Open Access
Add to BibSonomy
Abstract
Maintaining a high spatial resolution in photoacoustic microscopy (PAM) of deep tissues is difficult due to large aberration in an objective lens with high numerical aperture and photoacoustic wave attenuation. To address the issue, we integrate transmission-type adaptive optics (AO) in high-resolution PAM with a low-frequency ultrasound transducer (UT), which increases the photoacoustic wave detection efficiency. AO improves lateral resolution and depth discrimination in PAM, even for low-frequency ultrasound waves by focusing a beam spot in deep tissues. Using the proposed PAM, we increased the lateral resolution and depth discrimination for blood vessels in mouse ears.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Photoacoustic imaging (PAI) [1–3] is a hybrid modality combining the advantages of optical and ultrasound imaging at deep imaging depths. Optical-resolution photoacoustic microscopy (OR-PAM) with an objective lens particularly exhibits sub-micrometer and micrometer lateral resolution [4–10]. Microvasculature [7], cells [11], and melanoma [5,12] are some of the morphological characteristics visualized by OR-PAM. Tens of micrometers-diameter blood arteries in thick tissues and organs, such as the brain [8] and eye [13], require a reflection-type PAM be visualized in vivo. Reportedly, a high numerical aperture (NA) objective lens [14,15] and a high-frequency ultrasound transducer (UT) [16] have been used in PAM to improve the spatial resolution. A high-NA objective lens increases wavefront aberration, whereas a high-frequency UT limits the observation depth and impedes the photoacoustic signals.
The lateral resolution of OR-PAM is dictated by the optical focus diameter, ideally determined by the NA and wavelength. However, a wavefront aberration caused by mismatch of refractive indices in an optical route and a roughness of tissue surfaces expand the optical focus diameter inside tissues. With an increasing viewing depth, the wavefront distortion due to spherical aberration becomes considerable, especially for high-NA objective lenses [17,18], thereby reducing the spatial resolution of the PA image. A correction collar is frequently integrated into a high-NA objective lens to rectify the aberrations. However, to adjust the corrective collar, one must touch the objective lens causing a disruption, thus impacting the precision.
Similarly, several spherical aberrations can be corrected by changing the effective tube length [19] and the refractive index of the immersion medium [20]. Regardless of the observation depth, the aberrations can simply be rectified without lens adjustments or manual operation. However, these procedures are performed statically and difficult to change during the observation. Electrically-controlled adaptive optics (AO) approaches have reportedly circumvented this issue [17,18,21–25]. In prior work, we used a transmissive liquid-crystal AO element [26–29] in PAM with a high-NA objective lens [30,31]. Compared to the conventional reflection-type AO [32], transmissive AO can be simply implemented into PAM without significant changes to the optical apparatus. The electrically-controlled liquid-crystal AO element is small and cheap, using little power.
The detection frequency of the generated photoacoustic waves determines the depth resolution in PAM. The depth resolution increases as the detection frequency increases. However, if the detection frequency is increased, the detection effectiveness of photoacoustic waves decreases, whereas the depth resolution of PAM increases. As a result, high-frequency UTs are only useful for observing thin samples and cell cultures [16].
Even with low-frequency UT, we could increase the lateral resolution and depth discrimination using transmissive liquid-crystal AO in PAM in this work. The lateral resolutions and depth discrimination at deep layers in the proposed PAM were investigated by comparing PA images without and with AO correction using the USAF 1951 test target, gold wires, and blood vessels of mouse ear in vivo. The combination of AO and low-frequency UT allows for high-resolution PA imaging in deep tissues in vivo.
2. Experimental setup and methods
2.1 Reflection-type AO-PAM system with a high-NA objective lens and a narrow reflection plate
Figure 1(a) shows a schematic of the AO-PAM system with a high-NA objective lens and a narrow reflection plate. The AO element was placed near the back aperture of the objective lens. A wavelength-tunable nanosecond pulse laser was used as the laser source (NT342A-10-AW, EKSPLA; pulse duration, 5 ns; repetition rate, 10 Hz). The laser beam was collimated and subsequently expanded to have a diameter of 1.0 cm. The pulse energy was controlled with a half-wave retarder, a polarizer, and neutral density filters, and measured using a pyroelectric energy sensor (919E-0.1-12-25K, Newport). The laser beam passed through the transmissive liquid-crystal AO element and was focused on the target by the objective lens. The rising edge of the excitation pulses was used as a trigger for PA signal acquisition. The measured pulse energy normalized the acquired PA signals. The sample was zigzag-scanned in 2D using a stepping-motor stage (TAMM100-50C and HSC-103, OptoSigma).
Fig. 1. (a) The AO-PAM system with a high-NA objective lens and a narrow reflection plate; (b) The layout of the narrow reflection plate and high-NA objective lens. The size of the right-angle prism is 3 mm. The size of the reflection plate is 2 × 4 mm (width × length). AO: Adaptive optics; BS: Beam splitter (reflection:transmission = 90:10 in the range of 400–700 nm); CL: Collimator lens; M: Mirror; OL: Objective lens; PA: Photoacoustic; PES: Pyroelectric energy sensor; and UT: Ultrasound transducer.
The AO device comprised of three layers of liquid-crystal molecules sandwiched between two glass substrates coated with clear electrodes (ITO: Indium Tin Oxide) [28,33]. The AO device we used in this work is described in detail [33,34]. The liquid-crystal molecules were aligned parallel to the glass substrate. On applying an alternating voltage to the electrodes, the liquid-crystal molecules tilted in the direction of the electric field. Depending on the amplitude of the applied voltage, the tilt of the liquid-crystal molecules could be controlled because the liquid-crystal molecules have a dielectric anisotropy. In other words, the refractive index of the liquid-crystal layer (phase modulation) could be continuously controlled by the applied voltage. The three layers contained different concentric patterns of electrodes. The overall phase modulation created by three liquid-crystal layers was optimized to compensate for a spherical aberration resulting from the mismatched refractive indices between water and the target sample for a 40X water-immersion objective lens (LUMPLFLN 40XW, Olympus; NA 0.8; working distance, 3.3 mm). The response time of the AO modulator was about 100 msec.
