Abstract
Ultrahigh-resolution (UHR) spectral-domain optical coherence tomography (SD-OCT) and speckle noise-reduced SD-OCT instruments were compared with regard to their ability to visualize retinal microstructures and detect micropathologies in the same series of eyes in this hospital-based study. Both the instruments identically visualized normal retinal structures, except for the retinal ganglion cell layer, which was better delineated by the speckle noise-reduced SD-OCT instrument. Retinal pigment epithelium (RPE) and Bruch’s membrane were better delineated by UHR SD-OCT. Retinal and sub-RPE pathologies were also identically visualized by both the instruments. Layer differentiation for locating each pathology was better visualized by speckle noise-reduced SD-OCT.
©2009 Optical Society of America
1. Introduction
Optical coherence tomography (OCT) is an interferometric imaging technique that allows for in vivo high-resolution, cross-sectional imaging of biological tissue, which is clinically applied in ophthalmology [1]. Since biopsy of the retina is not possible in a routine clinical test, OCT has established a unique status in routine ophthalmologic examination as the only technique that can noninvasively obtain depth-resolved retinal structural information. Commercially available standard-resolution OCT systems can delineate individual retinal layers and aid in the diagnosis of various retinal diseases. However, some of the retinal structures and micropathologies are not evident in the cross-sectional images obtained by these conventional OCT systems, which suffer from drawbacks such as a relatively low-depth resolution of ~10 µm, noisy B-scan images, and a relatively slow acquisition speed. Therefore, improvement of B-scan images on laboratory basis has been one of the main purposes of research, which has been pursued by exploring ultrahigh-resolution (UHR) OCT, adaptive optics OCT, eye-tracking OCT, and changes in signal detection technique from time-domain (TD) detection to spectral-domain (SD) detection. SD detection technique has a higher potential for ophthalmic clinical use than TD detection technique, because it shows extremely high sensitivity (i.e., signal-to-noise ratio) [2–4] and image-acquisition speed, required for ophthalmic applications [5–7].
Numerous studies have demonstrated that high axial resolution of a laboratory OCT system allows the visualization of finer retinal structures and pathologies [8–16]. Axial resolution in OCT depends mainly on the width and center of the wavelength of the light source used. Since the commercially available OCT instruments use light sources with the central wavelength of approximately 800 nm, axial resolution depends on the width of the wavelength of the light source used. With regard to commercial instruments, Stratus OCT instrument (Carl Zeiss Meditec, Dublin, CA) achieves an axial resolution of ~10 µm, using a 20-nm superluminescent diode (SLD). Currently, the commercially available spectral-domain OCT (SD-OCT) instruments achieve axial resolutions of 5~7 µm by using approximately 35- ~50-nm SLDs. Since the SD technique shows a higher sensitivity (i.e., signal-to-noise ratio) than the TD technique, the cross-sectional images obtained by the currently available SD-OCT instruments have both a higher axial resolution and sensitivity than those obtained by Stratus OCT. Thus, the currently available SD-OCT instruments allow improved visualization of 3 lines in the sensory retina, which represent the external limiting membrane (ELM), the photoreceptor inner and outer segment junction (IS/OS), and probably the interdigitation of the photoreceptor outer segments and the RPE; all these 3 are diagnostically useful in understanding the pathologies and structural damages of the macula [9–12,15,16].
On commercial basis, further improvement of B-scan images has been achieved; by using UHR SD-OCT and speckle noise-reduction technology with SD-OCT. Many authors have previously shown that UHR SD-OCT using a femtosecond titanium-sapphire laser light source achieves a 3-µm B-scan image with a higher contrast than that of the image obtained by UHR TD-OCT, and allows highly detailed imaging of pathologies [17–19]. However, the expensive cost of the titanium-sapphire laser light source hinders wide commercial use of this technique. Another approach using a light-source system composed of 2 SLDs with different central wavelengths has been shown to be effective in achieving a 3-µm B-scan image in the laboratory-prototype system [20,21]. SOCT Copernicus HR (Optopol Technology S.A., Zawiercie, Poland), a commercial prototype of the ultrahigh-resolution imaging procedure, employs the 2-SLD light-source system and has achieved an axial resolution of 3 µm.
