Open Access
Add to BibSonomyAbstract
We present an application of in vivo anterior segment imaging of the human eye with an ultrahigh speed swept source OCT instrument. For this purpose, a dedicated OCT system was designed and constructed. This instrument enables axial zooming by automatic reconfiguration of spectral sweep range; an enhanced imaging range mode enables imaging of the entire anterior segment while a high axial resolution mode provides detailed morphological information of the chamber angle and the cornea. The speed of 200,000 lines/s enables high sampling density in three-dimensional imaging of the entire cornea in 250 ms promising future applications of OCT for optical corneal topography, pachymetry and elevation maps. The results of a preliminary quantitative corneal analysis based on OCT data free form motion artifacts are presented. Additionally, a volumetric and real time reconstruction of dynamic processes, like pupillary reaction to light stimulus or blink-induced contact lens movements are demonstrated.
©2009 Optical Society of America
1. Introduction
Anterior segment of the human eye consists of ocular structures like cornea, crystalline lens, conjunctiva, iris and others, which form an optical system of the eye and have direct impact on vision. Pathologic changes in this part of the eye can drastically decrease ability of normal vision, in some cases even leading to blindness. The remarkable advances in the field of refractive surgery have increased requirements for more accurate imaging and measurements of corneal curvature and topography. The development of imaging and measuring instruments for anterior segment of human eye has a long history. Instrumentation including slit lamp microscopy, gonioscopy, Javal keratometry or Placido disk-based corneal topography systems are presently in broadly use in ophthalmic clinics [1]. Quantitative information on the corneal thickness and topography is especially useful for diagnosis of diseases, precise location of lesions and planning of medical and surgical treatments (especially refractive surgery). Relatively new imaging systems, like scanning slit topography, rotating Scheimpflug imaging and ultrasound biomicroscopy are being applied for this purpose [2]. Commercially available systems (Orbscan, Pentacam) enable measuring thickness of the cornea and contour and shape for its both surfaces. However, due to insufficient acquisition speed these instruments provide results with low sampling density.
Recently, there have been several developments in imaging technologies that promise to extend our ability to image and evaluate the anterior segment of the human eye. One of the most rapid developments in this field is the application of optical coherence tomography (OCT). This noninvasive optical method, is very well suited for biomedical applications [3–5]. It has been most widely applied to imaging of the posterior segment of the human eye [6–9]. Nevertheless, potential use for structural imaging and dynamic analysis of anterior segment has been already recognized [10–18], including application of the polarization sensitive OCT [19–21]. An increasing development of OCT instrumentation dedicated for the anterior segment imaging finally led to the first commercial versions of OCT instruments introduced to the market in 2005/2006: Visante OCT (Carl Zeiss Meditec) and SL-OCT (Heidelberg Engineering) [22]. Both instruments use light of 1.3 µm central wavelength, which is well suited for anterior segment imaging, due to a relatively good penetration through the sclera. This feature enables imaging of irido-corneal angle, which is very important in diagnosis of glaucoma [23,24]. Moreover the eye safety is improved by a high water absorption of 1.3 µm radiation comparing to 800 nm light wavelengths. Visante OCT as well as SL-OCT are based on time domain version of OCT enabling imaging in large examination range covering the entire anterior segment. However, these systems can only collect a couple of thousands A-scans per second. Therefore, to acquire an image in reasonable time only a sparse sampling is possible. Under such conditions a quantitative analysis of the cornea can be strongly affected by the motion artifacts and patient adjustment (in case of asterix scan patterns). Significant improvement of the imaging speed would allow to increase the sampling density and to reduce an influence of motion artifacts, which can be crucial for improved repeatability and accuracy of the corneal topography.
In this manuscript we propose to increase the speed of the OCT system dedicated to image the anterior segment of the human eye by the application of a high speed wavelength tunable laser as a light source [25,26]. Use of swept laser sources is especially important for imaging in the 1.3 µm wavelength range, where low cost CCDs are not available. Fourier domain OCT employing wavelength tunable lasers (swept source OCT) exhibits improved imaging speed and detection sensitivity in comparison to traditional time domain detection techniques [27–29]. Considerable advancement in instrumentation was offered by Tomey Corp., who launched an instrument (Casia SS-1000) with 30 kHz acquisition speed [30]. In this paper we present a potential next generation system using swept laser, working with the repetition rate of 200 kHz lines per second.
The promising capabilities of frequency swept light sources for OCT imaging have gained intense interest in their development. There are three major concepts to achieve high speed tuning depending on the method used for wavelength selection inside the laser cavity of the wavelength swept source: one based on a fast rotating polygonal mirror [31,32], the second based on a diffraction grating on a mechanically resonant galvo-scanner [33] and the third using a fiber Fabry-Perot tunable filter (FFP-TF) [34,35]. For very high tuning speeds and to overcome limitations given by the buildup time of lasing in the cavity, the technique of Fourier Domain Mode Locking (FDML) has been introduced. High repetition rate up to 370 kHz have been already demonstrated for FDML swept sources [36–40].
An application of the high speed swept source system to the anterior segment OCT imaging is especially interesting, since it might have potential to provide significant complementary information either on particular anatomical details or a large scale architecture of the cornea, iris and the crystalline lens. The first can be useful in ophthalmic diagnosis for detailed monitoring of the corneal epithelium, Bowman’s membrane, stroma, capsule of the crystalline lens, intraocular lens surface, conjunctiva or the corneo-scleral junction. Meanwhile, the large scale imaging covering the depth of 6 mm and 12 × 12 mm of the transverse range enables to give morphometric parameters including the corneal thickness and topography, corneo-scleral angle, orientation of intraocular lenses, etc. In the case of a quantitative analysis of the corneal geometry it must be considered that due to the fixation lag some misalignment between an eye and the instrument may occur [41]. As a result, the cornea may be translated and rotated in the 3-D space, which will strongly influence its topographical representation. In order to perform an effective correction of this misalignment data should be collected with the use of a raster scanning protocol with reduced influence of motion artifacts.
