Advanced Imaging, Volume. 1, Issue 3, 032002(2024)
In vivo multi-scale clinical photoacoustic imaging for analysis of skin vasculature and pigmentation: a comparative review On the Cover
Figures & Tables(18)
![PAI modalities in skin imaging. (a) A schematic of the PAI principle (left) and the optical absorption spectra of HbO2, HbR, and melanin (right). The spectra data are from http://omlc.ogi.edu. (b) Conceptual schematics and example images of the vessel and melanin in the skin for each PAI modality. US, ultrasound; PA, photoacoustic; PAI, photoacoustic imaging; PAM, photoacoustic microscopy; PAMes, photoacoustic mesoscopy; PAT, photoacoustic tomography; HbR, deoxy-hemoglobin; HbO2, oxy-hemoglobin; TR, transducer; EP, epidermis; DR, dermis; ST, subcutaneous tissue; FOV, field of view. The figures were reproduced with permission from Refs. [100] (Copyright © 2021 Optical Society of America), [120] (Copyright © 2021 Optical Society of America), [112] (Copyright © 2022 by the authors. Licensee MDPI, Basel, Switzerland), and [122] (Copyright © by the authors).](/Images/icon/loading.gif){kind=icon}
Fig. 1. PAI modalities in skin imaging. (a) A schematic of the PAI principle (left) and the optical absorption spectra of
Fig. 2. PAM techniques for improved visualization of skin structures. (a) An OR-PAM system integrated with a photoplethysmography (PPG) sensor. Simultaneous real-time monitoring of PA and PPG signals from superficial dermal vessels in fingers accurately estimated the heartbeat rate (
Fig. 3. Photoacoustic microscopic biopsy (PAMB) technique for port wine stain (PWS) lesion imaging. Compared to normal skins, PWS skins have dilated vascular structure in the dermal layer. MAP, maximum amplitude projection; EP, epidermis; DR, dermis; SB, stratum basale; SC, stratum corneum. The figures were reproduced with permission from Ref. [100] (Copyright © 2021 Optical Society of America).
Fig. 4. (a) Cross-sectional RSOM image composed of two bandwidth signal images and a binary image of the high-frequency signal within ROI. The combined image of high-frequency signal (green, representing small vessels) and low-frequency signal (red, representing large vessels) shows the boundary between epidermis and dermis. The binary image is used to calculate the vessel fragmentation value. (b) The process of skin biomarker quantification in cross-sectional RSOM images. First, the EP and DR regions are divided, and the EP thickness and the EP signal density are calculated. Then, the vessel numbers are calculated by counting the vessel branches (red dots) in the DR area, and the total blood volume is calculated from the binary image. RSOM, raster scan optoacoustic mesoscopy; ROI, region of interest; EP, epidermis; DR, dermis; VS, vessel segmentation; BC, biomarker computation. The figures were reproduced with permission from Refs. [101] (Copyright © 2020 by the authors. Contact Dermatitis published by John Wiley & Sons Ltd.) and [107] [Copyright © 2023 by the author(s)].
Fig. 5. PAT techniques for clinical skin vasculature imaging. (a) MSOT principle and image processing step. (b) MSOT images of HbR,
Fig. 6. Switchable OR-AR PAM techniques for skin vasculature and pigment imaging. (a) Switchable OR-AR PAM technique that alters the focusing mode by axially translating the depth of the fixed focus. (b1)–(b3) Autofocusing PAM technique using a voltage-controlled liquid lens. (b1) The focal sizes of the optical and acoustic beams were controlled by the input voltage at the electrodes. (b2) Phantom imaging demonstrated that autofocusing improved the imaging depth compared to the fixed focus mode. (b3) A healthy volunteer’s palm skin was imaged in different layers, where SC/SB layers and Va layers contained the melanin and vasculature, respectively. OR, optical resolution; AR, acoustic resolution; EP, epidermis; DP, dermal papillae; RD, reticular dermis; Hy, hypodermis; SC, stratum corneum; SB, stratum basale; SVP, superficial vascular plexus; DVP, deep vascular plexus; E-D, epidermis-dermis junction; VN, vascular network; PVDF, polyvinylidene fluoride. The figures were reproduced with permission from Refs. [114] [Copyright © 2020 by the author(s)] and [115] (Copyright © 2022 Wiley-VCH GmbH).
