1. Introduction

Acute myocardial infarction, the leading cause of death worldwide, is frequently caused by the rupture of unstable coronary plaques. Detecting plaque vulnerability helps the diagnosis and plays an important role in choosing proper interventional techniques. The typical microstructural hallmarks of vulnerable plaques include the presence of a thin fibrous cap, a large necrotic core, and activated macrophages near the fibrous cap [1–3]. Current techniques for investigation of atherosclerotic plaques include angiography, intravascular ultrasound (IVUS) [4], magnetic resonance imaging (MRI) [5], optical coherence tomography (OCT) [6], optical spectroscopy (NIR, Raman, and fluorescence spectroscopy) [7,8], and photoacoustic imaging (PAI) [9]. However, most imaging systems focus on one or dual-modality imaging [4–13], which are not enough for an accurate characterization of three features. For example, it is difficult to identify whether there is a thin fibrous cap or not by using a combined IVUS and PAI system [12]. More complete information concerning various markers involved in plaque vulnerability and rupture is needed for improving plaque evaluation. Recently, a few tri-modality imaging systems have been reported, and these works represent a significant step forward for the estimation of atherosclerotic plaques, such as IVUS-OCT-PAI [14], IVUS-OCT-fluorescence [15], fluorescence-ultrasonic backscatter microscopy-PAI [16], and IVUS-PAI-photoacoustic elasticity imaging [17]. However, these systems cannot simultaneously characterize the thickness of fibrous cap, the size of lipid core, and the extent of inflammatory reaction, making it difficult to achieve a comprehensive detection of plaque vulnerability.

Here, we investigate a new detection technique that combines OCT, PAI, and photoacoustic viscoelasticity imaging (PAVEI) to obtain information about plaque scattering, absorption, and viscoelasticity (phase delay), thereby achieving the comprehensive characterization of three traits of atherosclerotic plaques. Specifically, OCT is a high-resolution intracoronary imaging modality that uses the infrared light to image arterial wall structure, and responds sensitively to the thickness of fibrous cap [6,18]. PAI is an emerging technique capable of providing the analysis of localization and quantification of lipid core based on the distinct optical absorption signature of lipids [9,19]. PAVEI, based on the photoacoustic (PA) phase delay technique [20–22], providing a high resolution, loading-free method for the detection of viscoelasticity of atherosclerotic plaques. The phase delay of PA signal is directly related to the viscosity-elasticity ratio of the medium [21]. It is worth noting that the biomechanical properties of plaques are not only related to plaque composition, but also to the content and distribution of macrophages [23]. Furthermore, the risk of plaque rupture is closely linked to the changes of viscoelastic properties [24]. By recording the phase delay of the PA signal, the viscoelasticity image of the sample can then be reconstructed [22]. In the early stage of atherosclerotic plaque, the accumulation of macrophages leads to the formation of lipids, and due to the relatively high viscosity coefficient of lipid, the phase delay of PA signal will be higher than that in normal regions. In the middle stage of development, the transfer of smooth muscle cells and the generation of collagen fibrous lead to the formation of fibrous cap. At this time, macrophages mainly act under the cap, and the phase delay is small than that of normal one due to the high Young's modulus of fibrous collagen. Finally, when a large number of macrophages act on the fibrous cap, the large-scale degradation of collagen fibrous at this time will make the plaque extremely unstable, and the phase delay may reach the maximum. Therefore, the detection technique combining OCT, PAI and PAVEI is expected to provide complementary information regarding features of atherosclerotic plaques, specifically those that are potential predictors of plaque rupture.

The aim of this study is to demonstrate whether OCT-PAI-PAVEI system allows characterization of the three traits in atherosclerotic plaques, so as to assess the plaque vulnerability. In the mice atherosclerotic model, the thickness of fibrous cap is measured by OCT, the area distribution of lipid core is provided by PAI, and the relationship between the content and distribution of macrophages and phase delay was revealed by PAVEI. These three imaging methodologies complement each other and may provide more complete pathophysiological assessment of atherosclerotic plaques.