The following procedure was used to optimize the electric potential difference applied to the AO element and the sample position for the observation target: First, the sample position and applied potential difference were roughly adjusted to maximize the PA signals of the structure at target depth; second, the sample position was adjusted to maximize the slope of the PA intensity profile at the target edge while keeping the applied potential difference constant; third, with the sample position fixed, the applied potential difference was adjusted to maximize the slope at the target edge; fourth, the second and third procedures were performed repeatedly. The position of the laser focus spot and the bright-field image of the sample were recorded using a CMOS camera (DCC1645C, Thorlabs).
A low-frequency UT (10K6.4I PF15, Japan Probe; focal length, 15 mm; center frequency, 10 MHz) was positioned 15 mm from the target. The PA waves were detected perpendicular to the optical axis in the reflection detection. The narrow acoustic reflector, consisting of a cover glass cut into a 2.0-mm rectangle, was inserted along the y-direction between the objective lens and the sample, as shown in Fig. 1(b). PA signals were monitored using a digital oscilloscope (DS-5654A, IWATSU) after passing through a low-pass filter (BLP-21.4+, Mini-Circuits; bandwidth, DC-22MHz) and an amplifier (AU-1647, MITEQ; bandwidth, 0.1 K–400 MHz; gain, 57 dB).
2.2 Evaluation of a beam diameter under slide glass using transmissive liquid-crystal AO
Figure 2 shows the experimental setup to evaluate a beam diameter using PA signals depending on the depth position. Here we evaluated the performance of our transmissive element of the liquid-crystal AO. UT was placed in the water tank on the optical axis opposite to the objective lens. The sample was scanned in the lateral and depth directions. To evaluate the beam diameter for PA signal generation, we measured the PA profile at the edge of a chrome line obtained by moving the USAF 1951 test target (R3L1S4P, Thorlabs) perpendicular to the optical axis depending on the depth position using 500-nm optical pulses [Fig. 2]. The measured pulse energy normalized the acquired PA signals. The PA profile was measured at 0.1-µm lateral and 10-µm depth intervals. A slide glass (S2441, Matsunami Glass; thickness, 1.2 mm; refractive index, 1.515) was placed on the test target to evaluate beam focusing due to the spherical aberration after passing through a medium with the refractive index different from that of water. The test target and slide glass were moved together [Fig. 2(i)]. Here, the obtained PA profile at each depth position is expressed as the convolution integral of step and Gaussian functions with a full-width-at-half-maximum (FWHM) indicating the beam diameter [Fig. 2(ii) and 2(iii)]. Depending on the depth position, the fitted Gaussian functions are shown in Fig. 2(iv) and the beam diameter is symmetric concerning the focal point for no spherical aberration [35]. If the spherical aberration is corrected at the observation depth, the beam diameter must be symmetric with the focal location depending on the depth position. Under the following situations, we compared the dependence of the beam diameter on depth position: the case without the AO element without the slide glass; the case without the AO element with the slide glass; and the case with both the AO element and the slide glass. Minimum diameters (lateral FWHM) were compared for each potential change in the AO element. The depth FWHM of the intensity profile was also estimated by the distance between positions at the half peak intensity for linearly-interpolated PA intensity profile.
Fig. 2. Experimental setup to determine the beam focusing as a function of depth position to generate PA signals with the transmissive element of liquid-crystal adaptive optics. Beam profiles at depth positions were achieved by measuring the PA profiles at the edge of the chrome patterns in the USAF 1951 test target. OL: objective lens; PA: photoacoustic; and UT:ultrasound transducer.
2.3 Determination of the lateral resolution using the USAF 1951 test target in reflection-type AO-PAM
We used USAF 1951 test target as the target sample to determine the lateral resolution of our reflection-type AO-PAM at different depths. Using 500-nm optical pulses without and with the AO element, the PA images were obtained for the chrome pattern of element 6 in group 7 on the deposition surface or through the glass substrate, cover glass (C24601, Matsunami Glass; thickness, 0.12 mm; refractive index, 1.5255) and slide glass (thickness: 1.2). The intensity of each pixel in obtained PA images is the value obtained by subtracting background (average PA intensity generated from the place without the test target) and measured PA intensity. The beam profile is assumed to have a Gaussian function with a lateral resolution of FWHM. The FWHM of the beam profile was calculated by fitting the PA intensity profile of the edge on the resultant image with the convolution function of the beam profile with the step function. We compared the lateral resolutions in the x and y directions to see how they were affected by the acoustic reflector using Welch's t-test. The information is presented as means and standard deviations (SD). Significant P-values were defined as those less than 0.05. The improvement in lateral resolution was estimated as the percentage improvement (|LRw−LRwo|/LRwo) from the lateral resolution without AO correction (LRwo) to that with AO correction (LRw).