In this paper we present an 200 kHz FDML laser with a FFP-TF and show possibility of an on-the-fly change of the wavelength sweep range, only by adjusting the amplitude of the voltage signal, which drives the Fabry-Perot tunable filter. In the presented OCT imaging system, the axial resolution can be traded off against ranging depth for its maximum flexibility. We demonstrate the application of this system for in vivo imaging of the anterior segment of the human eye. High imaging speed of the proposed system enables acquiring three-dimensional data sets within 250 ms with sampling density sufficient for quantitative analysis of the anterior segment of the eye [42,43]. Furthermore, we demonstrate that high speed imaging enables reconstructing morphometric data including corneal topography and thickness maps. For the first time, in vivo corneal topography, elevation BFS (Best Fit Sphere) maps and the corneal pachymetry is performed based on a dense raster scan OCT data with reduced number of motion artifacts. To provide corneal parameters that depend only on the actual geometry of the cornea, we performed mathematical alignment of the measured data. We also demonstrate the applicability of this new imaging system for a patient with corneal pathology and for a quantitative estimation of a blink-induced contact lens movement.
2. Experimental setup and Methods
2.1 The SSOCT instrument
The schematic diagram of the Swept Source OCT system, constructed at Nicolaus Copernicus University, is shown in Fig. 1 . The SSOCT system uses a custom made 200 kHz swept laser working in the FDML regime. A broadband semiconductor optical amplifier (SOA: Covega BOA 1132) is used as a laser gain medium. Two polarization controllers (PC) allow for matching the polarization state of light at the input and the output of the SOA. To force direction of light propagation in the ring cavity and to protect the SOA from back reflected light, two optical isolators (IO) are used. The Fiber Fabry-Perot tunable filter (FFP-TF) (Micron Optics) with 230 nm free spectral range (FSR) and a bandwidth of ~0.12 nm is tuned in resonance with the roundtrip time in the cavity. A fiber delay line (FDL) of 2 km determines laser’s effective sweep rate as 200 kHz. The laser output light is centered at 1300 nm with 135 nm maximum scanning range and an average output power of 6 mW.
Fig. 1 Schematic diagram of the Swept Source OCT system (SOA – Semiconductor Optical Amplifier, FFP-TF – Fiber Fabry-Perot Tunable Filter, FDL – Fiber Delay Line, OI – Optical Isolator, PC – Polarization Controller, FC – Fiber Coupler, C –Fiber Optical Circulator, NDF – Neutral Density Filter, DBPD – Dual-balanced Photo Diode).
About 70% of light propagating in the cavity is coupled out and delivered to two fiber Michelson interferometers. One interferometer, with the scanning head placed in the sample arm, is used for imaging. The second one is used to generate a reference fringe signal necessary for time to optical frequency remapping. Fringe signals from both interferometers are detected with dual-balance photodiodes (DBPD) (Thorlabs, 75 MHz and 100 MHz in imaging and reference interferometer, respectively) and then recorded with two channel acquisition card (Gage Compuscope 14200, 200 Ms/s, 14 bit resolution). The acquisition process and scanning protocols are controlled by a custom designed compact electronic driving unit. The AD converter card is triggered for each A-scan, synchronously with the signal, which drives the FFP-TF. The data acquisition process is clocked internally. Sensitivity of the system measured for 1.8 mW of optical power reaching an object is ~103 dB.
2.2 K-space resampling by phase analysis and inverse Fourier transformation (iFFT)
In the first approximation the voltage applied to the Fabry-Perot tunable filter is proportional to the change of a wavelength of transmitted light rather than to its optical frequency [44]. Furthermore, because the FFP-TF in the FDML laser is driven with a sinusoidal electrical signal, the function describing a distribution of optical frequencies over time is highly non-linear. Thus, the OCT imaging application requires remapping to equidistant spacing in optical frequency domain (wavenumber space) before the Fourier transformation is performed. For this purpose, an appropriate data processing has to be introduced. In order to resample data from the wavelength space into wavenumber one has to find fractional values of sample number corresponding to equidistantly distributed optical frequencies. This procedure is followed by the interpolation of the measured spectral fringes. Finally, new values of signal corresponding to the fractional indexes are returned.
We describe here a method for highly accurate time to optical frequency resampling using phase analysis of a calibration signal. This method is similar to this proposed by Yasuno et al. [45] the only difference is that we had to resample each acquired A-scan due to limited stability of the Fabry-Perot filter operating at 100 kHz scanning speed. Because this procedure is crucial for obtaining high quality OCT data we will describe it here, in detail. Let us introduce a variable q, that is derived from data represented in optical frequency domain. Values of q are linearly dependent on the wavenumber:
In order to resample the data into wavenumber space a functional relation between the sample number and q has to be found.
In contrast to the spectrometer based OCT systems, in the most of the swept source OCT setups the characteristic non-linearity in optical frequency is not know a priori, so it has to be measured [35]. The simplest way of finding this relation is to measure the spectral fringes in the Michelson or Mach-Zender interferometer with mirrors in both arms. In this case the measurable quantity that fulfils Eq. (1). (q variable) is unwrapped phase of the fringe signal, which is proportional to the distribution of equidistantly spaced optical frequency components. The relation is obtained by a polynomial fit to the retrieved phase. Next, the fractional indexes are easily found from the following relations:
where M is number of samples in the spectral fringe signal. In the above equations values of the quantity q are equidistant and so we can assume that they correspond to values of wavenumber (Eq. (1). In order to use efficiently all samples and not to extrapolate spectral fringes the extreme values of the quantity q are chosen according to the following relations:In our setup we use the unwrapped phase of the spectral interference fringe with the method mentioned above. The spectral fringe signal was measured by additional Michelson interferometer (Fig. 1). The optical path difference between the arms of the interferometer is chosen to be large enough, so that the image of the mirror and its complex conjugate counterpart do not overlap neither with themselves nor with the central peak (Fig. 2b ). Since we have well defined oscillatory signal, which is separated from its complex conjugate part we can easily retrieve its Real and Imaginary parts simply by multiplying the spectrum of the signal by Heaviside function (unit step function) to remove complex conjugate and DC signals (Fig. 2c). According to Goodman this procedure provides the complex representation of measured signal (Fig. 2d) [46]. To calculate the phase of the signal we used Fast Fourier Transform (Fig. 2e). As a result of remapping the spectral fringe signal linear in k-space is obtained (Fig. 2f) and thus its coherence function obtained after Fourier transformation has appropriate width (Fig. 2g).
