Ex vivo Cryoimaging as an Adjuvant for Plaque Validation by Optical Coherence Tomography and Characterization of Fibrous Cap Inflammation in Atherosclerotic Plaques

To the editor: Histopathology is gold-standard for validation of arterial plaque characterization by in-vivo methods. Volumetric comparison of images to histopathology however can be cumbersome and cost prohibitive. Cryo-imaging provides real-time, ultra high-resolution anatomical brightfield and fluorescence data by alternatively sectioning and imaging a tissue block face (1–2). Plaque morphology is minimally altered, permitting precise volumetric tissue characterization. Fibrous cap inflammation and elastin are important biomarkers that correlate to plaque stability. With accurate characterization of these constituents, cryo-imaging can detect a volumetric plaque transition to a vulnerable state.

In this pilot study, we sought to characterize fibrous cap inflammation and elastin in coronary plaques by cryo-imaging. We also introduce methodology for use of cryo-imaging as an adjuvant for plaque validation via a single-case volumetric comparison of cross-sections from intravascular optical coherence tomography (OCT) to cryo-imaging and histopathology.

A total of 18 atherosclerotic plaques were evaluated from 10 coronaries of male cadavers (mean age 50.2 ± 6.2 years). Vessels were explanted and stored at 4° C no longer than 48 hours. A vessel was fixed to a plastic cylindrical rig and infused with optimal cutting temperature gel. OCT imaging catheter was inserted through the ostial end of the vessel and images were captured every 200 um by an automated 10 mm/sec pull-back. Images were screened for fibroatheromatous and fibrous segments using consensus criterion (3). The start and end frames for plaque were bookmarked for analysis by cryo-imaging. The vessel was sectioned into 5 cm cylindrical blocks and stored at −80°C. Cryo-imaging was executed as previously described (2). Images were acquired in 5 µm intervals until reaching the plaque of interest. A 3 mm slab was sectioned and saved for histological analysis. Slabs were frozen (n=7) or fixed in 10% formalin and embedded in paraffin (n=11). Slices were sectioned 7µm thick and stained. Cryo-imaging was restarted until another plaque was reached or the vessel terminated.

Slides were digitally scanned and reviewed by one pathologist (NPZ). Fibrous caps were stratified into inflamed or non-inflamed groups based on two thresholds for percentage of capsule occupied by macrophage (0% and 10%). Images from cryo-imaging were split into color channels and average intensity was measured over the fibrous cap pixel area. Univariate and multivariate logistic regression with stepwise selection were performed. Odds Ratios (OR) and area under the receiver operating characteristic curves (AUC) were calculated. A P value < .05 was considered statistically significant.

Methodology and representative images from OCT, cryo-imaging, and histopathology are provided in Figure 1. A total of 3,246 OCT frames, 22,864 cryo-imaging frames, and 126 histological slides were evaluated. OCT accurately screened 81% (9/11) and 86% (6/7) of fibroatheromatous and fibrous plaques versus histopathology. Two cases were classified as fibrous by OCT, but fibroatheromatous by histopathology; and one case was fibroatheromatous by OCT, but fibrocalcific by histopathology.

Figure 1. Illustration of plaque evaluation methodology: volumetric co-registration of images from OCT, cryo-imaging, and histopathology.

(a) OCT was performed and images are screened for plaque type. This is a longitudinal OCT image with the fibroatheromatous plaque of interest (4mm to 15mm). Red dashed lines indicate start (7mm) and end frame (10mm) for subsequent histological analysis. (b) Schematic of cryo-imaging. A bookmarked 3mm segment was sectioned for histological analysis. Examples of co-registered images provided from (c) frames preceding the region of interest, (d) starting at the region of interest, and (e) within the region of interest. Images in (d) provide opportunity for frame comparison among all three modalities. Masson’s tri-chrome was used to assess the fibrous cap (blue), verhoeff van geison for elastin content (yellow-black), anti-CD68 for macrophage (brown), anti-smooth muscle actin for smooth muscle cells (dark brown). Lesions per row: (f) dim brown appearance of macrophage (open arrows) by fluorescence, and (g, h) bright green streaks of elastin (closed arrows) by fluorescence. (h) Less inflamed fibrous tissue appears homogenously dim by fluorescence and pink to pale white by brightfield (between closed arrows on right). * = target lipid plaque; slab = 3 mm plaque of interest

Macrophages appeared brown with low intensity by fluorescence, and pink to pale white by brightfield (Figure 1f). Elastin appeared as bright green streaks by fluorescence, and pink to pale white by brightfield (Figure 1g). Cryo-imaging green fluorescence was the best individual marker of inflammation at 0% (OR=0.93, p<0.05) and 10% (OR=0.95, p<0.05) macrophage with AUC of 0.958 (95% CI=0.869–1.000, p<0.001) and 0.790 (95% CI=0.516–1.000, p<0.05), respectively. Using green fluorescence intensity less than 60 as positive for inflammation (i.e. greater than 0% macrophage), sensitivity was 91.7% (p=0.003), specificity 100% (p=0.016), positive predictive value 100% (p<0.001), negative predictive value 85.7% (p=0.06), and accuracy 94.4% (p<0.001).

We demonstrate less cryo-imaging fluorescence correlates with greater density of macrophage and less elastin. These properties are consistent with studies using different excitation and emission wavelengths (4). While this trial was small, a volumetric approach for comparing OCT images to cryo-imaging and histopathology was successfully introduced. Early data yielded OCT sensitivities within range reported by literature. Errors in screening are typically related to misidentification of lipid versus fibrous and calcified plaques (5). Cryo-imaging is a promising adjuvant for plaque characterization based on high sensitivity and specificity for these constituents.

Clinical applications of volumetric tissue characterization by cryo-imaging are numerous. Volume rendering techniques are being developed to improve 3D co-registration between in-vivo modalities and cryo-imaging. Tissue volumes may be segmented and compared quantitatively to obtain more accurate results than by histopathology alone. Segmentation of elastin and inflammation during volume renderings may be analyzed via finite element analysis to understand effects of plaque deformation on genuine vessel architecture. Coupling to in-vivo imaging modalities provides potential for understanding real-time effects of therapeutic intervention (i.e. stent deployment) on fibrous cap stress. Fluorescence tagging prior to imaging allows volumetric quantification of desired cells or proteins.

In conclusion, this pilot study demonstrates cryo-imaging is capable of detecting fibrous cap inflammation with high sensitivity and specificity. Potential for macrophage quantification was demonstrated at two thresholds. Cryo-imaging is a promising adjunct to histopathology for studies aiming to validate OCT for plaque characterization.