1. Introduction

Imaging of macular disease using optical coherence tomography (OCT) has changed the understanding of many retinal diseases. Since its introduction in 1991, OCT has increasingly been used mainly to identify discrete morphologic features in vitreoretinal disease [1-3].

Before that time, it was impossible to visualize discrete alterations like retinal edema in a cross section [4]. Studies evaluating the presence of retinal edema in patients with diabetes mellitus using stereo fundus photography (SFA) and angiography show a significant lack of reproducibility in these examinations [5, 6]. Identification of distinct morphologic features like edema however is often essential for correct treatment indications. Failures in false positive detection of edema may result in invasive and often irreversible interventions such as argon laser photocoagulation.

Implementing advances of evolving technology, OCT model 2000 (Carl Zeiss Meditec) was the first OCT used for daily patient care in many clinics [7, 8]. This device was able to measure retinal thickness in an automated process. To acquire this information, the device interpolated retinal thickness data of six scans, which were performed in a radial pattern. Results were visualized in a modified grid according to the experience of the early treatment of diabetic retinopathy study (ETDRS).

By using this approach OCT 2000 and the later Stratus OCT, which allowed for superior scanning speed and usability [9], OCT maps could be compared to angiography or fundus photography. Moreover, therapeutic effects could now be quantified and helped to evaluate treatment effects, mainly of laser and the upcoming intravitreal therapies like triamcinolone, pegaptanib followed by bevacizumab and ranibzumab. Especially the availability of the second-generation antiangiogenic drugs targeting all isoforms of vascular endothelial growth factor (VEGF) changed the prognosis of many retinal diseases [10], since these drugs proved excellent clinical efficacy and safety, promoting the diagnostic value of OCT technology even further [11-14]. Offering these treatment options, OCT is recommended for follow-up examinations, treatment control and particularly re-treatment indications.

Meanwhile, clinical studies show the benefit of OCT-guided treatment in neovascular age related macular degeneration (AMD), which is the most common cause of irreversible vision loss in the developed world [15].

Slow scanning speed and limited resolution were the main drawbacks in conventional OCT technology. Detailed studies have shown that errors in retinal thickness analysis strongly compromise the data of time domain OCT [16-19]. Increasing evidence can be found in the literature, showing that interpolating retinal thickness data out of only six scans may be an inappropriate methodology, prone for segmentation artifacts.

Newer OCT generations utilizing spectral domain OCT (SD-OCT) technology may overcome these limitations of the time domain era, since SD-OCT allows for much higher scanning speed, better resolution and, presumably in the near future, deeper penetration and an improved delineation of the RPE [20-24]. Retinal thickness, photoreceptor- and nerve fiber layer thickness can now be evaluated in 3D datasets allowing for a realistic localization without interpolation [25-29]. Novel parameters like the deviation of the pigment epithelium layer, which is a critical structure for many diseases can be evaluated and quantified [30]. Studies are underway investigating the clinical benefit and implications of these parameters.

However, technological parameters do not necessarily lead to advances in clinical management of patients. Therefore, there is a need to identify clinically relevant prognostic factors and critical changes in retinal microstructure able to predict a patient’s individual course under a specific, often expensive, therapy or to actually identify the anatomical reason for visual deficits in patients with macular disease, implicating that current three-dimensional (3D) OCT technology is clearly an enriching diagnostic procedure.

It is the aim of this article to demonstrate the amenities of 3D macular imaging using highresolution spectral domain OCT technology in retinal disease.

2. Methods

2.1 Patients’ inclusion

2714 patients with various macular diseases representing the spectrum from age-related macular degeneration, retinal vascular disease, macular dystrophies and vitreoretinal disease were imaged using a Cirrus SD-OCT prototype between 2005 and 2008.

Before inclusion into this study, a written informed consent was obtained from all patients following an extensive discussion on scientific purposes, procedures and potential risks. The local ethics committee approved the procedures of the study. The protocol complied with the declaration of Helsinki. Examinations were performed at the Department of Ophthalmology, Medical University, Vienna, Austria. Typical cases were selected to demonstrate the clinical potential of current SD-OCT technology.