Fig. 2 Schematic diagram of remmapping procedure from wavelengths to equidistantly sampled wavenumbers: (a) spectral fringes registered by the calibration Michelson interferometer; (b) amplitude of the Fourier transform of a.; (c) signal b. multiplied by a window to remove complex conjugate and DC signals; (d) complex representation of a.; (e) unwrapped phase of signal d.; (f) signal a. remapped to the wavenumber space; (g) amplitude of the Fourier transform of signal f.
It has to be noted that instead of the phase it is possible to use the numbers of consecutive maxima and minima of the fringe signal as the values of q [35]. The advantage of the described method by iFFT compared to a simple peak picking is that a higher number of effective samples before FFT is obtained and only a minimum amount of information is lost. This method is also insensitive to spectral shape of the fringe signal. The technique described above can be also used in the Spectral OCT, where additional example of the q quantity are the positions of peaks in light spectrum filtered in the Fabry-Perot interferometer with narrow FSR (less than 1 nm) generating optical frequency comb [47].
For the preview mode, when patient is being adjusted, the calibration procedure is performed only once and then calibration coefficients are used for rescaling of every-A-scan. In this case the processing time is comparable with this proposed by Yasuno et al. [45]. However, due to the instabilities of the light source, to obtain the highest possible quality of the data, the K-space resampling procedure has to be applied for each A-scan in the post-processing. Total processing time including the callibration and the resampling steps of each A-scan takes approximately 6ms for the regular 2.4 GHz Athlon processor.
2.3 Imaging of the anterior segment using adjustable axial range
Imaging of the anterior segment of the human eye requires collecting data from a much larger volume than in the case of retinal OCT. Thus, if we keep relatively high axial resolution and high scanning density the size of data to be captured with the acquisition system and stored in the hard drives will increase significantly. Furthermore, in the swept source OCT, high axial resolution simultaneously combined with long ranging depths at high sweep speeds results in prohibitively high fringe frequencies up into the multi 100 MHz range and analog to digital conversion with high bit resolution becomes difficult. To avoid these problems, imaging range, axial resolution and imaging speed have to be compromised.
In the case of a swept source OCT instrument using an FDML laser the imaging speed is fixed and depends on the length of the optical fiber delay line placed in the laser cavity. Furthermore, a minimum imaging speed of 100,000 lines/s or more is necessary for fast, equidistantly sampled 3-dimensional OCT imaging. However, at constant number of sampling points N, we can trade-off axial resolution δz against axial imaging range ∆zmax, if we sweep the laser over a wider or narrower spectral range. Because of the limited sampling rate of currently available A/D converters with high bit-resolution, the only way to increase imaging depth is to reduce ∆λ on expense of in-depth resolution.
One of the most important technological advantages of the presented set-up is its high flexibility. The swept source OCT instrument allows for fast and convenient adjustment of working parameters – axial resolution and measurement range – to operate in different programmable regimes by setting the value of the FFP-TF driving signal amplitude. The lower the amplitude of the driving signal the narrower is the wavelength scanning range. In principle it is possible to choose any value of the axial resolution limited only either by the spectral bandwidth of SOA (highest resolution) or by the laser mode stability (lowest resolution). In our system axial zooming can be performed continuously by changing amplitude of FFP-TF driving signal (∆λ changes) on a function generator. However, it must be taken into account that for different scanning range the procedure of K-space resampling has to be repeated and thus the automation of this procedure is required. For the purpose of the anterior segment imaging we propose using two axial scanning modes: a high axial resolution mode with axial resolution of 9 µm and imaging range of 2 mm and an enhanced imaging range mode with an axial resolution of 25 µm and a high imaging range of 8 mm. Figure 3 shows sensitivity roll-off as a function of optical path difference for both operating modes. The sensitivity roll-off is higher for the high axial resolution mode, because by preserving the same number of sampling points and sampling frequency we average out more optical frequency components. This is similar effect to finite pixel size in spectrometer in Spectral OCT systems [47]. However, this effect should only cause 4 dB drop over the entire measurement range. The remaining 4 dB may be related to the effective laser line width.
Fig. 3 SSOCT sensitivity drop-off as a function of the double pass optical path difference for high axial resolution mode (a) and enhanced imaging range mode (b). Horizontal bar indicates −8dB level.
The instantaneous line width of an FDML laser is influenced by many parameters and there is no comprehensive theoretical description available, yet. However, it has been shown that a complex interplay between dispersion, timing jitter of the filter and total filter sweep speed governs the instantaneous linewidth and with it the instantaneous coherence length [48]. This means, that if in FDML configuration the filter is swept slower in optical frequency per time units, the instantaneous linewidth is narrower and the coherence length improves. This point should be emphasized in so far, because the situation is different for non-FDML swept laser sources. It has been shown, that in rapidly swept non-FDML lasers there is no time for mode competition and the instantaneous spectrum always corresponds to the filter bandwidth at high sweep rates. If in a regular wavelength swept laser, which is swept at very high speeds above the saturation limit and near the single roundtrip limit [35], the sweep range is reduced, usually no linewidth improvement is observed. This means the presented concept of adjustable sweep range is especially suited for FDML based OCT systems. Figure 2b shows, that with the described method, the 8 dB roll-off level can be increased from 3 mm up to ~6 mm (Fig. 3b).