2. Materials and methods

2.1 Experimental setup

Figure 1(a) shows the typical traits of most vulnerable plaques, including thin fibrous cap, large lipid core and severe inflammation reaction. These features could be characterized by OCT, PAI and PAVEI, respectively. A schematic of the OCT-PAI-PAVEI system is shown in Fig. 1(b). For the OCT subsystem, the light source is a broadband super luminescent diode (SLD1018S, Thorlabs, USA) with a bandwidth of 45 nm centered at 1310 nm, corresponding to 17 µm axial resolution (in air). Light from SLD passed through the isolator, which protected the light source from damage by reflected light. And then the light was divided into the sample arm and the reference arm by a fiber coupler. 90% of the light energy went to the sample arm, while 10% went to the reference arm. To optimize the axial resolution, two sets of polarization controllers were used in the sample arm and the reference arm, respectively, to manage the polarization mode dispersion. The interference between the light from the sample arm and reference arm was detected by a customized spectrometer. The output of the fiber was collimated (AC254-050-C, Thorlabs) and towards a transmission grating (1145 line/mm, Wasatch Photonics). An achromatic lens with f = 100 mm (AC508-0100-C, Thorlabs), focused the dispersed light onto a high-speed line scan charge-coupled device (2048L InGaAs Linescan Camera, Sensors Unlimited, NJ) with up 76 kHz line rate at 2048 pixels. Data was streamed from the camera to an acquisition computer through a Camera Link interface card (PCIe-1433, National Instrument, TX). During the signal processing, we applied the dispersion compensation algorithm to alleviate the dispersion mismatch between the sample arm and the reference arm.

 

Fig. 1. (a) Typical traits of most atherosclerotic plaques. (b) Schematic of the OCT-PAI-PAVEI system. (c) Principle of the PAVEI for detecting the viscoelasticity of plaques though the phase delay of PA signal. η, viscous coefficient; E, Young's modulus; δ, photoacoustic phase delay; ω, laser modulation frequency; I(t), intensity; t, time; C_1,2,3,4, collimator; CCD, charge-coupled device; DAS_1,2, data acquisition system; DM_1,2, dichroic mirror; FC, fiber coupler; MO, microscope objective; P, plaque; PC_1,2, polarization controller; SLD, superluminescent diode; SMF_1,2,3, single-mode fiber; UT_1,2, ultrasonic transducer.

Download Full Size | PDF

The PA technique included intensity (PAI) and phase (PAVEI) imaging. Previously, we proposed and validated simultaneous optical absorption and viscoelasticity imaging for atherosclerosis characterization using a quasi-continuous laser [22]. However, the optical absorption of lipid is lesser than collagen at 1064 nm. Therefore, for the PAI subsystem, an optical parametric oscillator (OPO) laser (NT200 Series Laser, Ekspla, Vilnius, Lithuania) delivers 8 ns laser pulses at a wavelength of 1700 nm was used as the radiation source with a repetition rate of 2.5 kHz. For the PAVEI subsystem, a quasi-continuous laser (DS20HE-1064D/R, Photonics) with pulse width of 22 ns and wavelength of 1064 nm operating at the repetition frequency of 50 KHz was used as the excitation source. The laser beams were co-aligned with the OCT beam using a dichroic mirrors and focused onto the target by a 4× microscope objective (numerical aperture, NA = 0.1). The laser energy density incident on the tissue surface was controlled below 15 mJ/cm² during the experiments. The PAI signals carried out by a custom-made hollow-focused ultrasound transducer 1 (UT₁, 75 MHz center frequency, 90% ∼6 dB bandwidth). The focusing radius was approximately 15 mm, the height was about 10 mm, and the diameter of the laser transmitting hole was about 3 mm. The signals were transferred to a wide-bandwidth low-noise amplifier (Ha2, Precision Acoustics Ltd.), digitized and collected by a high-speed data-acquisition card (M-3i. 4120, Spectrum) at a sampling rate of 250 MHz, and finally stored in the computer. For PAVEI, the center frequency of the customized ultrasonic transducer 2 (UT₂) for receiving signals was 50 KHz. The UT₂ was a cylinder with a diameter of 10 mm and a length of 30 mm, and was unfocused. Its sensing area was close to the focal point of laser on the tissue and placed at an angle of about 45 degrees to the plane. The pre-amplified PA signal was detected by a lock-in amplifier (SR830, Stanford Research Systems), which was used as a phase-sensitive detector. The phase delay of the PA signal was analyzed by the computer. During the experiment, the scanning step of the system was set to 10 µm. The system was operated via a custom-made LABVIEW program (National Instruments, Austin, TX). MATLAB software (Mathworks, Natick, MA) was used for image construction and index measurement.

