Visible light OCT-based quantitative imaging of lipofuscin in the retinal pigment epithelium with standard reference targets

Zahra Nafar; Rong Wen; Shuliang Jiao

doi:10.1364/BOE.9.003768

1. Introduction

Lipofuscin, a complex mixture of partially digested lipids and protein components in the retinal pigment epithelium (RPE) cells, is a major source of fundus autofluorescence (FAF) [1, 2]. It accumulates with age and is implicated in age-related macular degeneration (AMD) and Stargardt disease [3]. FAF hence is a natural biomarker that carries the information of lipofuscin content, and quantification of FAF signal could be used to assess the amount of lipofuscin in the RPE for diagnosis and monitoring disease progression [4–11]

FAF imaging has been used in ophthalmology clinics for many years. For example, hyper-autofluorescence is positively correlated with the progression of AMD and Stargardt’s macular dystrophy [12, 13]. In the case of geographic atrophy (GA), the late stage of dry AMD, advanced RPE alterations exhibit clinically recognizable patterns of hyper-autofluorescence [14], which is positively correlated with the rate of GA progression and can be analyzed semi-automatically with newly developed software by non-expert graders [15].

However, the currently available technologies are not capable to measure standardized FAF intensity [16]. The measured FAF signal by the currently available technologies could be affected by the excitation light intensity, detector sensitivity and gain of the instrument used. In addition and importantly, signals are attenuated by the ocular tissues anterior to the RPE, especially by the lens, and the attenuation cannot be measured directly. The ocular properties of tissues anterior to the RPE could be significantly different among individuals, and it changes over time in the same person. Thus, it is difficult to compare images obtained by the currently available technologies from the same person over time, or from different individuals, which hinders the clinical usefulness of FAF images [17, 18]. It is a challenge to obtain the absolute intensity of FAF.

To compare the two different FAF images taken at different times or from different individuals, the intensity of the FAF images should be standardized and free of the effects of instrumental variation and attenuation by the ocular tissues in the optical path before the RPE. Standardized FAF images can be achieved by measuring the absolute FAF intensities, or the intensity of each pixel in a FAF image is referenced to an intensity standard. The issue of FAF intensity calibration and instrument fluctuation compensation has been addressed by Delori et al. [18] by using a fluorescent target with known emission efficiency in the intermediate retinal imaging plane as a reference. However, how to accurately compensate for the signal attenuation by tissues in the optical path before the RPE remains a challenge.

We reasoned that compensation for signal attenuation by ocular tissues can be achieved by combining visible-light optical coherence tomography (VIS-OCT) and confocal FAF microscopy that share a common light source. Previously we explored simultaneous VIS-OCT and FAF imaging using first a frequency-doubled Ti: sapphire laser [19], and later a supercontinuum laser with a filtered output centered at 488 nm, an AF excitation wavelength used in current FAF imaging systems [20]. Recently, we demonstrated that in a dual-modal VIS-OCT and FAF system, the simultaneously acquired VIS-OCT images can be used to compensate the power variations and attenuation factors for the FAF images [21]. In such a dual-modal imaging system, the photons from the light source go through the same optical pathways and are attenuated by the same media to generate signals for OCT and FAF. Therefore, the OCT signals can be used to compensate both the illumination variation and the signal attenuation by the ocular tissues. A unique feature of the technique is that the OCT and FAF images are intrinsically registered because the images in both modalities are generated by the same group of photons. However, a missing link in this system is the relationship between the amplitude of the OCT signals and the reflectivity of the sample boundaries, e.g. the RPE of the retina.

Here, we report a new development of the VIS-OCT based multimodal imaging technology for quantitative FAF imaging. In the new system, we introduced two reference targets in the intermediate retinal imaging plane and thus the two imaging modalities are correlated quantitatively. One target has known reflectivity for OCT reference, and the other has known fluorescence efficiency for FAF reference. The fluorescence reference target has similar fluorescence characteristics of the target used by Delori [18]. The reflectivity reference has the same plastic substrate as the fluorescence target with known refractive index and emission spectrum out of the major detection wavelengths of the FAF. The system uses a single broadband light source with a center wavelength of 480 nm to provide simultaneous OCT and retinal AF imaging. True FAF signal intensities can be obtained by normalizing the OCT and FAF images from the eye to the corresponding signals from the reference standards to compensate the instrument fluctuations, including laser power and detector sensitivity changes, and the signal attenuation by the media anterior to the RPE can be compensated by AF/OCT calculation. The system has been successfully tested on phantoms, and in rat retinas in vivo.

