1. Introduction

As two important biomedical optical imaging modalities, optical coherence tomography (OCT) [1, 2] and autofluorescence (AF) microscopy [3, 4] image different yet complementary contrasts of biological tissues. In ophthalmic applications, spectral-domain OCT has become an indispensable imaging modality for the diagnosis of retinal diseases by providing high-resolution three-dimensional imaging capabilities based on mainly the scattering contrast. AF imaging maps the distribution of endogenous molecules based on detecting the AF emission when stimulated with photons of suitable wavelengths. AF imaging has been demonstrated to be able to detect the accumulation of lipofuscin within the retinal pigment epithelium (RPE), which is related to aging of the retina and the physiopathology of age-related macular degeneration (AMD) [5, 6].

Dual-modal ophthalmic imaging [7, 8] by combining OCT and AF could potentially be a valuable tool for the research and clinical diagnosis of AMD. Clinical ophthalmic imaging system manufacturers like Heidelberg and Topcon provide both OCT [9, 10] and AF [11, 12] instruments and have the option to combine these two imaging modalities. In non-ophthalmic imaging field, several groups have investigated the combination of OCT with AF imaging in different applications including cancer detection [13–16], heart imaging [17], morphological and biochemical tissues characterization [18, 19], lung tissue imaging [20], and tissue imaging with needle probes [7, 8, 13–21]. In these studies, two different light sources at different wavelengths were used for OCT and fluorescence excitation, i.e. OCT uses a light source in the near infrared (NIR), AF uses a visible (VIS) light source, e.g. 488 nm. Combined OCT and multiphoton microscopy using a single broadband ultrafast laser has also been reported for ex vivo and in vivo applications [22, 23]. Visible OCT (VIS-OCT) was first reported in 2002 without showing imaging of biological tissues [24]. Other groups have investigated the application of VIS-OCT in imaging the spectral contrast of the retinal nerve fiber layer [25], retinal oximetry [26], and molecular imaging [27].

In our previous study we achieved simultaneous dual-modal imaging of SD-OCT and AF microscopy with a single broadband light source in the visible spectrum (415 nm center wavelength and 8 nm bandwidth, depth resolution: ~12 μm) [25]. Since many different biomolecules fluoresce when illuminated with light at 415 nm, the system is not suitable for specific in vivo lipofuscin imaging. In addition, the depth resolution of the previous VIS-OCT was not satisfactory because the bandwidth of the light source was limited to 8 nm, which was achieved by frequency doubling of a broadband ultrafast laser.

For lipofuscin imaging, fundus AF signals can be excited between 430 to 600 nm, but the highest excitation efficiency is obtained with wavelengths from 430 to 510 nm. Commercial AF systems typically use an excitation wavelength of 488 nm [28]. As a result, to achieve higher lipofuscin AF excitation efficiency and be compatible with commercial retinal AF imaging systems, a center wavelength close to 488 nm should be selected for the dual-modal OCT and AF imaging system. In this paper, we report on our work on developing an improved AF-OCT system for dual-modal imaging of the retina. The goal of the study is to demonstrate that simultaneous OCT and AF imaging with good image quality can be achieved with a single broadband light source centered at 480nm, which has not been demonstrated before, and this system can be successfully applied to retinal imaging in vivo.

2. Methods

2.1. Imaging system

Figure 1(a) shows a schematic of the experimental system. A broadband supercontinuum laser (SuperK EXTREME, NKT Photonics) was used as the light source together with a band selection module (SuperK VARIA). Figure 1(b) shows the measured spectrum of the selected output from the light source that was used for the imaging system. The output from the light source was coupled into the source arm of a single-mode optical fiber-based (fused 2 × 2 fiber coupler with center wavelength of 514 nm, OZ Optics, Ottawa, Canada) Michelson interferometer. After exiting the sample arm, the light was collimated, reflected by a dichroic mirror (DMLP505, cut-on wavelength: 505 nm, Thorlabs), scanned by a X-Y galvanometer scanner, and then delivered to the retina by the combination of a relay lens and an ocular lens. The lateral resolution of the system is calculated to be 7.3 μm in the retina. In the reference arm, a glass slab is used to compensate group-velocity dispersion mismatch between the reference and sample arms. In the detection arm, the combined reflected light from the sample and reference arms were collimated and detected by a spectrometer, which consisted of a 1800 line/mm transmission grating, a multi-element imaging lens (f = 150 mm), and a line scan CCD camera (Aviiva-SM2-CL-2010, 2048 pixels with 10 μm pixel size operating in 12-bit mode, e2V). An image acquisition board (IMAQ PCI-1428, National Instruments) acquired the interfering spectrum captured by the camera and transferred it to a workstation (DELL Precision T7500, 4 GB memory) for signal processing and image display. The linear CCD camera in the OCT spectrometer was operated at a line rate of 24 kHz.