2.2 PAVEI principle

The basic principle of PAVEI is shown in Fig. 1(c). Light absorption by the tissue resulted in a sinusoidal temperature variation due to the non-radiative transition, and then caused thermal expansion and shrinkage as well as PA wave generation based on the thermo-elastic mechanism. Meanwhile, the cyclical heating in the local region induced the thermal stress, generating a strain in the form of force-produced PA wave, which had the same frequency with laser excitation, while a phase lag behind it existed owing to the damping effect caused by biological viscoelasticity. In the rheological Kelvin-Voigt model, the relationship between the phase delay δ and the viscosity-elasticity ratio (η/E) as [17]

2.3 Experimental protocol

Five Apolipoprotein E-knockout (ApoE) mice (C57BL/6 background) were used in the time-lapse experiment, and received 4, 8, 12, 20 or 24 weeks high-fat/high-cholesterol (HFC) diet respectively. At the end of each feeding stage, the mice were euthanized with an overdose of pentobarbital (3%, 120 mg/kg). The thoracic aorta of each mice was excised and flushed with PBS (PH = 7.4). Then the extracted aorta was cut open, paved and fixed on the surface of 3% agar, and was promptly examined with the OCT-PAI-PAVEI system. For the aorta of mice fed at 4, 8 or 12 weeks, we selected blood vessels with characteristic sites (that is, positions where plaques existed) for imaging, and the imaging length was about 14 mm. For the aortic vessels of mice fed for 20 or 24 weeks, the imaging length was about 17.5 mm. After each experiment, the characteristic plaque positions were selected, marked on the agar corresponding to the positions, and then cut along each mark. Each small section was moved from agar to formalin for fixation.

For all vessels, sections of atherosclerotic plaques were sliced near the center of the marked segment for lipid with Oil Red O, fibrous cap with Masson, and macrophages with F4/80 Immumohistochemical staining. Frozen sections were used for Oil Red O and paraffin sections were used for the other two. The thickness of Frozen and paraffin sections was 6 µm and 3 µm, respectively. The histology was performed by investigators blinded to the imaging results. In order to determine the target cross section in the image, the position of the marker could be roughly determined, and then reconstructed all the cross sections near the position. After obtaining the histological results, the cross section could be further confirmed according to the morphological and structural characteristics. In this way, it could be determined that the histological results correspond to the cross-sectional images used for analysis. The statistics of experimental data were presented as mean ± standard deviation.

3. Results

3.1 Characterization of system performance

Phantom experiment was conducted to evaluate the spatial resolution and imaging depth of the OCT-PAI-PAVEI system. A blade was embedded in a gelatin (48722, Sigma-Aldrich) phantom. The sample was imaged via the system, and the signal distribution of OCT and PAI are shown in Figs. 2(a) and 2(b), respectively. The phase delay distribution of the sample edge is shown in Fig. 2(c). The lateral resolutions of the OCT, PAI, and PAVEI were 11.4, 9.7, and 10.2 µm, respectively, which were defined as the full-width-at-half-maximum (FWHM) of the line spread function (LSF). Figures 2(d) and 2(e) show the axial resolution of the OCT and PAI, based on the FWHM of the profile, were estimated to be 18.5 and 22.8 µm, respectively. To demonstrate the imaging depth of the system, an iron wire was inserted into the fat for simulating the lipid. The iron wire diameter was about 0.3 mm. The measuring method was shown in the inserted schematic. The signal intensities of PAI and OCT were decreased with increasing the imaging depth, as demonstrated in Fig. 2(f). Obviously, the depth of PAI was significantly deeper than that of OCT, it could image over 2 mm in the fat. Results suggested that the OCT-PAI-PAVEI system had the ability to detect the mice atherosclerotic plaques.

