Spectral-domain Optical Coherence Tomography: A Real-world Comparison
IRENE A. BARBAZETTO, MD · SANDRINE A. ZWEIFEL, MD · MICHAEL ENGELBERT, MD, PhD · K. BAILEY FREUND, MD · JASON S. SLAKTER, MD
Over the last decade, optical coherence tomography has become an important imaging technology used in diagnosing and following macular pathologies. It has complemented, and in some cases replaced, fluorescein angiography in many instances, especially in the diagnosis and management of various retinal disorders, including macular edema and AMD.
ENTER FOURIER-DOMAIN OCT
The introduction of spectral-domain or Fourier-domain OCT (vs time-domain OCT, which characterized all previously commercially available machines) has taken image resolution, acquisition time, registration, and display options to new levels. As clinicians have a variety of new devices to choose from, it is worthwhile to take a closer look at the different features, including imaging capabilities, before making the decision of which system to purchase.
All devices currently available offer a high theoretical axial image resolution between 3 to 7 μm, with the 3D SD-OCT (Bioptigen) and Copernicus (Optopol Technology SA) marking the higher end of the spectrum and Spectralis (Heidelberg Engineering) representing the lower end, with 7-μm "optical" resolution (but 3.5-μm "digital" resolution). The transverse image resolution ranges between 10 and 25 μm. Greater variability can be found looking at the scanning speed with the Copernicus HR (Optopol Technology SA) acquiring 55 000 A-scans per second, the Spectralis acquiring 40 000, the Cirrus (Carl Zeiss Meditec, Inc.) acquiring 27 000 A-scans per second, and the 3D OCT-1000 (Topcon Medical Systems, Inc.) acquiring 18 000 A-scans per second.
Despite these improved acquisition times, motion error and poor image contrast remain problems. Thus, theoretical axial and transverse resolution specified by the manufacturer do not necessarily translate into crisper, more detailed images. Most devices improve contrast and reduce motion errors through frame averaging. Also the Spectralis allows for real-time eye tracking to eliminate motion artifacts.
Another variable is the devices' capability of transverse image registration (see, for example, the Topcon 3D OCT-1000, and the Heidelberg Spectralis), which allows for a direct, side-by-side comparison of different visits and easy monitoring of localized anatomic changes.
|Irene A. Barbazetto, MD, is a retina fellow at the Harkness Eye Institute of Columbia University. Sandrine A. Zweifel, MD, is a medical retina fellow at the Vitreous-Retina-Macula Associates of New York (VRMNY). Michael Engelbert, MD, PhD, is a clinical fellow at VRMNY. K. Bailey Freund, MD, and Jason S. Slakter, MD, are partners in Vitreous-Retina-Macula Consultants of New York. The authors report no financial interest in any products mentioned here. Dr. Slakter can be reached via e-mail at firstname.lastname@example.org.|
When comparing actual images taken with different devices, it is not surprising that the resolution and the contrast seem to be superior for machines with axial image resolution of 3 to 4 μm, higher scanning speed, and real-time eye tracking. However, these differences may not play as important a role for standard applications, since many of the images appear rather comparable when it comes to standard diagnostic applications (Figures 2, 3, and 4). Sayanagi et al. compared Stratus OCT images of AMD patients to studies obtained with spectral-domain devices for the detection of typical neovascular patterns and concluded that all 4 spectral-domain devices were superior in delineating sub-RPE fluid.1 However, they did find differences when comparing detection of subretinal fluid and intraretinal changes, including cysts, with the Spectralis and Cirrus performing best in their study.
Figure 1. Comparison of OCT images of a patients with neovascular age-related macular degeneration obtained with the RTVue OCT (OptoVue) and the Cirrus OCT (Zeiss Meditec, Inc.).
Figure 2. Comparison of images obtained with 3 different spectral-domain OCT devices (Topcon 3D OCT-1000, Zeiss Cirrus, Heidelberg Spectralis) of both eyes of the same patient with early AMD changes taken just minutes apart.
