Open Access
Add to BibSonomy
Abstract
Fluorescence lifetime imaging ophthalmoscopy (FLIO), a technique for investigating metabolic changes in the eye ground, can reveal the first signs of diseases related to metabolism. The fluorescence of the natural lens overlies the fundus fluorescence. Although the influence of natural lens fluorescence can be somewhat decreased with mathematical models, excluding this influence during the measurement by using hardware enables more exact estimation of the fundus fluorescence. Here, we analyze four 1-photon excitation hardware solutions to suppress the influence of natural lens fluorescence: aperture stop separation, confocal scanning laser ophthalmoscopy, combined confocal scanning laser ophthalmoscopy and aperture stop separation, and dual point confocal scanning laser ophthalmoscopy. The effect of each principle is demonstrated in examples. The best suppression is provided by the dual point principle, realized with a confocal scanning laser ophthalmoscope. In this case, in addition to the fluorescence of the whole eye, the fluorescence of the anterior part of the eye is detected from a non-excited spot of the fundus. The intensity and time-resolved fluorescence spectral data of the fundus are derived through the subtraction of the simultaneously measured fluorescence of the excited and non-excited spots. Advantages of future 2-photon fluorescence excitation are also discussed. This study provides the first quantitative evaluation of hardware principles to suppress the fluorescence of the natural lens during measurements of fundus autofluorescence.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Fluorescence lifetime imaging ophthalmoscopy (FLIO) is a technique to investigate metabolic changes at the eye ground. Since its development in the early 2000s [1–4], influencing factors have been investigated [5–12], and its clinical value has been demonstrated in several studies of ocular diseases, such as AMD [13–20], diabetes [21,22], Stargardt disease [23,24], macular telangiectasia type 2 [25–28], and others [29–34], as partly reviewed in [35–37]. The detected fluorescence in FLIO consists of the fluorescence of the anterior and posterior parts of the eye. The fluorescence of the natural lens and the cornea overlies the fundus fluorescence and has the longest fluorescence lifetime among that of all ocular structures [38–40]. For example, the mean fluorescence lifetime of the natural lens is approximately 2.8 ns [41], whereas that of the retina is in the range of 0.1–0.2 ns [41,42]. Furthermore, the fluorescence lifetime of the natural lens is prolonged in older people, especially in the presence of cataracts [43–46]. However, older patients are those likely to benefit most from FLIO. In some patients, the influence of the natural lens fluorescence can be as strong as that of the fundus fluorescence. Thus, the fluorescence of the natural lens should be considered in analysis of the fluorescence from the fundus.
A comparison of the fluorescence lifetime in participants before and after cataract surgery (inserting an artificial intraocular lens, which has no fluorescence), has demonstrated the effects of the long fluorescence lifetime of the crystalline lens [41,42,47]. The effects of natural lens fluorescence on fluorescence lifetime measurements of the eye ground can be incorporated in some software models. In the simplest and most often used case, the fluorescence decay of the eye is approximated by a triple-exponential model (Eq. (1)) that does not include the slope. The mean fluorescence lifetime, calculated from the short and middle component before surgery, corresponds well to the mean fluorescence lifetime in a triple-exponential approximation of the fluorescence decay, detected in eyes with artificial lenses [42] and is thus taken as the fundus fluorescence. In another triple-exponential model, which uses both the slope and the decay of the detected fluorescence, the third component, corresponding to the slow fluorescence decay of the natural lens, is shifted by a delay corresponding to the time required for the light to travel between the anterior and posterior parts of the eye [42]. The decay time of the fundus is again the mean fluorescence lifetime of the short and middle components. This model works well if the instrumental response function is very narrow (FWHM of ≤ 50 ps [47]). Another model includes the separately measured fluorescence decay of the natural lens, which can lead to even better correspondence of the mean fluorescence lifetime between phakic and pseudophakic eyes [42]. Although these models typically show good correspondence with the mean fluorescence lifetimes between eyes with natural versus artificial lenses, information on single exponential components (τi, αi) regarding pathological changes is lost [21].
A better approach than the subsequent correction of the anterior eye fluorescence by using approximation models directly avoids the detection of this artifact. Herein, we evaluate four hardware-based methods of 1-photon excitation (OPE) to decrease the influence of anterior eye fluorescence: aperture stop separation (APS) in fundus cameras, confocal scanning laser ophthalmoscopy (cSLO), combined APS and cSLO, and dual point confocal scanning laser ophthalmoscopy. We analyze the suppression of the fluorescence of the anterior eye in currently used ophthalmoscopes and explain the potential for higher suppression through feasible technical modifications. This study provides the first quantitative evaluation of these OPE hardware principles in suppressing the fluorescence of the natural lens during measurement of the fundus autofluorescence. We aimed to determine which of the four hardware principles is best suited to suppressing anterior eye fluorescence.
2. FLIO state of the art
2.1. Instrumentation
For FLIO, we used a demonstrator based on the SPECTRALIS confocal scanning laser system from Heidelberg Engineering GmbH (Heidelberg, Germany), which has been described in detail [9,12]. In brief, the fluorescence of endogenous fluorophores of the eye is excited by a pulsed laser diode (473 nm, 80 MHz). Several endogenous substances, such as flavin adenine dinucleotide (FAD), advanced glycation end products (AGEs), connective tissue (collagen and elastin), melanin, or nicotinamide adenine dinucleotide (NADH), emit fluorescence in a fairly short wavelength range, whereas the emission spectrum of lipofuscin is in a longer wavelength range. Therefore, fluorescence is detected in two spectral channels. The short wavelength channel (SSC) is between 498 nm and 560 nm, and the long wavelength channel (LSC) is between 560 nm and 720 nm. The fluorescence photons are detected by using a fiber with a diameter of 100 µm. Its area acts as a confocal field stop. After blocking of the excitation light and splitting of the fluorescence in both spectral channels, the photons are registered by hybrid photomultipliers (HPM-100-40, Becker & Hickl GmbH, Berlin, Germany). A fundus field of 30° is scanned at 9 Hz with 256 lines per image and 256 pixels per line. Thus, the maximal local resolution is approximately 35 × 35 µm2. The time-resolved fundus autofluorescence decays are acquired by a time-correlated single-photon counting (TCSPC) device (SPC-150, Becker & Hickl GmbH) for each spectral channel. To record a sufficient number of photons (e.g., 1000 photons in the macula), an acquisition time of 1–2 minutes is required. An internal fixation target is used to minimize eye movements during the measurement. In addition, automatic image registration, based on simultaneously detected infrared images, is applied to compensate for head and eye movements.
2.2 Data analysis
The fluorescence decay at each pixel is approximated by a multi-exponential model, convolved with the separately measured instrumental response function (IRF):
I0 is the number of photons at time 0
αi is the pre-exponential factor (amplitude) of component i
τi is the decay time (fluorescence lifetime) of component i
b is the background light/dark current
A triple-exponential model is optimal for FLIO, because it is able to obtain a goodness-of-fit value (reduced χ2) [48] near the theoretical ideal value, and the course of residuals is free of any maxima. Beyond the evaluation of the single fitting parameters of amplitude and fluorescence lifetime, the mean amplitude-weighted fluorescence lifetime τm is used for comparison between patients and healthy participants:
In some cases, using the relative contribution Qi is helpful:
In current FLIO applications, SPCImage software (Becker & Hickl GmbH) is often used to apply Eq. (1) to the measured data. In the software package FLIMX [49], various eye-specific models are additionally implemented, including a model using the separately measured fluorescence of the natural lens. Image segmentation and statistical analysis of the FLIO data are also integrated into FLIMX. To achieve a standardized comparison of FLIO data at the same locations in diseased and healthy eyes, the evaluation is often performed in sections of the EDTRS grid [50].
