Purpose
To compare the 6-line radial vs the 25-line raster spectral-domain optical coherence tomography (SD OCT) acquisition patterns at detecting intraretinal fluid, subretinal fluid, vitreomacular traction, and full-thickness macular hole (MH).
Design
Retrospective cross-sectional analysis.
Methods
Series of 365 eyes with neovascular age-related macular degeneration (AMD), diabetic macular edema (DME), central and branch retinal vein occlusion (CRVO/BRVO), central serous chorioretinopathy, vitreomacular traction, and full-thickness MH. Sequential 6-line radial and 25-line raster scans were evaluated for intraretinal/subretinal fluid and, when applicable, vitreomacular traction and MH.
Results
For neovascular AMD (133 scans), 7 25-line raster scans confirmed subretinal/intraretinal fluid not identified by the 6-line radial ( P = .02). For DME (140 scans) and central serous chorioretinopathy (91 scans), 25-line raster confirmed fluid in 4 scans ( P = .13) and 1 scan ( P = .32), respectively, that was not observed with the 6-line radial. For CRVO (123 scans) and BRVO (126 scans), 25-line raster confirmed fluid on 2 ( P = .25) and 4 scans ( P = .13), respectively, that was not detected by the 6-line radial. Conversely, for focal vitreomacular traction (70 scans) and full-thickness MH (82 scans), 25-line raster missed focal traction (<1500 μm) and MH in 5 scans ( P = .07) and 7 scans ( P = .02), respectively, that were identified using the 6-line radial.
Conclusions
The 6-line radial scan is statistically comparable to the 25-line raster at detecting fluid in DME, BRVO/CRVO, and central serous chorioretinopathy, but not neovascular AMD. Furthermore, it is superior to the 25-line raster pattern at detecting early MH formation, while demonstrating a positive trend in identifying focal vitreomacular traction.
The advent of spectral-domain optical coherence tomography (SD OCT) has enhanced our ability to image the preretinal, intraretinal, and subretinal space in vivo at near-histologic resolution. As such, this technology has revolutionized ophthalmic clinical practice and, in particular, the diagnosis and management of patients with retinal disease.
By providing a cross-sectional depiction of macular pathology as well as enabling quantification of retinal morphology (ie, central macular thickness), SD OCT now plays an instrumental role in guiding the physician’s treatment decisions. Furthermore, OCT-derived parameters have quickly become incorporated into clinical trials for the inclusion of study subjects and monitoring their response to intervention, specifically those evaluating macular edema secondary to neovascular age-related macular degeneration (AMD), diabetes, and retinal vein occlusion. With the ever-increasing use of anti–vascular endothelial growth factor (VEGF) pharmacotherapy for exudative macular disease, the recent introduction of ocriplasmin (Jetrea; Thrombogenics, Inc, Iselin, New Jersey, USA) for vitreoretinal interface pathology, and additional novel intravitreal therapeutics looming on the horizon, the number of patients with treatable retinal disorders is rapidly on the rise. Consequently, optimizing the efficient use of office-based imaging platforms during the management of these patients is becoming paramount.
Currently, the 2 most commonly used SD OCT acquisition scan patterns are the radial and raster scans. The radial scan is a series of B-scans at regular angular intervals, whereas the raster (rectangular) line scan is generated from a series of parallel B-scans. The aim of this study was to determine whether there is a difference between the standard 6-line radial and 25-line raster scans at detecting macular pathology in the following diseases routinely encountered in a retina clinic: neovascular AMD, diabetic macular edema (DME), branch retinal vein occlusion (BRVO) with associated macular edema, central retinal vein occlusion (CRVO) with associated macular edema, central serous chorioretinopathy, vitreomacular traction, and full-thickness macular hole (MH).
Methods
Following approval from the Institutional Review Board at Wills Eye Hospital, a retrospective observational case series of patients having undergone simultaneous 6-line radial and 25-line raster SD OCT testing for macular diseases was performed. Electronic billing records of office visits between July 1, 2010 and September 30, 2013 at Wills Eye Hospital were reviewed to identify patients with the following diagnoses: neovascular AMD (International Classification of Diseases, 9th revision code: 362.52), DME (362.07), CRVO (362.35), BRVO (362.36), central serous chorioretinopathy (362.41), vitreomacular traction (379.27), and full-thickness MH (362.54). Patients with multiple diagnoses were excluded from the study.