Another approach to improve B-scan images is to reduce speckle noise via multiple B-scan averaging [22]. In OCT B-scan images, the dominating noise that obscures the fine retinal structures is usually the speckle noise arising from the interference of coherent waves backscattered from nearby scatterers. The reduction of speckle noise in SD-OCT enables highly detailed imaging of pathologies in the photoreceptor and RPE layers [23]. Although high-speed image acquisition ability of SD-OCT allows effective acquisition of multiple B-scan images at the same location of interest, eye movement remains to be the greatest problem in ophthalmology, causing the failure of precise multiple B-scan averaging for speckle noise reduction. Spectralis HRA+OCT (Heidelberg Engineering, Heidelberg, Germany), which shows an axial resolution of only 7 µm, has enabled the most effective reduction of speckle noise by using a three-dimensional eye-tracking system, allowing more precise multiple B-scan averaging.
Currently, the 2 abovementioned SD-OCT instruments, UHR SD-OCT instrument SOCT Copernicus HR, and the speckle-noise-reduced SD-OCT instrument Spectralis HRA+OCT, seem to provide the best-enhanced B-scan images for visualizing fine retinal structures and detecting retinal pathologies. In this study, we compared the B-scan images of various retinal pathologies, which were obtained using the 2 abovementioned instruments characterized by UHR SD-OCT and speckle noise-reduced SD-OCT with standard axial resolution; the images were obtained from the same patients in our outpatient department. This comparison, which was based on whether the visualization of pathologies in each image is useful for clinical diagnosis, will facilitate the understanding of the impact of each approach on the visualization of retinal pathologies in B-scan images.
2. Method
2.1 Subjects
We examined 5 normal eyes and 27 eyes of 27 referral patients with various retinal diseases. These patients were examined in the Department of Ophthalmology at the Kyoto University Hospital between March and June 2008 (Table 1). All investigations conformed to the principles laid down by the Declaration of Helsinki. This study was approved by the Institutional Review Board and Ethics Committee of Kyoto University Graduate School of Medicine. Informed consent for the examination was obtained from all the patients.
For diagnosis, all the patients underwent a comprehensive ophthalmologic examination, including the measurement of best-corrected visual acuity (BCVA) by using Snellen charts, indirect and contact lens slit-lamp biomicroscopy, fundus photography, visual field testing, and fluorescein angiography (for selected cases).
2.2 Optical coherence tomography imaging
OCT imaging by using SOCT Copernicus HR and Spectralis HRA+OCT was performed on the same day after pupillary dilatation with eye drops containing 0.5% tropicamide and 2.5% phenylephrine. Two spectral, cross-hair OCT scans centered on the fovea were obtained.
2.3 Ultrahigh-resolution spectral-domain optical coherence tomography
SOCT Copernicus HR was used to obtain UHR B-scan images. This instrument is a commercial prototype that employs an 850-nm SLD light source with a bandwidth of 85 nm. The optical depth resolution of the instrument is ~3 µm. Each A-scan has a depth of 2 mm consists of 2,048 pixels, providing a digital depth sampling of 1 µm per pixel. In this study, high-definition B-scans were acquired, with each B-scan spanning a 6-mm region and consisting of 10,610 A-scans; each B-scan acquisition was performed in approximately 0.2 seconds and this provides a digital transverse sampling of 0.57 µm per pixel. The instrument scans the retina at 52,000 A-scans per second.
2.4 Speckle noise-reduced spectral-domain optical coherence tomography
At each location of interest on the retina, 50 SD-OCT images were acquired using a Spectralis™ HRA+OCT. The instrument uses broadband 870-nm SLD light source with a bandwidth of 35 nm. The optical depth resolution of the instrument is ~7 µm. Each A-scan has a depth of 2 mm and consists of 512 pixels, providing a digital depth sampling of 3.9 µm per pixel. In this study, each B-scan spanned 30° and consisted of 1,536 A-scans, providing a digital transverse sampling of 5 µm per pixel. The instrument scans the retina at 40,000 A-scans per second.
