Between November 2017 and March 2019, five patients (≥18 years old) with severe internal carotid artery (ICA) stenosis (70–90% diameter reduction) detected by digital subtraction angiography (DSA) and previous cervical radiotherapy were enrolled for pre-interventional OCT evaluation. Immediately after stenting, 4 patients underwent OCT evaluation. Since the patient experienced vasospasm and the blockage of the cerebral protection device, 1 patient did not receive post-interventional OCT evaluation. Only two patients underwent 6- or 13-month follow-up OCT evaluation since the other three patients refused another OCT evaluation due to the relatively expensive fee and the fear of invasive examination. Patients were from Jinling hospital in Nanjing, China. The study protocol was approved by the hospital’s ethics committee. Written informed consents were obtained from all patients. The relevant clinical and radiologic data were reviewed.

The intravascular frequency-domain OCT imaging system (ILUMIEN OPTIS, St. Jude Medical, Massachusetts, United States) and 2.7F OCT imaging catheters (Dragonfly Duo, Abbott Medical, Massachusetts, United States) were used for OCT image acquisition. With an 8F guide catheter placed in the common carotid artery (CCA), a cerebral protection device [FilterWire (Boston Scientific, Massachusetts, United States), Emboshield NAV6 (Abbott Vascular, California, United States), or ANGIOGUARD (Cordis Corporation, Florida, United States)] was advanced through the guide catheter and positioned in the ICA C1 segment. An OCT imaging catheter was then carefully advanced over the guidewire of the cerebral protection device and navigated through the carotid artery lesion. During the blood clearance by the automatic injection of 20 ml of undiluted iodixanol 320 (GE Healthcare Ireland Limited, County Cork, Ireland) with a velocity of 10 ml/s through the 8F guide catheter, the light mirror of the OCT imaging catheter was helically pulled back (18 or 36 mm/s) to get a series of cross-sectional OCT images of the vessel wall (Jones et al., 2014; Liu et al., 2015; Shi et al., 2020). OCT image acquisition was repeated immediately after stenting and during the follow-up DSA examination. The OCT imaging catheter was introduced into the ICA over a 0.014-inch microwire during the follow-up DSA examination.

Optical coherence tomography images were analyzed independently by two investigators (XX and FH) with extensive experience in reviewing OCT images. OCT images within 10 mm distal and proximal to the minimum lumen were assessed. The image was considered non-analyzable if the assessment of a continuous 270° arc was impaired by the intraluminal blood (Jones et al., 2014). Before quantitative measurement, manual OCT calibration was performed along with the entire pullback. Cross-sectional OCT images were analyzed at 0.1- or 0.2-mm intervals. The rates of tissue protrusion and stent strut malapposition were analyzed at 1-mm intervals (de Donato et al., 2013; Liu et al., 2015). The rates of tissue protrusion were calculated in the slice-based analysis, and the rates of stent strut malapposition were in the strut-based analysis.

Qualitative and quantitative OCT evaluations were performed precisely based on previously published criteria (Prati et al., 2010; Tearney et al., 2012). Lipid-rich plaque was defined as plaque with lipid content occupying more than one quadrant (Tearney et al., 2012). The calcific nodule was defined as single or multiple regions of calcium that protruded into the lumen (Tearney et al., 2012). A thin fibrous cap was present if the fibrous cap thickness was less than 65 μm (Tearney et al., 2012). Plaque rupture was defined as the discontinuity of the fibrous cap and the cavity formation, with or without a superimposed thrombus (Prati et al., 2010; Tearney et al., 2012). A thrombus was defined as a mass (≥250 μm) attached to the luminal surface or floating within the lumen (Jang et al., 2005; Tearney et al., 2012). Microvessels or neovascularization was defined as signal-voiding tubular structures (50–300 μm) present on at least three consecutive cross-sectional frames (Takano et al., 2009; Prati et al., 2010; Li et al., 2020). Dissection was defined as the presence of an intimomedial flap producing a double-lumen or an intramural hematoma formation (Alfonso et al., 2012). Intramural hematoma can be defined as a relatively homogeneous signal-rich material with variable attenuation separating the intima from the outer vessel wall (Alfonso et al., 2012). Tissue protrusion was defined as the prolapse of tissue into the lumen between adjacent stent struts (Tearney et al., 2012). It was divided into 3 groups, namely, smooth protrusion, irregular protrusion, and protrusion with attenuation (Funatsu et al., 2020). The distance from the surface of the stent strut to the lumen contour was measured. A stent strut was classified as “malapposed” (distance > 200 μm), “well apposed” (distance 10–200 μm), or “embedded” (distance < 10 μm) (de Donato et al., 2013; Liu et al., 2015). Follow-up OCT image with the minimum lumen was selected to measure the neointima burden. It was calculated as (stent area – lumen area)/stent area × 100% (Tearney et al., 2012).

