1. Introduction

Accidental invasion of foreign body into human eye is rather frequent, and penetrating injury is an important cause of visual loss. Foreign bodies in anterior chamber angle need prompt evaluation to determine the extent of injury, decision regarding surgical intervention to remove the foreign body and meticulous attention to establish the integrity of the globe is necessary. Failure to delineate the nature of foreign body and extent of injury can lead to vision threatening complication postoperatively if initial treatment is delayed. So it is important to rule out the intraocular foreign body with open eye after injury. Computed tomography (CT) is useful for detection of the metal intraocular foreign body, but it is invasive and small intraocular foreign body may be missed with the technique [1]. Magnetic resonance (MRI) can allow the detection and localization of nonmetallic intraocular foreign bodies, but it is not suitable for detecting metallic intraocular foreign body because it gives rise to strong interference artifacts [2, 3]. Ultrasound imaging is widely used in clinics because of its real-time display, zero ionizing radiation exposure, and affordable price. Unfortunately, it suffers from poor sensitivity (52.6%) and specificity (47.2%) in detecting small foreign bodies [4].

Photoacoustic imaging (PAI) is a hybrid modality capable of providing contrast similar to that of direct optical imaging techniques, but with increased penetration depth in turbid media by encoding the optical information as acoustic waves [5–7]. The technique utilizes a pulsed laser to diffusely illuminate optically absorbing objects in the turbid medium. Provided the laser pulse is sufficiently brief, thermal confinement criteria are met and cause the optically absorbing objects to most efficiently emit the absorbed optical energy as a transient bipolar pressure wave. Information related to the location, size, shape, and optical properties of the absorbing objects is present in the pressure waves. The pressure signals acquired over a number of transducers can be processed by a reconstruction algorithm such that images of the optically absorbing objects can be produced. In recent years, biomedical photoacoustic imaging has attracted considerable attention and made great progress in investigations of breast tumor tissue [8–11], brain vasculature [12–14], prostate [15], heterogeneous media [16–18], chemical tracers(e.g. carbon nanotubes, gold nanocages and nanorods) [19–21], and so on.

To be accepted as a diagnostic medical imaging technique, PAI method should provide images in a short time. Fast imaging cannot only reduce motion artifacts by respiration or heartbeat but also allow the doctor to find crucial regions of interest and to instantly diagnose the investigated region. To realize fast photoacoustic imaging, transducer array with multi-channel parallel acquisition system, and high repetition excitation sources with high-peak power are usually required. In the past decades, many researchers have focused on multi-element PAI. Oraevsky et al. applied a 128-channel photoacoustic imaging system for breast cancer diagnostics [22]. Xing et al. developed a plane transducer array with spatial phrase-controlled algorithm to obtain 3D photoacoustic images [23]. Wang et al. applied a full-ring transducer array for small animal and developed a handheld PAI imaging probe for mapping sentinel lymph nodes in real time [24, 25].

Photoacoustic tomography of foreign bodies in soft biological tissue using mechanical scanning has reported [26], and the literatures related to PAI ocular imaging also have been reported [27–29]. In this work, we present a PAI system for rapid, high-resolution photoacoustic detection of intraocular foreign bodies. To avoid the mechanical scan, an annular transducer array is designed, which is made of piezocomposite technology and consists of 256 elements arranged along a 300-deg arc with a 50-mm radius of curvature. To reduce the data acquisition time, an eight-channel data acquisition system is used to capture 256 serials signals using multiplexing with 32 laser firings. Due to limitation of receiving angle of the individual element of the array, a limited-view back projection algorithm (LFBP) is used to reconstruct the photoacoustic images. Experimental models of intraocular metal and glass foreign bodies were constructed on ex vivo pig's eyes and clear photoacoustic images of foreign bodies were obtained. Experimental results demonstrate that this photoacoustic imaging system holds the potential in clinic detection of intraocular foreign bodies.

