Fig. 1. PAM modalities and their biomedical applications. (a) PAM system^[16]; (b) realizations of PAM systems^[18], I represents transmissive OR-PAM, II represents reflective OR-PAM, III represents AR-PAM based on dark-field illumination; (c) photoacoustic imaging in biomedical^[14,17-18]

Fig. 2. Experimental demonstration of subwavelength focusing and imaging using a flat acoustic superlens^[25]. (a) Flat lens composed of 124 soda cans; (b) real part of the pressure field at f=417.5 Hz; (c) pressure intensity field after compensation of losses occurring during the propagation within the lens; (d) normalized amplitude of the field in the close vicinity of the output surface proves the existence of a focusing area of λ₀/15 (red line) while the source (blue line) is λ₀/5 wide and the source of the control experiment (black line), that is, without the lens, is λ₀/1.2 wide; (e)‒(g) as Fig.2 (b)‒(d), but for two sources that play sound out of phase to demonstrate super-resolution. It clearly demonstrates the same negative refraction results with a resolution of λ₀/7

Fig. 3. Reflective switchable subwavelength Bessel beam photoacoustic microscopic system and Gaussian beam photoacoustic microscopic system, simulation and experimental results of Bessel beam parameters. (a) Reflective switchable subwavelength Bessel beam photoacoustic microscopic system and Gaussian beam photoacoustic microscopic system^[37]; (b) (c) intensity distribution of the calculated Bessel beam at the z=0 plane in the x-y and x-z planes^[38]; (d) cross-sectional image at z=0^[38]; (e) (f) intensity distribution and experimental diameter of the Bessel beam along the z-axis^[38]

Fig. 4. Principle of wavefront shaping. (a) Before wavefront shaping^[44]; (b) after wavefront shaping^[44]; (c) experimental setup for wavefront shaping^[64]; (d) comparison between linear optimization and nonlinear optimization focusing effects^[64]

Fig. 5. Specific schematic diagrams and imaging results of the Fabry-Perot interferometer (FPI). (a) Schematic diagram of an all-optical micro-ultrasound sensor composed of a rigid optical fiber-coupled FPI and an open microcavity^[67]; (b) schematic illustration of optical illumination for photoacoustic microscopy with acoustic resolution on a hair model^[69]; (c) (d) images of three individuals with slight lateral displacement at different depths of hair obtained using a commercial piezoelectric sensor and a microfiber sensor^[69]; (e) photograph of a mouse head with the scalp removed^[69]; (f) MAP image of mouse brain vasculature^[69]; (g) B-scan image along the dotted line in the y-z plane^[69]

Fig. 6. Images of the dual piezoelectric transducer's imaging results^[74]. (a) Depth image of microvasculature in a tumor model; (b) optical image of the tumor model; (c) photoacoustic cross-sectional image corresponding to the dashed line in Fig.6 (a); (d) ultrasound image corresponding to the dashed line in Fig.6(a); (e) image stained with hematoxylin and eosin (H&E); (f) three-dimensional photoacoustic image of the tumor region; (g) three-dimensional ultrasound image of the tumor region

Fig. 7. Schematic diagram and imaging results of the micro-ring resonator. (a) Schematic diagram of a polymer MRR used for a PA endoscopic probe^[78]; (b) schematic representation of the packaged MRR on a glass cover slip with optical fiber^[79]; (c) polymer MRR installed on the front of a mouse for chronic cranial window^[79]; (d) illustration of optical scanning using a transparent polymer MRR^[79]; (e)‒(g) longitudinal PAM images of the mouse cerebral cortex vasculature over 28 days^[79]

Fig. 8. Effect of synthetic aperture algorithm compared with other methods. (a) Effects of different methods on mice in vivo^[86]; (b) photo of a cube (gelatin) model with nine linear targets (denoted as T1‒T9) spaced at 1.2 mm apart^[84]; (c) photo of a chicken breast with three pencils inserted at different depths as imaging targets^[84]; (d)‒(f) 2D cross-sectional images obtained using conventional technique, SAFT, and IntC-SAFT^[84]; (g)‒(i) 2D reconstruction images obtained using conventional technique, SAFT, and IntC-SAFT, with T1‒T3 representing different pencils^[84]

Fig. 9. Overview figures. (a) Neural network-enhanced OR-PAM's imaging speed and in-vivo imaging results in mice^[95]; (b) trade-off relationship between imaging depth and resolution, with OR-PAM having the highest resolution and the lowest imaging depth, and PACT having the lowest resolution and the highest imaging depth^[96]; (c) enhancement of image signal-to-noise ratio by neural networks

Fig. 10. Three-dimensional microvascular imaging system, segmentation method, and imaging results. (a) System diagram for imaging^[100]; (b) network training process^[101]; (c) imaging results^[101]

Fig. 11. WGAN results for mouse ear vasculature and results of an adaptive enhancement method with a deep CNN prior. (a)‒(c) Input AR-PAM image of the network, output image of the network, ground truth of OR-PAM image^[102]; (d)‒(i) AR-PAM imaging result, OR-PAM imaging result, result enhanced using the FDU-Net on Fig.11 (d), enhancement result using the total variation algorithm on Fig.11 (d), result enhanced using the proposed algorithm on Fig.11 (d), signal intensity distribution along the dashed line, scale is 1 mm^[103]