Figure 3. Comparison of images obtained with 3 different spectral-domain OCTs (Heidelberg Spectralis, Optovue RTVue, Topcon 3D OCT-1000) and with 1 time-domain OCT (Zeiss Stratus) of both eyes of the same patient with a history of central serous chorioretinopathy in both eyes.
Figure 4. The same set of images as shown in Figure 3 in pseudo color.
In general, software has become an equally, if not more important part of the package. These applications include, among others, retinal and macular thickness maps and 3D reconstruction of the vitreoretinal interface, a feature that can be helpful in surgical planning (Figure 5). Also, C-scan or "en face" representation of the retinal layers and choroid adds a new dimension to visualization. Curved C-scans may be set to conform to the contours of the internal limiting membrane, the retinal pigment epithelium, or a spherical approximation to the RPE surface called the RPE fit, depending on whether the structures of interest are located in the inner or outer retina (Figure 6). En face OCT images will likely be helpful in assessing progression of drusen or geographic atrophy in AMD or photoreceptor loss in retinal dystrophies or AZOOR.
Figure 5. Comparison of horizontal B-scan images and 3D images of a patient with neovascular age-related macular degeneration obtained with Heidelberg Spectralis, Zeiss Cirrus, Topcon 3D OCT-1000.
Figure 6. En face OCT (C-scan) through the choroid (top left) and corresponding B-scan (top right) of the same patient as shown in Figure 5. The double blue lines represent the curved optical slab, which is then overlaid onto the scanning laser ophthalmoscopy image. En face OCT (C-scan) through the outer nuclear layer of a patient with long-standing neovascular age-related macular degeneration (bottom left) with an interconnected network of tubules (outer retinal tubulation). The corresponding B-scan (bottom right) shows the location of these tubular structures within the outer retina.
One has to keep in mind that the measurements with different machines are not always interchangeable from device to device and that previous time-domain devices use different reference points for analysis with respect to the retina or pigment epithelium. A recent study by Wolf-Schnurrbusch et al. showed that, of six tested spectral-domain and time-domain OCT devices, the Cirrus and Spectralis showed similar values for serial measurements of central retinal thickness in "normal" patients, but those values were higher compared to other devices, which did not include the RPE layer into the segmentation algorithm when calculating retinal thickness.2 It may be of interest that, of all the devices tested in this study, the Spectralis had the lowest coefficients of variation for repeated measurements and allowed for the most accurate repeat measurements, possibly due to the combination of image registration and eye tracking.
The fact that time- and spectral-domain machines will differ both qualitatively and quantitatively has been corroborated by several studies.3-5 In "real life," this will mainly affect users who are in the process of upgrading their current equipment, as not only retinal thickness but also nerve-fiber layer analyses will no longer be comparable.
These issues are also relevant in clinical trials with OCT measurement endpoints and for practices using different imaging devices in different offices. However, if thickness determination is based on identical anatomical boundaries, thickness measurements may be comparable.6 This is of import for retrospective studies that follow patients over a long enough time span that parallels the evolution of OCT.7
Advanced 3D viewing for surgical planning and patient education may be as important for some users, particularly when this feature can be performed on a networked computer in an examination or consultation room. Most devices provide desktop review software to allow for varying degrees of 3D viewing, visit-to-visit comparisons, and advanced image analysis on networked computers within the office. If, instead, the practice chooses to upload representative B-scans, thickness maps, and change analyses to a pre-existing image database, they will lose many of these advanced features.
With regard to patient tolerance and office workflow, ease of use and acquisition times may vary substantially from device to device. Certain systems require considerably more training than others in order for the user to obtain high-quality images. Also, while useful for obtaining higher quality images, real-time eye tracking lengthens the acquisition time and is very difficult to perform in patients with poor fixation. The size or "footprint" of the systems vary considerably, making some units, most notably the Cirrus, more suited for small examination rooms.
Several companies offer combined application devices, most notably Heidelberg Engineering, which offers the Spectralis as part of their HRA package. It includes fluorescein and indocyanine green angiography, as well as fundus autofluorescence imaging — a space-saving approach, but of concern in case of any need for repair, as in all multifunction machines. The Topcon 3D OCT-1000 takes a nonmydriatic color fundus image of the posterior pole at the end of the scan; however, the limited resolution may not necessarily replace the need for a standard digital fundus camera in many instances. The Spectral OCT SLO (OPKO/OTI) offers additional microperimetry — a useful, but also time-consuming tool in assessing retinal function.