 

Fig. 2. The performance of OCT-PAI-PAVEI system. (a), (b), and (c) represent the lateral resolution of OCT, PAI, and PAVEI, respectively. (d) and (e) represent the axial resolution of OCT and PAI, respectively. (f) Imaging depth of the system. The inserted schematic shows the measuring method. ESF, edge spread function; LSF, line spread function.

Download Full Size | PDF

3.2 Characterization of lipid plaques

The OCT-PAI-PAVEI system was able to characterize the lipid plaques, which were the early lesions of atherosclerosis. There were no fibrous caps in the plaques at this stage. Figure 3(a) shows the en-face PAI and PAVEI images found in mice aortas of 4 to 12 weeks old. The strong signal areas in the PAI images gradually increased, reflecting the increase of the spatial distribution of lipids. The phase delay measured by PAVEI also became larger. Figure 3(b) shows the cross-section PAI and OCT images of plaques, as well as the histological results. Lipids presented a strong absorption in PAI, and OCT images showed a weak signal for the lipid region. The size of lipid core measured by PAI corresponded well to the Oil Red O staining (the area enclosed by the black dotted line in the images). The cross-section information of PAI also corresponded with the surface phase delay distribution of PAVEI. The formation of lipid core was the product of inflammatory reaction. The more macrophages were contained, the larger the lipid core was, the greater the phase delay was. The phase delay significantly expanded from nearly 46.26 ± 6.56 deg (normal blood vessels) to 58.47 ± 5.99 deg (lipid plaques) after 12 weeks of the HFC diet. These results indicated that PAI was sensitive to detect the distribution and size of lipid cores and PAVEI was positively responded to the macrophage aggregation and lipid accumulation.

 

Fig. 3. Lipid accumulating in the arterial wall characterized by OCT-PAI-PAVEI system. (a) En-face images of PAI and PAVEI at 4, 8 and 12 weeks; (b) Cross-section images of PAI, OCT and histology (Oil Red O and Masson staining). The area enclosed by the black dotted line indicates the presence of lipids.

Download Full Size | PDF

3.3 Characterization of fibroatheromas

Lipid plaques have further developed into fibroatheromas under the influence of various factors. Proliferation of vascular smooth muscle cells, a large number of collagen fibrous, a few elastic fibrous and proteoglycans to form a fibrous cap. Subsequently, under the action of inflammatory cells, the fibrous cap became thinner and the lipid core became larger, which made the plaques easily to rupture. For research convenience, fibroatheromas were artificially divided into thick cap fibroatheromas (fibrous cap thickness ≥ 200 µm) and thin cap fibroatheromas (fibrous cap thickness ≤ 65 µm). The OCT-PAI-PAVEI system was able to characterize these features of fibroatheromas. Figure 4 shows the imaging results (including en-face images of PAVEI and cross-section images of PAI, OCT and histology) of mice aortas of 20 and 24 weeks old. The size of lipid core measured by PAI corresponded well to the Oil Red O staining (the area enclosed by the black dotted line in the images). The thickness of fibrous cap detected by OCT also corresponded well to the Masson staning (the green and yellow arrows represent the location of thick and thin fibrous cap, respectively). Figures 4(a) and 4(b) show the imaging results of mouse aorta of 20 weeks old. The plaques had the small phase delay, some even smaller than the normal vessels, because most plaques at this time had the thicker fibrous caps and macrophages mainly accumulated under the fibrous cap. From the results of cross-section PAI and OCT images, there was a larger lipid core under the thick fibrous cap, as shown in regions I and III. However, results of OCT images showed that region III contained a slightly thinner fibrous cap, so the phase delay in the relative position would be slightly larger. Region II had a small lipid core and no lipid was presented in region IV. But they all had the thick fibrous caps, so the phase delay was small.