Fig. 7. Bifocal dual-wavelength PAM technique for port wine stain (PWS) characterization. (a) Two optical wavelengths, 532 and 1064 nm, were focused at different depths and superposed to provide elongated focal zone and deep imaging depth up to 3 mm. The epidermis (EP) and superficial vascular plexus (SVP) were imaged at 532 nm and deep vascular plexus (DVP) at 1064 nm. (b) Compared to normal skin, PWS skin expressed less melanin in the EP layer and more vasculature in the SVP and DVP layers. DE, dermis; Hy, hypodermis; SC, stratum corneum; ED, epidermal-dermal junction; PWS, port wine stain; PA, photoacoustic; and a.u., arbitrary unit. The figures were reproduced with permission from Ref. [118] (Copyright © 2021 European Academy of Dermatology and Venereology).
Fig. 8. Classification of melasma lesion types using PAM. (a) The PAM images are from a patient with epidermal M + V type melasma, showing denser melanin in the epidermal layer (
Fig. 9. RSOM images of a melanoma lesion and the surrounding skin tissue. (a) A photograph of a melanoma lesion. Arrows indicate the three scanning positions. The red rectangle depicts the FOV of
Fig. 10. UWSB-RSOM system configuration and cross-sectional USWB-RSOM images over 5 wavelengths. (a) Schematic of UWSB-RSOM and a structure of the imaging target (skin). (b1)–(b6) Cross-sectional RSOM images of skin over 5 wavelengths. (b1), (b2) Distribution of hemoglobin and melanin in the skin at 515 and 532 nm. Arrows indicate the position of hair. (b3) Melanin in the epidermis and hair at 650 nm. (b4) Subcutaneous fat, sebaceous glands (marked with arrows 1 and 2), and water content in the epidermis at 1210 nm; arrow 3, sebum content on the hair shaft. (b5) Water distributed in the epidermis at 1450 nm. (b6) Composite of images at all five wavelengths. ADC, analog-to-digital converter; AMP, 60 dBm amplifier; CL, collimating lens; D, dermis; EP, epidermis; FB, fiber bundle; FL, focusing lens; H, UWSB-RSOM holder; HPF, high-pass filter; HS, hair shaft; HB, hair bulb; IU, detachable interface unit; M, mirror; PBS, polarizing beam splitter; S, schematic of human skin; SF, subcutaneous fat; SG, sebaceous glands; UT, ultrasound transducer. The figures were reproduced with permission from Ref. [121] (Copyright © 2019 Optical Society of America).
Fig. 11. PAT techniques for melanoma imaging. (a) Clinical photograph of a melanoma. (b) 3D PA scan image of the black dashed region in (a). (c) Co-registration of PA and B-mode US images of
Fig. 12. PAT techniques for skin lesions. (a) Comparison of combined
Fig. 13. Specification plots relating the spatial resolution, imaging depth, and lateral FOV of clinical PA skin imaging systems. The overall relationship of the imaging depth versus spatial resolution (left) and the lateral FOV versus spatial resolution (right) is presented. Dotted lines show reference tradeoff lines with a given ratio, and the blue dotted line is obtained from least-square fitting including both axial (triangle marker) and lateral (circle marker) resolutions. PAM, photoacoustic microscopy; PAMes, photoacoustic mesoscopy; PAT, photoacoustic tomography; FOV, field of view; PA, photoacoustic. (Refs. PAM- [95,97,99,100,114
Table 1. Specifications of PAM Systems for Clinical Skin Imaging.
View tableView in ArticleTable 1. Specifications of PAM Systems for Clinical Skin Imaging.