2.3 Influence of the natural lens
In current FLIO measurements in patients with cataracts, the relative contribution of Q3 (Eq. (3)) of the longest component is between approximately 40% and 50% (see [42] supplement). This contribution influences the fluorescence lifetimes of the short and middle components, which correspond to the fundus fluorescence. The effects of the natural lens fluorescence on fundus fluorescence measurement were demonstrated for 29 patients before and after cataract surgery. In Table 1, the fitting parameter amplitudes and fluorescence lifetimes, according to Eq. (1), are given for those patients, on the basis of previously reported measurements [42]. The strongest change occurs in SSC, because the fluorescence spectrum of the natural lens is located predominantly in this spectral range. The amplitude and the fluorescence lifetime of the long component considerably decrease after surgery, but the parameters of the short and the middle components are also changed in SSC. For example, τ2 in SSC decreases from 717 ± 97 ps before cataract surgery to 524 ± 47 ps after cataract surgery. In LSC, the short component (α1, τ1) appears unchanged, the middle component is only slightly changed (τ2 decreased by 10%), and τ3 of the long component is moderately changed (decreased by approximately 1 ns). These demonstrations indicate the influence of the anterior eye on the fluorescence detected from the eye. This influence should be avoided if pathological processes are investigated at the fundus.
Table 1. Fluorescence lifetimes (τ), their corresponding amplitudes (α), and the relative contributions (Q) of a triple-exponential approximation from the outer ring of the ETDRS grid, averaged over 29 patients before and after cataract surgery. P values are given for this comparison (paired t-test, not Bonferroni corrected).
3. Hardware methods to suppress the fluorescence of the natural lens
3.1 Components of eye fluorescence
Both the anterior part (lens and cornea) and the posterior part (retina and choroid) of the eye are excited during fluorescence measurements of the eye ground. The fluorescence of the anterior part overlies the fundus fluorescence. Two components of the anterior part fluorescence must be considered. One component travels in the opposite direction from the excitation light (Flens,1). The other component (Flens,2) travels in the same direction as the excitation light and is detected after reflection at the fundus. The excited imaged field at the fundus acts as an aperture stop for the fluorescence of both components Flens,1 and Flens,2 of the anterior part. The measuring stop, optically conjugated to the iris plane of the eye, is the aperture stop of the light emitted from the fundus. These relationships are also valid for avoiding scattering light of the natural lens.
The values in Table 2 are assumed for the calculation of the ratio of fundus fluorescence Ffundus and the remaining overlying natural lens fluorescence.
Table 2. Description of optical parameters and their assumed values related to measurements in the eye.
3.2 Aperture stop separation
The first principle to decrease the effects of natural lens fluorescence in measurements of the fundus fluorescence is classical APS, invented by Helmholtz [55]. Excitation and emission light penetrate the aperture plane of the iris in separate regions (Fig. 1). The annular stop is used for the excitation, and the circular stop is used for the fluorescence. Generally, the fundus fluorescence is detectable through the detection aperture stop. The excited field, e.g., 20° in a fundus camera, acts as an aperture stop for the natural lens fluorescence.
Fig. 1. Schema of the aperture stop principle in the eye, used in fundus cameras. Solid angles for the detection of the fundus fluorescence, natural lens fluorescence, and excitation by the conjugated separate aperture stops are shown.
The detected fluorescence from the eye Iem,det, according to Eq. (4), is the sum of the natural lens fluorescence Flens,2 reflected from the fundus and the fundus fluorescence Ffundus. The natural lens fluorescence Flens,1, directed opposite to the excitation light, is suppressed by the separation of the aperture stops between the excitation and detection area.
3.3 Confocal scanning laser ophthalmoscopy
The suppression of the fluorescence of the natural lens in confocal scanning laser ophthalmoscopy [56] is demonstrated in Fig. 2. A major feature is the different sizes of the solid angles for detection of the fundus fluorescence and the natural lens fluorescence.
Fig. 2. Schema of confocal scanning laser ophthalmoscopy in the eye. Solid angles for detection of the fundus fluorescence and natural lens fluorescence are shown.
In a confocal scanning laser ophthalmoscope, the fluorescence detected from the eye is the sum of the natural lens fluorescence directed in the opposite direction from the excitation light (Flens,1), the natural lens fluorescence Flens,2, reflected from the fundus, and the fundus fluorescence Ffundus:
The complete area of the aperture stop in the plane of the iris can theoretically be used for the detection of the fundus fluorescence. The small excited laser spot at the fundus, equivalent to its confocal stop, acts as an aperture stop for the natural lens fluorescence, thus considerably decreasing the influence of natural lens fluorescence in the fluorescence signal detected from the eye.
3.4 Combined confocal scanning laser ophthalmoscopy and aperture stop separation
The effect of natural lens fluorescence on the fundus fluorescence can be further decreased by combining the confocal scanning laser principle and the APS principle (cSLO-APS), as shown in Fig. 3.
Fig. 3. Schema of combined confocal scanning laser ophthalmoscopy and aperture stop separation in the eye.
In this case, the excitation beam penetrates the natural lens centrally. The fundus fluorescence is detectable in a peripheral annular aperture plane. Thus, the part of the natural lens fluorescence Flens,1 is blocked, which is directed opposite to the excitation beam. The detected fluorescence is:
Only the natural lens fluorescence, reflected from the fundus (Flens,2), overlies the fundus fluorescence.
3.5 Dual point confocal scanning laser ophthalmoscopy
The basic idea of the dual point principle to suppress the natural lens fluorescence during fluorescence measurements of the fundus was realized by Delori through use of a spectrometer [57]. A 3° fundus field is excited, and the fluorescence is measured from a 2° field centered on the excited field. For the detection of the fluorescence of the natural lens, the excited field is shifted away from the detection field. In this way, no fundus fluorescence is detected. This so-called baseline spectrum is determined by only the fluorescence of the anterior eye. The pure fundus fluorescence is then the difference between the fluorescence of the measurement from the excited field, where the fundus fluorescence is superimposed by the fluorescence of the natural lens, and the measurement outside the excited location. This idea has been adapted for the confocal scanning laser technique [58].
The dual point principle for nearly complete suppression of the fluorescence of the anterior eye in a confocal scanning laser ophthalmoscope (DP-cSLO) is shown in Fig. 4. According to this principle, the fluorescence is measured at two spots: the excited fundus spot and an additional spot of identical size and shape in a non-excited fundus area. This additional spot should be located in the field imaged by cSLO.
The fluorescence, confocally detected from the excited fundus spot, is the sum of the fundus fluorescence (Ffundus) and of the components of the natural lens fluorescence (Flens,1) and (Flens,2) according to Eq. (5). The fluorescence from the non-excited fundus spot, according to Eq. (7), includes the natural lens fluorescence Flens,1 and Flens,2:
The pure fundus fluorescence is the difference between the fluorescence of the excited (Eq. (5)) and non-excited spots (Eq. (7)) in each channel.