2. Materials and methods

2.1 Imaging principle

When an annular transducer array is used to detect laser-induced photoacoustic signals, because individual transducer of the transducer array is of finite size, the directivity of transducer must be taken into account. In LFBP algorithm, the image value $s(i,j)$ in tissue pixel is defined as

2.2 Imaging system

Figure 1 depicts block diagram of the annular array photoacoustic system. A Ti sapphire (Symphotics TII, LS-2134) laser optically pumped with a Q-switched Nd:YAG laser delivers 8-12 ns pulses at 15Hz with the wavelength of 1064 nm, providing laser pulses to irradiate the sample for generating photoacoustic signals. The average laser pulse energy density is ~0.5 mJ/cm², which is within the American National Standards Institute limit for eye exposure [30]. The laser beam is positioned at the center of the transducer, homogenized by a circular profile diffuser and strikes the stage orthogonal to the imaging plane of the transducer for maximum uniformity. The transducer array consists of 256 elements arranged along a 300-deg arc with a 50-mm radius of curvature. The array was designed as building block for a 308-element closed-ring system. The array was custom fabricated by Doppler Electronic Technologies (Guangzhou, China) using piezocomposite technology for high sensitivity and signal-noise-ratio (SNR). The central frequency of the array is 7.5MHz and the bandwidth is beyond 75%. The individual element of the array features an elevation height of 8 mm and elementary pitch of 1.1mm and element inter space of 0.04 mm, and the relative pulse-echo sensitivity is −29 dB.

 

Fig. 1 Block diagram of the annular array photoacoustic system. Definitions: FPGA = field programmable gate array, MUX = multiplexer.

Download Full Size | PDF

The bottom of the annular array was sealed with elastic transparent membrane and the side face of it was sealed up with polyethylene materials to form a water tank. The samples were mounted on a sample stage and protruded into the water tank through the bottom of transducer array. A thin layer of ultrasonic coupling gel was applied to couple the sample and the membrane. The transducer array was mounted on a stepper motor to adjust the position up and down. By multiplexing the produced photoacoustic signals were captured by an eight-channel acquisition card (NI PCI5105, American). The card features a high-speed 12-bit analogue-to-digital converter with a sampling rate of 60 MHz. Due to the multiplexing 32 laser firings were required to generate a single 256-channel capture. The total time of data acquisition and transferring is within 3 s. After 256 series signals were transferred to the computer for further data processing using MATLAB (version 8.0, Mathworks, USA), images were reconstructed with the limited-field filtered back projection (LFBP) algorithm. In the experiments, the ultrasound velocity is assumed to be 1500 m/s for photoacoustic reconstruction.

3. Results and Discussion

The spatial resolution of the photoacoustic imaging system with the annular ultrasonic transducer array was tested by imaging a hair rod with diameter of 0.05mm, which was embedded in an artificial phantom (13% gelatine, 12.5% milk, and 74.5% water). The effective attenuation coefficient of the phantom is 1.2cm⁻¹, used to simulate the optical properties of human breast [6]. Figure 2(a) shows the photoacoustic reconstructed image of the phantom with the hair rod and the inserted is the photograph of phantom. Figure 2(b) is the normalization line profile of the reconstruction image shown in Fig. 2(a) with y = 8mm. The lateral resolution of the imaging system is estimated to be about 100 $\mu m$ according to the Rayleigh criterion [31].

 

Fig. 2 (a) The photoacoustic reconstructed image of the phantom with the hair rod, the inserted is the photograph of phantom; Fig. 2(b) is the normalization line profile of the reconstruction image shown in Fig. 2(a) with y = 8mm.

Download Full Size | PDF

The adult pig’s eye is a readily available well-established model in ophthalmology owing to its similar size to the human eye, approximately 22 mm in length compared to 24 mm in humans. So the experiments are performed on the ex vivo adult pig's eye. There are several advantages when the near-infrared (NIR) is used for detecting the intraocular foreign bodies. First, NIR light illumination is comfortable to the eye. Second, according to the ANSI safety standard, the NIR light has higher ocular damage threshold than visible light, i .e. the NIR light is much safer than the visible light for ocular imaging; third, the NIR light can provide better penetration than the visible light, which is helpful for detection of deep intraocular foreign bodies. So the wavelength of 1064 nm is employed in the experiments. In order to demonstrate the photoacoustic imaging system has the capacity of imaging the anatomical structure of the eye, the first experiment was performed. Figure 3 is the 2D photoacoustic image of the pig's eye and the inserted is the photograph of the pig's eye. The globe itself is a cyst-like structure filled with optically transparent media, the aqueous fluid and the gel-like vitreous, and the major chromophore of the ex vivo eye is the pigment, which absorption is significant at the wavelength of 1064 nm. So the photoacoustic image visualizes the distribution of pigments (melanin, and possibly lipofuscin and visual pigments) of the pig's eye.

