1.

INTRODUCTION

Photoacoustic imaging that detects ultrasonic waves induced by absorption of nanosecond laser pulses attracts attention as a new modality of medical diagnosis owing to the larger penetration depth than optical imaging methods and the higher resolution than ultrasonic ones^1,2. It is expected, specially, to observe blood capillaries located at relatively deep positions that cannot be seen with conventional modalities and this enables detection of cancer tissues³ and other lesions of cardiovascular system or digestive organs by combining with endoscopes. Wang et al. developed an endoscopic side-viewing probe that uses a rotating mirror for scanning an excitation laser beam in the radial direction and an ultrasonic transducer at the distal end of the probe for detecting the photo-acoustic signal. In addition, they developed a hybrid imaging probe combining a photo-acoustic device and an ultrasonic imaging probe^4-6. Zemp et al. reported development of an optical-fiber based, handheld photo-acoustic microscope. They used a bundle of optical fibers to obtain optical-resolution images; therefore, the system has no scanning mechanism at the distal end^7,8. Beard et al. developed an all-optical photoacoustic imaging probe with a coherent fiber bundle. For detecting acoustic waves, a polymer-film-based Fabry-Perot (FP) sensor is placed at the distal end. They combined the fiber bundle with a multimode optical fiber to obtain photoacoustic images without an ultrasonic transducer⁹.

Our group proposed an all-optical photoacoustic probe composed of a bundle of hollow-optical fibers for delivery of excitation laser light and a single-mode optical fiber with thin polymer film at the distal end to detect photo-acoustic signal. This probe utilized hollow-optical fibers with an inner diameter of 320 μm. Photoacoustic images were obtained by scanning excitation beam at the input end of the fiber bundle and this eliminated a scanning mechanism at the distal end. In addition, owing to the extremely small-numerical aperture of hollow optical fibers, a parallel output beam was obtained and this brought about acquisition of three dimensional images. We succeeded in obtaining 3-D images of a phantom of blood vessels with a diameter of 1 mm. However, the resolution was not enough for observation of blood capillaries because the imaging resolution that depends on the size of hollow optical fibers was limited to the sub millimeter order.

In this paper, we construct imaging probes with hollow optical fibers with an inner diameter of 100 μm to improve the resolution. In the photoacoustic imaging probe with hollow optical fibers, the resolution greatly depends on the fiber size because the divergence angle of the excitation beam emitted from the hollow optical fiber is very small. Furthermore the flexibility of the probe is expected to be largely improved by using the thin hollow optical fibers. For photoacoustic endoscopy, we developed imaging probes using a bundle of ultra-thin hollow optical fibers and performed imaging experiments using biological samples.

2.

DESIGN AND FABRICATION

Figure 1 shows a schematic of our photoacoustic imaging system that uses a bundle of hollow optical fibers to radiate an excitation laser beam on the sample. The input ends of the fiber are aligned so that, beam scanning is easily performed by using a linear-motion stage at the input end. As mentioned, a lens is not necessary at the output end because a nearly parallel beam is obtained from the hollow optical fibers, and this enables 3-D photoacoustic imaging. A fiber-optic probe with a polymer film that functions as a Fabry-Perot interferometer attached at the distal end is utilized to detect excited photoacoustic waves, and this comprises the all-optical photoacoustic imaging system.

Figure 1.

Schematic of the all-optical photoacoustic imaging probe with a hollow-optical fiber bundle.

We fabricated thin hollow optical fibers with an inner diameter of 100 μm by depositing silver thin film on the inside of glass capillaries using a liquid-phase deposition technique^11,12. To launch laser light into the thin hollow optical fiber with high coupling efficiency, we chose, as a laser source, a Q-switched, microchip laser (Hamamatsu L11038-12, 532-nm wavelength, 1.2-ns pulse width, and 100-pps repetition rate) that provides high beam quality and the beam is focused on the input end of fiber by a plano-convex lens with a focal length of 76 mm to excite the lowest order mode optimally in the hollow optical fibers¹³. The measured coupling loss was 1.8 dB for 100-μm core hollow optical fibers, and the transmission loss was 2.8 dB for the 27-cm long fiber.

First, we evaluated the lateral resolution of imaging systems based on hollow optical fibers. An almost parallel output beam was obtained from hollow optical fibers because of the extremely small numerical apertures (NA < 0.05). Therefore, assuming that the output beam has a Gaussian profile, beam radius w at distance z from the fiber’s output end is described as

where λ is the wavelength, and w₀ is the beam radius at the output end that is 64% of the fiber radius for the lowest order mode.

In the experiment, we placed a tungsten wire with a diameter of 50 μm at a depth of 10 mm in pure water and measured the photoacoustic amplitude by scanning the output end of a hollow optical fiber that was placed on the water surface. To detect acoustic waves, a PVDF needle hydrophone (Muller, Platte Needle Probe) was put inside the water. Figure 2 shows the measured amplitude profile of the photoacoustic waves as a solid line and a Gaussian curve fitted to the measured data as a dotted line. From this fitted data, the measured lateral resolution was found to be 106 μm at FWHM, which coincides well with the beam diameter calculated using Eq. (1).