 

Fig. 4. Fibroatheromas (aortas from the mice fed with HFC diet for 20 and 24 weeks) detected by the OCT-PAI-PAVEI system. (a) and (c) En-face image of PAVEI; (b) and (d) Cross-section images of PAI, OCT and histology (Oil Red O and Masson staining). The area enclosed by the black dotted line indicates the presence of lipids; the green and yellow arrows represent the location of thick and thin fibrous cap, respectively.

Download Full Size | PDF

Figures 4(c) and 4(d) show the imaging results of mouse aorta of 24 weeks old. The phase delay of fibroatheromas at 24-weeks (62.97 ± 24.71 deg) were generally significantly greater than those at 20-week (40.53 ± 13.50 deg), and even was the maximum in all times of atherosclerotic plaques. At this stage, the fibrous cap became thinner due to the rapid increase in the content of macrophages acting on the fibrous cap, and the plaques became unstable and might rupture. As shown by the imaging results, not all plaques had thin fibrous caps, such as regions II and IV, which had large lipid cores underneath. However, the phase delay of the region II was significantly larger than that of the region IV, indicating that the region II had a large number of macrophages acted on the fibrous cap. Not all thin fibrous caps had lipid cores underneath, such as region I. Region III contained the typical characteristics of vulnerable plaques, a large lipid core covered with a thin fibrous cap, and the phase delay was very large. The above experimental results indicated that vulnerable plaques might not simultaneously exhibit the three characteristics (thin fibrous cap, large lipid core, and severe inflammatory reaction).

3.4 Characterization of plaque inflammation

To demonstrate that PAVEI has a high sensitivity to characterize the extent of inflammatory reaction, the relationship between the the content and distribution of macrophages and phase delay was explored. The phase delay distributions of different types of plaques are shown in Fig. 5. Images I, II, III, and IV represented the images enlarged by the black dotted frame in F4 / 80 staining with a magnification of 4 times. Mice before 12 weeks of age only exhibited fat streaks on the luminal wall of the aorta, whereas most of the luminal surface remained largely intact. The macrophage content was steadily increasing, and the phase delay was also increasing. The phase delay of normal vessel (about 48 deg) differed significantly compared to the lipid plaque (about 57 deg), as shown in Figs. 5(a) and 5(b). The F4/80 Immumohistochemical staining results showed that the macrophages accumulated in lipid core. When the mice were at the age of 20 weeks, where the macrophage content did not increase much, but phase delay decreased significantly. Because smooth muscle cells and collagen began to proliferate, and macrophages mainly acted under the fibrous cap. At 24 weeks of age, mice had fibroatheromas accumulation over large areas of the luminal surface of the aortas. The fibrous cap of some plaques had become very thin under the action of macrophages (collagen degradation). Compared to the thick cap fibroatheroma (about 45 deg), higher average phase delay was found in the thin cap fibroatheroma (about 65 deg), results are shown in Figs. 5(c) and 5(d). Furthermore, Masson staining results revealed the decrease of collagen in thin cap fibroatheroma stage, but the content of macrophages acting on the fibrous cap was significantly increased (from the F4/80 Immumohistochemical staining results), so the phase delay reached the maximum among all the lesions. Results demonstrated that the phase delay (viscoelasticity) of plaques was not only related to the plaque compositions, but also to the content and distribution of the macrophages.

 

Fig. 5. Phase delay in different types of atherosclerotic plaques: (a) Normal vessel, (b) Lipid plaque, (c) Thick cap fibroatheroma, and (d) Thin cap fibroatheroma. Corresponding histological results revealed by F4/80 Immumohistochemical and Masson staining. Images I, II, III, and IV represent the images enlarged by the black dotted frame in F4/80 staining with a magnification of 4 times.