 

Fig. 3 2D photoacoustic image of the pig’s eye and the inserted is the photograph of the pig’s eye.

Download Full Size | PDF

In the second experiment, a pure sterile copper wire (0.3mm diameter, 8mm length) was inserted into the pig’s eye to construct a model of intraocular metal foreign body. Figure 4(a) is the photoacoustic image of the intraocular copper-wire foreign body and the inserted is the photograph of the copper wire. Figure 4(b) is the ultrasound image of the same sample. The white arrows pointed are the results of photoacoustic and ultrasound imaging respectively. From two images, we can see that the copper wire in the photoacoustic image can be easily distinguished and difficultly distinguished in ultrasound image because ultrasound trailing echo generated by ultrasonic penetration through the metal seriously affect the recognition and extraction of characteristic signals. This experiment indicates that photoacoustic imaging has the ability of detecting the intraocular metal foreign body.

 

Fig. 4 (a) is the photoacoustic image of the intraocular copper-wire foreign body and the inserted is the photograph of the copper wire. Figure 4(b) is the ultrasound image of the intraocular copper-wire foreign body.

Download Full Size | PDF

X-ray, CT is the mass screening tool for detecting radiopaque foreign bodies, but repeated x-ray exposure can be harmful to the human body despite using low doses of radiation. Most importantly, x-ray contrast is not appropriate for detecting radiolucent substances, such as wood, cloth, and plastic. So it is of great significance to develop a noninvasive imaging technique to detect the intraocular radiolucent foreign bodies with good contrast. In the third experiment, a piece of small black glass was carefully inserted into the pig’s eye to simulate the intraocular radiolucent foreign body. Figure 5(a) is the photoacoustic reconstructed image, and the inserted is the photograph of the glass; Fig. 5(b) is the ultrasound reconstructed image. The white circles marked are the results of photoacoustic and ultrasound imaging respectively. As we can see the photoacoustic image can be well agree with the sample, while the ultrasound image cannot. Moreover, photoacoustic imaging technique can directly display the glass foreign body in the photoacoustic image, while the ultrasound imaging technique needs the operator to hunt and judge. This experimental result demonstrates that photoacoustic imaging also has the capacity of detection of the small intraocular radiolucent foreign body.

 

Fig. 5 (a) Photoacoustic image of the black glass intraocular foreign body, and the inserted is the photograph of the glass; (b) ultrasound image of the black glass intraocular foreign body.

Download Full Size | PDF

We have demonstrated that the photoacoustic imaging system can be used as a tool of detection of intraocular foreign bodies. Photoacoustic imaging is sensitive to intrinsic and extrinsic optical contrasts and is a promising imaging technique that is partly useful for imaging optically deep intraocular foreign bodies without compromising the spatial resolution. Traditional imaging techniques of detecting intraocular foreign bodies, such as CT, MRI, ultrasound, have their superiorities, inevitably also have their drawbacks. Photoacoustic imaging also has its limitations. For example, it is difficult to detect the transparent or low optical absorption foreign bodies; it is also difficult to detect the deeply intraocular foreign bodies, and so on. Our goal is not to replace the traditional imaging technique of detection of intraocular foreign bodies with photoacoustic imaging, but to provide a possible approach as the supplement of traditional imaging technique. Combined images of photoacoustic and ultrasound can potentially increase the sensitivity and specificity for detecting intraocular small foreign bodies.

At present, we successfully applied a large size annular transducer array to detect the intraocular foreign bodies in pig’s eye ex vivo. But there was difficulty for this system in detecting the intraocular foreign bodies in vivo because of the large size and the configuration. In the future, we will design a new annular transducer array with inner diameter of 25mm, which is near to the size of the human eye. The transducer array can directly place on the eye. Applying 128 or 256-channel parallel acquisition system, the system could conveniently detect the intraocular foreign body, like the camera takes pictures with a photoflash.

4. Conclusion

We developed an annular array photoacoustic system optimized for detection of intraocular foreign bodies. The experimental results demonstrate photoacoustic imaging is a viable noninvasive technique for detecting the intraocular foreign bodies. To our knowledge, this is the first demonstration of photoacoustic imaging of intraocular foreign bodies. We hope that this work can stimulate further work in the area of photoacoustic ocular imaging.