2. Methods

2.1 Imaging system

A schematic of the multimodal imaging system is shown in Fig. 1. The system consists of two spectral-domain OCTs in the NIR and visible spectrum, respectively. The NIR-OCT is used only for alignment to reduce visible light exposure and avoid additional bleaching effects to the fluorophores. The VIS-OCT consists of a supercontinuum laser (SC, model: EXB-6, SuperK EXTREME, NKT Photonics, Denmark) with a variable band-pass filter (selected center wavelength: 480 nm, bandwidth: 30 nm). The output VIS light is coupled into the source arm of a single-mode optical fiber-based Michelson interferometer. The NIR-OCT uses a superluminescent diode as the light source (SLD-37-HP, center wavelength: 840 nm, bandwidth: 50 nm, Superlum, Russia). The NIR light is coupled to another fiber-based Michelson interferometer after passing through an optical-fiber isolator. After exiting the optical fibers in the sample arms, both the NIR and VIS light are collimated and combined by two dichroic mirrors (DM1: DMLP505, Thorlabs, and DM2: NT43-955, Edmund Optics). The combined light beam is scanned and delivered to the eye by a combination of a relay lens (L1, f = 75 mm, achromatic) and an ocular lens (L2, Volk lens, 60D). Two standard reference targets for fluorescence (Fluor-Ref, Microscopy Education, Texas Red) and reflectance (Fluor-Ref, Microscopy Education, DAPI) are placed in the intermediate retinal imaging plane covering about 14% of the lower part of the field of view. The fluorescence reference target is optically cemented to a Neutral Density filter (OD: 10, absorptive type) to reduce the reference fluorescence signal to the range of the FAF of rodent retina. There are two VIS-OCT reference arms for retinal and reference-target imaging, respectively. After exiting the optical fiber in the reference arm the VIS light is collimated and split into two beams by a non-polarizing beam splitter (NBS). The two light beams are reflected back by two mirrors, recombined by the NBS, and coupled back into the optical fiber. The two reference arms have different path lengths, the longer path length for imaging the retina and the shorter path length for imaging the reference target. A computer-controlled shutter (SH1, Thorlabs) was used in the reference arm with shorter path length to synchronize with the reference-target imaging. During each raster scan, the shutter was closed to block the reference light for the reference target when the retina was imaged. When the probe light was scanned on the reference target the shutter was opened to allow the reference target to be imaged with the VIS-OCT. The VIS light power was 500 μW before entering the eye while NIR light power is 600 μW.

Fig. 1 Schematic of the simultaneous VIS-OCT and AF imaging system: VIS-OCT (blue), NIR-OCT (red) and AF (green). SLD: Superluminescent Diod; SC: Supercontinuum; PMT: Photo Multiplier Tube; SPEC1-2: Spectrometer; ISO: Isolator; M1-3: Reference arm mirrors; IRIS1-2: iris; G1-2: BK7 glass plates; BS: beam splitter; FC1-2: fiber coupler; FP1-6: collimation fiber ports; PC1-2: polarization controller; GM: galvanometer scanner; DM1-2: dichroic mirrors; PH: pinhole, L1-3: lens; LPF: long-pass filter; ND filter (OD: 10).

Download Full Size | PDF

In the detection arms, reflected light from the sample and reference arms of the two OCT systems is detected by two spectrometers (SPEC1-2). The VIS-OCT and NIR-OCT use spectrometers with the same parameters described in our previous publication [21]. The axial resolution of the VIS-OCT in the air was measured to be 6 µm with sensitivity of 85 dB. The OCT signal roll-off was measured to be −8 dB in a depth of 2mm and was compensated in image processing.

FAF is detected by a PMT module (H10723-20, Hamamatsu). The fluorescence photons pass through the two dichroic mirrors as well as a long pass filter (FGL515M, cut-on wavelength: 515 nm, Thorlabs), and are then focused to a 25 µm pinhole by an achromatic doublet with a focal length of 30 mm. The outputs of the PMT are digitized by a multifunction data acquisition board (DAQ, PCIe-6361, National Instruments) at a sampling rate of 2Ms/s and a sampling length of 80 points. The amplitudes of the FAF signal of the 80 points were averaged to form a single pixel of the FAF image. Synchronization of the FAF data acquisition, scanning of the galvanometer scanner and the OCT image acquisition was controlled by the multifunction DAQ board. The imaging speed of both the OCT and FAF was set to 24,000 A-lines/s limited by the speed of the line-scan CCD camera in the spectrometer.

2.2 Mathematical model of OCT reflectance and AF signals

The OCT signal in the time domain from a sample can be expressed as

I = R_{r} I_{r} + R_{s} I_{s} + 2 \sqrt{R_{r} I_{r} R_{s} I_{s}} \int^{​} G (ν) \cos (Δ ϕ) d ν = I_{D C} + I_{A C},

where I_s and I_r are the incident light intensity in the sample and reference arms; R_s and R_r are the reflecting coefficients of the sample and the reference mirror; G is the spectral power density of the light source; I_DC and I_AC are the DC and AC components of the detected OCT signal:

I_{D C} = R_{r} I_{r} + R_{s} I_{s},

I_{A C} = 2 \sqrt{R_{r} I_{r} R_{s} I_{s}} \int_{0}^{\infty} G (ν) \cos (Δ ϕ) d ν .