Download Full Size | PDF

3.5 Characterization of different types of plaques

To prove that the tri-modal OCT-PAI-PAVEI imaging system had the ability to provide the complementary multi-parameter detection for atherosclerotic plaques, including scattering, absorption, and phase delay (viscoelasticity). 110 cross-sections were analyzed from 5 mice arteries, 43 lesions (39.1%) were classified as lipid plaques. Thick cap fibroatheromas were found in 26 plaques (23.6%), 17 lesions (15.5%) were classified as thin cap fibroatheromas, and the remaining 24 lesions (21.8%) contained no plaque. Examples of the intensity for normal vessel and different types of plaques measured with PAI and OCT are shown in Fig. 6(a). Plaque types include lipid plaques, thick cap and thin cap fibroatheromas. The results showed that lipid plaque had a higher PA signal intensity than normal vessel, whereas lipid optical scattering intensity was lower than normal vessel, because the normal vessel contains more collagen than lipid plaque. However, in the stage of fibroatheroma, the signal intensity of PAI and OCT was very large whether it was a thick cap or a thin cap. The morphological characteristic and scattering difference of vascular plaque could be obtained by PAI and OCT, so as to evaluate the size of the lipid core and the thickness of the fibrous cap. In addition, due to the changes in plaque composition, and macrophages acting at the different positions of plaque, a large change in viscoelasticity was caused, that is, the phase delay was significantly changed. Analysis of the phase delay of normal vessels and plaque lesions are shown in Fig. 6(b), from the results, compared with the phase delay of the normal vessels (46.26 ± 6.56 deg), the lipid plaques were slightly larger, about 56.49 ± 13.77 deg. The thick cap fibroatheromas had the small phase delay, about 39.01 ± 6.43 deg, while the thin cap fibroatheromas reached the maximum, about 65.61 ± 12.13 deg. Experimental results proved that the OCT-PAI-PAVEI system could achieve the comprehensive characterization of atherosclerotic plaque vulnerability.

 

Fig. 6. Characterization of different types of atherosclerotic plaques. (a) Distribution of PAI and OCT signal intensity. (b) Distribution of PAVEI phase delay. TCFA, thick cap fibroatheroma; ThCFA, thin cap fibroatheroma.

Download Full Size | PDF

In order to further explore the relationship between the phase delay and the lipid core area, the fibrous cap thickness and the macrophage content, the statistical analysis of experimental data were performed. Early in the development of atherosclerosis, the change in mechanical properties of plaques were mainly due to the accumulation of lipids. Hence, at the stage of lipid plaques, the relationship between the lipid core area and the phase delay (Lipid plaques, n = 30) were analyzed. The result is shown in Fig. 7(a), linear regression analysis demonstrated a high positive correlation between lipid core area and phase delay (R = 0.89, P < 0.001). The results showed that the phase delay increased with the accumulation of lipids during the lesions of lipid plaques. In the fibroatheromas stage, the viscoelasticity measured by PAVEI was mainly determined by the fibrous cap, so the relationship between the fibrous cap thickness and the phase delay (Fibroatheromas, n = 30) were analyzed. In Fig. 7(b), the phase delay is plotted against cap thickness. The cap thickness ranged from 20 µm (phase delay ∼ 77.5 deg) to 300 µm (phase delay ∼ 39.0 deg). Linear regression analysis demonstrated a high negative correlation between cap thickness and phase delay (R = -0.90, P < 0.001). The inflammatory response had been accompanied by the development of atherosclerosis, and the macrophage content had been steadily increasing, moreover the phase delay and the macrophage content were not a linear relationship. When the smooth muscle cells began to proliferate, collagen fibrous accumulated, the phase delay could be greatly reduced. This process occurred in the aorta of mouse at 20 weeks of HFC diet. Figure 7(c) shows the change trend of macrophage content and phase delay in different periods of HFC diet (5 plaque specimens were taken at each period for statistical analysis). The results showed that the phase delay increased with the growing of macrophage content in the stage of lipid plaques, and in the thin cap fibroatheromas, the macrophage content and phase delay both reach the maximum.

 

Fig. 7. Statistical analysis. (a) High positive correlation between the lipid core area and the phase delay (n = 30). (b) High negative correlation between the fibrous cap thickness and the phase delay (n = 30). (c) Distribution of the macrophage content and the phase delay in different periods.