2.4 In vivo measurements
For the purpose of this study 4 healthy eyes and two eyes with diagnosed keratoconus were examined without pupil dilation. The protocols have been approved by Internal Ethics Committee and the subjects signed a consent form. The power of light impinging on the eye was set to 2 mW, which means that it was kept much below the safety requirements of ANSI standards [49]. A various scan protocols designed for imaging of specific features in the anterior chamber were used. In this paper we are showing examples of two- and three-dimensional images and OCT movies. In the cross-sectional images presented in this contribution the intensity of light scattered and/or reflected from the internal structures within the sample is coded in a grey scale. For three-dimensional data visualization a commercial software (Amira 4.1, Visage Imaging, Inc., USA) was used.
2.5 Quantitative analysis of the cornea
In order to obtain conclusive information about the corneal radii of curvature and topography an appropriate scanning protocol has to be selected. To avoid artifacts due to eye movements the tradeoff between optical sampling density and examination time has to be made. The scanning protocol of 50 B-scans with 1000 A-scans each was chosen for this purpose. For the 200 kHz system the examination lasted 250 ms. Imaging depth of 8 mm was chosen to cover the entire cornea.
To perform the quantitative corneal analysis we segmented the anterior and the posterior corneal surfaces out of all cross-sectional images from the 3-D data set. Due to the fact, that all distances measured in OCT are optical, a conversion to the real geometry is necessary. For this purpose a correction for refraction on the air-cornea interface was performed [50]. To recalculate the in-depth axis the group refractive index of the cornea, was set according to the theoretical eye of LeGrand to be 1.377 [51]. To reduce effect of unequal sampling in x and y transverse directions an interpolation was introduced. Before calculating the corneal parameters it must be considered that the cornea is usually misaligned, tilted and inclined with respect to the instrument. This misalignment has to be corrected before any quantitative analysis is performed. High speed, three dimensional OCT examination yields information about the position of the iris and the anterior surface of the crystalline lens supplementary to the corneal cross-sections. To minimize the influence of misalignment, tilt and inclination we propose to refer the 3-D data set to the plane determined by the eye pupil (Fig. 4 ). We believe that this procedure will not provide a true spatial orientation of the corneal apex but at least it can increase the repeatability of topographic analyses by applying the same procedure to different measurements.
Fig. 4 Schematic drawing showing mathematical correction of the data set for misalignment of the eye in respect to the instrument: M – axis normal to the pupil plane, P1 – central point of the pupil, M’ – axis after rotation on 3-D, P2 – origin of the spherical coordinate system.
To find the pupil plane we delineated the pupil based on the OCT en-face view and then we determined axial positions of the points corresponding to the pupil contour. Then we introduced an axis (M-axis) normal to this plane and positioned in the centre of the pupil (P1). The data set may be then rotated in 3-D to set the M-axis as the one of the main axes of the Cartesian coordinate system (Fig. 4). After these steps the topographic maps presenting elevation of the anterior and posterior corneal surfaces in respect to the iris plane may be easily calculated. In order to calculate thickness maps we introduced spherical coordinate system with the origin placed on the M axis, in the constant distance of 7.8 mm from the cornea (P2). This distance corresponds to the radius of curvature of the anterior corneal surface of normal eye according to the LeGrand model. Then the thickness were calculated along the radial coordinate. To calculate elevation maps a sphere was fitted for each corneal surface. The distances between data points and fitted spheres were measured along the radial coordinate. Finally, additional transverse averaging was performed to reduce numerical effects caused by the segmentation.
3. Results and discussion
3.1 Structural imaging of the anterior chamber
Using high resolution mode it is possible to visualize detailed morphology in different areas of the anterior segment. In Fig. 5 an example of a high quality cross-sectional image of the junction between the cornea and sclera of an eye with a contact lens is presented. The OCT image (B-scan) consist of 10,000 A-scans. Each tomogram line was acquired within 5 μs. Resolution of the system is sufficient to distinguish epithelium and delineate the position of the Bowman’s membrane.
Fig. 5 Cross-sectional image of the periphery of the cornea with contact lens acquired in high axial resolution mode.
One of the most important features of the system is its capability to measure the entire anterior chamber together with the cornea and anterior surface of the crystalline lens (Fig. 6 ). The cross-section consisting of 14,000 A-scans was collected within 70 ms (14,000 × 5 µs). Imaging with a short exposure times leads to lower sensitivity and finally to a moderate quality of the image. However, this quality is still sufficient to delineate structural elements including the cornea, limbus, iris, the lens and the ciliary body.
Fig. 6 Cross-sectional image of the anterior chamber together with cornea and anterior surface of crystalline lens acquired in enhanced imaging range mode.
Another clinically interesting anatomical structure in the human anterior segment is irido-corneal angle, located where base of the iris is attached to the peripheral cornea and sclera (Fig. 7a ). In this site the aqueous humor produced by the ciliary body is drained out of the eye. If the dynamic equilibrium is disturbed, due to the improper drainage of the aqueous humor, an increased intraocular pressure may occur causing optic nerve damage. Two- and three-dimensional cross-sectional information on structure and geometry and of the irido-corneal junction may be very helpful for early diagnosis of glaucoma. In Fig. 7b the region of the anterior sclera along with the ciliary body and the iris root is visualized. Additionally, a volumetric rendering of 3-D data comprising 300 B-scans with 1000 A-scans and 1024 pixels (1000 × 300 × 1024) is presented in Fig. 7c.
Fig. 7 Enhanced imaging range mode: cross-sectional images of irido-corneal angle (a) and anterior sclera (b); (Media 1) movie demonstrating 3-D rendering (c) generated from the 3-D data set (1000 × 300 × 1024 voxels).