We usually have:

\begin{array}{l} R_{s} I_{s} ≪ R_{r} I_{r}, \\ I_{D C} \approx R_{r} I_{r} . \end{array}

The OCT image, which is generated from the AC portion of the detected signals, is proportional to the square root of the sample reflectance (Eq. (3)). We thus have:

R_{O C T} = I_{A C}^{2} / I_{D C} \propto R_{s} I_{s}

where R_OCT represents OCT reflectance signal. The OCT signals from the RPE can be extracted by using image segmentation. For the OCT signals of the RPE, I_s in Eq. (5) is the light intensity incident onto the RPE layer, which contains the attenuation information from the anterior segments of the eye down to the retinal tissues before the RPE. The same incident light generates both the AF and the OCT signals. According to the ocular reflectance model proposed by van de Kraats et al, reflectance from the retina can be categorized into reflectance from layers of the pre-RPE, the RPE, and the post-RPE [22, 23]. Through a detailed analysis of the OCT and AF signals and in reference to the previous studies [18, 21], we can calculate the following ratio:

\frac{I_{F A F}}{I_{R A F}} = \frac{I_{0} τ_{p r e}^{2} (λ) {[1 - ρ_{p r e} (λ)]}^{2} ξ_{R P E} A_{d} \frac{π}{4} α^{2}}{I_{0} ξ_{R} A_{d} \frac{π}{4} α'^{2}} = \frac{τ_{p r e}^{2} (λ) {[1 - ρ_{p r e} (λ)]}^{2} ξ_{R P E} \frac{π}{4} α^{2}}{ξ_{R} \frac{π}{4} α'^{2}},

\frac{R_{O C T - R P E}}{R_{O C T - R}} = \frac{I_{0} τ_{p r e}^{2} (λ) {[1 - ρ_{p r e} (λ)]}^{2} ρ_{R P E} (λ) \frac{π}{4} α^{2}}{I_{0} ρ_{R} (λ) \frac{π}{4} α'^{2}} = \frac{τ_{p r e}^{2} (λ) {[1 - ρ_{p r e} (λ)]}^{2} ρ_{R P E} (λ) \frac{π}{4} α^{2}}{ρ_{R} (λ) \frac{π}{4} α'^{2}},

where, I_FAF and I_RAF are the fluorescence intensities of the retina and the reference target, respectively. R_OCT-RPE and R_OCT-R are the OCT signal intensities of the RPE and the reference targets, respectively, calculated with Eq. (5).

I_{0}

is the light intensity incident into the eye. τ, ρ,

ξ

and

A_{d}

represent transmittance, reflectance, fluorescence efficiency and detector sensitivity, respectively. Subscribes pre, RPE and R denote the pre-RPE media, the RPE layer and, the reference standard targets.

\frac{π}{4} α^{2}

and

\frac{π}{4} α'^{2}

are the solid angles that comprise the light reaching to the pinhole and hence the detector, respectively, from the retina and reference targets. The fluorescence efficiency

ξ

is defined as the product of the fluorophore concentration (C), the molecular quantum yield (Q), and extinction coefficient (

ε

), (

ξ = C \times Q \times ε \times d

) [24]. Assuming the same attenuation of the pre-RPE media for the excitation and emission wavelengths, we can calculate the ratio:

\frac{I_{F A F} / I_{R A F}}{R_{O C T - R P E} / R_{O C T - R}} = \frac{C_{L} Q_{L} ε_{L} d_{R P E} / C_{R} Q_{R} ε_{R} d_{R}}{ρ_{R P E} / ρ_{R}} .

where C, Q, ɛ, and d are concentration, quantum yield, extinction coefficient and the effective detection thickness, respectively. Subscribes L and R, respectively represent lipofuscin present in RPE and Texas Red embedded in the reference target. With the known properties of the reference target, we can see that the calculated ratio is only dependent on the optical properties of the RPE and the lipofuscin concentration present in the RPE.

2.3 Animal experiments

To assess the capability of the imaging technique and quantification method, we imaged both albino Sprague Dawley rats (SD rats, two age groups: five 2-months-old and six 14-months-old) and pigmented Long Evans rats (five 12-months-old). The experiments were conducted in agreement 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.

Intraperitoneal injection of a cocktail of ketamine (54 mg/kg body weight) and xylazine (6mg/kg body weight) was used to anesthetize the animals. Then the pupil was dilated with a drop of 10% phenylephrine solution. To prevent cornea dehydration and cataract formation a hard contact lens was put on the eye. The rat was restrained in an animal mount, which was placed in a five-axis platform.

2.4 Phantom study

To validate the theoretical model, we need to check the linearity of R_OCT/ R_OCT-R vs R_s, and the linearity of I_FAF/I_RAF vs fluorescence efficiency. The validation was accomplished by imaging a model eye consisting of an aspherical lens (f = 15mm) and a master reference target (Microscopy Education, Fluor Ref slide: Green excitation (FITC)), which was used to simulate the retina. A set of five neutral density (ND) filters (NEK01, Thorlabs,) with optical densities of OD = 0.1, 0.2, 0.3, 0.4, and 0.5 was used to simulate the attenuations before the retina. Figure 2(d) shows a schematic of the model eye in the phantom study. Imaging was performed with different powers of the light source and different detector sensitivities.