Download Full Size | PDF

4. Discussion

Our data demonstrate the feasibility of OCT-PAI-PAVEI system for multi-parameter characterization of atherosclerotic plaques, thereby identifying high-risk lesions. Compared with other tri-modality imaging techniques [14–17], the advantage of this system can comprehensively characterize fibrous cap thickness, lipid core size, and inflammation extent, and also can measure the viscoelastic properties of atherosclerotic plaques. OCT and PAI are two relatively mature and widely used imaging methods, while PAVEI is a new technique that has been developed in recent years. Compared with other biomechanical imaging [24, 25–28], PAVEI is a loading-free method to obtain the distribution of viscoelasticity with a high resolution through the detection of phase delay. Moreover, PAVEI is the comprehensive response imaging of the two parameters (viscous coefficient and Young's modulus) of atherosclerotic plaques, and has more judgment basis than a single parameter detection.

The OCT-PAI-PAVEI system was used to characterize the two major stages of atherosclerotic lesions (lipid plaques and fibroatheromas). In the stage of lipid plaques, the increase of lipid cores is due to the accumulation of macrophages [29], which leads to the increase of averaged phase delay. While in the stage of fibroatheromas, the change of lipid core area is not very obvious, and the formation of fibrous cap has a protective effect on plaque. During the thick cap fibroatheroma stage, macrophages mainly act under the fibrous cap, where phase delay is small, about 39.01 ± 6.43 deg. When a large number of macrophages begin to act on the fibrous cap (thin cap fibroatheroma stage), the cap becomes thinner [30], where the lipid core area reaches the largest and the phase delay reaches a maximum, about 65.61 ± 12.13 deg. In particular, vulnerable plaques may not satisfy all the three features simultaneously. Thin caps may be present in the collagen proliferation stage, where the inflammatory reaction is not severe and the plaque is stable. There may be no lipid core under the thin cap, but most macrophages act on the fibrous cap and the plaque is vulnerable. Therefore, the OCT-PAI-PAVEI detection method combines the advantages of the respective imaging modes to obtain multi-parameter information to comprehensively detect the plaque vulnerability and provide a more accurate basis for the diagnosis of atherosclerosis.

For the future development, first, a fully integrated system with a high spatial resolution endoscopic probe will be used in the future for in vivo detection of atherosclerotic plaques. The specific improvements include: a) The structure of the probe is designed reasonably based on the original intention of keeping the system stable. b) Signal post-processing (multiple averaging and median filtering) can be performed on the collected PA data in the blood environment, which will improve the signal-to-noise ratio (SNR) to a certain extent. c) As a sheath for the probe, the use of optical and acoustically transparent material not only protects the blood vessel during imaging, but also reduces attenuation of signals to a minimum and induces minimal artifacts. d) The increase of imaging speed will be achieved by improvements in real-time image processing and display algorithms (C programming language). Second, using the spatial light modulator to selectively focus the laser onto different depths of plaques for achieving the depth-resolved PAVEI. Third, current results were based on mice atherosclerotic plaques, and further studies should be undertaken on human plaques to assess the capabilities and potential benefits of OCT-PAI-PAVEI detection method. There are also some substances that play an important role in the development of human atherosclerosis, such as cholesterol crystals, cholesterol esters and calcifications. OCT can image and distinguish cholesterol crystals and calcifications from the morphology and structure. Both of them are hard substances with high elastic modulus, which will lead to the small phase delay. The specific differences still need to be further explored through PAVEI. Cholesteryl esters can be identified by PAI at the specific wavelengths (absorption peaks).

5. Conclusions

In conclusion, all the results demonstrate that the OCT-PAI-PAVEI system has potential to visualize the three characteristics of atherosclerotic plaque, thereby achieving the comprehensive assessment of plaque vulnerability. Specifically, the thickness of fibrous cap (OCT), the area of lipid core (PAI), and the content and distribution of macrophages (PAVEI) directly affect the stability of atherosclerotic plaques. Therefore, the technique combined with OCT, PAI and PAVEI imaging methods provides a more comprehensive and accurate basis for the detection of plaque vulnerability from the perspective of plaque structure, composition and mechanical properties.