The high speed of the SSOCT system enables high definition imaging of the large area. As can be seen in Fig. 8 the area of 24 × 24 mm can be measured with a high pixel density of 390 Mvoxels in 1.8 s.
Fig. 8 Large-scale three-dimensional reconstruction of the anterior segment of human eye in vivo: en face view (a), (Media 2) movie presenting 3-D rendering (b) and exemplary cross-sectional image is presented (c). Images are reconstructed from 1200 × 300 × 1024 voxel data set. The size of the imaged volume is 24 × 24 × 8 mm.
In this example the cornea and the anterior chamber along with the upper and lower eyelids and eyelashes are clearly visible at both en-face projection (Fig. 8a) and 3D rendering (Fig. 8b). The same 3-D data set was used to generate both images. Individual B-scan is presented in Fig. 8c.
3.2. Quantitative analysis of cornea in 3D
Figure 9a shows a rendering of the 3-D data set used for a quantitative analysis of the corneal curvature and its thickness in a healthy eye. As a result of a trade-off between the optical sampling density and the total acquisition time, the scanning protocol of 50 B-scans with 1000 A-scans was chosen. Ultrahigh speed swept source OCT system enables performing the entire measurement within only 250 ms. Corneal segmentation from the cross-sectional images (Fig. 9b) was followed by the correction of refraction (Fig. 9c).
Fig. 9 Data acquired in 250 ms with the scanning protocol 1000 × 50 × 1024 voxel selected in order to qualitatively analyze corneal curvature. (Media 3) The movie showing three-dimensional reconstruction (a) and exemplary cross-sectional image before (b) and after refraction correction (c) are presented.
Corneal surfaces were used for calculating the topography maps (Fig. 10a ) and thickness map (Fig. 10b). From the latter it may be seen that the thickness of the cornea changes from the 530 µm in the center to 680 µm at the periphery. Figure 10c presents the elevation BFS (Best Fit Sphere) maps for a normal eye.
Fig. 10 Results of the quantitative corneal analysis of 3-D OCT data, normal eye: topography of anterior and posterior corneal surfaces (a), thickness map (b) and elevation BFS (Best Fit Sphere) maps with radius of the fitted sphere of anterior and posterior corneal surfaces (c). Diameter of the most outer circle in thickness and elevation maps equals 10 mm.
The elevation BFS map presents differences in distance between the sphere and the eye surface calculated in the spherical coordinates. Green color corresponds to the best fit between eye surface and sphere. Areas under the fitted sphere are represented in blue, whereas red color identify areas above the ideal sphere. Irregularities visible in the Fig. 10c are due to the segmentation and averaging procedure.
The quantitative analysis was also performed for the patient with keratoconus. Figure 11a shows the 3-D rendering of OCT data set consists of 300 B-scans with 1000 A-scans each. Such a scanning protocol enables visualizing of anatomical details of anterior segment in 3-D. For quantitative analysis different scanning protocol characterized with lower sampling density but with the shorter acquisition time of 250 ms was chosen. Figure 11b presents one cross-sectional image taken from such 3-D data set. A decrease of the corneal thickness in the para-central region is evident.
Fig. 11 (Media 4) The movie showing reconstruction of three-dimensional OCT data (1000 × 300 × 1024) of the eye with keratoconus (a). An exemplary cross-sectional image taken form 3-D data set acquired in 250 ms with the scanning protocol 1000 × 50 × 1024 voxel (b).
Results of the quantitative corneal analysis of eye with keratoconus are shown in Fig. 12 . Overall the corneal thickness is much lower in the case of keratoconus than in the case of the normal eye. In the center of the cornea the thickness equals only to 240 µm. As can be seen from Fig. 12a,b the posterior corneal surface is highly conical and an apex of this surface is shifted to the left in respect to the anterior surface. The elevation BFS maps (Fig. 12c) differ significantly from those presented earlier for the normal eye. Aspherical shape of both anterior and posterior surfaces may be identified on the elevation maps by the regions characterized by large values of the distance between the corneal surface and the fitted sphere. In this particular case the distances were larger than the standard color scale span.
Fig. 12 Results of the quantitative corneal analysis of 3-D OCT data, keratoconus: topography of anterior and posterior corneal surfaces (a), thickness map (b) and elevation BFS (Best Fit Sphere) maps with radius of the fitted sphere of anterior and posterior corneal surfaces (c). Diameter of the most outer circle in thickness and elevation maps equals 10 mm.
3.3 Imaging of dynamic processes in the anterior chamber
The ultrahigh speed performance of the SSOCT system enables also video-rate monitoring of various dynamic processes in the anterior segment. With speed improvement two- and also three-dimensional imaging can be repeated in time. As an example a pupillary reaction to light stimulus is presented in Fig. 13 . The measurement of 12 volumetric data sets was performed in 3.6 s. Each 3-D data set consists of 400 B-scans, comprising 100 A-scans with 1024 pixels in line.
Fig. 13 (Media5) OCT movie showing pupil reaction on light stimulus, 400 × 100 × 1024 pixels (12 volumes), 15 × 15 mm, single line exposure time 5 μs.
Another potential application of the ultrahigh speed SSOCT instrument is a quantitative evaluation of a blink-induced contact lens movement. It must be noticed that such a measurement can be also performed with a slit lamp, and can be registered and measured with a digital camera. However, the major advantage of OCT is the ability to provide cross-sections that allow to assess how the contact lens alignment changes during its movements on the eye surface. Measurement of the lens movement is important in the lens fitting procedure. Previously we demonstrated video-rate two-dimensional cross-sectional OCT imaging of contact lens movement with the aid of Spectral OCT system [52,53]. Here we can either extend measurement protocol to three dimensions registered in real time or we can increase frame rate in two-dimensional imaging.