The CAS procedure was performed through the femoral approach through guide catheters. All patients underwent systemic anticoagulation with 4,000 IU unfractionated heparin after femoral artery puncture and an additional 2,000 IU/h. Pre-dilatation of the carotid lesion with a 4.0–5.0-mm balloon catheter was performed after OCT image acquisition. Stenting was then performed with one of the following open-cell stents: Acculink (Abbott Vascular, California, United States), Precise (Cordis, Florida, United States), and Protégé (Medtronic, Minnesota, United States). The selection of the stent diameter and length was decided by our experienced interventional neurologists according to the lesion feature. Post-dilatation was performed with a 5.0–6.0-mm balloon catheter when the remaining stenosis was > 30%.

The degree of diameter stenosis, residual stenosis, and in-stent stenosis was calculated accurately according to the North American Symptomatic Carotid Endarterectomy Trial criteria (Barnett et al., 1998). Technical success was defined as the residual stenosis < 30% on DSA. In-stent restenosis was defined as ≥ 50% stenosis within or at the edge of the stent. Angiograms and OCT images used co-registration on the basis of the landmarks, such as the bifurcation of the CCA and the stent edge, to ensure that they were at identical sites.

A total of 5 male patients [mean (standard deviation) age, 64.2 (8.5) years] with previous cervical radiotherapy underwent OCT evaluation of carotid artery stenosis successfully (Table 1). The tumor types indicating cervical radiotherapy were nasopharyngeal carcinoma (NPC) (n = 3, 60%), cervical esophageal carcinoma (n = 1, 20%), and cervical lymphoma (n = 1, 20%). The radiation dose was 70 Gy in three patients. The median interval from the radiotherapy to the diagnosis of RICS was 8 years (range 4–36 years). Interestingly, all patients had not more than one risk factor for atherosclerosis but 60% of them had severe bilateral carotid disease (70–100% diameter reduction). Except for one patient diagnosed with severe right coronary artery stenosis, all patients were free of coronary artery disease and peripheral artery disease.

As for the evaluated carotid artery lesion, the mean degree of diameter stenosis was 76% (range 70–90%) on DSA and the mean minimum lumen area was 3.10 mm² (range 1.57–5.23 mm²) on OCT (Table 2). The morphological characteristics of RICS in three NPC patients were different (Cases 1–3). In the NPC patient with a 4-year interval from radiotherapy, heterogeneous signal-rich tissue was observed in consecutive 30 frames (6 mm) and occupied almost three-quarters of the vessel at the minimum lumen (Case 1). Atypical lesion such as dissection in the ICA C1 segment was detected in the other two NPC patients with longer time intervals (Cases 2 and 3). In the NPC patient with an 8-year interval, multiple cavity formations were present besides the dissection (Case 2). In the NPC patient at a 36-year interval, advanced atherosclerosis such as ruptured lipid-rich plaque, ruptured calcific nodule, and thrombosis coexisted with the dissection (Case 3). The morphological characteristics of RICS in patients with esophageal carcinoma or lymphoma were similar to atherosclerosis (Cases 4 and 5). Lipid-rich plaque was present in both cases. Neovascularization and thrombosis were revealed in the patient with longer time intervals (Case 5).