2.4 Experiment for gold wires in silicone imitating blood vessels
A silicone block (SYLGARD 184, Dow; refractive index, 1.44) was embedded with gold wires with a diameter of 30 µm (AU-171095, Nilaco) that is nearly the same as the diameter of blood capillaries. A slide glass (thickness: 1.0 mm) was mounted on the silicone block, and the block with the slide glass was placed beneath a water tank made by wrapping the dish's bottom with a wrap film. An identical experimental setup was used for in vivo imaging. 450-nm optical pulses and reflection-type acoustic detection of PA waves were used to observe the overlaying gold wires. The lateral resolution was numerically confirmed to be proportional to the reciprocal of the slope of the PA intensity profile at the edge of a thin target. Based on this fact, we evaluated the percentage improvement in the lateral resolution as the increase of slope at the edge from without AO correction to with AO correction in percentage. The slopes of the lateral profiles in the perpendicular direction to the gold wires were estimated using linear least-square fitting using data between 30% and 70% of the maximum value in the normalized PA intensity profiles. For normalized PA intensity, the measured PA intensity in the absence of gold wire was subtracted from that before normalization by the profile's maximum value. The improvement in lateral resolution was measured by comparing the average slopes at the gold wire edges without and with AO correction. To evaluate the depth discrimination, the widths of normalized PA intensity depth profiles in the perpendicular direction to the gold wires were also calculated by fitting with a Gaussian function. We also compared the PA images of the gold wires when the applied potential to AO was optimized for gold wires at the shallow and deep positions.
2.5 Visualization of blood vessels
We employed the AO-PAM technique for the PA imaging of blood vessels in mouse ears in vivo. In in vivo measurement, the whole mouse was fixed below a wrap film with the ear placed flat. We thus observed the blood vessels under the skin through the wrap film. The Saga University Animal Care and Use Committee accepted this study (Permission number: 30-058-0), and it was conducted in accordance with Saga University's Regulation on Animal Experimentation. Anesthetics were made using a mixture of 0.75 mg/kg medetomidine hydrochloride (Domitor, ZENOAQ), 4.0 mg/kg midazolam (Midazolam Injection, SANDOZ), and 5.0 mg/kg butorphanol tartrate (Butorphanol, Meiji Seika Pharma). The anesthetic solution was administered to a 5-week-old male Jc1: ICR mouse (CLEA Japan) by an intraperitoneal injection. Using 577-nm optical pulses and reflection-type acoustic detection of PA waves, blood vessels in mouse ears were observed. PA images evaluated the slopes at the edges of blood vessels and depth discrimination. The slopes were determined by the linear least-square fitting using the data between 30% and 70% of the maximum value in the normalized PA intensity profiles. Here, to obtain the normalized PA intensity, the measured intensity in the absence of blood vessel is subtracted from that before being normalized by the maximum profile value. The percentage improvement in the lateral resolution from without AO correction to with AO correction was evaluated.
3. Experimental results and discussion
3.1 Evaluation of beam focusing using the transmissive element of AO
The beam generating PA signals for various potential differences applied to the AO element is shown in Fig. 3(a)–3(e). For negligible spherical aberration for an appropriately designed water-immersion objective lens, the beam profile focusing in the depth direction was symmetric relative to the focal point [Fig. 3(a)]. The obtained lateral and depth FWHMs without any obstacles (glass plates) between the objective lens and the test target were 0.59 µm and 10.1 µm, respectively. Obtained lateral and depth FWHMs of our PAM system were different from the mathematical lateral and depth resolutions using 0.51λ/NA and 1.8λ/NA2 [36] (lateral resolution: 0.32 µm, depth resolution: 1.40 µm), respectively. This is because the incident beam did not completely fill the back aperture of the objective lens to effectively use the pulse energy. We estimated the beam diameter (FWHM) at the back aperture (diameter: 8.0 mm) of the objective lens as 3.9 mm using knife-edge method [37]. The effective NA was reduced by almost half [38], because of which, the lateral and depth FWHMs nearly (or more than) doubled and quadrupled, respectively. The beam profile was asymmetric relative to the focal point when the test target was placed just below the slide glass with a thickness of 1.2 mm. Due to the spherical aberration, the PA intensity distribution in the depth direction shifted towards the UT [Fig. 3(b)]. The profile of the beam focusing became symmetrical at a potential difference of 0.4 Vrms [Fig. 3(d)]. The beam profile was asymmetric with respect to the focal position at the potential difference was 0.2 Vrms. As seen in Fig. 3(c), the PA intensity distribution was also skewed towards the UT. The PA intensity distribution was curved towards the objective lens at the applied potential difference of 0.6 Vrms with an asymmetric beam focusing profile with respect to the focal point, as illustrated in Fig. 3(e). These findings suggest that the AO element used a reasonable potential difference to correct the spherical aberration (0.4 Vrms).
Fig. 3. (a) Obtained beam focusing on generating PA signals without the slide glass and adaptive optics (AO) element; (b)–(e) Beam focused on generating PA signals through the slide glass (thickness: 1.2 mm) with AO element at the applied potential differences in (b) 0 Vrms, (c) 0.2 Vrms, (d) 0.4 Vrms, and (e) 0.6 Vrms; (f) Lateral and (g) depth FWHMs and (h) peak PA intensities function the potential difference applied to the AO element. The solid orange lines show the optimum values estimated from measured beam focusing on the case without a slide glass, as shown in (a). SG: Slide glass.
The minimum beam diameters (depth position = 0 µm) for applied potential differences between 0, 0.2, 0.4, and 0.6 Vrms are shown in Fig. 3(f). For the test target without the slide glass, the lateral FWHM of the beam profile was 0.59 ± 0.15 µm. Without AO correction, the lateral FWHM of the test target with the slide glass (thickness: 1.2 mm) was 0.94 ± 0.26 µm. For applied potential differences of 0.4 Vrms and 0.6 Vrms, the lateral FWHMs were much smaller than 0.2 Vrms.
For the applied potential differences at 0, 0.2, 0.4, and 0.6 Vrms, Fig. 3(g) illustrates the depth FWHM of the beam profile at the center (lateral position = 0 µm). For the test target without slide glass, the depth FWHM was 10.1 ± 0.4 µm. Without AO correction, the depth FWHM of the test target with the slide glass (at a depth of 1.2 mm) was 24.7 ± 7.5 µm. The depth FWHM was improved to 9.8 ± 0.8 µm when the applied potential difference of AO was 0.4 Vrms.