Fig. 12. Network structure and denoising effect^[105]. (a) Network structure; (b) demonstration of denoising effects on mouse ear vasculature and zebrafish pigment

Fig. 13. Optical paths of different systems and corresponding imaging results. (a) Flowchart of the switchable AR-OR-PAM scanning head system, where the beam splitter divides the optical path into two parts, and each passing through a multimode optical fiber and a filter device, ultimately enabling acoustic and optical resolution photoacoustic microscopy. Manual mode switching is required^[113]; (b) flowchart of the MEMS-based system^[115]; (c) schematic of the S-OR-ARPAD system, with illustrative diagrams of optical excitation and photoacoustic signal generation in skin multi-layer structures under both OR and AR modes of the optoacoustic visual objective^[114]; (d) (e) in vivo photoacoustic images of a mouse ear obtained by AR-PAM and OR-PAM^[113]; (f) (g) maximum intensity projection (MIP) images of the mouse ear obtained by OR-PAM and AR-PAM, and the scale in the enlarged region within the dashed box is 500 μm^[115]; (h)‒(k) in vivo comparison under OR and AR modes of the S-OR-ARPAD system among continuous cross-sectional PA images of the palm region, lateral maximum amplitude projection (MAP) of the depicting slices from 0‒180 μm below the skin surface, lateral MAP of the epidermis-dermis junction from 220‒550 μm beneath the surface, and lateral MAP of the vascular network located at depths of 460‒1800 μm below the epidermis-dermis junction^[114]

Fig. 14. Microvascular detection imaging. (a) Mouse ear before injecting 4T1 tumor cells (Day 0) ^[121]; (b) maximum amplitude projection microvascular image of the mouse ear on Day 0. The dashed box represents the region of interest^[121]; (c) depth-encoded microvascular image of the mouse ear on Day 0^[121]; (d) mouse ear on Day 4^[121]; (e) maximum amplitude projection microvascular image of the mouse ear on Day 4^[121]; (f) depth-encoded microvascular image of the mouse ear on Day 4^[121]; (g) mouse ear on Day 7^[121]; (h) maximum amplitude projection microvascular image of the mouse ear on Day 7^[121]; (i) depth-encoded microvascular image of the mouse ear on Day 7^[121]; (j) (k) images of the mouse ear's microvasculature system obtained using the RUT module and traditional photoacoustic combination module^[123]; (l) visible light NB-PAM of brain vasculature in a mouse with a skull, showing the depth-encoded brain vascular network^[28]

Fig. 15. Applications of PAM at the level of skin structure, diseases, and cells. (a) Cross-sectional image of healthy skin imaging, clearly showing the layered epidermal structure (EP) and blood vessels in the dermis (DR)^[124]; (b) cross-sectional image of psoriasis skin imaging, with capillaries (shown in green) interlacing with the epidermal structure (EP)^[124]; (c) top photograph of a psoriasis skin area and photoacoustic imaging results in the dashed region (above), and a photograph of healthy skin area and photoacoustic imaging results (below), showing superficial skin depressions^[124]; (d) photoacoustic images acquired using Gaussian beam photoacoustic microscopy (GB-PAM) and extended depth-of-focus photoacoustic microscopy (E-DOF-PAM)^[125]; (e) photoacoustic microscopic image of fibroblast cells^[18]; (f) mixed FDOM/multi-photon imaging of a mouse ear after injection of melanoma cells (MC)^[128]

Fig. 16. Applications of PAM in tumor research. (a) OR-PAM images of hemoglobin concentration, oxygen saturation, blood flow speed, depth, diameter, and tortuosity in the tumor region^[14]; (b) in vivo OR-PAM image of oxygen saturation with ultrafast dual-wavelength excitation. The first and the second pulses are 532 nm and 558 nm^[130]; (c) mammograms of the affected breasts, depth-encoded photoacoustic angiogram of whole cancerous breasts, and sagittal plane image across tumor region^[131]; (d) a representative x-y maximum amplitude projection (MAP) image of a mouse brain vasculature over the entire cortex acquired by UFF-PAM at 532 nm, and the oxygen saturation of hemoglobin (sO₂) map of the same mouse brain acquired with dual-wavelength measurements at 532 nm and 558 nm^[132]

Fig. 17. Applications of PAM in cells and tissues. (a) Merged photoacoustic image of microvasculature and nanoparticles distribution in a complete mouse skull state^[133]; (b) five-wavelength OR-PAM images of the blood and lymphatic vessels in the mouse ear^[131]; (c) head-restrained PAM of cerebral CHb, sO₂, and blood flow speed in the absence (OFF) and presence (ON) of isoflurane^[121]; (d) percentage change of PA amplitude of the mouse ear after epinephrine injection and sO₂ images of the mouse ear after epinephrine injection^[134]

Fig. 18. Applications of PAM technology in cerebrovascular medicine. (a) From left to right:in vivo imaging of the total hemoglobin concentration in the mouse ear, linear sO₂ image of the mouse ear, nonlinear sO₂ image of the mouse ear^[135]; (b) high-speed imaging of the mouse brain obtained under hypoxia challenge, showing reduced blood oxygenation^[136]; (c) simultaneously acquired high-resolution maps of CHb and sO₂, respectively^[137]; (d) baseline image of cerebral oxygen saturation (sO₂) before simulated metabolic acidosis and image of cerebral sO₂ after simulated metabolic acidosis^[138]