Ref. Modality Laser wavelength (nm) US transducer US frequency FOV, lateral × elevation × axial (mm × mm × mm) Spatial resolution (μm) Scan time Vasculature imaging [ 95 ]OR-PAM 532 Focused Center: 40 MHz BW: 110% Lateral: 5 Axial: 31.2 per B-scan [ 96 ]OR-PAM + PPG 532 Flat ring transducer Center: 15 MHz Lateral: 1 Axial: 1 Lateral: 2.5 Axial: 68.4 0.02 s per B-scan [ 97 ]OR-PAM + USM 532 Dual (coaxial annular, outer-PZT, inner-PVDF), focused US excitation (PZT) Center: 33.4 MHz BW: 19.5–47.2 MHz PA/US reception (PVDF) Center: 35 MHz BW: 11.8–58.2 MHz Lateral: Axial: 1.5 (PA), 5 (US) Lateral: 45 (US), 6.8–7.1 (PA) Axial: 57.8 (US) NA [ 98 ]OR-PAM + Deep learning 532 Focused Center: 20 MHz, BW: 100% NA NA [ 99 ]Bessel-beam OR-PAM 532 Focused ring transducer Center: 40 MHz BW: 120% Lateral: 6.5 (1 mm DOF) Axial: 35 NA [ 100 ]OR/AR-PAM 532 Focused annular PVDF Center: 42 MHz Lateral: 10–40 (0.2–1.5 mm depth) Axial: 34 3–5 min per C-scan Vasculature and pigmentation imaging [ 114 ]Switchable OR/AR-PAM 532 Focused Center: 45 MHz BW: 7–83 MHz Lateral: 4.4 (OR), 47 (AR) Axial: 35 0.5 s per B-scan, per C-scan [ 115 ]Switchable OR/AR-PAM 532 Autofocusing PVDF-based transducer Center: 32.8 MHz BW: 45.4 MHz Lateral: 9.8 (OR, variable with the applied voltage) Axial: NA [ 116 ]OR-PAM 532 Focused PVDF BW: 10–75 MHz Lateral: 4.2 Axial: 41 per B-scan [ 117 ]OR/AR-PAM 532 Focused annular PVDF Center: 45 MHz BW: 5–85 MHz Lateral: 3.8, 2.3, 1.5 (for , , objective lens) Axial: 34 NA [ 118 ]Dual-wavelength bifocal OR-PAM 532, 1064 Focused ring transducer Center: 30 MHz Lateral: Axial: 0–0.72 (532 nm), 0.72–3 (1064 nm) Lateral: 6.6 (532 nm), 46.8 (1064 nm) Axial: 50.6 NA [ 119 ]Commercially available PAM (PASONO-SKIN) 532 Ring transducer Center: 35 MHz BW: 114% (15–55 MHz) Lateral: 7 Axial: 50 NA
Table 2. Specifications of PAMes Systems for Clinical Skin Imaging.
View tableView in ArticleTable 2. Specifications of PAMes Systems for Clinical Skin Imaging.
Ref. Modality Laser wavelength (nm) US transducer US frequency (MHz) FOV, lateral × elevation × axial (mm × mm × mm) Spatial resolution (µm) Scan time Vasculature imaging [ 101 ]RSOM 532 Focused transducer Center: 55 BW: 10–120 Lateral: 30Axial: 8 70 s per C-scan [ 102 ]RSOM 532 Focused transducer Center: 55 BW: 10–120 Lateral: 30, Axial: 7 80 s per C-scan [ 104 ]RSOM 532 Focused transducer Center: 55 BW: 10–120 Lateral: 30, Axial: 8 70 s per C-scan [ 107 ]RSOM 532 Focused transducer Center: 55 BW: 10–120 Lateral: 30 Axial: 8 70 s per C-scan [ 108 ]FP-RSOM 532 Focused hole transducer Center: 50 BW: 10–120 Lateral: 30 Axial: 7 (same as RSOM) per B-scan 15 s per C-scan Vasculature and pigmentation imaging [ 120 ]FRSOM 532, 555, 579, and 606 Focused hole transducer Center: 50 BW: 10–120 Lateral: 30 Axial: 7 (same as RSOM) 15 s (single) 60 s (multi) per C-scan [ 121 ]UWSB-RSOM 650, 1210, 1450, 515, and 532 Two spherically focused piezoelectric transducers Center: 23.9, 55 Lateral: 4 Axial: 2 At 23.9 MHzLateral: 53 MHzAxial: 33 At 55 MHz Lateral: 29.6 Axial: 8.6 per scan
Table 3. Specifications of PAT Systems for Clinical Skin Imaging.
View tableView in ArticleTable 3. Specifications of PAT Systems for Clinical Skin Imaging.
Ref. Modality Laser wavelength (nm) US transducer US frequency FOV, lateral × elevation × axial (mm × mm × mm) Spatial resolution (µm) Scan time Vasculature imaging [ 110 ]MSOT 700 to 850 Linear array transducer Center: 4 MHz BW: 52% Lateral: 30 mm Axial: 30 mm NA [ 112 ]MSOT NA NA NA NA NA [ 113 ]LED based PAI 850 Linear array transducer Center: 9 MHz Axial: 38 mm a Lateral: 460 a Axial: 220a 20 s per B scan 1–3 min per C-scan Vasculature and pigmentation imaging [ 122 ]PAISU 680, 700, 750, 850, and 900 Linear array transducer Center: 40 MHz BW: 55% NA NA (single) (multi) per C-scan [ 125 ]PA/US 700, 756, 796, 866, and 900 Linear array transducer Center: 8.5 MHz BW: 3–12 MHz Lateral: 1000 Axial: 200 (single) (multi) per C-scan [ 126 ]PA/US 430 Single crystal transducer Center: 18 MHz BW: 67% Axial: 5 Lateral: 300 Axial: 200 per B-scan [ 127 ]PA/US Single: 430 or 530 Single crystal transducer Center: 18 MHz BW: 67% Axial: 5 Lateral: 300 Axial: 200 per C scan [ 128 ]MSOT NA NA NA NA NA NA [ 129 ]vMSOT 700, 715, 730, 760, 780, 790, 800, 825, 850, and 900 Spherical matrix array transducer NA 80 (isotropic) per C scan [ 130 ]OCT, RCM, and MSOT 700, 730, 760, 780, 800, 850, and 875 Hemispherical array transducer Center: 8 MHz BW: 60% NA per C scan [ 131 ]HFUS, RCM, OCT, and PAI NA NA NA NA NA
Table 4. PAI Analysis Features of the Skin Vasculature.