Fig. 2 Quantitative AF signals from the model eye and the standard reference at different imaging conditions. a) AF signals vs OD value of the ND filter in front of the model eye at 90% light source power and different PMT control voltage (0.50V – 0.56V); b) AF signals vs OD value of the ND filter in front of the model eye at fixed PMT control voltage (0.56V) and different light source power (75% – 90% output power); c) The calculated log₁₀(FAF/RAF) vs OD value of the ND filter in front of the model eye at all the different imaging conditions; d) Schematic of the model eye used in the experiments.

Download Full Size | PDF

The R_OCT signals of the master reference in the model eye were calculated from their OCT images and then divided by R_OCT-R calculated from the OCT images of the reflectance reference target. The calculated R_OCT/R_OCT-R is expected to be independent on the power of the light source.

Similarly, the AF signals of the master reference in the model eye were normalized to the AF signals of the fluorescence reference target. The result is expected to be independent on the detector sensitivity and the light power.

Finally, the ratio of normalized AF (qAF) to normalized OCT (qOCT), qAF/qOCT, is calculated, which is expected to be independent on the light power, the sensitivity of the detector, and the attenuation before the model eye.

2.5 Standard reference targets characterization

The fluorescence efficiency of the reference target was measured with both a commercial fluorescence microscope and our AF-OCT system by using a fluorescein solution with known concentration. The quantum yields of Texas Red and fluorescein we used for the characterization are 0.66 and 0.92, respectively, according to equivalent Alexa Fluor dyes (ThermoFisher Scientific). We can calculate the fluorophore concentration of the reference target by using $\frac{ξ_{F l u o r e s c e i n}}{ξ_{T e x a s R e d}} = \frac{C_{F} Q_{F} ε_{F} d_{f}}{C_{T} Q_{T} ε_{T} d_{T}}$ . The effective thickness in our measurement is limited by the axial resolution of confocal fluorescent microscope, which remains the same for both Fluorescein and Texas Red imaging. The calculated concentration of Texas Red is 0.01mg/ml.

Reflectance of the standard reference for OCT was also measured. The target was made of PMMA, which has a reported refractive index of n $≅$ 1.49, resulting in a theoretical reflectance of R $≅$ 0.039. The actual refractive index was measured with OCT by calculating the ratio of $O p t i c a l D i s t a n c e / A c t u a l t h i c k n e s s$ . Our measurements showed that n = 1.53 resulting in a theoretical reflectance of R = 0.044.

Reflectance of the target was also measured by using a beam splitter, a lens (f = 20mm) and a power meter. The intensity ratio of the reflected light to the incident light was measured to be 0.043, which agrees with the theoretical reflectance values with a 2.5% difference.

3. Results

3.1 System calibration

In the phantom experiments, we first investigated the influence of detector sensitivity of the system on the AF signal detection. The fluorescent intensities were measured at different detector gains (by varying the detector control voltage) to verify the capability of the technique for canceling the effects of detector sensitivity. In Fig. 2(a), the averaged FAF signals measured from the model eye and signals from the reference fluorescent target (RAF) were plotted against the OD values of the ND filters. At a given gain, the logarithm of the fluorescent signals from the master reference in the model eye linearly decreases with the OD values of the ND filters at a slope of −2, which agrees with the round-trip attenuation of light through the ND filter and the listed transmission data of the ND filters. The signals from the standard fluorescent reference target are independent of the attenuation of the ND filters and remain constant. The measured signal intensities of FAF and RAF decreased with the PMT control voltage accordingly (Fig. 2(a)).

To verify the capability of the technique to compensate for the effects of light power, we changed the power of the light source at a constant detector gain to simulate power fluctuation. Figure 2(b) shows the linear decrease of the logarithm of the FAF signal with respect to OD values of the ND filter. As the incident light power was lowered, both the FAF and RAF signals decrease accordingly. Figure 2(c) shows the calculated I_FAF/I_RAF (qAF) for all the data shown in Fig. 2(a and b). We can see that all the effects of detector sensitivity and light power are successfully eliminated. However, the attenuation by the ND filters is not compensated in qAF calculation. This result showed that conventional FAF imaging techniques cannot eliminate the attenuation effects caused by tissues anterior to the RPE, even with a fluorescence reference target placed in the intermediate retinal imaging plane.

The simultaneously acquired OCT image of the master reference in the model eye was processed and the averaged values for each acquired image are shown in Fig. 3. The calculated log₁₀ (R_OCT-RPE), corresponding to the OCT signal intensities of the master reference, decreased linearly with the OD values of the ND filters (Fig. 3(a)). Similarly, the calculated log₁₀ (R_OCT-RPE/R_OCT-R) also decreases linearly with the OD values but are independent of the light power (Fig. 3(b)). The slope of −2 of the fitted line agrees with the round-trip attenuation of the ND filters. However, the qAF/qOCT value remains relatively constant over all the experimental conditions including the different attenuation values of the ND filter (Fig. 3(c)). The results proved that the system is capable of compensating the attenuation of the ND filter, thus the attenuation by the media anterior to the RPE for quantifying FAF intensities.

