Cholesterol crystal-induced coronary inflammation: Insights from optical coherence tomography and pericoronary adipose tissue computed tomography attenuation

Vascular inflammation is a key driver of atherogenesis and atherothrombosis.¹ Free cholesterol accumulation within the arterial wall formulates cholesterol crystals, which incite a local inflammatory response.¹ Intracoronary optical coherence tomography (OCT) enables in vivo visualization of cholesterol crystals, shown to associate with features of plaque vulnerability.² Recent evidence demonstrates that coronary inflammation propagates “inside-out” to pericoronary adipose tissue (PCAT), with proinflammatory cytokines inhibiting adipocyte lipid accumulation.³ This can be detected on coronary computed tomography angiography (CCTA) as an increase in CT attenuation of PCAT surrounding coronary lesions³ and a reference region of the proximal right coronary artery⁴ (RCA). We evaluated the association of cholesterol crystals identified on OCT with PCAT CT attenuation.

We prospectively studied 17 vessels in 13 patients with stable coronary artery disease (CAD) (10 males, mean age 59 ± 8 years) who underwent serial (median interval 373 days; IQR 301–455) invasive coronary angiography, OCT and CCTA within a two-week period at each time point. The study had institutional ethics approval and all patients provided written informed consent. Frequency-domain OCT imaging was performed in a standard fashion (ILUMIEN OPTIS system, St Jude Medical) and analysis performed by a core laboratory. A cholesterol crystal was defined as a thin, linear region of high-signal intensity within a lipid plaque.² CCTA was performed on a 320-detector row CT scanner. OCT-determined lesions were matched to those on CCTA by an independent cardiologist (DW), using anatomical landmarks (e.g. side branches and bifurcations) as fiduciary points. PCAT attenuation was measured around coronary lesions³ (PCAT_lesion), lesion segments³ (proximal, middle and distal segments; PCAT_segment), and the proximal 10–50 mm of the RCA as the most standardized method⁴ (PCAT_RCA). Using semi-automated software (Autoplaque 2.0), an expert reader (AL) blinded to OCT findings contoured the outer coronary wall. PCAT was automatically sampled in 3D layers, moving radially outwards from the vessel wall in 1 mm increments. Adipose tissue was defined as all voxels between −190 and −30 Hounsfield units (HU), and PCAT attenuation was defined as the average CT attenuation of adipose tissue within the volume of interest.³ As the average vessel diameter of all analyzed segments was 3.2 mm, we considered PCAT attenuation within a radial distance of 3 mm from the vessel wall. ΔPCAT was PCAT_lesion at follow-up minus that at baseline. Our institutional intra- and inter-observer agreement for PCAT attenuation is high (intraclass correlation coefficient 0.98 [p < 0.001] and 0.95 [p < 0.001], respectively).

We identified 7 plaques (4 in LAD, 2 in LCx, 1 in RCA) in 5 patients (mean low-density lipoprotein cholesterol [LDL-C] 3.5 ± 0.8 mmol/L, all statin naïve) which contained at least 1 cholesterol crystal on OCT at baseline. There were 10 plaques in 8 patients with no cholesterol crystals. PCAT_lesion was significantly higher around plaques with cholesterol crystals compared to plaques without cholesterol crystals (−77.2 ± 5.0 HU vs. −89.7 ± 8.2 HU; p = 0.019). Similarly, PCAT_segment was higher in lesion segments with (n = 11) versus without (n = 40) cholesterol crystals (−76.5 ± 5.8 HU vs. −87.1 ± 6.2 HU; p = 0.005). PCAT_RCA was higher in patients with cholesterol crystals compared to patients without cholesterol crystals (−82.9 ± 2.4 HU vs. −91.6 ± 4.9 HU; p = 0.008). PCAT_lesion increased with the number of intraplaque cholesterol crystals (range 0–3; standardized β = 0.963, p = 0.009); the same association was not observed with PCAT_RCA. PCAT_lesion had a receiver operator characteristic area under the curve of 0.90 to identify the presence of cholesterol crystals, with an optimal cutoff of −81.6 HU as determined by Youden’s J index. All patients were commenced on moderate-to-high intensity statin therapy at the time of baseline imaging as clinically indicated. On follow-up, there was a mean reduction in PCAT_lesion (ΔPCAT = −6.2 ± 2.8 HU). Mean lowering of LDL-C was by 1.6 ± 0.4 mmol/L. There was no significant difference in ΔPCAT between plaques with resolution (n = 4) versus persistence (n = 3) of cholesterol crystals (−8.77 ± 2.88 HU vs. −2.72 ± 1.64 HU, p = 0.12). PCAT_lesion did not associate with thin-cap fibroatheroma (p = 0.70) or intraplaque macrophage infiltration (p = 0.56), nor correlate with fibrous cap thickness (r = 0.26, p = 0.45). Fig. 1 demonstrates ‘inflamed’ versus ‘non-inflamed’ per-lesion PCAT on CCTA.

Fig. 1.

Association of cholesterol crystals on OCT with PCAT CT attenuation.

A. OCT image of cholesterol crystals (arrow) within a thin-cap fibroatheroma in the LAD; B. CCTA curved view and cross-section of ‘inflamed’ PCAT (−71.2 HU) around this lesion. Hounsfield unit scale inset, with colour map ranging from dark red (−30 HU) to bright yellow (−190 HU). C. The same LAD lesion on follow-up (412 days) is a fibrous plaque with no cholesterol crystals; D. Surrounding ‘non-inflamed’ PCAT (−93.7 HU).

This small, single center, proof-of-concept analysis is the first to assess cholesterol crystal-induced coronary inflammation using CCTA. We found a higher PCAT CT attenuation around plaques and plaque segments containing cholesterol crystals compared to those without cholesterol crystals on OCT. Furthermore, PCAT attenuation surrounding the proximal RCA, a standardized measurement for per-patient PCAT analysis, was higher in the presence versus absence of cholesterol crystals anywhere in the coronary tree. Cholesterol crystals, present in all stages of atherogenesis, activate the NLRP3 (NOD-like receptor protein 3) inflammasome in macrophage foam cells, leading to secretion of the highly pro-inflammatory interleukin-1β.¹ Our findings suggest that this local inflammatory response in the coronary vasculature can be detected using a novel, validated, non-invasive imaging biomarker. Measurement of PCAT attenuation complements current CCTA-based plaque analysis and may enhance identification of high-risk plaque. From a treatment perspective, the CANTOS trial showed that specific targeting of interleukin-1β with the monoclonal antibody canakinumab reduced in recurrent cardiovascular events.⁵ If confirmed in a larger study, the ability of PCAT CT attenuation to detect cholesterol crystal-induced coronary inflammation may distinguish patients who would benefit from intensification of medical therapy, and can help guide future trials of targeted treatments with novel anti-inflammatory agents.