1.

Introduction

The microcirculation serves key functions in the body, e.g., regulate blood pressure and body temperature, exchange nutrients and metabolic waste to body, etc. Structural and functional changes within the microcirculation have been associated with various diseases including cancer, diabetes, psoriasis, capillary malformation, and Raynaud’s disease.^1–4 Microcirculation imaging can provide early indication of disease prior to clinical suspicion.⁵ The importance of noninvasive imaging techniques to get a better understanding of the vascular involvement in such diseases is critical. There are various techniques available for in vivo imaging of blood vessels within human skin. Capillaroscopy, videocapillaroscopy, laser Doppler perfusion imaging, and dynamic laser speckle imaging are commonly used, but all these techniques are limited to imaging vessels close to the surface of the skin.^6–10 Optical coherence tomography¹¹ can be combined with novel flow contrast schemes^12,13 to obtain high resolution microvascular morphology but with a low imaging depth.

Photoacoustic imaging (PAI) breaks through the optical diffusion limit and can provide microvasculature information at a high penetration depth with resolution superior than pure optical techniques by taking advantage of the low acoustic scattering in the tissue. In PAI, image contrast is dominated by the strong optical absorption of hemoglobin; therefore, vasculature can be imaged effectively. In the last few years, there has been a huge interest in the development of PAI techniques with the applications explored in dermatology,¹⁴ oncology,^15,16 vascular biology,^17,18 cardiology,^19,20 ophthalmology,^21,22 neurology,²³ and gasteronology.²⁴ The most commonly used PAI systems employ either a tomographic^25,26 or planar geometry with a linear transducer array.^27,28 In conventional photoacoustic tomography (PAT), an entire region of interest is excited using full field illumination and the photoacoustic (PA) waves are simultaneously detected either using single ultrasound (US) detector or an array of detectors. Then an acoustic back propagation algorithm is used to reconstruct a three-dimensional (3-D) image. Linear-array-based PAI systems detect PA waves from limited angles around the object using an array of detectors. PAT systems suffer from low frame rates due to the need for hundreds to thousands of laser pulses per frame. Linear-array-based PAI systems allow images to be acquired with just a few laser pulses and provide much higher frame rates which make them more suitable for clinical imaging applications. A variety of PAI systems have been developed based on various scanning configurations and reconstruction algorithms to get the optimal resolution, imaging depth, and contrast. Although PAT scanners based on spherical and cylindrical detection geometries offer large angular aperture for data collection and an accurate image reconstruction, they are not well suited for imaging highly superficial features such as the skin microvasculature for clinical imaging applications.²⁹ Moreover, the commonly used single element PAI systems cannot satisfy the requirement of real-time data acquisition and imaging, which is a prerequisite in the clinical scenario. Linear-array-based PAI is an alternative option, particularly for clinical imaging of skin and subcutaneous morphologies. In this study, PAI based on a high-frequency multielement linear-array transducer combined with a multichannel collecting system was used for volumetric structural and functional imaging within human skin. In vivo 3-D microcirculation, total concentration of hemoglobin (HbT), and the hemoglobin oxygen saturation (${\mathrm{sO}}_2$) maps of the human forearm skin were obtained. The high-frequency linear-array transducer probes used in this study are similar in style, shape, and use to regular hand-held clinical US probes, which can easily be acoustically coupled to the skin and moved around while imaging in real time.

2.

Materials and Methods

A schematic of the experimental setup used in this study is shown in Fig. 1. A combined PA and US imaging system were operated with a linear-array transducer probe. The key elements of the PAI system (Vevo LAZR, Fujifilm VisualSonics) are: tunable PA excitation laser system (optical parametric oscillator pumped by frequency-doubled Nd:YAG laser with a repetition rate of 20 Hz, pulse duration of 4 to 6 ns, spot size of $24{\mathrm{mm}}^2$, and step size of 2 nm), multielement linear-array transducer, amplifier, and a digitizer. Each linear-array transducer probe used in this study consisted of 256 elements, which were divided into four quadrants, each with 64 elements. Pulsed laser light was focused into the tissue through two fiber optic bundles ($20\times 1.25\mathrm{mm}$) mounted on each side of the acoustic aperture of the transducer probe, emitting two laser beams at an angle of 30 deg relative to the imaging plane. The generated PA waves propagated back to the transducer probe were coupled through US gel and acquired by the transducer array. For each laser pulse, the PA signals were captured by one quadrant of the transducer array. Since four pulses were required for each full width image, the frame rate was one-fourth of the laser repetition rate (i.e., 5 Hz). The PA information was passed onto a computer through an amplifier and a digitizer where it was processed into a 3-D image.

Fig. 1

Schematic of the experimental setup used in this study. Combined photoacoustic (PA) and high-frequency ultrasound (US) imaging within human forearm skin using linear-array transducer probe.