All patients received balloon pre-dilatation and carotid stenting. The rate of technical success was 100% on DSA. Four patients underwent OCT examination immediately after stenting to evaluate the stent–vessel relationship in RICS patients (Cases 1, 2, 4, and 5). A total of 67 cross-sectional OCT images [mean (standard deviation), 17 (5)] were analyzed to assess the rates of tissue protrusion and stent strut malapposition. The mean rates of tissue protrusion and stent strut malapposition were 34.9% (range 18.2–57.1%) and 7.0% (range 3.3–13.8%), respectively. Besides, there were the highest rates of tissue protrusion and stent strut malapposition in Case 4. Most of the tissue protrusion was smooth tissue protrusion. Small dissection may appear sometimes (Cases 2 and 4). Two patients underwent follow-up OCT examination to evaluate the pattern of neointimal hyperplasia in RICS patients (Cases 3 and 4). There were homogeneous signal-rich neointima and signal-poor regions around stent struts in both follow-up OCT images. The neointima burden of Cases 3 and 4 was 31.6 and 48.9%, respectively. Microvessels were observed in the thicker 6-month neointima of Case 4. Detailed data are summarized in Table 2.

The patient presented with a history of NPC 4 years had been treated with cervical radiotherapy (70 Gy) and chemotherapy (cisplatin, docetaxel, and bevacizumab). The patient presented with dizziness for 2 months. DSA showed 70% of stenosis at the left ICA (LICA) sinus (Figure 1A). The patient underwent balloon pre-dilatation (5 × 30 mm) and stent implantation (Acculink, 7–10 × 40 mm). The residual stenosis was 20% (Figure 1B). The 4-month follow-up DSA detected that there was no in-stent restenosis (Figure 1C). Pre-interventional OCT examination of the distal lesion with mild stenosis revealed macrophage accumulations at 2 o’clock, and there was no clear three-layered vessel structure (Figure 1D). More proximally, nearly half of the vessel was occupied by a heterogeneous signal-rich tissue located close to the luminal surface (Figure 1E). There was a clear demarcation between the tissue and the outer fibrous tissue. Moreover, there was a microvessel at 12 o’clock (Figure 1E). Nearly three-quarters of the vessel was occupied by the heterogeneous tissue at the minimum lumen, and there was irregular signal-poor tissue accumulating at 5–7 o’clock (Figures 1F,G). At the proximal site of the lesion, there were multiple layers of heterogeneous signal-rich tissue, forming the onion-like structure (Figure 1H). The post-interventional OCT evaluation revealed excellent stent strut apposition and smooth tissue protrusion at the image with the minimum lumen (Figure 1I).

The patient presented with a history of NPC 8 years had been treated with cervical radiotherapy (70 Gy) and chemotherapy. The patient presented with paroxysmal vertigo for 3 days. Magnetic resonance imaging (MRI) revealed acute cerebral infarction in the right frontal lobe, occipital lobe, and semi-oval center. DSA showed 80% stenosis with an irregular surface at the LICA C1 segment (Figure 2A) and occlusion at the left external carotid artery origin. The patient underwent balloon pre-dilatation (4 × 30 mm) and stent implantation (Precise, 8 × 40 mm) at the LICA lesion. The residual stenosis was 20% (Figure 2B). Pre-interventional OCT examination of the distal lesion revealed that the crescent-shaped material separated the intima from the outer vessel wall at 5–9 o’clock (Figure 2C). More proximally, nearly half of the vessel was occupied by the relatively homogeneous signal-rich material with variable attenuation that separated the intima from the outer vessel wall, indicating intramural hematoma (Figure 2D). The intramural hematoma expanded to almost three-quarters of the vessel at the minimum lumen (Figure 2E). In addition, the intimal tear at 5 o’clock, intimomedial flap floating, and double-lumen formation were well visualized by OCT (Figures 2F,G), demonstrating dissection. Multiple cavity formations and macrophage accumulations were detected at the proximal normal-appearing vessel (Figure 2H). The post-interventional OCT evaluation disclosed excellent stent strut apposition, smooth tissue protrusion, and a residual small dissection (Figures 2I–K). The dissection corresponded to the intimal tear mentioned above (Figure 2J).