The PA intensity at the focal point with the applied potential difference of 0.4 Vrms became the strongest among those for the other AO applied potential differences [Fig. 3(h)]. These results show that by optimizing the potential difference applied to the AO element for the maximized PA intensity, the lateral and depth FWHMs can be improved near the focal point.
3.2 Evaluation of lateral resolutions under glass plates in reflection-type AO-PAM
The PA images without and with the AO correction are compared in Fig. 4(a)–4(e). USAF 1951 test target with a resolution of element 6 in group 7 was used as a sample, and placed under cover (thickness: 0.12 mm) and slide (thickness: 1.2 mm) glasses. The PA image’s field of view (FOV) was 32 × 18 µm with a pixel size of 0.5 µm. The value for each pixel represents the maximum envelope value of the 8-times-averaged PA signal by subtracting background (average PA intensity generated from the place without the test target). Figure 4(f)–4(k) show the average edge PA profiles of the test target along the x-direction (solid blue line) and y-direction (solid orange line) in the PA image, respectively. The estimated lateral resolutions in the x and y directions and attenuation for PA signals obtained without and with the AO correction depending on the imaging depth are summarized in Table 1. The estimated lateral resolutions in the x and y directions at the surface of the test target without the glass plates (depth: 0 mm) were 0.59 ± 0.19 µm and 0.51 ± 0.15 µm, respectively without the AO correction [Fig. 4(f) and 4(g)].
Fig. 4. PA images of USAF 1951 test target (element 6 of group number 7) at a wavelength of 500-nm: (a) without cover and slide glasses; (b) with cover glass (thickness: 0.12 mm) without AO correction; (c) with cover glass with AO correction; (d) with slide glass (thickness: 1.2 mm) without AO correction; (e) with slide glass with AO correction. (f) and (g) respectively denote the average PA intensity profiles along with the blue (x-direction) and orange (y-direction) lines shown in (a); (h) and (i) respectively denote the average PA intensity profiles along the blue and orange lines shown in (b) without and (c) with AO correction; (j) and (k) respectively denote the average PA intensity profiles along the blue and orange lines shown in (d) without and (e) with AO correction. Scale bars are 5 µm.
Table 1. The estimated lateral resolutions in the x and y directions and PA signal attenuation without and with AO aorrection under glass plates.a
The lateral resolutions in the x and y directions were as low as 0.91 ± 0.48 µm and 1.57 ± 1.14 µm, respectively, when the cover glass was placed on the test target (depth: 0.12 mm). The lateral resolutions in the x and y directions were improved to 0.72 ± 0.18 µm and 0.61 ± 0.21 µm, respectively, when the optimum potential difference (0.06 Vrms) was applied to the AO element [Fig. 4(h) and 4(i)]. The test target patterns behind the cover glass were more clearly visible with the AO correction than without [Fig. 4(b) and 4(c)]. The average intensity of PA signals generated from chromium patterns was nearly 80% higher with the AO correction than without.
The average PA intensity without AO correction significantly reduced when the slide glass was placed on the test target (depth: 1.2 mm) [Fig. 4(d)]. When the applied potential difference was optimized at 0.42 Vrms for the test target under 1.2 mm slide glass, the AO correction increased the PA intensity by about 240% [Fig. 4(e)]. The lateral resolutions in the x and y directions were similarly improved with AO correction, from 1.20 ± 0.42 µm and 1.64 ± 1.48 µm without AO correction to 0.52 ± 0.17 µm and 0.79 ± 0.25 µm, respectively [Fig. 4(j) and 4(k)]. Here we note that the optimum applied potential difference to the AO (0.42 Vrms) differed from the 0.4 Vrms obtained experimentally, as mentioned in section 3.1 to evaluate the beam focusing using an identical 1.2-mm thick slide glass. This is because a reflector plate was not used in the beam focusing experiment, while it was inserted between the objective lens and the sample in the reflection-type AO-PAM.
The AO correction for spherical aberration according to the thickness of the glass plate improved the lateral resolutions in both x and y directions, as well as the PA intensity. However, at a depth of 1.2 mm, the lateral resolutions attained by the AO correction were slightly different in the x and y directions due to the difference in wavefront distortions in the x and y directions. To reflect the PA waves, a narrow reflection plate was put between the objective lens and the sample, as illustrated in Fig. 1(b). Because the diagonally placed reflecting plate affects wavefront differently in the x and y directions, the wavefront distortions differed in the x and y directions. Noncentrosymmetric aberrations cannot be corrected by the AO element, compensating for centrosymmetric spherical wavefront distortion. Devices that adjust for various noncentrosymmetric wavefront aberrations, such as coma and astigmatism, have been developed for transmissive liquid-crystal AO element in two-photon excitation laser scanning microscopy with AO correction [29]. In future reflection-type AO-PAM, combining these transmissive liquid-crystal AO elements could improve lateral resolution.