3.6 Grading the suppression of natural lens fluorescence
The effect of the suppression of the overlaid natural lens fluorescence is calculated for each principle as:
4. Results
4.1 Suppression of the natural lens fluorescence in fundus images with a fundus camera vs. scanning laser ophthalmoscopy according to the state of art
According to the data from Table 2, the ratio of fluorescence originating from the fundus and fluorescence originating from the natural lens for a fundus camera (APS) is Reye/lens (APS, real) 4.35.
If a 7 mm pupil diameter is assumed to act as an aperture stop in cSLO, the maximal ratio of fundus fluorescence and natural lens fluorescence is Reye/lens (cSLO, max) 288.2. This theoretical value is not realistic in practice, because the areas of the scanner mirrors limit the theoretical aperture of an eye with a 7 mm pupil diameter. A realistic diameter of the aperture stop is approximately 3 mm. If this diameter is assumed, then the ratio is Reye/lens (cSLO, real) 52.9. By using parts of the natural lens fluorescence and the fundus fluorescence in Eq. (5) as well as individual geometric data for aperture stops, the suppression of the natural lens fluorescence can be calculated for a specific ophthalmoscope.
Scanning laser ophthalmoscopy thus achieves much better suppression of natural lens fluorescence in patients with cataracts than separation of excitation and emission aperture stops in a fundus camera, as demonstrated in Fig. 5. Fluorescence intensity images from a healthy participant are shown in A1–A3, and fluorescence intensity images of a patient with an early cataract are shown in B1–B3. The images in A1 and B1 were taken with a fundus camera, and A2, A3, B2, and B3 were taken with cSLO. In the fundus camera images, the vessel structure is only weakly visible because of the low contrast. Clinicians usually look for local changes in the fluorescence intensity to evaluate their local extension. Normalization of the fundus fluorescence intensity with respect to that of an internal fluorescence standard in the device cannot sufficiently decrease the measurement variability [59], because the fluorescence intensity detected from the eye greatly depends on the adjustment of the device to the unfixed eye.
Fig. 5. Fluorescence intensity images of a healthy participant (A1–A3) and a patient with cataracts (B1–B3). Left column: image taken by a fundus camera (A1, B1), middle line: cSLO image in the SSC (A2, B2), right column: cSLO image in the LSC (A3, B3). The low contrast in B1 and B2 is caused by the fluorescence of the natural lens, which is increased in cataract and superimoses with the fundus autofluorescence.
For the fundus camera, the fluorescence excitation was between 475 nm and 610 nm, and the emission was detected in the range of 630 nm to 760 nm. For cSLO, the fluorescence was excited at 473 nm and detected in two spectral channels (Section 2.1). The fluorescence emission of the natural lens is predominantly in the SSC, together with several other fluorophores [40,60]. The fluorescence in the LSC is dominated by lipofuscin. Therefore, the structure of the retinal vessels appears to have higher contrast in LSC than in SSC, in which the fluorescence of the natural lens superimposes with the fundus fluorescence.
4.2 Combined confocal scanning laser ophthalmoscopy and aperture stop separation
After adding an annular detection aperture with a diameter of 4–7 mm in cSLO with an excitation beam of 3 mm diameter (Section 3.4), the maximum ratio of fundus fluorescence and natural lens fluorescence increases to Reye/lens (cSLO-APS, max) 9898. The 7 mm maximal diameter of the detection aperture stop is a theoretical value, because the scanning mirrors limit the detection aperture. The calculation is more realistic with the assumptions that the diameter of the excitation beam is 0.5 mm, and the annular aperture stop is between 1.5 mm and 3 mm. Thus, a more realistic suppression ratio is Reye/lens (cSLO-APS, real) 1999. Therefore, the suppression of the natural lens fluorescence is approximately 38 times better than that when the cSLO principle alone is applied. Here, the outer diameter of the annular aperture stop is assumed not to be further limited by the size of the scanning mirrors. In the case of the theoretical values calculated with a 7 mm aperture diameter, the suppression of the natural lens fluorescence is 34 times better with the combined confocal scanning laser principle and APS than the confocal scanning laser principle alone. However, the fundus fluorescence intensity is decreased by the additional annular stop.
The effect of combined confocal scanning laser ophthalmoscopy and APS was demonstrated in an experimental setup [61]. Figure 6 shows the principle of the experimental setup, adapted from [61].
Fig. 6. Scheme of the experimental setup to investigate combined cSLO and the separation of excitation and emission aperture stops (adapted from [61]).
To determine the quantitative effect of APS in a confocal system separately for lens fluorescence and fundus fluorescence, we used two substances with different emission spectra. The fluorescence in the eye was modeled with two cuvettes. One cuvette, containing sodium fluorescein (emission at λmax = 514 nm), was a proxy for the natural lens fluorescence. The other cuvette, containing Rhodamine B (emission at λmax = 576 nm), was a proxy for fundus fluorescence located in the position of the fundus. The fluorescence was excited by a laser at 446 nm, which penetrated the model eye centrally. Excitation light was blocked in the detection beam. The resulting emission spectrum was detected by a spectrometer via an annular aperture stop in combination with a confocal field stop.
The effect of additional blocking of natural lens fluorescence by APS together with the confocal principle is demonstrated in Fig. 7.
Fig. 7. Spectra of simulated natural lens fluorescence (peak at 514 nm) and simulated retina fluorescence (peak at 576 nm) without (A) and with (B) the confocal principle alone or in combination with aperture stop separation after filtering and normalization. The gray line marks the emission maxima of both fluorophores. Adapted from [61].
Figure 7(A) shows the resulting emission spectrum if the confocal field stop in front of the detector and the annual aperture stop are removed. This setup simulates a non-confocal scanning laser ophthalmoscope. The blue line in Fig. 7(B) shows the emission spectra of both dyes in confocal detection when the annular aperture stop in front of lens 8 in Fig. 6 is removed. Here, the fluorescence intensity of sodium fluorescein considerably decreased, and the spectrum of Rhodamine B was visible. The ratio of the fluorescence maximum of Rhodamine B and sodium fluorescein was 1.37. If the additional annual aperture stop (inner diameter of the ring: 1 mm; outer diameter of the ring: 7 mm) was applied, the detected fluorescence intensity of sodium fluorescein further substantially decreased (orange line). The ratio of the maxima of Rhodamine B and sodium fluorescein increased to 5.4. Although the fluorescence intensity of Rhodamine B, a proxy for fundus fluorescence, also decreased, combined confocal excitation with emission detection by an annular aperture stop is well suited for the suppression of the natural lens fluorescence in FLIO measurements.
4.3 Dual point confocal scanning laser ophthalmoscopy
The dual point principle was tested with a confocal scanning laser ophthalmoscope in conjunction with a spectrometer. The principle is demonstrated in Fig. 8. No scanning components are shown.
In a first step, the fluorescence was excited in a 20° fundus field of a participant by a laser at 446 nm. The exit fiber of the confocal scanning laser ophthalmoscope was connected to a spectrograph, in which the fluorescence of all excited points was accumulated. The summed fluorescence of the fundus and the natural lens was detected during these measurements. In the second step, a tilted plane plate was inserted in front of the exit of the confocal scanning laser ophthalmoscope. During this measurement, the fundus was also excited, but the detected signal originated from a location outside the excited spot. In these measurements, the isotropically distributed fluorescence of the cornea and the natural lens was detected for both the fluorescence opposite to the excitation (Flens,1) and after reflection at the fundus (Flens,2). The difference between the confocally detected sum spectrum and the spectrum of the anterior part of the eye, detected from the non-excited spot, resulted in a separate fluorescence spectrum of the fundus. The fluorescence spectrum of the anterior part of the eye, the summed fluorescence of the fundus and the natural lens, and the resulting separate fundus fluorescence are displayed in Fig. 9.