Fig. 3 OCT reflectance signals from the model eye at different OD values of the ND filter and from the standard reference at different imaging conditions; a) The source power was changed from 75% to 90%; b) OCT reflectance of the model eye normalized to the reference target (qOCT) for the data in (a); c) qAF, qOCT, and qAF/qOCT ratio. d) qAF, qOCT, and qAF/qOCT ratio for the model eye with the TexasRed slide at the retinal plane.

Download Full Size | PDF

To verify the accuracy of Eq. (8), we repeated the experiment with another fluorescent slide (Microscopy Education, Texas Red) with the same reflectance as the green master slide used before, and the results are consistent with that of the FITC slide (Fig. 3(d)). The measured excitation and emission spectra of the red and green slides are shown in Fig. 4. Considering the spectrum of the supercontinuum laser (465nm– 496nm), the detection spectral range (>515 nm), and the quantum yields of the two fluorescent dyes (Q_TexasRed = 0.66, Q_FITC = 0.92), we have

\frac{C_{R} \int_{465}^{495} ε_{R} (λ) Q_{R} η_{R} d λ}{C_{G} \int_{465}^{495} ε_{G} (λ) Q_{G} η_{G} d λ} = 0.72,

where

η_{R}

and

η_{G}

represent the detection efficiency calculated as

\frac{\int_{515}^{800} G (λ) d λ}{\int_{346}^{800} G (λ) d λ}

, where

G (λ)

is the power spectral density of the emission spectrum of the fluorescent dye. The ratio of

\frac{{(\frac{q A F}{q O C T})}_{R}}{{(\frac{q A F}{q O C T})}_{G}} = \frac{0.10}{0.13} = 0.77

agrees to the theory with less than 7% deviation. However, the accurate interpretation of the values of qAF/qOCT in relation to the optics of the system still needs further investigation. We hypothesize that the qAF/qOCT value with the master reference in the model eye will be able to serve as a compensation factor for in vivo applications, which also need further verification.

Fig. 4 a) Absorption spectrums of green and red fluorescent slides. b) Fluorescence emission spectrums of green and red fluorescent slides.

Download Full Size | PDF

3.2 In vivo simultaneous VIS-OCT and retina FAF imaging

The system was tested for in vivo imaging on 3 groups of rats: 2-months-old albino, 14-months-old albino, and 12-months-old pigmented rats. During imaging, the retinal OCT image was acquired with the shutter closed to block the light from the reference arm for the OCT reference target. The shutter was opened to image the reflectance reference target when the probe light was scanned to the corresponding location. AF images from the retinas of a young and an old albino rats were shown in Fig. 5(a) and 6(a), respectively. The intensities of the retinal AF images were normalized to that of the fluorescence reference target, the image of which is shown at the bottom right of each image. Since the OCT reference target is also fluorescent, although, off the peak detection range, it also showed up in the AF images (bottom left in each image). OCT fundus images (projection of the 3D OCT data on the X-Y plane or en face view) [25] normalized to the reflectance reference target are shown in Fig. 5(b) and 6(b), respectively. Using manual segmentation on the OCT B-scans we calculated the reflectance projection of the RPE layer (Fig. 5(c) and 6(c)). For segmentation, several points on the RPE boundaries on several B-scans were manually selected and the boundary surfaces were estimated using interpolation. Average of the RPE reflectance was calculated after removing the optic disc and within a square window with controlled distance from the optical disc. The window was selected within the region without the blood vessel shadows. This calculation is used later in comparison studies (Fig. 7). Figures 5(d) and 6(d) show the OCT B-scans at the location marked with yellow dashed lines on the fundus projections.

Fig. 5 Simultaneous fundus AF and OCT images of a 2-months-old albino rat. a) AF image normalized to the fluorescence reference. b) Fundus OCT projection normalized to the reflectance reference. c) qOCT projection of the segmented RPE. d) OCT B-Scan at the location marked with a yellow dashed line. Bar: 200µm.

Download Full Size | PDF

Fig. 6 Simultaneous fundus AF and OCT images of a 14-months-old albino rat. a) AF image normalized to the fluorescence reference. b) Fundus OCT projection normalized to the reflectance reference. c) qOCT projection of the segmented RPE. d) OCT B-Scan at the location marked with a yellow dashed line. Bar: 200µm.

Download Full Size | PDF

Fig. 7 Comparison of the qAF, qOCT, and qAF/qOCT in three groups of albino and pigmented rats.