PAI in this work was performed using three transducer probes of center frequencies: 15, 21, and 40 MHz. The 15 MHz probe (broadband frequency: 9 to 18 MHz) provides an axial resolution of $100\mu \mathrm{m}$, imaging depth up to 36 mm, and imaging width up to 32 mm. The 21 MHz probe (broadband frequency: 13 to 24 MHz) provides an axial resolution of $75\mu \mathrm{m}$, imaging depth up to 20 mm, and imaging width up to 23 mm. The 40 MHz probe (broadband frequency: 32 to 55 MHz) provides an axial resolution of $40\mu \mathrm{m}$, imaging depth up to 15 mm, and imaging width up to 14.1 mm. The lateral resolution of each transducer probe was measured by scanning a 1951 United States Air Force (USAF) resolution test standard and determining the largest pattern that cannot be discerned. Figures 2(a)–2(c) show the maximum intensity projection (MIP) images of the 1951 USAF target scanned by 40, 21, and 15 MHz frequency transducer probes, respectively, along with the intensity profiles and the fitted Gaussian functions. The full width at half maximum lateral resolutions for 15, 21, and 40 MHz frequency transducer probes were found to be 314, 158, and $140\mu \mathrm{m}$, respectively.

Fig. 2

The maximum intensity projection (MIP) images of the 1951 USAF target scanned by (a) 40 MHz, (b) 21 MHz, and (c) 15 MHz frequency transducer probes along with the intensity profiles and the fitted Gaussian functions.

This study was approved by National University of Ireland, Galway Research Ethics Committee and written informed consent was obtained from the volunteers. All the experimental procedures were in accordance with the Helsinki declaration of 1975, as revised in 2008. In vivo images of the subcutaneous vasculature in the human forearm were acquired using 15, 21, and 40 MHz frequency transducer probes. The forearm skin was acoustically coupled to the transducer probe head through US gel and successive PA and US scans were acquired. 3-D data sets were collected by linearly translating the transducer (with integrated optical fibers) with a stepper motor over a region of interest, while capturing each two-dimensional (2-D) image of the 3-D stack. For a single 3-D PA/US scan, 300 frames (B-scans) were acquired over a region of 30 mm with a step size of 0.1 mm. The data acquisition time was 60 s.

3.

Results and Discussion

Figure 3 shows in vivo images of the human forearm acquired using 40, 21, and 15 MHz frequency transducer probes at 800-nm wavelength. The fluence was below the safe maximum permissible exposure of $20\mathrm{mJ}/{\mathrm{cm}}^2$ for human skin.³⁰ The 800-nm wavelength was used to obtain sufficient tissue penetration depth because of the lowest light absorbance in the tissue components such as melanin, oxy- and deoxyhemoglobin, and lipid and water in the near-infrared (NIR) wavelength range (600 to 1000 nm). The photograph taken from the subject showing the forearm skin examined by PA/US is shown in Fig. 3(a). Figures 3(b), 3(d), and 3(f) show fused PA/US single vertical ($x–y$) slices (B-scans) of the forearm skin acquired using 40, 21, and 15 MHz frequency transducer probes, respectively. The US image (gray scale) shows the layered skin morphology. The PA data (red) shows several blood vessels distributed throughout the dermis and the underlying subcutaneous tissue. Figures 3(c), 3(e), and 3(g) show MIP images through the PA volumes obtained using 40, 21, and 15 MHz frequency transducer probes, respectively. These figures demonstrate the ability of the system to detect the PA signal from the microvasculature as a series of 2-D images rendered in 3-D. High imaging depths can be achieved using low frequency transducer probes but with a lower resolution, as a tradeoff, due to the decreasing attenuation of US with frequency. Although the ultimate resolution limit is defined by acoustic attenuation, other factors such as element size, detector bandwidth, and aperture can be limiting factors in practice.

Fig. 3

In vivo PA/US images of the human forearm acquired using 40, 21, and 15 MHz frequency transducer probes at 800-nm wavelength: (a) photograph taken from the subject showing the forearm skin examined by PA and high-frequency US imaging (b, d, and f) fused PA/US vertical ($x–y$) slices (B-scans) of the forearm skin acquired using 40, 21, and 15 MHz frequency transducer probes, respectively (c, e, and g) MIP images through the PA volumes of the human forearm acquired using 40, 21, and 15 MHz frequency transducer probes, respectively.

Figure 4 shows in vivo images of the human forearm acquired using 40, 21, and 15 MHz frequency transducer probes at a 1064-nm wavelength. The fluence was once again below the safe maximum permissible exposure of $20\mathrm{mJ}/{\mathrm{cm}}^2$ for human skin. The photograph taken from the subject showing the forearm skin examined by PA/US is shown in Figure 4(a). Figures 4(b), 4(d), and 4(f) show fused PA/US single vertical ($x–y$) slices (B-scans) of the forearm skin acquired using 40, 21, and 15 MHz frequency transducer probes, respectively. Figures 4(c), 4(e), and 4(g) show MIP images through the PA volumes obtained using 40, 21, and 15 MHz frequency transducer probes, respectively. The lower optical attenuation by blood at 1064 nm compared to 800 nm resulted in a higher penetration depth than was obtained in Fig. 3.