 

Fig. 1 (a) Schematic of the experimental system. C1-C3: Collimator; L1-L3: Lens; DM: Dichroic mirror; LPF: Long-pass filter; PC: Polarization controller; PMT: Photomultiplier; (b) The measured spectrum of selected output from the light source.

Download Full Size | PDF

The VIS-OCT has a calibrated depth resolution of 5.8 μm and an imaging depth of 1.6 mm in the air when we selected a center wavelength of 480 nm and a Full-width-at-half-maximum (FWHM) bandwidth of 20 nm as shown in Fig. 1(b). The light power in front of the eye was measured to be 545 μW, which was below the ANSI safety limits for eye imaging [29]. When raster scan was performed for imaging the rat retina the maximum scanning angle was ± 9°, the scanning range was estimated to be 1.9 mm × 1.9 mm. The OCT sensitivity was measured to be ~85 dB, which is not as high as those NIR OCT system using a more stable SLD light source.

For AF imaging the back-traveling fluorescent photons emitted from the sample passed through the dichroic mirror and a long-pass filter (FGL515M, cut-on wavelength: 515 nm, Thorlabs), and was then focused into a 25 μm pinhole by an achromatic doublet with a focal length of 30 mm (L3). The AF photons were detected by a PMT module (PMM02, Thorlabs). The outputs of the PMT were digitized by a multifunction data acquisition board (DAQ, PCIe-6361, National Instruments) at a sampling rate of 1M/s. At each scanning position on the retina a total of 40 points corresponding to a sampling length of 40 µs were acquired. The amplitudes of the AF signal of these 40 points were averaged to form one pixel of the AF image. Synchronization of the AF data acquisition, scanning of the galvanometer scanner and the OCT image acquisition was controlled by the multifunction DAQ board.

2.2. Animal imaging

To test the capability of the dual-modal imaging technique we imaged the retina of albino rat (Sprague Dawley, Taconic). All the experiments were performed in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and with the guidelines of the Florida International University’s Institutional Animal Care and Use Committee.

The animals were anesthetized by intraperitoneal injection of a cocktail containing ketamine (54 mg/kg body weight) and xylazine (6mg/kg body weight). The pupil was dilated with 10% phenylephrine solution. A powerless contact lens was applied to the eye to prevent cornea dehydration and cataract formation. The rat was restrained in an animal mount, which was placed in a multi-axis animal positioning system.

3. Results and discussion

Figure 2 shows a typical VIS-OCT cross-sectional image of the rat retina acquired with the system. The image consists of 2048 A-lines (depth scans). Compared to the images acquired with the system at 415 nm [25] the cross-sectional image shown in Fig. 2 has better depth resolution (improved from 12μm to 5.8μm) resulting better visualization of the retinal layers. In addition, stronger signals from the deeper layers like the choroid can be observed. Compared to OCT images in the NIR [30], the blood vessels of VIS-OCT image cast much clearer shadows on the retinal layers behind them due to much higher hemoglobin absorption in the visible spectrum.

 

Fig. 2 VIS-OCT cross-sectional image of a rat retina consisting of 2048 depth scans. Bar: 200 μm.

Download Full Size | PDF

Figure 3 shows the simultaneously acquired VIS-OCT [Fig. 3(a) and 3(c)] and AF [Fig. 3(b)] images of the rat retina. Figure 3(a) is the OCT fundus image generated from the 3D VIS-OCT data set [31]. Figure 3(c) is the VIS-OCT B-scan image, the location of which is marked on Fig. 3(a) and 3(b) as a white line. The data set consists of 512 (horizontal) × 128 (vertical) A-scans. Since the OCT and AF images are generated from the same group of photons we can see that Fig. 3(a) and Fig. 3(b) are precisely registered. At 480 nm the dominant fluorophore in the eye is lipofuscin [32]. Lipofuscin is a product of photoreceptor phagocytosis and is accumulated in the RPE layer [33], which is behind the retinal vessels in reference to the direction of propagation of light entering the eye. The dark appearance of the retinal blood vessels means the fluorescent signals were from retinal layers behind the retinal vessels, so we can confidently believe that Fig. 3(b) represents the distribution of the concentration of lipofuscin in the RPE cells. To test the capabilities of the imaging system for longitudinally monitoring lipofuscin accumulation in the RPE cells, we followed three 10-week-old normal rats for 4 weeks. During this period the rats grew from a mean weight of 206g to 251g. We imaged the same area of the retina bi-weekly. To make sure a fair comparison among the different time point we did our best to keep the imaging conditions the same, including the focusing and the power of the probe light, the position of the reference arm, and the depth of the retina in the OCT image.