Download Full Size | PDF

qAF was calculated by normalizing the averaged AF signal of the retina to the AF signal of the reference target. Similarly, qOCT was obtained by averaging the OCT reflectance signal of the RPE layer and normalizing it with that of the reflectance standard target. The ratio qAF/qOCT was then calculated as a quantity only dependent on the optical properties of the RPE layer. The qAF/qOCT ratios were calculated for the images of the three groups of rats. As shown in Fig. 7, qAF (p = 0.0063) and qAF/qOCT (p = 0.0325) levels in the older albino rats are significantly higher than the corresponding quantities in the younger albino rats, consistent with lipofuscin accumulation with aging [21]. On the other hand, the qAF level in the pigmented rats is lower than that of the 14-months old albino rats (p = 0.0019), which agrees with previous studies [26]. However, the OCT reflectance signal of the segmented RPE layer in pigmented rats is also significantly lower than that of the 14-months old albino rats. As a result, the calculated qAF/qOCT ratio in the pigmented rats is not significantly different from the ratio in the albino rats (p = 0.8021). Since melanin is distributed at a higher concentration in the apical portion of RPE cells, and lipofuscin is more concentrated in the basal portion [27], melanin could significantly attenuate the AF signals. These results suggest that the qAF/qOCT ratio is independent of melanin in the RPE, and thus the system is capable to compensate signal attenuation by RPE melanin. This feature of the system could have great clinical significance.

Acquired OCT fundus images from some of the old rats revealed noticeable dark regions around the optic disc (marked in Fig. 6(b)). The corresponding cross-sectional OCT images showed signs of photoreceptor loss or impairment (arrows in Fig. 5(d)). These findings may relate to the natural degeneration of the photoreceptors with age in albino rats, which were not further investigated in the current study.

4. Discussion

We have demonstrated that the VIS-OCT based multimodal imaging technology is capable to obtain both OCT reflectance and AF images simultaneously. The intensities of the OCT reflectance from the RPE can serve as an internal reference to compensate signal attenuation by ocular tissues anterior to the RPE, including melanin in the RPE cells. The OCT and fluorescence signal intensities from the two reference standards placed in the intermediate retinal imaging plane can be used to not only eliminate the effects of light source fluctuation and detector sensitivity for OCT and AF imaging, but also quantify the FAF signal to a value that is proportional to the absolute concentration of lipofuscin.

In our calculation, the calculated ratio, according to Eq. (8), is dependent on the RPE reflectance in addition to fluorescence efficiency. How to determine or measure the RPE reflectance needs further investigations. Weersink and associates [28] studied the Fluorescence/Reflectance ratio using the diffuse reflection of the excitation light in a turbid media and used the ratio as a correction method to reduce the dependence of the fluorescence signal on tissue optical properties (µa, µs’). Their results showed a linear correlation between Fluorescence/Reflectance and fluorophore concentration. Although their results cannot be applied directly to our VIS-OCT and FAF imaging system, their experiments inspire us in the next step phantom studies.

In the current study, we used two separate reference targets for OCT and FAF because we used a high optical density ND filter to reduce the fluorescence signal of the reference target to the level of FAF. The high OD value of the ND filter (200 dB roundtrip attenuation) made it impossible for OCT to detect the reflectance from the reference target. In the future, a single customized standard reference target can be used for both FAF and OCT if the fluorophore concentration in the reference can be reduced so that the fluorescence signal from the target is in the level of FAF.

The constant values of qAF/qOCT in Fig. 3(c) and 3(d) do not agree with the prediction of Eq. (8), but the ratios agree with that of the fluorescent efficiencies of the two fluorescent dyes in the red (Texas Red) and green (FITC) slides. The results thus indicate that qAF/qOCT has a constant scaling factor for the values as predicted in Eq. (8). This constant scaling factor may be related to the geometric parameters of the model eye. Investigations with different parameters of the model eye could find the detailed interpretation of the scaling factor. It is important to determine this scaling factor for human eye imaging to optimize the system for the clinical application.

Our in vivo results show that the AF intensities from pigmented rat retinas, when normalized to the OCT reflectance as qAF/qOCT, are not significantly different from those in albino rats of similar ages. These results strongly suggest that our system is capable of compensating signal attenuation by melanin in the RPE cells. These findings could be of great clinical significance. Further investigations are warranted to correlate qAF/qOCT ratios of pigmented and albino animals at different ages to the actual lipofuscin levels from the same retinas determined chemically.

The accuracy of OCT reflectance from the RPE is important for determining the qAF/qOCT ratio. We manually segmented the OCT reflectance of the RPE layer in our in vivo experiments to provide proof-of-concept in the present work. Automatic segmentation by available software or new algorism could be used to improve the accuracy and to avoid bias and human error [29].

The assumption that the same attenuation factors apply to excitation and emission wavelengths was based on the analysis of the total transmission of the ocular media. Total transmission of the ocular media was reported by Boettner and Wolter which includes the loss due to the normal incident light reflection at the tissue boundaries, scattering, and absorption. We have illustrated this attenuation effect by τ_pre (1-ρ_pre) in Eq. (6) and 7. Total transmittance of ocular media is relatively constant from 450nm to 600nm with less than 10% increase with wavelength [30].

