Combined near infrared photoacoustic imaging and ultrasound detects vulnerable atherosclerotic plaque

Abstract

Atherosclerosis is an inflammatory process resulting in the deposition of cholesterol and cellular debris, narrowing of the vessel lumen and clot formation. Characterization of the morphology and vulnerability of the lesion is essential for effective clinical management. Here, near-infrared auto-photoacoustic (NIRAPA) imaging is shown to detect plaque components and, when combined with ultrasound imaging, to differentiate stable and vulnerable plaque. In an ex vivo study of photoacoustic imaging of excised plaque from 25 patients, 88.2% sensitivity and 71.4% specificity were achieved using a clinically-relevant protocol. In order to determine the origin of the NIRAPA signal, immunohistochemistry, spatial transcriptomics and spatial proteomics were co-registered with imaging and applied to adjacent plaque sections. The highest NIRAPA signal was spatially correlated with bilirubin and associated blood-based residue and with the cytoplasmic contents of inflammatory macrophages bearing CD74, HLA-DR, CD14 and CD163 markers. In summary, we establish the potential to apply the NIRAPA-ultrasound imaging combination to detect vulnerable carotid plaque and a methodology for fusing molecular imaging with spatial transcriptomic and proteomic methods.

Keywords: Atherosclerosis; NIR biomarker; Photoacoustic imaging; Spatial transcriptomics; Vulnerable plaque.

Figure 1.. Photoacoustic imaging of the near-infrared auto-photoacoustic (NIRAPA) biomarker in human carotid plaque.

A) Schematic of human carotid endarterectomy (CEA) sample. B) Human carotid plaque under white light. C-E) Longitudinal anatomic C) NIRAPA, D) computed tomography (CT) and E) color flow ultrasound (US) images of human carotid plaque. The dashed lines on the longitudinal image in (C) represent the imaging locations of the axial images in columns i, ii and iii in (F). F) Comparison of the NIRAPA signal (680-700nm) with NIRAF and Masson’s trichrome, CD68, bilirubin, H&E and picrosirius red staining for the three tissue section locations, i, ii and iii, indicated in Fig. 1C. Scale bar, bottom right, applies to all panels in F.

Figure 2.. Measurements of fibrous cap thickness and plaque volume by histology and imaging establish sensitivity and specificity of the imaging technique.

A) Measurements and correlation of fibrous cap thickness on NIRAPA (680-700 nm)-US images and picrosirius red (PSR) histopathology with r²=0.99. B) Measurements and correlation of the vulnerable plaque volume (red outline) on NIRAPA (680-700 nm)-US images and picrosirius red histopathology with r²=0.86. C) Photoacoustic (PA) signal intensity of NIRAPA (680-700 nm) averaged in vulnerable (red outline) and stable (green outline) plaque areas. Picrosirius red histopathology shows correlating areas. p<0.0001. D) Representative PA NIRAPA-US images of asymptomatic stable, asymptomatic vulnerable (unstable) and symptomatic vulnerable plaque cases. Correlating Masson’s trichrome stains. E) Summary of diagnostic accuracy of fibrous cap thickness as measured by PA imaging compared with histological classification. Vulnerable plaque defined as fibrous cap not existing or <65 μm. PPV: positive predictive value, NPV: negative predictive value, L: lumen, ****, p<0.0001. n=24 patients.

Figure 3.. NIRAF signal correlates with CD68 and bilirubin.

A) Overview of representative H&E, Masson’s trichrome and picrosirius red images used to localize the image features. CD68, bilirubin and αSMA images provided for reference. If NIRAF was detected, results were used to calculate the Pearson’s Coefficient (termed NIRAF True Positive, pink highlighted columns). L: Lumen, F: Fibrous Cap, N: Necrotic. If NIRAF was not detected, the percentage pixel area was calculated (termed NIRAF True Negative, gray highlighted columns). B) Pearson’s Coefficient results based on the colocalization of CD68, bilirubin and αSMA with respect to the NIRAF image. C) Difference in the analyzed pixel area of CD68, bilirubin or αSMA, when NIRAF was not detected. ****, p<0.0001, n=25 patients.

Figure 4.. Spatial transcriptomic analysis of stable plaque specimen identifies specific macrophage populations that spatially correlate with the NIRAF and NIRAPA signals.