Fig. 9. Summed spectrum of the fluorescence of the eye and fluorescence of the anterior eye, the fundus fluorescence as the difference between both, and the lipofuscin component A2E (normalized).
The effects of the fluorescence of the natural lens and the cornea are removed. Because the fundus spectrum was detected in an eye with a natural lens, the shape of this spectrum was influenced by the transmission of the ocular media. After correction for ocular transmission according to Ref. [62], good correspondence was observed between the normalized in vivo detected fundus fluorescence spectrum and the emission spectrum of synthetic A2E [40], the component VIII of lipofuscin [63], as depicted in Fig. 9.
The emission maxima of FAD and AGEs, which are relevant in metabolism, are in SSC [40]. Therefore, pathologic alterations in these endogenous fluorophores will be detectable predominantly in SSC. The deviation of the fundus spectrum from the A2E spectrum above 600 nm indicates that fluorophores other than A2E emit at the fundus.
5. Discussion
5.1 Summary
Fluorescence lifetime imaging ophthalmoscopy is transitioning from a research method to a clinical application. Nonetheless, the FLIO hardware can and should be enhanced to decrease artifacts to improve sensitivity in detecting retinal diseases in early stages. Recent studies [41,42] have demonstrated that the fluorescence of the natural lens causes strong artifacts in FLIO data, especially with increasing age or cataract formation. This work reviews four hardware principles to suppress the natural lens fluorescence. The aim was to demonstrate the degree to which the effects of natural lens fluorescence can be decreased in clinically applied ophthalmoscopes. Our findings indicate feasible technical changes that can increase the suppression of this disruptive fluorescence.
For each hardware principle, the suppression of the natural lens fluorescence is estimated and summarized in Table 3.
Table 3. Ratio of the fluorescence, detected from the eye and natural lens fluorescence. A higher value implies stronger suppression of natural lens fluorescence.
Regardless of how the natural lens fluorescence is avoided, the pure fundus fluorescence can be evaluated through the phasor approach [64] or by the decay of exponential functions. Approximating the measured fluorescence decay by exponential functions according to Eq. (1) is time consuming. The information from such fitting is suited to the analysis of the pathological process, because fluorescence lifetime data of the involved fluorophores are available in the literature. Knowledge regarding the fluorophores involved in pathological process is helpful in therapy development. Fluorescence lifetimes can also be determined through machine learning [65–67] but have not been demonstrated for FLIO data. Other machine learning applications related to fluorescence lifetime analysis [68–71] might also be promising for FLIO. The calculation time of the phasor diagram is extremely short. The results for healthy volunteers and of patients with specific diseases are located in certain positions in the phasor diagram. These different positions may be used for ophthalmic diagnostic applications by means of machine learning, e.g., similarly to the analysis of stained cell cultures [72].
The best hardware principle to suppress the natural lens fluorescence by using OPE is the dual point principle, because it allows for almost complete removal of the anterior eye fluorescence from the measured signal. In the case of complete suppression of the anterior eye fluorescence, the fluorescence lifetimes and their corresponding amplitudes describe only the time-resolved behavior of fundus endogenous fluorophores, thus potentially enabling better and more detailed interpretation of pathological alterations than can be achieved by comparing only the mean fluorescence lifetime [21].
5.2 Limitations
Because the fluorescence of the anterior eye is below 600 nm, FLIO measurements with fluorescence detection above 600 nm would be much less influenced by natural lens fluorescence. Unfortunately, the emission spectra of important fundus fluorophores (e.g., FAD, AGE, and NADH) [40] are at wavelengths < 600 nm. Thus, minimizing the effects of natural lens fluorescence is a matter of great interest. Therefore, the fluorescence detection in FLIO is separated into two spectral channels: SSC (498–560 nm) and LSC (560–720 nm).
The four hardware principles could not be compared according to fluorescence lifetime measurements, because of the lack of capable hardware. A fundus camera for fluorescence lifetime measurements has not yet been built, because the excitation of such an extended field would exceed the maximal permissible exposure [73]. A FLIO device in its current state could not be modified to implement combined cSLO and APS, or the dual point principle. Therefore, we performed spectral measurements, partly in a model eye. We assumed that the influence of the natural lens fluorescence on measurements of the time-resolved fundus autofluorescence was equivalent to the influence of the natural lens in measurements of fundus fluorescence spectra. The described hardware-based principles suppress not only natural lens fluorescence, but also the scattering in the natural lens as well as the cornea reflection.
The basic conditions for suppressing the fluorescence of the anterior eye are detection of fundus fluorescence through the measuring aperture stop, while the excited field acts as an aperture stop for the natural lens fluorescence. Together with the blocking of Flens,1 or Flens,2, the different sizes of the measuring aperture stop and of the excited field determine the suppression of the natural lens fluorescence by the hardware. The data in Table 2, which were used for the calculations, were taken according to the anatomy of the eye and under the assumption of a possible technical implementation. An identical quantum efficiency was assumed for both the natural lens fluorescence and the fundus fluorescence, because no relevant data were found in the literature. Delori [57] has reported that the fluorescence effectivity of the natural lens is much higher than that of the fundus fluorescence. Some compensation of this difference might be caused by the lower irradiance of the natural lens in comparison to the fundus, where the excitation light is focused. Furthermore, the attenuation of the excitation light by its transmission through the ocular media is not considered. Consequently, the absolute values for the ratio of fundus fluorescence and natural lens fluorescence would be changed, but the relation of the suppression effect on the natural lens fluorescence would still be the same between the analyzed hardware principles.
5.3 Aperture stop separation
The classic principle of APS in fundus cameras enables imaging of a large fundus field. Unfortunately, the contrast in fundus images is very low if a considerable amount of autofluorescence or scattering of the natural lens is present. Because the excited fundus field acts as an aperture stop for the natural lens fluorescence, the effect of natural lens fluorescence increases with the size of the excited field.
The advantage of APS is the observation of a larger fundus field than that in the commonly used cSLO. The detection of the time-resolved autofluorescence in the time domain (TCSPC) or the frequency domain is theoretically possible with a two-dimensional matrix detector. Beyond the very low suppression of the effects of the anterior part of the eye, the maximal permissible exposure [73] excludes the excitation of an extended fundus field [1]. These aspects hinder the application of fundus cameras for measurement of time-resolved autofluorescence.
The differences in the excitation and emission ranges between a fundus camera and cSLO are caused by the applied light sources. The main goal of a fundus camera is obtaining extended color images from the reflected light in one flash. To do so, light sources are used with a broad emission spectrum. The light density in the small short wavelength range is not sufficiently high for the excitation of fundus fluorescence. Therefore, a broad spectral interval is used for excitation in a fundus camera. In contrast, a short wavelength laser is used in cSLO for excitation, focused on the fundus. Consequently, the emission range can be shifted to shorter wavelengths.
5.4 Confocal scanning laser ophthalmoscopy
The confocal scanning laser technique is the optimal technical principle for the detection of the autofluorescence of the fundus. This technique dominates in clinical applications. Considering the limitations of the maximal permissible exposure, a small spot can be excited by a high intensity. This excitation results in a sufficiently strong fluorescence signal. Furthermore, a laser can be used with a wavelength specifically adapted to the absorption spectrum of targeted fluorophores.