5. Conclusion

We have developed a new multimodal imaging system with two reference targets placed in the intermediate retinal imaging plane for quantitative imaging of RPE lipofuscin. The system employs a single light source to acquire AF and VIS-OCT images of the retina simultaneously. Since the same group of photons is responsible for both AF and OCT imaging, the OCT image intensities can be used to compensate signal attenuation by media anterior to the RPE, including melanin in RPE cells. The reference standards are used not only to eliminate the influence of fluctuation in illumination power and in detector sensitivity but also to calculate the qAF to qOCT ratio of the RPE layer. This work is a major step towards determining the absolute AF intensity and correlating it to the lipofuscin content in the RPE.

Funding

Institutes of Health (NIH) (R01EY026643, R01EY018586, and P30-EY014801).

Disclosures

The authors declare that there are no conflicts of interest related to this article.

References and links

1. J. R. Sparrow and M. Boulton, “RPE lipofuscin and its role in retinal pathobiology,” Exp. Eye Res. 80(5), 595–606 (2005). [CrossRef] [PubMed]

2. S. R. Sadda, “Fundus autofluorescence imaging: Principles and applications,” Retin. Physician. 2, 1 (2011).

3. N. P. Boyer, D. Higbee, M. B. Currin, L. R. Blakeley, C. Chen, Z. Ablonczy, R. K. Crouch, and Y. Koutalos, “Lipofuscin and N-retinylidene-N-retinylethanolamine (A2E) accumulate in retinal pigment epithelium in absence of light exposure: their origin is 11-cis-retinal,” J. Biol. Chem. 287(26), 22276–22286 (2012). [CrossRef] [PubMed]

4. R. Afridi, A. Agarwal, M. A. Sadiq, M. Hassan, D. V. Do, Q. D. Nguyen, and Y. J. Sepah, “Fundus Autofluorescence Imaging in Posterior Uveitis,” in Sen H. Read R. (eds) Multimodal Imaging in Uveitis, Springer (2018).

5. G. L. Wing, G. C. Blanchard, and J. J. Weiter, “The topography and age relationship of lipofuscin concentration in the retinal pigment epithelium,” Invest. Ophthalmol. Vis. Sci. 17(7), 601–607 (1978). [PubMed]

6. L. An, P. Li, T. T. Shen, and R. Wang, “High speed spectral domain optical coherence tomography for retinal imaging at 500,000 A‑lines per second,” Biomed. Opt. Express 2(10), 2770–2783 (2011). [CrossRef] [PubMed]

7. F. G. Holz, C. Bellman, S. Staudt, F. Schütt, and H. E. Völcker, “Fundus autofluorescence and development of geographic atrophy in age-related macular degeneration,” Invest. Ophthalmol. Vis. Sci. 42(5), 1051–1056 (2001). [PubMed]

8. A. Bindewald, A. C. Bird, S. S. Dandekar, J. Dolar-Szczasny, J. Dreyhaupt, F. W. Fitzke, W. Einbock, F. G. Holz, J. J. Jorzik, C. Keilhauer, N. Lois, J. Mlynski, D. Pauleikhoff, G. Staurenghi, and S. Wolf, “Classification of fundus autofluorescence patterns in early age-related macular disease,” Invest. Ophthalmol. Vis. Sci. 46(9), 3309–3314 (2005). [CrossRef] [PubMed]

9. S. Schmitz-Valckenberg, A. Bindewald-Wittich, J. Dolar-Szczasny, J. Dreyhaupt, S. Wolf, H. P. Scholl, and F. G. Holz, “Correlation between the area of increased autofluorescence surrounding geographic atrophy and disease progression in patients with AMD,” Invest. Ophthalmol. Vis. Sci. 47(6), 2648–2654 (2006). [CrossRef] [PubMed]

10. Y. Yonekawa, J. W. Miller, and I. K. Kim, “Age-related macular degeneration: advances in management and diagnosis,” J. Clin. Med. 4(2), 343–359 (2015). [CrossRef] [PubMed]

11. M. Yung, M. A. Klufas, and D. Sarraf, “Clinical applications of fundus autofluorescence in retinal disease,” Int J Retina Vitreous 2(1), 12 (2016). [CrossRef] [PubMed]

12. J. R. Sparrow and T. Duncker, “Fundus autofluorescence and RPE lipofuscin in age-related macular degeneration,” J. Clin. Med. 3(4), 1302–1321 (2014). [CrossRef] [PubMed]

13. M. Boulton and P. Dayhaw-Barker, “The role of the retinal pigment epithelium: topographical variation and ageing changes,” Eye (Lond.) 15(3), 384–389 (2001). [CrossRef] [PubMed]

14. M. Rudolf, S. D. Vogt, C. A. Curcio, C. Huisingh, G. McGwin Jr, A. Wagner, S. Grisanti, and R. W. Read, “Histologic basis of variations in retinal pigment epithelium autofluorescence in eyes with geographic atrophy,” Ophthalmology 120(4), 821–828 (2013). [CrossRef] [PubMed]

15. M. J. Allingham, Q. Nie, E. M. Lad, D. J. Izatt, P. S. Mettu, S. W. Cousins, and S. Farsiu, “Semiautomatic segmentation of rim area focal hyperautofluorescence predicts progression of geographic atrophy due to dry age-related macular degeneration,” Invest. Ophthalmol. Vis. Sci. 57(4), 2283–2289 (2016). [CrossRef] [PubMed]