Figure 14 shows real time three dimensional OCT registration of blink induced contact lens movement of a rigid gas permeable contact lens (BC 8,4, DIA 9,6, Power −2,75 D, Hecht Contactlinsen GmbH) on a healthy human eye. Movie consists of 12 frames each 100 × 400 × 1024 pixels corresponding to 13 × 13 × 10.5 mm volumes. Frame rate of this volumetric movie is 5 frames per second.
Fig. 14 (Media 6) En-face projection of 3-D data set (100 × 400 × 1024 pixels, 13 × 13 mm) of the blink-induced vertical movement of a rigid gas permeable contact lens.
To demonstrate a feasibility of the high speed OCT imaging applied to the contact lens fit we have performed another experiment in which vertical cross-sectional images were taken at the same horizontal position. Figure 15a shows a real time movie presenting consecutive 100 B-scans comprising 4000 A-scans acquired in 2 s (frame rate of 50 frames per second). The movement of the contact lens can be assessed by choosing a marker, which in our case was a point corresponding to a cross-section of the bottom edge of the lens. The position of the marker can be plotted versus time (Fig. 15b). Such diagram provides information about a relative shift of the lens. In this particular case the lens slips downward by 2.8 mm in 0.9 s after the blink.
Fig. 15 An analysis of the blink-induced vertical movement of a rigid gas permeable contact lens: (a) single frame (cross-sectional image) from a 2 s movie (Media 7) and (b) plot representing the relative blink-induced movement of the inferior lens edge as a function of frame number or time.
7. Conclusions
In this study we demonstrate the high speed Fourier domain Optical Coherence Tomography working with the wavelength swept laser source at 1310 nm applied to ophthalmic imaging of the human anterior segment. An effective axial scan rate of 200 kHz was achieved, enabling three-dimensional imaging of anterior segment with reduced number of motion artifacts. Application of Fabry-Perot filter ascertain high flexibility of the system, in terms of adjustable axial imaging range. This feature is especially important in the case of corneal imaging. The high speed performance of the SSOCT system shortens an examination time and thus reduces influence of inevitable motion artifacts. The preliminary images obtained with this system illustrate the possibility of ultrahigh speed and high resolution imaging of large volumes associated with the anterior chamber of the human eye. We were also able to demonstrate corneal topography based on dense raster OCT scan with reduced motion artifacts.
Acknowledgements
This work was supported by Research Grant of Polish Government (years 2006-2009), the Grant N N402 084435 and EuroHORCs-European Science Foundation EURYI Award EURYI-01/2008-PL (M.W.). Project was operated within the Foundation for Polish Science Ventures Programme co-financed by the EU European Regional Development Fund (M.G.) and START Programme (M.S.). R. Huber acknowledges support from the Emmy Noether program of the German Research Foundation (DFG-HU 1006/2-1). Authors would like to acknowledge help, patience and support of dr Kevin Hsu and Clyde Story from Micron Optics.
References and links
1. A. Konstantopoulos, P. Hossain, and D. F. Anderson, “Recent advances in ophthalmic anterior segment imaging: a new era for ophthalmic diagnosis?” Br. J. Ophthalmol. 91(4), 551–557 (2007). [PubMed]
2. A. C. Cheng, S. K. Rao, S. Lau, C. K. Leung, and D. S. Lam, “Central corneal thickness measurements by ultrasound, Orbscan II, and Visante OCT after LASIK for myopia,” J. Refract. Surg. 24(4), 361–365 (2008). [PubMed]
3. D. Huang, E. A. Swanson, C. P. Lin, J. S. Schuman, W. G. Stinson, W. Chang, M. R. Hee, T. Flotte, K. Gregory, C. A. Puliafito, and J. G. Fujimoto, “Optical coherence tomography,” Science 254(5035), 1178–1181 (1991). [PubMed]
4. J. F. de Boer, T. E. Milner, M. J. C. van Gemert, and J. S. Nelson, “Two-dimensional birefringence imaging in biological tissue by polarization-sensitive optical coherence tomography,” Opt. Lett. 22(12), 934–936 (1997). [PubMed]
5. A. F. Fercher, W. Drexler, C. K. Hitzenberger, and T. Lasser, “Optical coherence tomography-principles and applications,” Rep. Prog. Phys. 66, 239–303 (2003).
6. M. Wojtkowski, R. Leitgeb, A. Kowalczyk, T. Bajraszewski, and A. F. Fercher, “In vivo human retinal imaging by Fourier domain optical coherence tomography,” J. Biomed. Opt. 7(3), 457–463 (2002). [PubMed]
7. N. Nassif, B. Cense, B. H. Park, S. H. Yun, T. C. Chen, B. E. Bouma, G. J. Tearney, and J. F. de Boer, “In vivo human retinal imaging by ultrahigh-speed spectral domain optical coherence tomography,” Opt. Lett. 29(5), 480–482 (2004). [PubMed]
8. M. Wojtkowski, V. Srinivasan, J. G. Fujimoto, T. Ko, J. S. Schuman, A. Kowalczyk, and J. S. Duker, “Three-dimensional retinal imaging with high-speed ultrahigh-resolution optical coherence tomography,” Ophthalmology 112(10), 1734–1746 (2005). [PubMed]
9. U. Schmidt-Erfurth, R. A. Leitgeb, S. Michels, B. Povazay, S. Sacu, B. Hermann, C. Ahlers, H. Sattmann, C. Scholda, A. F. Fercher, and W. Drexler, “Three-dimensional ultrahigh-resolution optical coherence tomography of macular diseases,” Invest. Ophthalmol. Vis. Sci. 46(9), 3393–3402 (2005). [PubMed]
10. J. A. Izatt, M. R. Hee, E. A. Swanson, C. P. Lin, D. Huang, J. S. Schuman, C. A. Puliafito, and J. G. Fujimoto, “Micrometer-scale resolution imaging of the anterior eye in vivo with optical coherence tomography,” Arch. Ophthalmol. 112(12), 1584–1589 (1994). [PubMed]
11. J. J. Kaluzny, M. Wojtkowski, and A. Kowalczyk, “Imaging of the anterior segment of the eye by spectral optical coherence tomography,” Opt. Appl. 32, 581–589 (2002).