The patient presented with a history of NPC of 36 years had been treated with cervical radiotherapy. The patient presented with recurrent weakness and numbness of the right limb, and slurred speech for 1 month. MRI revealed acute cerebral infarction in the bilateral frontal and parietal lobes. DSA showed 90% of stenosis with an irregular surface at the LICA C1 segment (Figure 3A) and occlusion at both the left external carotid artery and the right CCA (RCCA) origin. Balloon pre-dilatation (4 × 30 mm), stent implantation (Acculink, 9 × 40 mm), and balloon post-dilatation (5 × 20 mm) were performed at the LICA lesion. The residual stenosis was 20% (Figure 3B). The 13-month follow-up DSA showed no in-stent restenosis (Figure 3C). Pre-interventional OCT examination of the lesion revealed that the crescent-shaped material separated the intima and the outer vessel wall at 9–1 o’clock, indicating intramural hematoma (Figure 3D). In addition, intraluminal thrombus was observed and shadowed the underlying vessel wall (Figure 3D). At the minimum lumen, there was a ruptured lipid-rich plaque and a ruptured calcific nodule with an overlying thrombus next to the intramural hematoma (Figures 3E,F). More proximally, the double-lumen with the fenestration communicating the true lumen and the false lumen was identified as the sign of dissection (Figures 3G,H). In addition, the calcification with overlying thrombus (Figures 3G,H) and the ruptured calcific nodule mentioned above were consecutive. The length of the calcification reached 6.8 mm. Cholesterol crystals and cavity formations due to intimal disruption were frequently detected at the proximal site of the lesion (Figure 3I). The 13-month follow-up OCT examination revealed mild in-stent neointimal hyperplasia (Figures 3J–L). The neointima was homogeneous signal-rich, and there were signal-poor regions near the stent struts (Figures 3J,K). Some struts in the cavity mentioned above were not covered by the neointima (Figure 3L).

The patient presented with a history of cervical esophageal carcinoma 4 years had been treated with cervical radiotherapy (70 Gy) and chemotherapy (nedaplatin, docetaxel, and raltitrexed). The patient presented with slurred speech and paroxysmal unconsciousness for 3 days. MRI revealed acute cerebral infarction in the left frontal and temporal lobes. DSA showed a long lesion with 70% stenosis at both the right ICA (RICA) sinus and C1 segment (Figure 4A) and occlusion at the LICA sinus. The patient underwent balloon pre-dilatation (5 × 30 mm) and stent implantation (Protégé, 8 × 60 mm) at the RICA lesion. The residual stenosis was 20% (Figure 4B). The 6-month follow-up DSA showed no in-stent restenosis (Figure 4C). Pre-interventional OCT examination of the lesion disclosed focal signal-poor tissue in the fibrous tissue (Figure 4D). More proximally, there was a lipid-rich plaque with a thick fibrous cap at 6–10 o’clock (Figure 4E). At the minimum lumen, the intima thickened with mainly fibrous tissue and linear signal-rich cholesterol crystals were detected at 6 o’clock (Figure 4F). The cholesterol crystals were detected at 10 adjacent cross sections. The post-interventional OCT evaluation disclosed the small dissection, smooth tissue protrusion, and irregular tissue protrusion at the stented segment (Figures 4G–I). Besides, its rates of tissue protrusion and stent strut malapposition were the biggest among the four patients. The 6-month follow-up OCT examination revealed evident in-stent neointimal hyperplasia, signal-poor regions around stent struts, and neovascularization (Figures 4J–L). Stent struts were all covered by the neointima, and the small dissection was healed (Figure 4J). There were some microvessels near the signal-poor regions (Figure 4K). The neointima with the minimum lumen was fibrotic and homogeneous signal-rich (Figure 4L).

The patient presented with a history of cervical lymphoma for 15 years had been treated with cervical radiotherapy and chemotherapy. The patient presented with recurrent vision loss in the right eye for more than 1 year and paroxysmal vertigo for 2 months. DSA showed 70% stenosis at the bifurcation of the left CCA (LCCA) (Figure 5A) and occlusion at the RCCA origin. Balloon pre-dilatation (4 × 30 mm), stent implantation (Precise, 8 × 40 mm), and balloon post-dilatation (6 × 30 mm) were performed at the LCCA lesion. The residual stenosis was 20% (Figure 5B). Pre-interventional OCT examination of the lesion revealed the intraluminal thrombus floating at 12–3 o’clock (Figure 5C). There was a fibrous plaque with the intraluminal thrombus at the minimum lumen (Figure 5D). A lipid-rich plaque with macrophage accumulations was visualized at the proximal site of the lesion (Figures 5E,F). In addition, there was a microvessel connecting with the edge of the lipid plaque. More proximally, the lipid plaque with the neovascularization still existed (Figure 5G). The length of the lipid plaque reached 5.5 mm. The post-interventional OCT evaluation disclosed fine stent strut apposition, and some of the vessel walls were out of the imaging range (Figure 5H).