16. C. Bellmann, G. S. Rubin, S. A. Kabanarou, A. C. Bird, and F. W. Fitzke, “Fundus autofluorescence imaging compared with different confocal scanning laser ophthalmoscopes,” Br. J. Ophthalmol. 87(11), 1381–1386 (2003). [CrossRef] [PubMed]

17. N. Lois, A. S. Halfyard, C. Bunce, A. C. Bird, and F. W. Fitzke, “Reproducibility of fundus autofluorescence measurements obtained using a confocal scanning laser ophthalmoscope,” Br. J. Ophthalmol. 83(3), 276–279 (1999). [CrossRef] [PubMed]

18. F. Delori, J. P. Greenberg, R. L. Woods, J. Fischer, T. Duncker, J. Sparrow, and R. T. Smith, “Quantitative measurements of autofluorescence with the scanning laser ophthalmoscope,” Invest. Ophthalmol. Vis. Sci. 52(13), 9379–9390 (2011). [CrossRef] [PubMed]

19. C. Dai, X. Liu, and S. Jiao, “Simultaneous optical coherence tomography and autofluorescence microscopy with a single light source,” J. Biomed. Opt. 17(8), 080502 (2012). [CrossRef] [PubMed]

20. M. Jiang, T. Liu, X. Liu, and S. Jiao, “Simultaneous optical coherence tomography and lipofuscin autofluorescence imaging of the retina with a single broadband light source at 480nm,” Biomed. Opt. Express 5(12), 4242–4248 (2014). [CrossRef] [PubMed]

21. Z. Nafar, M. Jiang, R. Wen, and S. Jiao, “Visible-light optical coherence tomography-based multimodal retinal imaging for improvement of fluorescent intensity quantification,” Biomed. Opt. Express 7(9), 3220–3229 (2016). [CrossRef] [PubMed]

22. J. van de Kraats, T. T. Berendschot, and D. van Norren, “The pathways of light measured in fundus reflectometry,” Vision Res. 36(15), 2229–2247 (1996). [CrossRef] [PubMed]

23. J. I. Morgan and E. N. Pugh Jr., “Scanning laser ophthalmoscope measurement of local fundus reflectance and autofluorescence changes arising from rhodopsin bleaching and regeneration,” Invest. Ophthalmol. Vis. Sci. 54(3), 2048–2059 (2013). [CrossRef] [PubMed]

24. A. Schwartz, L. Wang, E. Early, A. Gaigalas, Y. Z. Zhang, G. E. Marti, and R. F. Vogt, “Quantitating fluorescence intensity from fluorophore: The definition of MESF assignment,” J. Res. Natl. Inst. Stand. Technol. 107(1), 83–91 (2002). [CrossRef] [PubMed]

25. S. Jiao, R. Knighton, X. Huang, G. Gregori, and C. Puliafito, “Simultaneous acquisition of sectional and fundus ophthalmic images with spectral-domain optical coherence tomography,” Opt. Express 13(2), 444–452 (2005). [CrossRef] [PubMed]

26. X. Zhang, H. F. Zhang, C. A. Puliafito, and S. Jiao, “Simultaneous in vivo imaging of melanin and lipofuscin in the retina with photoacoustic ophthalmoscopy and autofluorescence imaging,” J. Biomed. Opt. 16(8), 080504 (2011). [CrossRef] [PubMed]

27. J. J. Weiter, F. C. Delori, G. L. Wing, and K. A. Fitch, “Retinal pigment epithelial lipofuscin and melanin and choroidal melanin in human eyes,” Invest. Ophthalmol. Vis. Sci. 27(2), 145–152 (1986). [PubMed]

28. R. Weersink, M. S. Patterson, K. Diamond, S. Silver, and N. Padgett, “Noninvasive measurement of fluorophore concentration in turbid media with a simple fluorescence /reflectance ratio technique,” Appl. Opt. 40(34), 6389–6395 (2001). [CrossRef] [PubMed]

29. J. Tian, B. Varga, G. M. Somfai, W. H. Lee, W. E. Smiddy, and D. C. DeBuc, “Real-time automatic segmentation of optical coherence tomography volume data of the macular region,” PLoS One 10(8), e0133908 (2015). [CrossRef] [PubMed]

30. E. A. Boettner and J. R. Wolter, “Transmission of the ocular media,” Invest. Ophthalmol. Vis. Sci. 1(6), 776–783 (1962).

Visible light OCT-based quantitative imaging of lipofuscin in the retinal pigment epithelium with standard reference targets

Abstract

1. Introduction

2. Methods

2.1 Imaging system

2.2 Mathematical model of OCT reflectance and AF signals

2.3 Animal experiments

2.4 Phantom study

2.5 Standard reference targets characterization

3. Results

3.1 System calibration

3.2 In vivo simultaneous VIS-OCT and retina FAF imaging

4. Discussion

5. Conclusion

Funding

Disclosures

References and links

Cited By

Figures (7)

Equations (9)

Biomedical Optics Express