Other additional features make some of the devices more suitable for research, like the 3D SD-OCT (Bioptigen, Inc.), which has specific applications for scanning small animal eyes, anterior-segment examinations, and tissue cultures. The machine has two engines, one at 840 nm for imaging the retina in small animals (rodents, rabbits), larger animals (such as dogs and pigs), and humans, and one at 1310 nm, which is suited for tissue imaging, anterior segment imaging, small animal, external, and ex vivo imaging.
Several devices allow for retinal nerve fiber layer analysis and can be used to monitor glaucoma patients and their progression. Not all devices, however, offer a normative database. Others combine not only retina and glaucoma applications but also allow for anterior-segment visualization and corneal topography like the RTVue OCT (Optovue) and the 3D SD-OCT. Some of the most recent devices broadened the applications even further and measure blood flow using a Doppler OCT.
Most recently approved among OCT devices is the OPKO/OTI Spectral OCT SLO Combination Imaging System, which received FDA 510(k) clearance in January 2009 and is manufactured in Canada. Similar to the Spectralis, this machine is a combination of OCT and scanning laser ophthalmoscope (SLO). Its features include real-time 3D imaging and an optional anterior-segment OCT module. In August 2008, Optopol Technology received market approval in Japan for its SPOCT Copernicus HR spectral computed tomography system, which includes retinal- and RNFL-thickness maps, ocular motion tracking, and optional Doppler imaging.
In summary, spectral-domain OCTs allow not only for greater image resolution at much higher speeds than the previous generation time-domain devices. They also have the potential to improve patient care and surgical planning, as well as to broaden the understanding of pathologies and their mechanisms. Although differences in image quality can be found between devices of different manufactures, it is safe to say that all images obtained with SD-OCT devices exceed the previous standard by far. Therefore, one should carefully review the individual technical specifications and analyze how the devices will be most efficiently used in the individual office or research lab with its specific needs before making the investment, which might be quite substantial in some cases. Software updates, technical support, image storage, and networking capabilities, as well as ease of use and available ancillary imaging capacities, may prove to be as or even more important than small differences in picture quality. RP
- Sayanagi K, Sharma S, Yamamoto T, Kaiser PK. Comparison of spectral-domain versus time-domain optical coherence tomography in management of age-related macular degeneration with ranibizumab. Ophthalmology. 2009;116:947-955.
- Wolf-Schnurrbusch UE, Ceklic L, Brinkmann CK, et al. Macular thickness measurements in healthy eyes using six different optical coherence tomography instruments. Invest Ophthalmol Vis Sci. 2009 Feb 21. [Epub ahead of print]
- Leung CK, Cheung CY, Weinreb RN, Lee G, Lin D, Pang CP, Lam DS. Comparison of macular thickness measurements between time domain and spectral domain optical coherence tomography. Invest Ophthalmol Vis Sci. 2008;49:4893-4897.
- Forooghian F, Cukras C, Meyerle CB, Chew EY, Wong WT. Evaluation of time domain and spectral domain optical coherence tomography in the measurement of diabetic macular edema. Invest Ophthalmol Vis Sci. 2008;49:4290-4296.
- Forte R, Cennamo GL, Finelli ML, de Crecchio G. Comparison of time domain Stratus OCT and spectral domain SLO/OCT for assessment of macular thickness and volume. Eye. 2008 Dec 12. [Epub ahead of print]
- Engelbert M, Zweifel S, Imamura Y, Fisher YL. Spectral and time domain OCT measure identical retinal thickness if identical boundaries are selected for analysis. Eye. 2009 Apr 17. [Epub ahead of print]
- Engelbert M, Zweifel S, Freund KB. "Treat and extend" dosing of intravitreal anti-vascular anti-endothelial growth factor agents for retinal angiomatous proliferation (RAP/type 3 neovascularization. Retina, accepted for publication.