The combination of high-resolution scanning laser imaging of the retina and SD-OCT enables real-time tracking of eye movements and real-time averaging of multiple OCT B-scans, thus reducing the speckle noise of OCT images. In this study, 50 B-scans were averaged to reduce the speckle noise. Theoretically, the time required to complete 50 B-scans is 1.92 seconds. However, this instrument pauses the scanning procedure during the periods when the patient is blinking. Therefore, the actual scanning time was higher than the theoretical value.
3. Results
3.1 Normal eye: border
In the fovea, 3 highly reflective lines representing the ELM, IS/OS, and the interdigitation of the cone photoreceptor outer segments and the RPE were identically visualized by both the UHR SD-OCT and speckle noise-reduced SD-OCT (Fig. 1). ELM reflection is thought to result from the border (tight junctions) between the Müller cells and the photoreceptors. IS/OS backreflection is thought to result from the border of the inner and outer segments. Thus, the 3 structures are only the boundaries of different structures. The 3 lines seem to reflect the abrupt change in the optical index of refraction at the respective boundaries. Therefore, the resolution of each line seems to be determined not by the lines themselves, but by whether the axial resolution of an instrument is higher than the axial distance between the neighboring lines. This is supported by a previous report in which a B-scan image obtained at an axial resolution of 20 µm showed the ELM line after reduction of speckle noise [22].
In the extrafoveal region, the line representing the interdigitation of the cone photoreceptor outer segments and the RPE seemed to end and another line appeared at the level closer to the RPE line in both the SD-OCT instruments; this line probably represents the interdigitation of the rod photoreceptor outer segments and the RPE [24].
3.2 Normal eye: Differentiation of the intensity of neighboring layers
Each layer has its own intensity of backreflection. The contrast between neighboring layers seemed to be higher in speckle noise-reduced SD-OCT images than in UHR SD-OCT images (Fig. 1). This comparison was particularly evident in the visualization of the ganglion cell layer (GCL). The GCL has lower reflectivity than the inner plexiform layer (IPL), but the difference in backreflection between the 2 layers is small as compared to that between other boundary pairs such as retinal nerve fiber layer (RNFL)-GCL, IPL-inner nuclear layer (INL), INL-outer plexiform layer (OPL), and OPL-outer nuclear layer (ONL). The GCL-IPL borderin the speckle-noise-reduced SD-OCT image (c) was much clearer than that in the single scan SD-OCT image (d), when the images were obtained using the same instrument and the same number of axial scans. In UHR SD-OCT images, the GCL-IPL border was not clear even if the number of axial scans was increased to 10,610 (a). Because each layer is considerably thicker than shown by the axial resolution of both the instruments, speckle noise seems to have more influence on border visualization between layers than a low axial resolution. In contrast, the Bruch’s membrane was well differentiated in most of the UHR SD-OCT images, but rarely in speckle-noise-reduced SD-OCT images. This is probably because the 2 lines representing the RPE and Bruch’s membrane are too close to be resolved by the optical depth resolution (~7 µm) of the speckle-noise-reduced SD-OCT.
3.3 Retinal pigment epithelium
The RPE layer seemed to be better resolved into two highly-reflective layers representing the RPE and Bruch’s membrane in UHR-SD-OCT than speckle-noise reduced SD-OCT (Fig. 1).
3.4 Pathological retina: Posterior hyaloid membrane
Both the instruments clearly visualized the posterior hyaloid membrane (PHM) in the eyes with vitreoretinal traction syndrome (VRT), epiretinal membrane (ERM), and macular hole (MH). Although the detached PHM was clearly visualized by TD-OCT instruments with poor axial resolution, the PHM attached to the retina in the eyes with ERM or VRT could not be clearly visualized by the conventional TD-OCT instruments. In contrast, the PHM attached to the retina and the gap between the membrane and the inner surface of the retina were clearly depicted by both UHR SD-OCT and the speckle-noise-reduced SD-OCT.