View tableView in ArticleTable 4. PAI Analysis Features of the Skin Vasculature.
Ref. Imaging target Analysis metrics Key features & findings PAM [ 95 ]Volunteer’s opisthenar and palm areas Visual assessment of the number of capillaries, depth distribution, and vessel diameter Imaging depth was greater in the palm than in the opisthenar due to higher melanin concentration in the epidermis [ 96 ]Volunteer’s finger Vascular movements and changes in the blood volume The vascular movements and the heartbeat rates measured with both OR-PAM and PPG well matched between the two modalities [ 97 ]Volunteer’s palm PA: Epidermal (0–90 µm), epidermal-dermal junction (90–210 µm), dermis (210–350 µm), and vessel plexus (350–1500 µm). US: big vessel (1.5–2.6 mm) and fibrous tissue (2.6–5 mm) The PA and US microscopic images characterized the microvasculature information and full structures of skin in vivo, respectively [ 98 ]Volunteer’s palm Axial resolution and deep vessel visibility Deep learning model improved the epidermal layer differentiation and visibility of deep vessels [ 99 ]Traumatized skin SC layer thickness and PA amplitude, SB layer PA amplitude, and dermal vessel diameter Increased thickness and PA amplitude in the SC layer, disappearance of signals in the SB layer, and development of small and dense microvasculature in the dermal layer [ 100 ]PWSpre- and post-PDT treatment Density, depth, and diameter of dermal vessels Density, depth, and diameter of lesion vessels were reduced after PDT where the effects were more significant in superficial dermis than in deep dermis PAMes [ 101 ]Allergic and irritant skin reactions in patch testing Blood volume, vessel fragmentation, and frequency content ratio Allergic reactions showed higher vessel fragmentation and lower frequency ratios compared to irritant reactions [ 102 ]UV radiation region in volunteers’ skin Vessel visibility and vessel diameter RSOM imaging revealed UV-induced recruitment of previously non-perfused vessels and vasodilation, visible as an increase in vessel diameter [ 104 ]Patient with moderate atopic dermatitis Epidermis thickness, total blood volume, and vessel diameter 32% reduction in epidermis thickness, 10% decrease in total blood volume, and 26% reduction in vessel diameter post-treatment [ 107 ]Pretibial region of the participants with diabetes mellitus and age-matched healthy volunteers Epidermal thickness, signal density of epidermis layer, dermal blood volume, vessel diameter, small vessel number, and large vessel number The analysis revealed significant differences in microvascular parameters between healthy individuals and diabetes patients, with smaller vessels ( in diameter) showing the most marked differences [ 108 ]Healthy individuals, smokers, and CVD patients Vessel diameter, blood volume, and vessel density Significant differences in microvascular ED biomarkers among healthy individuals, smokers, and CVD patients, providing a novel tool for early diagnosis and monitoring of cardiovascular risks PAT [ 110 ]Fingers of patients with SSc and healthy volunteers HbR, , and HbT Assessing microvascular dysfunction in SSc. Patients with progressive SSc had significantly lower MSOT values compared to patients with stable disease and healthy volunteers [ 112 ]Healthy ventral forearm with RH , HbR, HbT, and PAT observations of the RH identified superficial and deeper vascular structures parallel to the skin surface as part of the human skin vascular plexus. PAT revealed that the suprasystolic occlusion impacts both plexus differently [ 113 ]Patients with facial PWS “PWS level,” indicating the severity of PWS lesions The decrease in PWS levels after treatment based on PAI-association indicates better repeatability in assessing response to treatment compared to dermatologist-assigned clearance scores. It is a new quantitative assessment method for PWS severity and treatment response
Table 5. PAI Analysis Features of the Skin Vasculature and Pigmentation.