12. G. Baikoff, E. Lutun, C. Ferraz, and J. Wei, “Static and dynamic analysis of the anterior segment with optical coherence tomography,” J. Cataract Refract. Surg. 30(9), 1843–1850 (2004). [PubMed]
13. C. K. Leung, W. M. Chan, C. Y. Ko, S. I. Chui, J. Woo, M. K. Tsang, and R. K. Tse, “Visualization of anterior chamber angle dynamics using optical coherence tomography,” Ophthalmology 112(6), 980–984 (2005). [PubMed]
14. B. J. Kaluzny, B. J. Kaluzy, J. J. Kałuzny, A. Szkulmowska, I. Gorczyńska, M. Szkulmowski, T. Bajraszewski, M. Wojtkowski, and P. Targowski, “Spectral optical coherence tomography: a novel technique for cornea imaging,” Cornea 25(8), 960–965 (2006). [PubMed]
15. Y. Li, R. Shekhar, and D. Huang, “Corneal pachymetry mapping with high-speed optical coherence tomography,” Ophthalmology 113(5), 792–799, e2 (2006). [PubMed]
16. S. Radhakrishnan, J. See, S. D. Smith, W. P. Nolan, Z. Ce, D. S. Friedman, D. Huang, Y. Li, T. Aung, and P. T. K. Chew, “Reproducibility of anterior chamber angle measurements obtained with anterior segment optical coherence tomography,” Invest. Ophthalmol. Vis. Sci. 48(8), 3683–3688 (2007). [PubMed]
17. V. Christopoulos, L. Kagemann, G. Wollstein, H. Ishikawa, M. L. Gabriele, M. Wojtkowski, V. Srinivasan, J. G. Fujimoto, J. S. Duker, D. K. Dhaliwal, and J. S. Schuman, “In vivo corneal high-speed, ultra high-resolution optical coherence tomography,” Arch. Ophthalmol. 125(8), 1027–1035 (2007). [PubMed]
18. I. Grulkowski, M. Gora, M. Szkulmowski, I. Gorczynska, D. Szlag, S. Marcos, A. Kowalczyk, and M. Wojtkowski, “Anterior segment imaging with Spectral OCT system using a high-speed CMOS camera,” Opt. Express 17(6), 4842–4858 (2009). [PubMed]
19. E. Götzinger, M. Pircher, M. Sticker, A. F. Fercher, and C. K. Hitzenberger, “Measurement and imaging of birefringent properties of the human cornea with phase-resolved, polarization-sensitive optical coherence tomography,” J. Biomed. Opt. 9(1), 94–102 (2004). [PubMed]
20. E. Götzinger, M. Pircher, I. Dejaco-Ruhswurm, S. Kaminski, C. Skorpik, and C. K. Hitzenberger, “Imaging of birefringent properties of keratoconus corneas by polarization-sensitive optical coherence tomography,” Invest. Ophthalmol. Vis. Sci. 48(8), 3551–3558 (2007). [PubMed]
21. M. Yamanari, S. Makita, and Y. Yasuno, “Polarization-sensitive swept-source optical coherence tomography with continuous source polarization modulation,” Opt. Express 16(8), 5892–5906 (2008). [PubMed]
22. C. K. Leung, H. Li, R. N. Weinreb, J. Liu, C. Y. Cheung, R. Lai, C. P. Pang, and D. S. Lam, “Anterior Chamber Angle Measurement with Anterior Segment Optical Coherence Tomography (OCT) - A Comparison Between Slit Lamp OCT and Visante OCT,” Invest Ophthalmol Vis Sci (2008).
23. S. Radhakrishnan, J. Goldsmith, D. Huang, V. Westphal, D. K. Dueker, A. M. Rollins, J. A. Izatt, and S. D. Smith, “Comparison of optical coherence tomography and ultrasound biomicroscopy for detection of narrow anterior chamber angles,” Arch. Ophthalmol. 123(8), 1053–1059 (2005). [PubMed]
24. M. Miura, K. Kawana, T. Iwasaki, T. Kiuchi, T. Oshika, H. Mori, M. Yamanari, S. Makita, T. Yatagai, and Y. Yasuno, “Three-dimensional anterior segment optical coherence tomography of filtering blebs after trabeculectomy,” J. Glaucoma 17(3), 193–196 (2008). [PubMed]
25. S. Radhakrishnan, A. M. Rollins, J. E. Roth, S. Yazdanfar, V. Westphal, D. S. Bardenstein, and J. A. Izatt, “Real-time optical coherence tomography of the anterior segment at 1310 nm,” Arch. Ophthalmol. 119(8), 1179–1185 (2001). [PubMed]
26. C. Kerbage, H. Lim, W. Sun, M. Mujat, and J. F. de Boer, “Large depth-high resolution full 3D imaging of the anterior segments of the eye using high speed optical frequency domain imaging,” Opt. Express 15(12), 7117–7125 (2007). [PubMed]
27. R. Leitgeb, C. K. Hitzenberger, and A. F. Fercher, “Performance of fourier domain vs. time domain optical coherence tomography,” Opt. Express 11(8), 889–894 (2003). [PubMed]
28. J. F. de Boer, B. Cense, B. H. Park, M. C. Pierce, G. J. Tearney, and B. E. Bouma, “Improved signal-to-noise ratio in spectral-domain compared with time-domain optical coherence tomography,” Opt. Lett. 28(21), 2067–2069 (2003). [PubMed]
29. M. A. Choma, M. V. Sarunic, C. H. Yang, and J. A. Izatt, “Sensitivity advantage of swept source and Fourier domain optical coherence tomography,” Opt. Express 11(18), 2183–2189 (2003). [PubMed]
30. S. Fukuda, K. Kawana, Y. Yasuno, and T. Oshika, “Anterior ocular biometry using 3-dimensional optical coherence tomography,” Ophthalmology 116(5), 882–889 (2009). [PubMed]
31. S. H. Yun, C. Boudoux, G. J. Tearney, and B. E. Bouma, “High-speed wavelength-swept semiconductor laser with a polygon-scanner-based wavelength filter,” Opt. Lett. 28(20), 1981–1983 (2003). [PubMed]
32. S. Yun, G. Tearney, J. de Boer, N. Iftimia, and B. Bouma, “High-speed optical frequency-domain imaging,” Opt. Express 11(22), 2953–2963 (2003). [PubMed]
33. R. Huber, M. Wojtkowski, J. G. Fujimoto, J. Y. Jiang, and A. E. Cable, “Three-dimensional and C-mode OCT imaging with a compact, frequency swept laser source at 1300 nm,” Opt. Express 13(26), 10523–10538 (2005). [PubMed]
34. M. A. Choma, K. Hsu, and J. A. Izatt, “Swept source optical coherence tomography using an all-fiber 1300-nm ring laser source,” J. Biomed. Opt. 10, 044009 (2005).