3.3 Depth discrimination estimated by PA imaging of gold wires embedded in silicone
As a target sample, we embedded gold wires in silicone about 100 µm deep from the surface of the silicone block placed under a slide glass of 1.0 mm thickness. The cylinder-shaped gold wires with a diameter of 30 µm, resemble blood vessels in shape. Figure 5(a) and 5(b) show the maximum amplitude projection (MAP) PA image of gold wires in the silicone block smoothed with a Gaussian filter without and with AO element, respectively. The average background intensity in the absence of the gold wires was subtracted from the intensity of each pixel in the MAP image. The focal point was set on the gold wire placed at a deeper position. The applied potential difference (0.2 Vrms) to the AO element was optimized to maximize the PA signals generated from the deeper gold wire at the closed circle indicated by the arrow in Fig. 5(a) and 5(b). The applied potential difference was fixed until the image was obtained. The PA peak intensity with AO correction was about 57% higher than that without AO correction. The FOV of the PA images was 500 × 500 µm with a pixel size of 5 µm. The laser pulse energy at the surface of the wrap film was about 840 nJ. The pulse energy per unit area was calculated as 110 J/cm2 or less using the fact that the beam diameter at focus under the glass plate (thickness: 1.2 mm) was estimated to be about 0.6 µm for PAM with AO, as shown in Section 3.1. Figure 5(c) and 5(d) show the cross-section images along the dashed line of each PA MAP image shown in Fig. 5(a) and 5(b). Under the assumption that the propagation speed of PA waves in silicone block is 1000 m/s, the PA magenta-colored images shown in Fig. 5(e) and 5(f) are created using the maximum values of the PA signals in the region of 940–960 nsec corresponding to a depth region of 1.07–1.09 mm, indicated by the magenta dashed rectangle in Fig. 5(c) and 5(d). The PA green-colored images shown in Fig. 5(g) and 5(h) are created using the maximum values of the PA signals in the region of 1020–1040 nsec corresponding to the depth region of 1.17–1.19 mm, indicated by the green dashed rectangle in Fig. 5(c) and 5(d). The PA intensity was increased at the depth used in the optimization by AO correction [Fig. 5(g) and 5(h)]. Merged images without and with AO correction are shown in Fig. 5(i) and 5(j), respectively.
Fig. 5. PA images of gold wires in the silicone block under a glass plate (thickness: 1.0 mm): PA MAP images (a) without and (b) with AO correction. The PA intensity at the closed circle indicated by the arrow was maximized to optimize the applied voltage of the AO element; (c) and (d) respectively denote the cross-section image of the dashed line in PA MAP images (a) and (b); (e) and (f) respectively denote the PA MAP intensities (a) without and (b) with AO correction at the depth regions of 1.07–1.09 mm shown in magenta; (g) and (h) respectively denote the PA MAP intensities (a) without and (b) with AO correction at the depth regions of 1.17–1.19 mm shown in green; Merged PA images (i) without and (j) with AO correction of PA image color-coded by depth; Graphs (k)–(m) show the PA intensity profiles along the solid line (I–III) of each PA image without and with AO correction of PA image color-coded by depth; Graph (n) shows the PA intensity profile along the solid white line (IV) indicated by the black triangle in cross-section images (c) and (d). The values in (k)–(n) show the diameters of gold wires by Gaussian fit. MAP: maximum amplitude projection. Scale bars are 50 µm.
The PA intensity profiles along the solid lines [I–III] of each PA image in Fig. 5(e)–5(h) are shown in Fig. 5(k)–5(m). The black and red curves indicate the PA profiles without and with AO correction, respectively. The diameters of gold wires obtained from PA images in Fig. 5(k)–5(m) were estimated by the fit using a Gaussian function. Compared to the diameters of the PA signals from the gold wires without and with AO correction in Fig. 5(k), the estimated diameter with AO correction came close to the actual diameter of gold wires. However, Fig. 5(l) and 5(m) show that the diameter of the gold wire was smaller for the deeper position than that at the shallower position. This is because only an upper part of the cross-section of the gold wire was strongly irradiated when the beam focus was on the top surface of the gold wire. The average value of the slopes estimated the improvement in a lateral resolution at the edges of PA profiles at ten locations [Table 2]. The slope at the edge of the gold wire at the depth of 1.07–1.09 mm improved from 0.029 ± 0.007 µm−1 without AO correction to 0.037 ± 0.005 µm−1 with AO correction [Fig. 5(e), 5(f), and 5(k)]. The slope at the edge of the gold wire in the depth of 1.17–1.19 mm also improved from 0.047 ± 0.008 µm−1 without AO correction to 0.059 ± 0.006 µm−1 with AO correction [Fig. 5(g), 5(h), 5(l), and 5(m)].
Table 2. The estimated alope at the edge of gold wires without and with AO correction and estimated percentage improvement of lateral resolution.a
The profiles of PA intensity along the solid white line [IV] indicated by the black triangle in Fig. 5(c) and 5(d) are shown in Fig. 5(n). The average width of the PA profile for ten places in the depth direction for the gold wire at deeper position was narrowed by 11.6% from 87.0 ± 13.5 µm without AO correction to 76.9 ± 2.9 µm with AO correction. Without AO correction, the merged image at the upper half of the deeper gold wire appeared magenta [Fig. 5(i)]. Whereas, with AO correction, the combined image at the upper half of the deeper gold wire appeared green [Fig. 5(j)]. This indicates that AO correction differentiated the gold wires at different depths with the two gold wires being around 100 µm apart. The detectable distance using a low-frequency 10-MHz UT was around 100 µm considering the velocity of sound speed in the silicone block. AO correction improved depth discrimination, even though the proposed PAM used a low-frequency 10-MHz UT. These findings suggest that even for low-frequency UT, AO correction in PAM improves lateral resolution and depth discrimination. The proposed reflection-type AO-PAM can provide precise information in deep layers.