3.5 Pathological retina: Intraretinal pathologies
Intraretinal pathologies such as cystoid macular edema (CME) and hard exudates (HE) were observed in macular edema associated with branch retinal vascular occlusion (BRVO) and diabetic maculopathy (DM). CME was identically visualized by both the instruments, whereas HEs appeared to be better visualized by speckle noise-reduced SD-OCT, although it is possible that the visualization of tiny hard exudates will be quite different between the images from the two instruments if the scan lines of the images are not identical (Fig. 2). It was easy to locate each pathological feature by speckle noise-reduced SD-OCT, probably because of the high contrast between the layers. Identification of the thickening of retinal layers that occurs in ERM was also easy by speckle noise-reduced SD-OCT. The shadowing effects observed in the numerous cystoid spaces are responsible for the loss of ELM- and IS/OS-associated signals in both the instruments; therefore, it was difficult to correctly interpret the disrupted ELM and IS/OS lines.
3.6 Pathological retina: 3 lines in elevated retina
The two lines representing the IS/OS and interdigitation of the cone photoreceptor outer segments and the RPE become invisible in the elevated retina due to macular hole and serous retinal detachment in central serous chorioretinopathy (CSC) (Fig. 3). In contrast, the ELM line is clearly visible in the elevated retina. The decrease in the former two lines’ reflection has been attributed to the altered orientation of the respective boundaries, which is caused by the elevation of the photoreceptor tissue away from the RPE [11]. This retina-specific optical nature of the two lines was identically observed in the OCT images obtained by both the instruments.
The three highly reflective lines have been thought to be indicators of the photoreceptor layer integrity or impairment. In retinitis pigmentosa, the three lines indicate the region where the photoreceptor layer is preserved. Both instruments showed the region clearly (Fig. 4).
3.7 Pathological retina: Subretinal pigment epithelium pathologies
Pathological changes in RPE morphology, such as pigment epithelium detachment and occult choroidal neovascularization (CNV), occur in exudative age-related macular degeneration (AMD) and CSC. The common feature of pathological RPE is the undulation of the RPE line and the underlying thin straight line representing the outer part of Bruch’s membrane, which could not be visualized by Stratus OCT [25]. The outer part of Bruch’s membrane was identically visualized by both the instruments (Fig. 5).
In exudative AMD, moderately reflective lesions were observed between the undulating RPE line and the Bruch’s membrane line, which represent intra-Bruch’s membrane CNV. This feature of CNV was identically delineated by the two instruments (Fig. 5).
4. Discussion
Numerous articles published in ophthalmology journals have reported that UHR OCT using TD and SD detection techniques enables improved visualization of pathological microstructures in various macular diseases [8–19]. A UHR SD-OCT instrument is soon expected to be commercially available. Some of the commercially available SD-OCT instruments with a high resolution of 5~7 µm employ speckle noise reduction technology to improve B-scan images via multiple B-scan averaging. Although SD-OCT allows high-speed imaging at 17,000~52,000 Hz, ocular movements hinder the acquirement of multiple B-scans in the same angle, at the same depth, and at the same location of interest. Thus, precise averaging of multiple B-scans is not possible in the case of many patients. Spectralis employs a three-dimensional eye-tracking system, which has been proved to be effective in reducing the influence of eye movements. The three-dimensional eye-tracking system allows more precise acquisition of as many as 50 B-scans at the same location of interest for removing the speckle noise. Therefore, in the present study, the best commercial instrument currently available for speckle noise reduction was compared with a prototype instrument with UHR. However, it is not certain whether the three-dimensional eye-tracking system can allow precise averaging in diseased eyes with poor fixation. Although the possibility remains that the B-scan images obtained by both the instruments will be improved by improving the systems involved, we believe that our study has successfully revealed the advantages of each technology in visualizing retinal microstructures and detecting retinal pathologies, which are useful for ophthalmic clinical use.
5. Conclusion
Comparison of B-scan images obtained from the same patients on the same day by using UHR SD-OCT and speckle noise-reduced SD-OCT showed that both instruments can delineate the retinal microstructure with fine detail, and each instrument has different strengths in the imaging of various intraretinal and sub-RPE structures. From the clinical point of view with regard to retinal structures and pathologies, the resolution of OCT B-scans based on SD-OCT seems to be identically enhanced by ultrahigh axial resolution and speckle noise reduction. Thus, in the future, both the technologies should be combined in order to achieve a B-scan image that can resolve pathological structures to the maximum possible extent.
