular imaging may allow the physician to make a real-time intraprocedural risk assessment and aid in guiding therapy that could reduce the potential for restenosis. In addition to stenting and drug-coated balloon angioplasty, which have conventionally been the options for treatment of endothelial damage, the finding of thrombus formation lends credence to the use of more aggressive antiplatelet and antithrombotic therapy post procedure in these patients. Such a regime may reduce platelet recruitment and further inflammation, which are precursors to restenosis. Further studies with such histopathologic risk assessment may be necessary to validate its utility in planning the treatment strategy for complex PAD.