Figure 6 compares the PA images of the gold wires with the optimized applied potential at the shallow position and that at the deep position. The PA images and intensity profiles wherein the point on the gold wire at the shallow position was used to optimize the applied potential are shown in Fig. 6(a)–6(f). When a potential difference of 0.4 Vrms was applied to the AO element, the PA intensity at the open circle indicated by the arrow in Fig. 6(a) and 6(b) was maximized. The peak of the PA intensity profile along the solid line (I) with AO correction was about 125% higher than that without AO correction [Fig. 6(e)]. Figure 6(c) and 6(d) show the cross-section images along the dashed line of PA MAP images shown in Fig. 6(a) and 6(b), respectively. The profiles of PA intensity along the solid line (II) indicated by the black triangle in Fig. 6(c) and 6(d) are shown in Fig. 6(f). Compared to the depth PA profiles of cross-section images without and with AO correction, the width of the PA profile for the gold wire at the shallow position was narrowed by 10.3% from 96.7 µm without AO correction to 86.7 µm with AO correction [Fig. 6(f)]. Conversely, the PA images and intensity profiles for the case where the point for the optimization was set to the gold wire at the deep position are shown in Fig. 6(g)–6(l). When a potential difference of 0.46 Vrms was applied to the AO element, the PA intensity at the open circle indicated by the arrow in Fig. 6(g) and 6(h) was maximized. The peak of the PA intensity profile along the solid line (III) with AO correction was about 132% higher than that without AO correction by the arrow in Fig. 6(k). Figure 6(i) and 6(j) show the cross-section images along the dashed line of PA MAP images shown in Fig. 6(g) and 6(h), respectively. The profiles of PA intensity along the solid line (IV) indicated by the black triangle in Fig. 6(i) and 6(j) are shown in Fig. 6(l). Compared to the depth PA profiles of cross-section image without and with AO correction, the width of the PA profile for the gold wire at the deep position was narrowed by 17.5% from 106.7 µm without AO correction to 88.0 µm with AO correction [Fig. 6(l)]. From these results, we inferred that AO improves the spatial resolution of PA image of the gold wire at selected depth. The optimization of the potential difference of AO not only improves the depth discrimination but also selectively improves the PA intensity of the target placed at selected depth.
Fig. 6. PA images of gold wires in the silicone block under a glass plate for different correction positions. In (a)–(d) and (g)–(j), AO correction was optimized for the gold wires at shallow and deep positions, respectively. (a) and (g) PA MAP images without AO correction; (b) and (h) PA MAP images with AO correction; (c) and (i) Cross-sectional images along the dashed lines in images (a) and (g) without AO correction, respectively; (d) and (j) Cross-sectional images along the dashed lines in images (b) and (h) with AO correction, respectively; Graphs (e) and (k) PA intensity profiles along the solid lines [(I) and (III)] in images without [(a) and (g)] and with [(b) and (h)] AO correction. Graphs (f) and (l) PA intensity profiles along the solid lines [(II) and (IV)] indicated by the black triangle in cross-section images without [(c) and (i)] and with [(d) and (j)] AO correction. The values in (f) and (l) show the diameters of gold wires by Gaussian fit. MAP: maximum amplitude projection. Scale bars are 50 µm.
3.4 Imaging of blood vessels in mouse ear
Figure 7(a) and 7(b) show the PA MAP image of blood vessels in in vivo mouse ears without and with AO correction. The average background value in the absence of the gold wires was subtracted from the MAP image. The FOV of the PA images was 750 × 600 µm with a pixel size of 5 µm. The laser pulse energy at samples was about 84 nJ. The laser pulse energy per unit area was estimated to 30 J/cm2 or less at focus, assuming that the beam diameter at focus is 0.6 µm. During the measurement, the applied voltage (0.4 Vrms) to the AO element was fixed. When a potential difference of 0.4 Vrms was applied to the AO element, the PA intensity at the closed circle indicated by the arrow in Fig. 7(a) and 7(b) was maximized. The PA peak intensity with AO correction was about 20% higher than that without AO correction. The PA images were smoothed with a Gaussian filter.
Fig. 7. Images of blood vessels in a mouse ear in vivo: PA MAP images (a) without and (b) with AO correction. The PA intensity at the closed circle indicated by the arrow was maximized to optimize the applied voltage of the AO element; (c) and (d) respectively denote the PA MAP intensities (a) without and (b) with AO correction at the depth regions of 110–140 µm shown in magenta; (e) and (f) respectively denote the PA MAP intensities (a) without and (b) with AO correction at the depth regions of 140–215 µm shown in red; (g) and (h) respectively denote the PA MAP intensities (a) without and (b) with AO correction at the depth regions of 215–260 µm shown in green; Merged PA images (i) without and (j) with AO correction of PA image color-coded by depth; (k) and (l) respectively denote the enlarged PA image of the dashed rectangle in the merged PA images (i) without and (j) with AO correction; Graphs (m)–(p) show the PA intensity profiles along the solid line (I–IV) of each PA image without and with AO correction of PA image color-coded by depth. MAP: maximum amplitude projection. Scale bars are 100 µm.
The PA MAP images at the depths of 110–140 µm (972–992 nsec), 140–215 µm (992–1042 nsec), and 215–260 µm (1042–1072 nsec) were respectively colored by magenta [Fig. 7(c) and 7(d)], red [Fig. 7(e) and 7(f)], and green [Fig. 7(g) and 7(h)], under the assumption that the propagation speed of PA waves in in vivo was set to 1500 m/s. The applied potential difference for AO correction was optimized at the depth of 110–140 µm [Fig. 7(c) and 7(d)]. Merged images are shown in Fig. 7(i) and 7(j). Figure 7(k) and 7(l) respectively denote the enlarged PA images of the dashed rectangle in the merged PA images [Fig. 7(i) and 7(j)]. From the regions surrounded by dashed circles in Fig. 7(k) and 7(l), blood vessels colored by magenta are clearly visible in the merged image with AO correction, compared with that without AO correction. This means that the blood vessel placed at different depths (depth separation of about 110 µm) is correctly distinguished by the PA image with AO correction even with the low-frequency UT.