The small excited confocal spot at the fundus is the aperture stop of the natural lens fluorescence, thus resulting in much better suppression of the natural lens fluorescence in cSLO than a fundus camera, as demonstrated in Fig. 5(A2, A3; B2, B3). The time-resolved autofluorescence of each excited fundus pixel can be detected with an excitation power far below the maximal permissible exposure [9]. In this way, measurements are also possible through the un-dilated pupil. The serial image formation is disadvantageous for cSLO, because eye movements during the image acquisition can perturb image structure.
Unfortunately, both the directly detected natural lens fluorescence (Flens,1) and the contribution after reflection at the fundus (Flens,2) overlie the fundus fluorescence in currently used confocal scanning laser ophthalmoscopes. This overlaid fluorescence is present predominantly in the short spectral channel, which is the spectral range of the emission spectra of important fundus fluorophores, as described above.
5.5 Combined cSLO and aperture stop separation
By applying the principle of APS in conjunction with the confocal scanning laser ophthalmoscopy principle [61], the part of the natural lens fluorescence (Flens,1) directed opposite to the excitation, is blocked. Only the part of the natural lens fluorescence after reflection at the fundus still superimposes with the fundus fluorescence. Under realistic conditions, the suppression of the natural lens fluorescence may be more than 30 times better than that by cSLO alone. Because the annular stop also blocks a small proportion of the fundus autofluorescence, the acquisition time must be increased to compensate, depending on the parameters of the annular stop. The annual aperture stop should be located in the detection beam, conjugated to the iris plane of the eye. This technical modification is relatively easy.
5.6 Dual point confocal scanning laser ophthalmoscopy
The dual point principle has already been used to decrease the influence of scattering or absorption in the natural lens during the measurement of fundus reflection spectra at single locations [54]. The correction of the fluorescence, detected from the eye, by the fluorescence of a non-excited location, results in a fundus fluorescence spectrum of lipofuscin [74].
This principle can be achieved through two parallel detection pathways, or only one detection pathway and a quick switch between the signals from the excited and the non-excited points. The parallel detection enables simultaneous acquisition of excited and non-excited points, but requires separate detectors and TCSPC hardware for the non-excited point detection, thus substantially increasing the cost of such a FLIO device. A quick switch could be less expensive and would have the advantage of using an identical detection path for the fluorescence signals of the excited and non-excited points, at the cost of longer acquisition time. Figure 10 shows a schema for a FLIO device using the dual point principle based on an optical switch.
On the basis of the assumption of isotropic fluorescence of the anterior eye, detecting its fluorescence at one suitable single point during the scanning of an image is sufficient. After the accumulation of several images, the photons in each time channel of the signal of the non-excited point will be subtracted from the photons in the corresponding time channels in each pixel of the scanned image. In this way, the time-resolved fundus fluorescence can be fitted, e.g., with Eq. (1), and its model parameters can be directly compared between healthy and diseased eyes. Because the time-resolved fluorescence of the natural lens is detected separately in the non-excited pathway, pathologic alterations of the natural lens can be investigated in addition to changes at the fundus from a single acquisition. For example, changes in the fluorescence of the natural lens in diabetes have been described [60].
5.7 Comparison of one-photon excitation and two-photon excitation
The OPE in cSLO systems is currently dominant in clinical applications. No problems exist in relation to the maximal permissible exposure. In combination with adaptive optics, investigation of single cells is possible [75]. As described above, integrating combined cSLO and APS, or the dual-point principle can be achieved with little additional cost.
A promising arrangement is the two-photon excitation (TPE) of the fundus. Short infrared pulses result in second harmonic generation at the fundus. Because nearly no absorption in the cornea and natural lens occur with infrared excitation, no fluorescence is generated there. Two major problems have hindered the development of TPE in ophthalmology in the past: thermal and biochemical damage at the fundus. According to experiments on rabbit eyes [76], the difference between the excitation intensity required to obtain a useful fluorescence signal and the threshold for damage of the absorbing RPE is only a factor of 2 to 3. Another problem is that the photon energy E = h·ν can break bonds between atoms, e.g., C-S, C-O, or C-N [77,78], if the excitation wavelength is shorter than 400 nm. The ocular transmission of the eye protects the fundus of photons with such a wavelength. Any damage to the photoreceptors in the retina must be avoided under all circumstances.
In comparison with OPE, no confocal stop is required in the excitation beam, because the two-photon effect occurs only in small volumes. Because of the absence of fluorescence excitation in the anterior eye, no confocal stop is required in the detection path. Thus, a stronger fluorescence signal is detectable. Only one detection system is necessary. A TPE scanning laser ophthalmoscope based on the HRA system from Heidelberg Engineering was developed [79], but studies have examined only human fundus samples. By combining adaptive optics and two-photon excitation, time-resolved in vivo fundus fluorescence measurements have been achieved in mice and non-human primate eyes [80,81]. The first in vivo application of TPE in human eyes was demonstrated by Boguslawski et al. [82]. A detailed safety analysis was calculated for this highly sophisticated device. The generated second harmonic signal was 390 nm. Because the two-photon effect occurs in only small volumes, the TPE enables fluorescence excitation in single fundus layers. The axial resolution was 130 µm. No time-resolved measurements were collected. Thus, no fluorescence lifetime information could be obtained.
Despite these advantages, the TPE technique is highly expensive. Future developments should enable translation to clinical application.
6. Conclusion
The combination of cSLO and APS, or the dual point principle implemented in a FLIO device almost completely suppresses the natural lens fluorescence, which is the cause of strong artifacts in older patients. The former method requires longer acquisition times, whereas the latter can be implemented without any additional stress for patients.
Suppression of the natural lens fluorescence would greatly increase the quality of the FLIO technique by increasing its sensitivity and thus should be applied in future FLIO devices. Such FLIO devices will be able to detect much smaller fluorescence lifetime changes in the fundus than the current FLIO devices. Hence, diseases could be detected in earlier stages. Accordingly, the detection of metabolic alterations as first signs of the onset of pathologic processes might be possible through time-resolved autofluorescence measurements at the fundus.
Generally, the dual point principle enables investigation of the fluorescence in a single layer of a multi-layer structure.
Funding
Thüringer Aufbaubank (2019 FGR 0083); Deutsche Forschungsgemeinschaft (Ha 2899/23-2); Technische Universität Ilmenau (Open Access Publication Fund).
Disclosures
The authors declare that there are no conflicts of interest related to this article.
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. D. Schweitzer, A. Kolb, M. Hammer, and E. Thamm, “τ mapping of the autofluorescence of the human ocular fundus,” in EOS/SPIE/ELA European Biomedical Optics Week – EBiOS, (2000), 79–89.
2. D. Schweitzer, A. Kolb, M. Hammer, and E. Thamm, “Basic investigations for 2-dimensional time-resolved fluorescence measurements at the fundus,” Int Ophthalmol 23(4/6), 399–404 (2001). [CrossRef]
3. D. Schweitzer, A. Kolb, and M. Hammer, “Autofluorescence lifetime measurements in images of the human ocular fundus,” European Conferences on Biomedical Optics (SPIE, 2001), Vol. 4432.