Figure 2.

Measured amplitude profile of photoacoustic signal in lateral direction from a 50-μm tungsten wire placed 10 mm in water.

We also evaluated the imaging resolution in biomedical tissues using a 1%-intralipid solution having almost the same scattering coefficient as that of biomedical tissues^2,14,15. Figure 3 shows the amplitude profile of the photoacoustic waves from a 50-μm diameter tungsten wire in the solution at depths of 200 and 400 μm. These results show that the lateral resolution does not significantly change based on the depth, although the amplitude decreases due to absorption and scattering. From these results, we can expect that a photoacoustic imaging system based on hollow-optical fibers will be useful for visualizing blood vessels in the submucosal layer with diameters of around 100-μm.

Figure 3.

Measured amplitude profiles of photoacoustic signals in the lateral direction from a 50-μm tungsten wire placed at depths of 200 and 400 μm in a 1%-intralipid solution.

3.

PHOTOACOUSTIC IMAGING

For photoacoustic imaging, we fabricated a bundle of 37 hollow-optical fibers with a diameter of 100 μm—for which the output end is shown in Figure 4. The outer diameter of the bundle was 1.2 mm, and the minimum bending radius was around 10 mm. We used copper wires with a diameter of 150 μm as samples, and two wires dipped in pure water in an “X” shape, as shown in Figure 5(a). There was a gap of 700 μm between the wires in depth direction. The pulse energy of the excitation laser was around 100 μJ/pulse, and the acoustic signal was detected using the PVDF needle hydrophone.

Figure 4.

Output end of a bundle composed of 37 hollow optical fibers with an inner diameter of 100 μm.

Figure 5.

Photoacoustic images of copper wires of 150-μm diameter placed with a gap of 700 μm: (a) outlook of the sample, (b) C-mode image, (c) B-mode image, and (d) three-dimensional image.

Figure 5(b) shows a C-mode image of the wire placed in front of another one. This image was obtained by mapping the amplitude of photoacoustic signals from each fiber element onto closely packed hexagonal lattices after applying digital filtering by using MATLAB software. Figure 5(c) is a B-mode image that was calculated using envelope demodulation of the C-mode images. Figure 5(d) shows a three-dimensional image created by reconstructing superimposed B-mode images using Image J software. The image resolution was improved from that of our previous system by using the bundle of a thin hollow optical fibers, and we succeeded in identifying two metal wires placed at different depths.

Next, we tried to obtain photoacoustic images of blood vessels in biological samples. In the experiment, a piezoelectric zirconate titanate (PZT) probe that has higher sensitivity than PVDF probes was used with a preamplifier for reducing the excitation laser power so as not to cause damage to the biological samples. Also, prior to photoacoustic imaging experiments, the positions of the output ends of each fiber element were observed and recorded by using a conventional CCD camera. The amplitude of the photoacoustic signals was set at the recorded positions so as to obtain images that exactly reflect the excitation patterns of the laser beam. In addition, to suppress the effect of variation of fibers’ transmission losses, we corrected photoacoustic signals measured for samples beforehand using a reference amplitude pattern that was obtained for a uniform medium such as a colored plastic tape.

Figure 6(a) shows outlook of ovarian membrane of fish (cod roe). The imaging area is shown as dot line. Figure 6(b) is a C-mode photoacoustic image. Figure 6(c) shows the image after smoothing one shown in Figure 6(b). The blood vessels whose thicknesses is approximately 250 μm was clearly observed. It was also seen that the changes in the photoacoustic amplitude reflect the differences in hemoglobin concentration of the vessel. In this experiment, the pulse energy of the radiated laser beam was as low as 0.2 μJ/pulse. This corresponds to the fluence of 2.8 mJ/cm², which was much lower than the American National Standards Institute (ANSI) safety limit (20 mJ/cm²)¹⁶. Therefore, we found that our system can be safely used in the photoacoustic imaging of biological tissues.

Figure 6.

Photoacoustic images of blood vessels in the ovarian membrane of fish (cod roe) of 250-μm diameter: (a) outlook of the sample, (b) C-mode image, and (c) After smoothing C-mode images.

4.

CONCLUSION

We developed a photoacoustic forward-viewing imaging probe using a bundle of thin hollow optical fibers to improve the image resolution. The diameter of each hollow fiber was 100 μm, and a measured lateral resolution of 106 μm was obtained owing to a parallel output beam from hollow optical fibers with an extremely small numerical aperture. Experiments conducted using a 1%-intralipid solution found that the resolution does not change substantially in highly diffusing biomedical tissues. We performed three-dimensional imaging of copper wires of 150-μm diameter placed 700 μm deep by using a thin-hollow fiber bundle with 37 elements and we also successfully obtained photoacoustic images of blood vessels in the membrane of cod roe as biomedical tissues. By combining the hollow fiber bundle with an optical-fiber-based acoustic probe, an all-optical photoacoustic probe with high flexibility is expected for endoscopic imaging. We are currently working on improving field of view of probe and the sensitivity of the fiber-based acoustic probe. The results of the combined system will be reported elsewhere.