View tableView in ArticleTable 5. PAI Analysis Features of the Skin Vasculature and Pigmentation.
Ref. Imaging target Analysis metrics Key features & findings PAM [ 114 ,115 ]Volunteer’s opisthenar and palm areas Melanin concentration of the SC (0–100 µm) and the SB (100–200 µm), vascular density below the SB (200–2000 µm) Melanin concentration was significantly smaller, and vascular density was higher in the palm skin than in the opisthenar skin [ 116 ]CALM before and after Q-switched ruby laser treatment PA amplitude of the SC and SB, epidermal thickness, and melanin concentration of CALM patient Melanin concentration and epidermal thickness were significantly higher in the CALM area than in normal skin. Melanin concentration was significantly reduced after laser treatment [ 117 ]CALM, PWS CALM: melanin concentration, SB thickness; PWS: microvessel density, vascularization increase CALM skin had a higher PA signal intensity and thicker SB than that of healthy skin. Dermal vessels had greater diameters and were denser in appearance in PWS skin than in healthy skin [ 118 ]PWS Melanin density in epidermal layer, vessel density and diameter in superficial dermal, and deep dermal layers In PWS, the epidermal melanin density was lower, the superficial dermal vessels were significantly increased, deep dermal vessels had more branches, and vascular diameter in superficial dermis tended to be thicker [ 119 ]Melasma PA amplitude and depth of epidermal melanin, mean dermal vascular diameter, and mean dermal vascular density The PA amplitude of the epidermis, mean vascular diameter, and density in lesional skin were significantly higher than that in non-lesional skin. Mean vascular diameter and density were important in classifying melasma types M and M + V PAMes [ 120 ]10 melanoma lesions and 10 benign nevi Vascular density, complexity, and tortuosity The FRSOM system effectively distinguished melanoma from benign nevi, providing valuable biomarkers for lesion characterization and improving diagnostic accuracy [ 121 ]Forearm skin of healthy volunteers Fat, sebaceous glands, hair follicles, and microvasculature UWB-RSOM distinguished between melanin and melanoidins in the epidermis, identified lipid distributions in the SWIR range, and visualized water content at 1450 nm PAT [ 122 ]Patients with pigmented cutaneous lesions suspicious of melanoma Tumor thickness The photoacoustically measured lesion thickness gave a high correlation with the histological thickness measured from resected surgical samples [ 125 ]Melanoma patients Tumor depth The 3D multispectral photoacoustic imaging not only provides well-measured depth and sizes of various types of melanomas but also visualizes the metastatic type of melanoma [ 126 ]Patient with skin nevi Maximal nevi depth Developed a single-head co-localized US/PAT imaging system allowing for structural and depth measurement with the application for non-invasive diagnosis of skin cancer. The C-mode measured dermal structure and melanocytic depth correlated well with the histology [ 127 ]Patients who had suspicious skin lesions identified by dermatologists as potentially indicative of skin cancer Thickness of the lesion at its thickest position The results procured in our study underscore the potential of combined ultrasound and photoacoustic tomography as a promising non-invasive 3D imaging approach for evaluating human nevi and other skin lesions [ 128 ]Patients presented with lesions suspicious of non-melanoma carcinomas Lesion dimensions The dimensions were then correlated from the measurements acquired from histology, showing a good correlation via the intraclass correlation coefficient [ 129 ]Patients with lesions suspicious of non-melanoma skin cancer Tumor depth and length A statistically significant correlation for both tumor depth and length was found between vMSOT and histologic analysis [ 130 ]Patients with skin tags , HbR, and melanin MSOT supplemented with spatial distribution of melanin and that indicated all skin tags were benign with no infiltration of vessels inside the melanin signal [ 131 ]Patients with pigmented skin lesions , HbR, and melanin OCT, RCM, and PAI in combination enable image-guided bedside evaluation of suspicious pigmented skin tumors
Tools
Get Citation
Copy Citation Text
Junho Ahn, Minseong Kim, Chulhong Kim, Wonseok Choi, "In vivo multi-scale clinical photoacoustic imaging for analysis of skin vasculature and pigmentation: a comparative review," Adv. Imaging 1, 032002 (2024)
Paper Information
Category: Review Article
Received: Jul. 15, 2024
Accepted: Sep. 20, 2024
Published Online: Oct. 29, 2024
The Author Email: Kim Chulhong (chulhong@postech.edu), Choi Wonseok (wonseok.choi@catholic.ac.kr)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20