4. D. Schweitzer, M. Hammer, F. Schweitzer, R. Anders, T. Doebbecke, S. Schenke, E. R. Gaillard, and E. R. Gaillard, “In vivo measurement of time-resolved autofluorescence at the human fundus,” J. Biomed. Opt. 9(6), 1214–1222 (2004). [CrossRef]
5. L. Sauer, A. S. Vitale, C. M. Milliken, N. K. Modersitzki, J. D. Blount, and P. S. Bernstein, “Autofluorescence lifetimes measured with fluorescence lifetime imaging ophthalmoscopy (FLIO) are affected by age, but not by pigmentation or gender,” Trans. Vis. Sci. Tech. 9(9), 2 (2020). [CrossRef]
6. M. Klemm, L. Sauer, S. Klee, D. Link, S. Peters, M. Hammer, D. Schweitzer, and J. Haueisen, “Bleaching effects and fluorescence lifetime imaging ophthalmoscopy,” Biomed. Opt. Express 10(3), 1446–1461 (2019). [CrossRef]
7. J. B. Lincke, C. Dysli, D. Jaggi, R. Fink, S. Wolf, and M. S. Zinkernagel, “The influence of cataract on fluorescence lifetime imaging ophthalmoscopy (FLIO),” Trans. Vis. Sci. Tech. 10(4), 33 (2021). [CrossRef]
8. C. Dysli, M. Dysli, J. Lincke, D. Jaggi, S. Wolf, and M. S. Zinkernagel, “Imaging artifacts in fluorescence lifetime imaging ophthalmoscopy,” Retina 41(11), 2378–2390 (2021). [CrossRef]
9. C. Dysli, G. Quellec, M. Abegg, M. N. Menke, U. Wolf-Schnurrbusch, J. Kowal, J. Blatz, O. La Schiazza, A. B. Leichtle, S. Wolf, and M. S. Zinkernagel, “Quantitative analysis of fluorescence lifetime measurements of the macula using the fluorescence lifetime imaging ophthalmoscope in healthy subjects,” Invest. Ophthalmol. Vis. Sci. 55(4), 2106–2113 (2014). [CrossRef]
10. S. Kwon, E. Borrelli, W. Fan, A. Ebraheem, K. M. Marion, and S. R. Sadda, “Repeatability of fluorescence lifetime imaging ophthalmoscopy in normal subjects with mydriasis,” Trans. Vis. Sci. Tech. 8(3), 15 (2019). [CrossRef]
11. S. R. Sadda, E. Borrelli, W. Y. Fan, A. Ebraheem, K. M. Marion, and S. Kwon, “Impact of mydriasis in fluorescence lifetime imaging ophthalmoscopy,” Plos One, 13 (2018).
12. M. Klemm, A. Dietzel, J. Haueisen, E. Nagel, M. Hammer, and D. Schweitzer, “Repeatability of autofluorescence lifetime imaging at the human fundus in healthy volunteers,” Curr Eye Res 38(7), 793–801 (2013). [CrossRef]
13. D. Schweitzer, S. Quick, S. Schenke, M. Klemm, S. Gehlert, M. Hammer, S. Jentsch, and J. Fischer, “Comparison of parameters of time-resolved autofluorescence between healthy subjects and patients suffering from early AMD,” Ophthalmologe 106(8), 714–722 (2009). [CrossRef]
14. C. Dysli, R. Fink, S. Wolf, and M. S. Zinkernagel, “Fluorescence lifetimes of Drusen in age-related macular degeneration,” Invest Ophthalmol Vis Sci 58(11), 4856–4862 (2017). [CrossRef]
15. C. Dysli, S. Wolf, and M. S. Zinkernagel, “Autofluorescence lifetimes in geographic atrophy in patients with age-related macular degeneration,” Invest. Ophthalmol. Vis. Sci. 57(6), 2479–2487 (2016). [CrossRef]
16. L. Sauer, C. B. Komanski, A. S. Vitale, E. D. Hansen, and P. S. Bernstein, “Fluorescence lifetime imaging ophthalmoscopy (FLIO) in eyes with pigment epithelial detachments due to age-related macular degeneration,” Invest Ophthalmol Vis Sci 60(8), 3054–3063 (2019). [CrossRef]
17. L. Sauer, R. H. Gensure, K. M. Andersen, L. Kreilkamp, G. S. Hageman, M. Hammer, and P. S. Bernstein, “Patterns of fundus autofluorescence lifetimes in eyes of individuals with nonexudative age-related macular degeneration,” Invest Ophthalmol Vis Sci 59(4), AMD65–AMD77 (2018). [CrossRef]
18. J. B. Lincke, C. Dysli, D. Jaggi, Y. Solberg, S. Wolf, and M. S. Zinkernagel, “Longitudinal foveal fluorescence lifetime characteristics in geographic atrophy using fluorescence lifetime imaging ophthalmoscopy,” Retina 41(11), 2391–2398 (2021). [CrossRef]
19. L. Sauer, M. Klemm, S. Peters, D. Schweitzer, J. Schmidt, L. Kreilkamp, L. Ramm, D. Meller, and M. Hammer, “Monitoring foveal sparing in geographic atrophy with fluorescence lifetime imaging ophthalmoscopy - a novel approach,” Acta Ophthalmol. (2017).
20. M. Hammer, R. Schultz, S. Hasan, L. Sauer, M. Klemm, L. Kreilkamp, L. Zweifel, R. Augsten, and D. Meller, “Fundus autofluorescence lifetimes and spectral features of soft Drusen and hyperpigmentation in age-related macular degeneration,” Trans. Vis. Sci. Tech. 9(5), 20 (2020). [CrossRef]
21. D. Schweitzer, L. Deutsch, M. Klemm, S. Jentsch, M. Hammer, S. Peters, J. Haueisen, U. A. Müller, and J. Dawczynski, “Fluorescence lifetime imaging ophthalmoscopy in type 2 diabetic patients who have no signs of diabetic retinopathy,” J. Biomed. Opt 20(6), 061106 (2015). [CrossRef]
22. J. Schmidt, S. Peters, L. Sauer, D. Schweitzer, M. Klemm, R. Augsten, N. Muller, and M. Hammer, “Fundus autofluorescence lifetimes are increased in non-proliferative diabetic retinopathy,” Acta Ophthalmol. 95(1), 33–40 (2017). [CrossRef]
23. Y. Solberg, C. Dysli, P. Escher, L. Berger, S. Wolf, and M. S. Zinkernagel, “Fluorescence lifetime patterns of retinal pigment epithelium atrophy in patients with stargardt disease and age-related macular degeneration,” Ophthalmologica 243(3), 195–206 (2020). [CrossRef]
24. C. Dysli, S. Wolf, K. Hatz, and M. S. Zinkernagel, “Fluorescence lifetime imaging in Stargardt disease: potential marker for disease progression fluorescence lifetime imaging in Stargardt disease,” Invest. Ophthalmol. Vis. Sci. 57(3), 832–841 (2016). [CrossRef]
25. L. Sauer, A. S. Vitale, N. K. Modersitzki, and P. S. Bernstein, “longitudinal fluorescence lifetime imaging ophthalmoscopy analysis in patients with macular telangiectasia type 2 (MacTel),” Retina 41(7), 1416–1427 (2021). [CrossRef]
26. Y. Solberg, C. Dysli, S. Wolf, and M. S. Zinkernagel, “Fluorescence lifetime patterns in macular telangiectasia type 2,” Retina 40(1), 99–108 (2020). [CrossRef]
27. L. Sauer, R. H. Gensure, M. Hammer, and P. S. Bernstein, “Fluorescence lifetime imaging ophthalmoscopy: a novel way to assess macular telangiectasia type 2,” Ophthalmol Retina 2(6), 587–598 (2018). [CrossRef]
28. L. Sauer, A. S. Vitale, K. M. Andersen, B. Hart, and P. S. Bernstein, “Fluorescence lifetime imaging ophthalmoscopy (flio) patterns in clinically unaffected children of macular telangiectasia Type 2 (Mactel) patients,” Retina 40(4), 695–704 (2020). [CrossRef]
29. K. M. Andersen, L. Sauer, R. H. Gensure, M. Hammer, and P. S. Bernstein, “Characterization of retinitis pigmentosa using fluorescence lifetime imaging ophthalmoscopy (FLIO),” Translational Vision Science & Technology 7(3), 20 (2018). [CrossRef]
30. D. Jaggi, Y. Solberg, C. Dysli, A. Ebneter, S. Wolf, and M. S. Zinkernagel, “Fluorescence lifetime imaging ophthalmoscopy: findings after surgical reattachment of macula-off rhegmatogenous retinal detachment,” Retina (2019).