Acknowledgments
The use of the prototype instrument SOCT Copernicus HR was kindly allowed by Optopol Technology S.A. (Zawiercie, Poland).
References and links
1. 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, 1178–1181 (1991). [CrossRef] [PubMed]
2. R. Leitgeb, C. K. Hitzenberger, and A. F. Fercher, “Performance of fourier domain vs. time domain optical coherence tomography,” Opt. Express 11, 889–894 (2003). [CrossRef] [PubMed]
3. J. F. de Boer, B. Cense, B. H. Park, M. C. Pierce, G. J. Tearney, and B. E. Bouma, “Signal to noise gain of spectral domain over time domain optical coherence tomography,” Opt. Lett. 28, 2067–2069 (2003). [CrossRef] [PubMed]
4. M. A. Choma, M. V. Sarunic, C. Yang, and J. A. Izatt, “Sensitivity advantage of swept source and fourier domain optical coherence tomography,” Opt. Express 11, 2183–2189 (2003). [CrossRef] [PubMed]
5. 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 ultra high-speed spectral domain optical coherence tomography,” Opt. Lett. 29, 480–482 (2004). [CrossRef] [PubMed]
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, 457–463 (2002). [CrossRef] [PubMed]
7. M. Wojtkowski, T. Bajraszewski, P. Targowski, and A. Kowalczyk, “Real-time in vivo imaging by high-speed spectral optical coherence tomography,” Opt. Lett. 28, 1745–1747 (2003). [CrossRef] [PubMed]
8. W. Drexler, U. Morgner, R. K. Ghanta, F. X. Kärtner, J. S. Schuman, and J. G. Fujimoto, “Ultrahigh-resolution ophthalmic optical coherence tomography,” Nat. Med. 7, 502–507 (2001). [CrossRef] [PubMed]
9. W. Drexler, H. Sattmann, B. Hermann, T. H. Ko, M. Stur, A. Unterhuber, C. Scholda, O. Findl, M. Wirtitsch, J. G. Fujimoto, and A. F. Fercher, “Enhanced visualization of macular pathology with the use of ultrahigh-resolution optical coherence tomography,” Arch. Ophthalmol. 121, 695–706 (2003). [CrossRef] [PubMed]
10. T. H. Ko, J. G. Fujimoto, J. S. Duker, L. A. Paunescu, W. Drexler, C. R. Baumal, C. A. Puliafito, E. Reichel, A. H. Rogers, and J. S. Schuman, “Comparison of ultrahigh- and standard-resolution optical coherence tomography for imaging macular hole pathology and repair,” Ophthalmology 111, 2033–2043 (2004). [CrossRef] [PubMed]
11. E. Ergun, B. Hermann, M. Wirtitsch, A. Unterhuber, T. H. Ko, H. Sattmann, C. Scholda, J. G. Fujimoto, M. Stur, and W. Drexler, “Assessment of central visual function in Stargardt’s disease/fundus flavimaculatus with ultrahigh-resolution optical coherence tomography,” Invest. Ophthalmol. Vis. Sci. 46, 310–316 (2005). [CrossRef]
12. T. H. Ko, J. G. Fujimoto, J. S. Schuman, L. A. Paunescu, A. M. Kowalevicz, I. Hartl, W. Drexler, G. Wollstein, H. Ishikawa, and J. S. Duker, “Comparison of ultrahigh- and standard-resolution optical coherence tomography for imaging macular pathology,” Ophthalmology 112, 2033–2043 (2005). [CrossRef]
13. M. G. Wirtitsch, E. Ergun, B. Hermann, A. Unterhuber, M. Stur, C. Scholda, H. Sattmann, T. H. Ko, J. G. Fujimoto, and W. Drexler, “Ultrahigh resolution optical coherence tomography in macular dystrophy,” Am. J. Ophthalmol. 140, 976–983 (2005). [CrossRef] [PubMed]