The profiles of PA intensity along the solid lines (I–IV) of each PA image in Fig. 7(c)–7(h) are shown in Fig. 7(m)–7(p). The PA profiles with AO correction show that the lateral resolution improved because of the steeper edge and lower noise as compared to that without AO correction, as shown in Fig. 7(m)–7(p). The PA profile without AO correction [black line in Fig. 7(n) and 7(p)] has two peaks due to the low depth discrimination. The left peak is generated by the PA signals of the blood vessel at the shallower region (110–140 µm), which cannot be distinguished from the blood vessels at the deeper region without AO correction. However, with the AO correction, blood vessels at shallower and deeper regions are identified corresponding to the depth [red line in Fig. 7(n) and 7(p)]. Table 3 shows the slopes averaged at ten locations at the edges of the blood vessels without and with AO correction at different depths. Table 3 shows that at the depths of 110–140 µm, 140–215 µm, and 215–260 µm, the slopes at the edge became steeper from 0.020 ± 0.011 µm−1, 0.026 ± 0.007 µm−1, and 0.028 ± 0.012 µm−1 without AO correction to 0.043 ± 0.020 µm−1, 0.035 ± 0.009 µm−1, and 0.039 ± 0.006 µm−1 with AO correction, respectively. Conclusively, AO-PAM can improve the depth discrimination and lateral resolution for vascular imaging in vivo, even using the low-frequency UT.
Table 3. The estimated slope at the edge of blood vessels without and with AO correction and estimated percentage improvement of lateral resolution.a
4. Conclusion
We have shown that transmissive liquid-crystal AO improves the depth discrimination and lateral resolution in PAM even using low-frequency ultrasound transducer for a USAF 1951 test target under glass plates, gold wires in silicone block under a glass plate, and blood vessels in a mouse ear in vivo. The PA intensity at the target depth was improved by AO correction. The use of low-frequency UT in AO-PAM enables the observation of deep tissues in vivo with great depth discrimination.
Funding
Kao Corporation (Kao Melanin Workshop); Japan Society for the Promotion of Science (15H03036, 19K12787).
Disclosures
The authors declare no conflict of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. P. Beard, “Biomedical photoacoustic imaging,” Interface Focus. 1(4), 602–631 (2011). [CrossRef]
2. L. V. Wang and S. Hu, “Photoacoustic tomography: In vivo imaging from organelles to organs,” Science 335(6075), 1458–1462 (2012). [CrossRef]
3. J. Xia, J. Yao, and L. V. Wang, “Photoacoustic tomography: Principles and advances,” Prog. Electromagn. Res. 147, 1–22 (2014). [CrossRef]
4. H. F. Zhang, K. Maslov, G. Stoica, and L. V. Wang, “Functional photoacoustic microscopy for high-resolution and noninvasive in vivo imaging,” Nat. Biotechnol. 24(7), 848–851 (2006). [CrossRef]
5. J. T. Oh, M. L. Li, H. F. Zhang, K. Maslov, G. Stoica, and L. V. Wang, “Three-dimensional imaging of skin melanoma in vivo by dual-wavelength photoacoustic microscopy,” J. Biomed. Opt. 11(3), 034032 (2006). [CrossRef]
6. S. Hu, K. Maslov, V. Tsytsarev, and L. V. Wang, “Functional transcranial brain imaging by optical-resolution photoacoustic microscopy,” J. Biomed. Opt. 14(4), 040503 (2009). [CrossRef]
7. S. Hu, K. Maslov, and L. V. Wang, “In vivo functional chronic imaging of a small animal model using optical-resolution photoacoustic microscopy,” Med. Phys. 36(6), 2320–2323 (2009). [CrossRef]
8. J. Yao and L. V. Wang, “Sensitivity of photoacoustic microscopy,” Photoacoustics 2(2), 87–101 (2014). [CrossRef]
9. M. W. Schellenberg and H. K. Hunt, “Hand-held optoacoustic imaging: A review,” Photoacoustics 11, 14–27 (2018). [CrossRef]
10. A. B. E. Attia, G. Balasundaram, M. Moothanchery, U. S. Dinish, R. Bi, V. Ntziachristos, and M. Olivo, “A review of clinical photoacoustic imaging: Current and future trends,” Photoacoustics 16, 100144 (2019). [CrossRef]
11. A. Danielli, K. Maslov, A. Garcia-Uribe, A. M. Winkler, C. Li, L. Wang, Y. Chen, G. W. Dorn, and L. V. Wang, “Label-free photoacoustic nanoscopy,” J. Biomed. Opt. 19(8), 086006 (2014). [CrossRef]
12. C. Zhang, K. Maslov, and L. V. Wang, “Subwavelength-resolution label-free photoacoustic microscopy of optical absorption in vivo,” Opt. Lett. 35(19), 3195–3197 (2010). [CrossRef]
13. S. Jeon, H. B. Song, J. Kim, B. J. Lee, R. Managuli, J. H. Kim, and C. Kim, “In vivo photoacoustic imaging of anterior ocular vasculature: A random sample consensus approach,” Sci. Rep. 7(1), 4318 (2017). [CrossRef]
14. W. Song, W. Zheng, R. Liu, R. Lin, H. Huang, X. Gong, S. Yang, R. Zhang, and L. Song, “Reflection-mode in vivo photoacoustic microscopy with subwavelength lateral resolution,” Biomed. Opt. Express 5(12), 4235–4241 (2014). [CrossRef]
15. C. Liu, J. Liao, L. Chen, J. Chen, R. Ding, X. Gong, C. Cui, Z. Pang, W. Zheng, and L. Song, “The integrated high-resolution reflection-mode photoacoustic and fluorescence confocal microscopy,” Photoacoustics 14, 12–18 (2019). [CrossRef]