2.2 Optical coherence tomography

Three-dimensional imaging was performed using a prototype of the Cirrus high resolution OCT (Carl Zeiss Meditec). This device is a spectral domain, high-resolution optical coherence tomography system with an axial resolution of 6 µm and a scanning speed of 25000 AScans/ second. The scanning area covered 6*6 mm within the central retina. Scan depth was 2 mm. Pixel density was 200*200*1024 (number of A-scans/number of B-scans/ pixel number for 2mm of axial depth), 512*128*1024 and 4096*5*1024. All scans were selected for comparison of different scanning patterns.

Patients were scanned with the macular raster scan, before a high density scan comprising 5 scans (4096*5*1024) of the central macula was performed. The 5 Line Raster protocol includes a horizontal scan through the foveal center, with two further horizontal scans on each side. By default, each line is 6 mm in length and separated by 250 µm, so that together the five lines cover 1 mm in height. Patients were encouraged to fixate a target. Scans were repeated when imaging quality was reduced due to significant eye movement or other reasons of compromised image quality. All B-scans were checked for quality using a video function once the scanning process was completed. Patients were scanned at baseline and follow up visits if applicable.

For daily analysis, automatic segmentation and volumetric analysis were performed using previously described [30], proprietary software, designed to automatically delineate retinaland RPE-layers. Carl Zeiss Meditec specially developed this software for the Cirrus SD-OCT.

2.3 Data export and visualization

Datasets comprising all B-Scans of a macular raster scan were exported from the system using bitmap file format after the datasets were corrected for motion artefacts using public domain Image J software (Version1.4d, http://rsb.info.nih.gov/ij). Only data for Interactive Science Publication (ISP) presentation was corrected for motion artefacts. 3D-models were created using current 3D-rendering software (OSA ISP version 2.0, Kitware Inc., kitware@kitware.com). This software allows for free manipulation and visualization of 3Ddatasets by the reader and is a novel approach to publish original data. Please note that arrows, belonging to a single B-scan image only, frequently mark two-dimensional data. Please scroll the dataset to find the image, matching to the landmarks (the correct image numbers are given in the comments section for each landmark).

Raw data of the acquired scans were saved in a database including patients’ personal data, diagnosis, and -if appropriate- considered treatment. For evaluation and documentation of patient data, Excel databases (Microsoft) were used.

3. Results

Case 1: Macular Pucker

This case of a 58-year-old woman’s right eye demonstrates a macular pucker with central macular edema and folds due to retinal traction. Epiretinal membranes are a relatively common disorder with a prevalence of approximately 2% in people aged 50 years and 20% in people older than 75 years of age. Slicing the dataset in the z-axis precisely shows the traction lines, which can be followed from the center to the arcades peripherally. Cysts become visible within the central edematous zone when subsequent B-scans are evaluated indicating that traction forces have already lead to significant intraretinal edema. A broader adhesion can be observed in the temporal lower part.

 

Fig. 1. Macular Pucker. The configuration of retinal traction and folds can be visualized in 2d and 3d images. The retinal pigment epithelium (RPE) and the external limiting membrane (ELM) are of regular shape. 3D data is available in View 1.

Download Full Size | PDF

Case 2: Macular hole

This SD-OCT examination of a 61 years old male patient presenting with sudden onset central scotoma revealed a full-thickness macular hole. Macular holes were lately shown to have a prevalence of 0,17% in a population in south India [31]. A symmetric pattern of cystic spaces can be visualized surrounding the central gap. The external limiting membrane is not affected in the periphery of the lesion and can be followed until close to the center, where the macular hole results in a circular defect in the ELM. Even though the defect in the ELM is clearly circumscribed it is significantly larger than the defect in the tissue itself. 3D datasets allow measuring and visualization of this defect with previously unknown accuracy. Most interestingly, cysts can only be found below the level of the inner nuclear layer in this patient. An operculum can be visualized above the inner surface of the macular hole. The precise morphologic configuration of the hole can be displayed three-dimensionally and shows a biconcave structure with its thinnest diameter at the anatomical level of the outer plexiform layer.

 

Fig. 2. Macular hole. Details are visualized in View 2. Figure 2a shows the 3D-image, whereas 2b shows the symmetric arrangement of cysts around the macular hole. 3c nicely shows the operculum attached to the vitreus and the central gap in the intraretinal structure. High resolution of SD-OCT imaging allows visualizing discrete features like the external limiting membrane (ELM).