14. A. J. Witkin, T. H. Ko, J. G. Fujimoto, J. S. Schuman, C. R. Baumal, A. H. Rogers, E. Reichel, and J. S. Duker, “Redefining lamellar holes and the vitreomacular interface: an ultrahigh-resolution optical coherence tomography study,” Ophthalmology 113, 388–397 (2006). [CrossRef] [PubMed]
15. L. S. Schocket, A. J. Witkin, J. G. Fujimoto, T. H. Ko, J. S. Schuman, A. H. Rogers, C. Baumal, E. Reichel, and J. S. Duker, “Ultrahigh-resolution optical coherence tomography in patients with decreased visual acuity after retinal detachment repair,” Ophthalmology 113, 666–672 (2006). [CrossRef] [PubMed]
16. T. H. Ko, A. J. Witkin, J. G. Fujimoto, A. Chan, A. H. Rogers, C. R. Baumal, J. S. Schuman, W. Drexler, E. Reichel, and J. S. Duker, “Ultrahigh-resolution optical coherence tomography of surgically closed macular holes,” Arch. Ophthalmol. 124, 827–836 (2006). [CrossRef] [PubMed]
17. 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, 3393–3402 (2005). [CrossRef] [PubMed]
18. M. Wojtkowski, V. Srinivasan, J. G. Fujimoto, T. H. Ko, J. S. Schuman, A. Kowalczyk, and J. S. Duker, “Three-dimensional retinal imaging with high-speed ultrahigh-resolution optical coherence tomography,” Ophthalmology 112, 1734–1746 (2005). [CrossRef] [PubMed]
19. V. J. Srinivasan, M. Wojtkowski, A. J. Witkin, J. S. Duker, T. H. Ko, M. Carvalho, J. S. Schuman, A. Kowalczyk, and J. G. Fujimoto, “High-definition and 3-dimensional imaging of macular pathologies with high-speed ultrahigh-resolution optical coherence tomography,” Ophthalmology113, 2054–2065.e3 (2006). [CrossRef] [PubMed]
20. T.H. Ko, D.C. Adler, JG. Fujimoto, D. Mamedov, V. Prokhorov, V. Shidlovski, and S. Yakubovich, “Ultrahigh resolution optical coherence tomography imaging with a broadband superluminescent diode light source,” Optics Express 12, 2112–2119 (2004). [CrossRef] [PubMed]
21. T.C. Chen, B. Cense, M.C. Pierce, N. Nassif, B.H. Park, S.H. Yun, B.R. White, B.E. Bouma, G.J. Tearney, and J.F. de Boer, “Spectral domain optical coherence tomography: ultra-high speed, ultra-high resolution ophthalmic imaging,” Arch. Ophthalmol. 123, 1715–1720 (2005). [CrossRef] [PubMed]
22. B. Sander, M. Larsen, Thrane L L., J. L. Hougaard, and T. M. Jorgensen, “Enhanced optical coherence tomography imaging by multiple scan averaging,” Br. J. Ophthalmol. 89, 207–212 (2005). [CrossRef] [PubMed]
23. A. Sakamoto, M. Hangai, and N. Yoshimura, “Spectral-domain optical coherence tomography with multiple B-scan averaging for enhanced imaging of retinal diseases,” Ophthalmology 115, 1071–1078.e7 (2008). [CrossRef]
24. V. J. Srinivasan, B. K. Monson, M. Wojtkowski, R. A. Bilonick, I. Gorczynska, R. Chen, J. S. Duker, J. S. Schuman, and J. G. Fujimoto, “Characterization of outer retinal morphology with high-speed, ultrahighresolution optical coherence tomography,” Invest. Ophthalmol. Vis. Sci. 49, 1571–1579 (2008). [CrossRef] [PubMed]
25. Y. Ojima, M. Hangai, A. Sakamoto, A. Tsujikawa, A. Otani, H. Tamura, and N. Yoshimura, “Improved visualization of polypoidal choroidal vasculopathy lesions using spectral-domain optical coherence tomography,” Retina 29, 52–59 (2009). [CrossRef]