16. E. M. Strohm, E. S. Berndl, and M. C. Kolios, “High frequency label-free photoacoustic microscopy of single cells,” Photoacoustics 1(3–4), 49–53 (2013). [CrossRef]
17. J. M. Girkin, S. Poland, and A. J. Wright, “Adaptive optics for deeper imaging of biological samples,” Curr. Opin. Biotechnol. 20(1), 106–110 (2009). [CrossRef]
18. N. Matsumoto, T. Inoue, A. Matsumoto, and S. Okazaki, “Correction of depth-induced spherical aberration for deep observation using two-photon excitation fluorescence microscopy with spatial light modulator,” Biomed. Opt. Express 6(7), 2575–2587 (2015). [CrossRef]
19. C. J. R. Sheppard, M. Gu, K. Brain, and H. Zhou, “Influence of spherical aberration on axial imaging of confocal reflection microscopy,” Appl. Opt. 33(4), 616–624 (1994). [CrossRef]
20. D. S. Wan, M. Rajadhyaksha, and R. H. Webb, “Analysis of spherical aberration of a water immersion objective: application to specimens with refractive indices 1.33-1.40,” J. Microsc. 197(Pt 3), 274–284 (2000). [CrossRef]
21. M. J. Booth, M. A. Neil, R. Juskaitis, and T. Wilson, “Adaptive aberration correction in a confocal microscope,” Proc. Natl. Acad. Sci. U. S. A. 99(9), 5788–5792 (2002). [CrossRef]
22. D. R. Williams, “Imaging single cells in the living retina,” Vision Res. 51(13), 1379–1396 (2011). [CrossRef]
23. J. Tang, R. N. Germain, and M. Cui, “Superpenetration optical microscopy by iterative multiphoton adaptive compensation technique,” Proc. Natl. Acad. Sci. U. S. A. 109(22), 8434–8439 (2012). [CrossRef]
24. D. Scoles, Y. N. Sulai, C. S. Langlo, G. A. Fishman, C. A. Curcio, J. Carroll, and A. Dubra, “In vivo imaging of human cone photoreceptor inner segments,” Invest. Ophthalmol. Visual Sci. 55(7), 4244–4251 (2014). [CrossRef]
25. K. Wang, W. Sun, C. T. Richie, B. K. Harvey, E. Betzig, and N. Ji, “Direct wavefront sensing for high-resolution in vivo imaging in scattering tissue,” Nat. Commun. 6, 7276 (2015). [CrossRef]
26. K. Otomo, T. Hibi, Y. Kozawa, M. Kurihara, N. Hashimoto, H. Yokoyama, S. Sato, and T. Nemoto, “Two-photon excitation STED microscopy by utilizing transmissive liquid crystal devices,” Opt. Express 22(23), 28215–28221 (2014). [CrossRef]
27. M. Reddikumar, A. Tanabe, N. Hashimoto, and B. Cense, “Optical coherence tomography with a 2.8-mm beam diameter and sensorless defocus and astigmatism correction,” J. Biomed. Opt. 22(2), 026005 (2017). [CrossRef]
28. A. Tanabe, T. Hibi, S. Ipponjima, K. Matsumoto, M. Yokoyama, M. Kurihara, N. Hashimoto, and T. Nemoto, “Correcting spherical aberrations in a biospecimen using a transmissive liquid crystal device in two-photon excitation laser scanning microscopy,” J. Biomed. Opt. 20(10), 101204 (2015). [CrossRef]
29. A. Tanabe, T. Hibi, S. Ipponjima, K. Matsumoto, M. Yokoyama, M. Kurihara, N. Hashimoto, and T. Nemoto, “Transmissive liquid-crystal device for correcting primary coma aberration and astigmatism in biospecimen in two-photon excitation laser scanning microscopy,” J. Biomed. Opt. 21(12), 121503 (2016). [CrossRef]
30. Y. Notsuka, M. Kurihara, N. Hashimoto, Y. Harada, E. Takahashi, and Y. Yamaoka, “Photoacoustic microscopy with transmissive adaptive optics using liquid crystal,” Proc. SPIE 10685, 106853N (2018). [CrossRef]
31. Y. Notsuka, M. Kurihara, N. Hashimoto, E. Takahashi, and Y. Yamaoka, “In vivo visualization of blood vessels in mouse ear by photoacoustic microscopy with transmissive liquid-crystal adaptive optics,” Proc. SPIE 11240, 1124039 (2020). [CrossRef]
32. M. Jiang, X. Zhang, C. A. Puliafito, H. F. Zhang, and S. Jiao, “Adaptive optics photoacoustic microscopy,” Opt. Express 18(21), 21770–21776 (2010). [CrossRef]
33. A. Tanabe, T. Hibi, K. Matsumoto, M. Yokoyama, M. Kurihara, S. Ipponjima, N. Hashimoto, and T. Nemoto, “Transmissive liquid crystal device correcting the spherical aberrations in laser scanning microscopy,” Proc. SPIE 9335, 933502 (2015). [CrossRef]
34. N. Hashimoto, Optical Applications of Liquid Crystals, L. Vicari ed., (CRC Press, 2003), pp. 150–200.
35. M. Baranski, S. Perrin, N. Passilly, L. Froehly, J. Albero, S. Bargiel, and C. Gorecki, “A simple method for quality evaluation of micro-optical components based on 3D IPSF measurement,” Opt. Express 22(11), 13202–13212 (2014). [CrossRef]
36. J. Yao and L. V. Wang, “Photoacoustic microscopy,” Laser Photonics Rev. 7(5), 758–778 (2013). [CrossRef]
37. A. H. Firester, M. E. Heller, and P. Sheng, “Knife-edge scanning measurements of subwavelength focused light beams,” Appl. Opt. 16(7), 1971–1974 (1977). [CrossRef]
38. L. Wawrezinieck, P-F. Lenne, D. Marguet, and H. Rigneault, “Fluorescence correlation spectroscopy to determine diffusion laws: Application to live cell membranes,” Proc. SPIE 5462, 92–102 (2004). [CrossRef]