31. C. Dysli, K. Schurch, E. Pascal, S. Wolf, and M. S. Zinkernagel, “Fundus autofluorescence lifetime patterns in retinitis pigmentosa,” Invest. Ophthalmol. Vis. Sci. 59(5), 1769–1778 (2018). [CrossRef]
32. C. Dysli, L. Berger, S. Wolf, and M. S. Zinkernagel, “Fundus autofluorescence lifetimes and central serous chorioretinopathy,” Retina 37(11), 2151–2161 (2017). [CrossRef]
33. A. S. Vitale, L. Sauer, N. K. Modersitzki, and P. S. Bernstein, “Fluorescence lifetime imaging ophthalmoscopy (FLIO) in patients with choroideremia,” Trans. Vis. Sci. Tech. 9(10), 33 (2020). [CrossRef]
34. L. Goerdt, L. Sauer, A. S. Vitale, N. K. Modersitzki, M. Fleckenstein, and P. S. Bernstein, “Comparing fluorescence lifetime imaging ophthalmoscopy in atrophic areas of retinal diseases,” Trans. Vis. Sci. Tech. 10(7), 11 (2021). [CrossRef]
35. L. Sauer, A. S. Vitale, N. K. Modersitzki, and P. S. Bernstein, “Fluorescence lifetime imaging ophthalmoscopy: autofluorescence imaging and beyond,” Eye 35(1), 93–109 (2021). [CrossRef]
36. L. Sauer, K. M. Andersen, C. Dysli, M. S. Zinkernagel, P. S. Bernstein, and M. Hammer, “Review of clinical approaches in fluorescence lifetime imaging ophthalmoscopy,” J. Biomed. Opt. 23(09), 1–20 (2018). [CrossRef]
37. C. Dysli, S. Wolf, M. Y. Berezin, L. Sauer, M. Hammer, and M. S. Zinkernagel, “Fluorescence lifetime imaging ophthalmoscopy,” Eye 35(1), 93–109 (2021). [CrossRef]
38. S. Peters, M. Hammer, and D. Schweitzer, “Two-photon excited fluorescence microscopy application for ex vivo investigation of ocular fundus samples,” Proc. SPIE 8086, 808605 (2011). [CrossRef]
39. D. Schweitzer, S. Jentsch, S. Schenke, M. Hammer, C. Biskup, and E. Gaillard, “Spectral and time-resolved studies on ocular structures,” European Conferences on Biomedical Optics (SPIE, 2007), Vol. 6628.
40. D. Schweitzer, S. Schenke, M. Hammer, F. Schweitzer, S. Jentsch, E. Birckner, W. Becker, and A. Bergmann, “Towards metabolic mapping of the human retina,” Microsc. Res. Tech. 70(5), 410–419 (2007). [CrossRef]
41. J. L. Brauer, R. Schultz, M. Klemm, and M. Hammer, “Influence of lens fluorescence on fluorescence lifetime imaging ophthalmoscopy (FLIO) fundus imaging and strategies for its compensation,” Trans. Vis. Sci. Tech. 9(8), 13 (2020). [CrossRef]
42. D. Schweitzer, J. Haueisen, J. L. Brauer, M. Hammer, and M. Klemm, “Comparison of algorithms to suppress artifacts from the natural lens in fluorescence lifetime imaging ophthalmoscopy (FLIO),” Biomed. Opt. Express 11(10), 5586–5602 (2020). [CrossRef]
43. J. Dillon and S. J. Atherton, “Time resolved spectroscopic studies on the intact human lens,” Photochem Photobiol 51(4), 465–468 (1990). [CrossRef]
44. L. Kessel, S. Kalinin, R. H. Nagaraj, M. Larsen, and L. B. A. Johansson, “Time-resolved and steady-state fluorescence spectroscopic studies of the human lens with comparison to argpyrimidine, pentosidine and 3-OH-kynurenine,” Photochem. Photobiol. 76(5), 549–554 (2002). [CrossRef]
45. J. E. Roberts, S. J. Atherton, E. R. Gaillard, and J. Dillon, “Age-related-changes in the human lens as monitored by detection of porphyrin excited-states,” Photochem. Photobiol. 62(2), 339–341 (1995). [CrossRef]
46. J. A. Van Best and E. V. Kuppens, “Summary of studies on the blue-green autofluorescence and light transmission of the ocular lens,” J. Biomed. Opt. 1(3), 243–250 (1996). [CrossRef]
47. W. Becker, A. Bergmann, and L. Sauer, “Shifted-Component Model Improves FLIO Data Analysis,” (Becker & Hickl GmbH, 2018).
48. J. R. Lakowicz, Principles of Fluorescence Spectroscopy, 3rd ed. (Springer, New York, 2006), p. 954.
49. M. Klemm, D. Schweitzer, S. Peters, L. Sauer, M. Hammer, and J. Haueisen, “FLIMX: a software package to determine and analyze the fluorescence lifetime in time-resolved fluorescence data from the human eye,” PLoS ONE 10(7), e0131640 (2015). [CrossRef]
50. “Grading diabetic retinopathy from stereoscopic color fundus photographs–an extension of the modified Airlie House classification. ETDRS report number 10. Early Treatment Diabetic Retinopathy Study Research Group,” Ophthalmology 98, 786-806 (1991).
51. A. Gullstrand, Allgemeine Theorie der monochromatischen Aberrationen und ihre nächsten Ergebnisse für die Ophthalmologie (Berling, 1900).
52. A. M. Rosen, D. B. Denharn, V. Fernandez, D. Boria, A. Ho, F. Matins, J. M. Parel, and R. C. Augusteyn, “In vitro dimensions and curvatures of human lenses,” Vision Res. 46(6-7), 1002–1009 (2006). [CrossRef]
53. C. E. Myers, B. E. K. Klein, S. M. Meuer, M. K. Swift, C. S. Chandler, Y. J. Huang, S. Gangaputra, J. W. Pak, R. P. Danis, and R. Klein, “Retinal thickness measured by spectral-domain optical coherence tomography in eyes without retinal abnormalities: the Beaver Dam eye study,” Am. J. Ophthalmol. 159(3), 445–456.e1 (2015). [CrossRef]
54. F. C. Delori and K. P. Pflibsen, “Spectral reflectance of the human ocular fundus,” Appl. Opt. 28(6), 1061–1077 (1989). [CrossRef]
55. H. V. Helmholtz, Beschreibung eines Augen-Spiegels zur Untersuchung der Netzhaut im lebenden Auge (A. Förstner'sche Verlagsbuchhandlung, 1851), pp. 43 S., [44] Bl. ; 22 cm.