Download Full Size | PDF

Case 3: Age-related macular degeneration (AMDI)

Neovascular age-related macular degeneration is the leading cause of severe vision loss in the US with a prevalence of ~6,4% in individuals aged 65-74 in the Framingham eye study. Neovascular AMD is one form of AMD typically associated with rapid progression of vision loss and may present with pigment epithelial detachment (PED), serous retinal detachment due to subretinal fluid (SRF) and retinal thickening with or without presence of intraretinal cysts in OCT. Eye movement corrected SD-OCT facilitates insights into the underlying threedimensional structure and topographic correlations between these findings. The irregular basis of the PED can be visualized in this datasets helping to define the zone of subpigment epithelial fluid extravasations due to the underlying neovascular complex. Discrete retinal structures like the outer limiting membrane or the photoreceptor layer can be visualized and analyzed for focal defects in tissue integrity. Not the zones of focal RPE atrophy which can be visualized in three-dimensions as well. These lesions are functionally important since they show the coincidence of neovascular and atrophic components in neovascular AMD.

 

Fig. 3. This figure shows a large detachment of the retinal pigment epithelium (PED). The basis of this PED is of irregular shape (View 3). Zones of significant pigment epithelial atrophy can be observed in View 4.

Download Full Size | PDF

Case 4: Age-related macular degeneration (AMDII)

This case of neovascular AMD shows less detachment to the pigment epithelium and less atrophic zones indicating less severe damage to the retina. However, typical alterations can be found in this example, too. A broad hyperreflective signal at the level of the outer plexiform/outer nuclear layer is clearly present and typical for close presence of subretinal fluid. Diffuse hyperreflectivity of the outer photoreceptor elements can also be observed near subretinal fluid and indicate stress on one of the critical structures of the photoreceptor.

 

Fig. 4. This figure demonstrates the typical intraretinal changes in neovascular age-related macular degeneration (nAMD). Subretinal fluid (SRF) has been proven to be a relevant marker for prognosis of nAMD disease and can be observed in distinct locations. 3D-data is available in View 5.

Download Full Size | PDF

4. Discussion

This manuscript presents complete and full functioning 3D-SD-OCT datasets to demonstrate the potential of current SD-OCT technology to enhance our pathophysiologic understanding of macular disease. Features like subretinal fluid which are known to have prognostic impact on functional outcome can clearly be identified in the data [32].

Diseases of the vitreoretinal interface like macular pucker could be visualized with great morphologic precision. Retinal folds were identified and showed a distinct traction orientation, which might be important for the surgical approach and the peeling process. Moreover, photoreceptor damage can be evaluated in detail allowing the vitreoretinal surgeon to analyze the severity of the disease on an objective anatomical basis All location mapping of the retinal surface prevents to overlook small macular holes in pericentral localization, which could easily be missed using the former macular mapping mode comprising only six radial scans to represent the entire macular region [33].

As photoreceptor morphology and associated function plays an important role in retinal disease [34], information regarding the status of the retinal pigment epithelium (RPE) is necessary to understand macular disease like age-related macular degeneration since the RPE is crucial for photoreceptor metabolism and survival. Subtle changes within the photoreceptor and RPE layer can be visualized in simplified models by novel 3D segmentation algorithms. However, relevant information might be overlooked, as some of the relevant information is lost during the segmentation process. Clinical studies are underway to evaluate, which of these factors are really relevant for the functional prognosis and outcome of these patients. This is particularly interesting for early age-related disease since identifying the patients threatened by late complications like geographic atrophy or neovascular AMD would be of great clinical importance to optimize preventive treatment and to establish regular control visits, before irreverible vision loss has taken place.

In exudative central retinal disease like central serous chorioretinopathy, precise SD-OCT imaging helped to identify pathophysiologic findings like PED which appear to be much more frequent than suggested in the pre-OCT or even time-domain OCT era [22]. Such distinct morphologic features could be identified and helped to better classify and understand the underlying disease process.

Meanwhile, alternative imaging techniques combining advantages of a scanning laser ophthalmoscope with angiography, measurements of fundus autofluorescence and current SDOCT technology have also become available and have contributed to a better understanding of macular disease [35, 36].

5. Conclusions

In conclusion, precise 3D retinal imaging in modern SD-OCT technology was shown to optimize correct diagnosis recently. Further developments of this evolving technology will consequently lead to optimized therapeutic strategies such as minimally invasive, targetorientated and thereby cost-efficient treatment options for patients with macular pathology.