35. R. Huber, M. Wojtkowski, K. Taira, J. Fujimoto, and K. Hsu, “Amplified, frequency swept lasers for frequency domain reflectometry and OCT imaging: design and scaling principles,” Opt. Express 13(9), 3513–3528 (2005). [PubMed]
36. R. Huber, M. Wojtkowski, and J. G. Fujimoto, “Fourier Domain Mode Locking (FDML): A new laser operating regime and applications for optical coherence tomography,” Opt. Express 14(8), 3225–3237 (2006). [PubMed]
37. R. Huber, D. C. Adler, and J. G. Fujimoto, “Buffered Fourier domain mode locking: Unidirectional swept laser sources for optical coherence tomography imaging at 370,000 lines/s,” Opt. Lett. 31(20), 2975–2977 (2006). [PubMed]
38. M. Y. Jeon, J. Zhang, and Z. Chen, “Characterization of Fourier domain mode-locked wavelength swept laser for optical coherence tomography imaging,” Opt. Express 16(6), 3727–3737 (2008). [PubMed]
39. G. Y. Liu, A. Mariampillai, B. A. Standish, N. R. Munce, X. Gu, and I. A. Vitkin, “High power wavelength linearly swept mode locked fiber laser for OCT imaging,” Opt. Express 16(18), 14095–14105 (2008). [PubMed]
40. Y. Mao, C. Flueraru, S. Sherif, and S. Chang, “High performance wavelength-swept laser with mode-locking technique for optical coherence tomography,” Opt. Commun. 282, 88–92 (2009).
41. R. Navarro, L. González, and J. Hernández, “Optics of the average normal cornea from general and canonical representations of its surface topography,” J. Opt. Soc. Am. A 23, 219–232 (2006).
42. M. V. Sarunic, S. Asrani, and J. A. Izatt, “Imaging the ocular anterior segment with real-time, full-range Fourier-domain optical coherence tomography,” Arch. Ophthalmol. 126(4), 537–542 (2008). [PubMed]
43. L. Plesea and A. G. Podoleanu, “Direct corneal elevation measurements using multiple delay en face optical coherence tomography,” J. Biomed. Opt. 13(5), 054054 (2008). [PubMed]
44. C. M. Eigenwillig, B. R. Biedermann, G. Palte, and R. Huber, “K-space linear Fourier domain mode locked laser and applications for optical coherence tomography,” Opt. Express 16(12), 8916–8937 (2008). [PubMed]
45. Y. Yasuno, V. D. Madjarova, S. Makita, M. Akiba, A. Morosawa, C. Chong, T. Sakai, K. P. Chan, M. Itoh, and T. Yatagai, “Three-dimensional and high-speed swept-source optical coherence tomography for in vivo investigation of human anterior eye segments,” Opt. Express 13(26), 10652–10664 (2005). [PubMed]
46. J. Goodman, Statistical Optics (Wiley New York, 1985).
47. T. Bajraszewski, M. Wojtkowski, M. Szkulmowski, A. Szkulmowska, R. Huber, and A. Kowalczyk, “Improved spectral optical coherence tomography using optical frequency comb,” Opt. Express 16(6), 4163–4176 (2008). [PubMed]
48. B. R. Biedermann, W. Wieser, C. M. Eigenwillig, T. Klein, and R. Huber, “Dispersion, coherence and noise of Fourier domain mode locked lasers,” Opt. Express . 17, 9947-9961 (2009). [PubMed]
49. A. N. S. Institute, “American National Standard for Safe use of Lasers,” (American National Standards Institute, Orlando, FL, 2000).
50. V. Westphal, A. M. Rollins, S. Radhakrishnan, and J. Izatt, “Correction of geometric and refractive image distortions in optical coherence tomography applying Fermat’s principle,” Opt. Express 10(9), 397–404 (2002). [PubMed]
51. Y. Le Grand, and S. El Hage, Physiological Optics (Springer-Verlag, Berlin, 1980).
52. B. J. Kałuzny, J. J. Kaluzny, A. Szkulmowska, I. Gorczyńska, M. Szkulmowski, T. Bajraszewski, P. Targowski, and A. Kowalczyk, “Spectral optical coherence tomography: a new imaging technique in contact lens practice,” Ophthalmic Physiol. Opt. 26(2), 127–132 (2006). [PubMed]
53. B. J. Kaluzny, W. Fojt, A. Szkulmowska, T. Bajraszewski, M. Wojtkowski, and A. Kowalczyk, “Spectral optical coherence tomography in video-rate and 3D imaging of contact lens wear,” Optom. Vis. Sci. 84(12), 1104–1109 (2007). [PubMed]