A-B) NIRAPA (A) and NIRAF (B) images of a carotid plaque cross section. Annotations indicate strong (white) and weaker (yellow) NIRAF signal. C) Illustration summarizing the stable plaque features and the location of the NIRAPA signal. D) Histological sections of the carotid endarterectomy (CEA) plaque specimen stained with CD68 and bilirubin. E) Overlay and Uniform Manifold Approximation and Projection (UMAP) cluster projection of spatial transcriptomics on carotid plaque H&E. Based on their gene expression, clusters have been assigned to macrophage, myofibroblast and smooth muscle cell (SMC) intermediate cell types. F) Overall heatmap of the general immunological signatures that differentiate the macrophage, myofibroblast and SMC intermediate clusters. G) Key genes that differentiate macrophage, myofibroblast, and SMC intermediate populations and their spatial intensity on the CEA specimen. H) Spatial deconvolution and UMAP cluster projection of the macrophage cluster in CD74⁺ and SPP1⁺ regions and the spatial location on the H&E-stained plaque cross section. UMAP projection of macrophage high resolution subtype clustering shows CD74⁺ and SPP1⁺ populations. I) Pearson’s correlation between genes within the macrophage clusters and the NIR signal. J) Heatmap of macrophage-specific gene signatures that differentiate the CD74⁺ and SPP1⁺ macrophage subpopulations. K) Key genes differentiating inflammatory (CD74⁺) and foamy (SPP1⁺) macrophages and their spatial location on the CEA specimen. P-value cutoff is 0.005. L: Lumen. Log₂FC cutoff is 2.

Figure 5.. Spatial proteomic analysis of resolved single cells and correlation with overlaid NIRAF signal in a stable plaque.

A) H&E cross section overview with black box showing CODEX region. B) CODEX images showing signal intensities of CD68, CD14, HLA-DR and CD163 on the same tissue section. C) CODEX image of NIRAF signal intensity on the same tissue section with the high-NIRAF signal region highlighted and the boxed region studied further in F. D) Pearson’s correlation between the NIRAF signal, genes within the macrophage clusters and Collagen IV. E) Venn diagram of the expression of HLA-DR, CD163 and CD14 based on segmented cells from CODEX images. F) Region of interest (ROI) images of CODEX showing DAPI, NIRAF, Collagen IV, HLA-DR, CD14, CD163 and CD31. G) Representative manually-segmented individual cell fluorescence examples from CODEX.

Figure 6.. Spatial transcriptomic and proteomic analysis of vulnerable plaque delineates spatially-dependent macrophage populations.

A-B) NIRAPA (A) and NIRAF (B) images of a carotid plaque cross section. C) Illustration summarizing the vulnerable plaque features and the location of the NIRAPA signal. D-E) Histological sections of the carotid endarterectomy (CEA) plaque specimen stained with CD68 (D) and bilirubin (E); corresponding Pearson’s correlation coefficients indicating correlation of each stain with NIRAF signal are reported below each image. F) Overlay and Uniform Manifold Approximation and Projection (UMAP) cluster projection of spatial transcriptomics on carotid plaque H&E. Based on their gene expression, clusters have been assigned to macrophage, myofibroblast and smooth muscle cell (SMC) intermediate cell types. G) Overall heatmap of general immunological signatures that differentiate the macrophage, myofibroblast and SMC intermediate clusters. H) Spatial deconvolution and UMAP cluster projection of the macrophage cluster in inflammatory CD74⁺, foamy SPP1⁺, and APOE⁺ regions and the spatial location on the H&E-stained plaque cross section. UMAP projection of macrophage high resolution subtype clustering shows CD74⁺, SPP1⁺, and APOE⁺ populations. I) Heatmap of macrophage-specific gene signatures that differentiate the CD74⁺, SPP1⁺, and APOE⁺ macrophage subpopulations. J) Key genes differentiating inflammatory (CD74⁺) and foamy (SPP1⁺) macrophages and their spatial location on the H&E-stained CEA specimen. A region with enhanced CD74 expression is outlined in a black box overlay and investigated further in (K). K) CODEX imaging of DAPI nuclear DNA stain, NIRAF, CD31, CD14, and HLA-DR in two areas of the vulnerable plaque region showing enhanced CD74 expression. P value cutoff is 0.005. Log₂FC cutoff is 2.