 

Fig. 3 VIS-OCT and AF images simultaneously acquired from a rat retina in vivo. (a) OCT fundus image; (b) AF image; (c) OCT B-scan image. Bar: 200 μm. The white line in the OCT fundus image marks the location of the OCT B-scan image.

Download Full Size | PDF

Figure 4 shows the OCT fundus images, AF images and the histogram of AF intensity distributions of the same eye of one rat from 10 weeks old to 14 weeks old. The patterns of retinal blood vessels look not exactly the same because the rat body is soft and it is impossible to place the head in the same position for imaging at different times. To help locate the same vessels we added numbers (1, 2 and 3) for three vessels in both OCT fundus images and AF images.

 

Fig. 4 OCT fundus images (a, d, g), AF images (b, e, h) and histograms of the AF intensity distributions (c, f, i) of one rat at different ages: (a) – (c) 10 weeks, (d) – (f) 12 weeks, (g)-(i) 14 weeks. Numbers 1, 2 and 3 mark the corresponding vessels at different time point.

Download Full Size | PDF

From Fig. 4, we can see that the intensity of AF signals increased strongly with aging during the four week period while the intensity of OCT fundus image remains relatively constant. Table 1 shows the mean intensities of both the OCT and AF images of the rat retina shown in Fig. 4. From the table, we can see that the mean intensity of the OCT image changed among the different time points, but the percentage change from the previous measurements is very small (under 5%). In contrast, the mean intensity of the AF image increased strongly (36% and 37%).

Table 1. The mean intensities of both the OCT and AF images of the rat retina shown in Fig. 4

View Table

From the histograms in Fig. 4, we can clearly see that more and more pixels have higher intensity counts with aging although at each age point the maximum is still located at lower intensity values. The reason that the maximum of each histogram locates at lower intensity values is the dominant number of dark pixels located in the optic disk and the retinal blood vessels.

The intensity of the OCT fundus images depends on the probing light intensity and the fundus reflectivity. The relatively constant intensity of the OCT fundus image showed us that the probing light intensity is constant during the imaging process. The AF signal intensity is proportional to the probing light intensity and the concentration of the fluorophore. Since the probing light intensity is constant during the imaging process, the increase of the AF signal intensity means the increase of lipofuscin concentration. As a result, we can conclude that Fig. 4 demonstrates the increase of lipofuscin concentration with aging in the rat retina.

Figure 5 shows the mean AF intensity counts and standard deviations calculated over the entire imaged area of the 3 rats at different ages. All the three rats showed the same trend of lipofuscin accumulation over time although there are individual variations in the mean AF intensity. Interestingly, all the three rats had similar mean AF intensities at the beginning of the study. The increased differences in the mean AF intensity in the following weeks among the three rats may be caused by the difference in light exposure due to the different cage locations in the housing facility. The rapid accumulation of lipofuscin in the relatively young animals may be caused by the lack of protection from melanin in the RPE cells of the albino rats.

 

Fig. 5 The mean AF intensity plus standard deviation of the 3 rats at different ages.

Download Full Size | PDF

4. Conclusion

We have accomplished simultaneous dual-modal in vivo AF and OCT imaging with a single broadband light source at 480nm, which can provide perfectly registered structural and lipofuscin autofluorescence imaging of the retina. The system was successfully tested on monitoring lipofuscin accumulation over time in the RPE cells of rat retina. Comparing to our previous work of VIS-OCT working at 415 nm center wavelength, which was served as a proof of concept for simultaneous OCT and AF imaging with a single broadband VIS light source, the current system is suitable for real ophthalmic imaging applications, which has been demonstrated in the animal study. By providing more comprehensive retinal imaging, this technique is potentially a powerful tool in the research and clinical diagnosis of age-related macular degeneration. Our next step will be applying this technique in human imaging upon IRB approval.