56. R. H. Webb, G. W. Hughes, and O. Pomerantzeff, “Flying spot TV ophthalmoscope,” Appl. Opt. 19(17), 2991–2997 (1980). [CrossRef]
57. F. C. Delori, “Spectrophotometer for noninvasive measurement of intrinsic fluorescence and reflectance of the ocular fundus,” Appl. Opt. 33(31), 7439–7452 (1994). [CrossRef]
58. D. Schweitzer, “Ophthalmic applications of FLIM,” in Fluorescence Lifetime Spectroscopy and Imaging: Principles and Applications in Biomedical, 1st ed., L. Marcu, P. M. W. French, and D. S. Elson, eds. (CRC Press/Taylor & Francis Group, Boca Raton, 2014), pp. xv, 554 pages.
59. J. P. Greenberg, T. Duncker, R. L. Woods, R. T. Smith, J. R. Sparrow, and F. C. Delori, “Quantitative fundus autofluorescence in healthy eyes,” Invest. Ophthalmol. Vis. Sci. 54(8), 5684–5693 (2013). [CrossRef]
60. D. Schweitzer, “Autofluorescence diagnostics of ophthalmic diseases,” in Natural Biomarkers for Cellular Metabolism: Biology, Techniques, and Applications, 1 ed., V. V. Ghukasyan and A. A. Heikal, eds. (CRC Press/Taylor & Francis Group, 2015), pp. 317–344.
61. M. Klemm, J. Blum, D. Link, M. Hammer, J. Haueisen, and D. Schweitzer, “Combination of confocal principle and aperture stop separation improves suppression of crystalline lens fluorescence in an eye model,” Biomed. Opt. Express 7(9), 3198–3210 (2016). [CrossRef]
62. J. van de Kraats and D. van Norren, “Optical density of the aging human ocular media in the visible and the UV,” J. Opt. Soc. Am. A 24(7), 1842–1857 (2007). [CrossRef]
63. G. E. Eldred and M. L. Katz, “Fluorophores of the human retinal-pigment epithelium - separation and spectral characterization,” Exp. Eye Res. 47(1), 71–86 (1988). [CrossRef]
64. M. A. Digman, V. R. Caiolfa, M. Zamai, and E. Gratton, “The phasor approach to fluorescence lifetime imaging analysis,” Biophys. J. 94(2), L14–L16 (2008). [CrossRef]
65. G. Wu, T. Nowotny, Y. L. Zhang, H. Q. Yu, and D. D. U. Li, “Artificial neural network approaches for fluorescence lifetime imaging techniques,” Opt. Lett. 41(11), 2561–2564 (2016). [CrossRef]
66. R. Y. Yao, M. Ochoa, P. K. Yan, and X. Intes, “Net-FLICS: fast quantitative wide-field fluorescence lifetime imaging with compressed sensing - a deep learning approach,” Light: Sci. Appl. 8(1), 26 (2019). [CrossRef]
67. S. X. Guo, A. Silge, H. Bae, T. Tolstik, T. Meyer, G. Matziolis, M. Schmitt, J. Popp, and T. Bocklitz, “FLIM data analysis based on Laguerre polynomial decomposition and machine-learning,” J. Biomed. Opt. 26(02), 022909 (2021). [CrossRef]
68. M. Ochoa, A. Rudkouskaya, R. Yao, P. Yan, M. Barroso, and X. Intes, “High compression deep learning based single-pixel hyperspectral macroscopic fluorescence lifetime imaging in vivo,” Biomed. Opt. Express 11(10), 5401–5424 (2020). [CrossRef]
69. J. T. Smith, M. Ochoa, and X. Intes, “UNMIX-ME: spectral and lifetime fluorescence unmixing via deep learning,” Biomed. Opt. Express 11(7), 3857–3874 (2020). [CrossRef]
70. L. Mukherjee, M. A. K. Sagar, J. N. Ouellette, J. J. Watters, and K. W. Eliceiri, “Joint regression-classification deep learning framework for analyzing fluorescence lifetime images using NADH and FAD,” Biomed. Opt. Express 12(5), 2703–2719 (2021). [CrossRef]
71. J. Unger, C. Hebisch, J. E. Phipps, J. L. Lagarto, H. Kim, M. A. Darrow, R. J. Bold, and L. Marcu, “Real-time diagnosis and visualization of tumor margins in excised breast specimens using fluorescence lifetime imaging and machine learning,” Biomed. Opt. Express 11(3), 1216–1230 (2020). [CrossRef]
72. A. Vallmitjana, B. Torrado, and E. Gratton, “Phasor-based image segmentation: machine learning clustering techniques,” Biomed. Opt. Express 12(6), 3410–3422 (2021). [CrossRef]
73. A. N. S. Institute and L. I. O. America, American National Standard for Safe Use of Lasers (ANSI Z136.1) (Laser Institute of America, 2007).
74. F. C. Delori, C. K. Dorey, G. Staurenghi, O. Arend, D. G. Goger, and J. J. Weiter, “In-vivo fluorescence of the ocular fundus exhibits retinal-pigment epithelium lipofuscin characteristics,” Invest. Ophthalmol. Vis. Sci. 36, 718–729 (1995).
75. J. A. H. Tang, C. E. Granger, K. Kunala, K. Parkins, K. T. Huynh, K. Bowles-Johnson, Q. Yang, and J. J. Hunter, “Adaptive optics fluorescence lifetime imaging ophthalmoscopy of in vivo human retinal pigment epithelium,” Biomed. Opt. Express 13(3), 1737–1754 (2022). [CrossRef]
76. B.-G. Wang, “[Multiphoton microscopy and laser nanosurgery of the cornea using near-infrared nanojoule femtosecond laser pulses],” Dissertation (University Hospital Jena, Jena, 2006).
77. S. J. Blanksby and G. B. Ellison, “Bond dissociation energies of organic molecules,” Acc. Chem. Res. 36(4), 255–263 (2003). [CrossRef]
78. Y.-R. Luo and Y.-R. Luo, Comprehensive Handbook of Chemical Bond Energies (CRC Press, 2007), p. 1655
79. O. La Schiazza and J. F. Bille, “High-speed two-photon excited autofluorescence imaging of ex vivo human retinal pigment epithelial cells toward age-related macular degeneration diagnostic,” J. Biomed. Opt. 13(6), 064008 (2008). [CrossRef]
80. J. A. Feeks and J. J. Hunter, “Adaptive optics two-photon excited fluorescence lifetime imaging ophthalmoscopy of exogenous fluorophores in mice,” Biomed. Opt. Express 8(5), 2483–2495 (2017). [CrossRef]
81. S. Walters, J. A. Feeks, K. T. Huynh, and J. J. Hunter, “Adaptive optics two-photon excited fluorescence lifetime imaging ophthalmoscopy of photoreceptors and retinal pigment epithelium in the living non-human primate eye,” Biomed. Opt. Express 13(1), 389–407 (2022). [CrossRef]
82. J. Boguslawski, G. Palczewska, S. Tomczewski, J. Milkiewicz, P. Kasprzycki, D. Stachowiak, K. Komar, M. J. Marzejon, B. L. Sikorski, A. Hudzikowski, A. Gluszek, Z. Laszczych, K. Karnowski, G. Sobon, K. Palczewski, and M. Wojtkowski, “In vivo imaging of the human eye using a 2-photon-excited fluorescence scanning laser ophthalmoscope,” J. Clin. Invest. 132(2), e154218 (2022). [CrossRef]