Fig. 4

In vivo PA/US images of the human forearm acquired using 40, 21, and 15 MHz frequency transducer probes at 1064-nm wavelength: (a) photograph taken from the subject showing the forearm skin examined by PA and high-frequency US imaging (b, d, and f) fused PA/US vertical ($x–y$) slices (B-scans) of the forearm skin acquired using 40, 21, and 15 MHz frequency transducer probes, respectively (c, e, and g) MIP images through the PA volumes of the human forearm acquired using 40, 21, and 15 MHz frequency transducer probes, respectively.

PAI provides an integrated platform for structural and functional imaging by combining high contrast and spectroscopic-based specificity of optical imaging with high spatial resolution of US imaging. Figure 5 shows coregistered PA and US images of the human forearm acquired using 21 MHz frequency transducer probe. The acquired scans measured $40(\text{length})\times 22(\text{width})\times 20(\text{depth}){\mathrm{mm}}^2$. Figure 5(a) shows a B-scan of the fused PA and US image of the forearm skin for a $22\times 20{\mathrm{mm}}^2$ region acquired at 800-nm wavelength. Figure 5(b) shows the volume rendered representation of the coregistered PA and US data of the forearm skin for a $40\times 22\times 20{\mathrm{mm}}^3$ region. The high optical contrast coregistered with high resolution US imaging allows real-time in vivo imaging of deep tissues with detailed anatomical analysis. A supporting movie (Fig. 6) is provided to present the rotating structure of the PA volume which illustrates the network of blood vessels that has been detected. Oxygenated hemoglobin (${\mathrm{HbO}}_2$) has different absorption characteristics than deoxygenated hemoglobin (Hb) so an estimate of HbT and ${\mathrm{sO}}_2$ can be derived and displayed as a parametric map by imaging with different wavelengths of light. For the HbT and ${\mathrm{sO}}_2$ estimates, PA images were collected at 750- and 850-nm wavelengths. Figure 5(c) shows a B-scan of the fused PA (HbT) and US image of the forearm skin for a $22\times 20{\mathrm{mm}}^2$ region. Figure 5(d) shows the volume rendered representation of the coregistered PA (HbT) and US data of the forearm skin for a $40\times 23\times 20{\mathrm{mm}}^3$ region. Figure 5(e) shows a B-scan of the fused PA (${\mathrm{sO}}_2$) and US image of the forearm skin for a $22\times 20{\mathrm{mm}}^2$ region. Figure 5(f) shows the volume rendered representation of the coregistered PA (${\mathrm{sO}}_2$) and US data of the forearm skin for a $40\times 23\times 20{\mathrm{mm}}^3$ region.

Fig. 5

In vivo coregistered PA and ultrasound (HFUS) images of the human forearm acquired using 21 MHz frequency transducer probe: (a) fused PA and US vertical ($x–y$) slice (B-scan) of the forearm skin for a $22\times 20{\mathrm{mm}}^2$ region acquired at 800-nm wavelength. (B) Volume rendered representation of the coregistered PA and US data of the forearm skin for a $40\times 22\times 20{\mathrm{mm}}^3$ region acquired at 800-nm wavelength. (c) Fused PA (HbT) and US B-scan of the forearm skin for a $23\times 20{\mathrm{mm}}^2$ region acquired at 750- and 850-nm wavelengths. (d) Volume rendered representation of the coregistered PA (HbT) and US data of the forearm skin for a $40\times 23\times 20{\mathrm{mm}}^3$ region acquired at 750- and 850-nm wavelengths. (e) Fused PA (${\mathrm{sO}}_2$) and US B-scan of the forearm skin for a $23\times 20{\mathrm{mm}}^2$ region acquired at 750- and 850-nm wavelengths. (d) Volume rendered representation of the coregistered PA (${\mathrm{sO}}_2$) and US data of the forearm skin for a $40\times 23\times 20{\mathrm{mm}}^3$ region acquired at 750- and 850-nm wavelengths.

Fig. 6

Rotating structure of the rendered PA volume (Video 1, MPG 4.31 MB) [URL: http://dx.doi.org/10.1117/1.JBO.20.5.051021.1].

The linear-array-based PAI has been found promising in terms of resolution, imaging depth, and imaging speed for in vivo microcirculation imaging within human skin. However, significant challenges remain, particularly with the imaging depth. The presented results clearly show the feasibility of linear-array-based PAI as a clinical tool for in vivo volumetric imaging of the blood microcirculation, HbT, and ${\mathrm{sO}}_2$ within human tissue. The 3-D microcirculation, HbT, and ${\mathrm{sO}}_2$ maps obtained will be useful for clinical imaging applications such as management of cancer including screening, diagnosis, treatment planning, therapy monitoring, and accurate measurement of metabolic rate during early diagnosis and treatment of various skin and subcutaneous tissue disorders. We believe that the reflection type probe used in this study is most likely to succeed in real clinical applications. Its advantages include ease of use, speed, and familiarity for radiographers and clinicians.