Optical Coherence Tomography Acquisition/Scanning Protocol
All SD OCT acquisitions were performed by certified ophthalmic photographers using the commercially available Heidelberg Spectralis unit (Spectralis; Heidelberg Engineering, Heidelberg, Germany). The standard imaging protocol at Wills Eye Hospital consists of 2 scanning patterns centered at the fovea, the 6-line radial, and the 25-line raster scan. Both of these pattern types were performed regardless of the patient’s diagnosis. If eccentric fixation was detected during the SD OCT acquisition, the photographers were instructed to manually center both scan patterns on the fovea. Internal software used an averaging system to calculate the central macular thickness (CMT) as the distance between the retinal pigment epithelium and the internal limiting membrane by preset algorithms.
Optical Coherence Tomography Grading
SD OCT images were interpreted independently by 2 experienced reviewers (E.R., N.R.). Inter-observer differences were resolved by a third interpreter (J.H.). For each eye, the reviewers evaluated serial SD OCT scans, beginning with the 6-line radial, then followed by the corresponding 25-line raster from the same examination date. Scans were analyzed for the presence of intraretinal fluid, subretinal fluid, and, when applicable, vitreomacular traction and full-thickness MH.
If present, the latter 2 were graded by a classification scheme recently described by the International Vitreomacular Traction Study Group. Briefly, vitreomacular traction was characterized by detectable retinal anatomic changes noted on the SD OCT scan, with concurrent vitreous status denoting a perifoveolar posterior vitreous detachment. The measured amount of vitreoretinal adhesion determined whether the vitreomacular traction was classified as focal (<1500 μm) or broad (>1500 μm). Next, full-thickness MH was defined as an anatomic defect in the fovea featuring interruption of all neural retinal layers from the internal limiting membrane to the retinal pigment epithelium (RPE). OCT-based measurement of the minimum hole width, or aperture size, determined the hole size: small (<250 μm), medium (>250 μm and <400 μm), and large (>400 μm). Using the OCT software caliper function, the aperture size was measured at the narrowest hole width in the mid-retina, as a line drawn roughly parallel to the RPE.
Up to 5 examination dates were included where the aforementioned findings were detected by at least 1 of the 2 acquisition patterns. Scans that did not detect macular pathology by either pattern type were excluded. Additionally, scans with poor image resolution precluding adequate visualization of all retinal layers and the underlying RPE were excluded.
Statistical Analysis
Statistical analysis was carried out using GraphPad software (GraphPad, La Jolla, California, USA). Categorical variables were reported as proportions. Binomial confidence intervals (Clopper-Pearson method) were calculated for missed proportions of intraretinal/subretinal fluid, vitreomacular traction, and full-thickness MH. The McNemar test was used to determine if there was a statistically significant difference between 6-line radial and 25-line raster scan acquisition patterns in detecting the presence of intraretinal/subretinal fluid, vitreomacular traction, and full-thickness MH. McNemar testing was selected as it is most suitable for comparing nominal data that are also paired proportions.
Results
A total of 365 eyes that underwent Heidelberg SD OCT image acquisition using both 6-line radial and 25-line raster scan patterns were included. Fifty eyes were reviewed for each of the following disorders: neovascular AMD, DME, CRVO, BRVO, central serous chorioretinopathy, and vitreomacular traction; and 65 eyes were reviewed with a diagnosis of full-thickness MH.
Neovascular Age-Related Macular Degeneration
For eyes with neovascular AMD, 133 SD OCT scans were interpreted (mean: 2.66 scans per eye). The 25-line raster detected the presence of intraretinal/subretinal fluid in all 133 scans. The 6-line radial detected intraretinal/subretinal fluid in all but 7 scans from 5 patients (5.3%; 95% CI 2.1%–10.5%), resulting in a sensitivity of 94.7% ( Figure 1 ). This difference in fluid detection between the 2 acquisition patterns for neovascular AMD was found to be statistically significant ( P = .02). Owing to the observed significance, subgroup analysis was performed to determine whether CMT may influence the ability of the 6-line radial scan pattern to detect fluid in eyes with neovascular AMD. The average CMT for scans that successfully detected intraretinal/subretinal fluid using both acquisition patterns was 334 μm, compared to 296 μm in the 7 scans in which the 6-line radial failed to identify fluid. This difference in average CMT was found to be statistically significant ( P = .01).
Diabetic Macular Edema
A total of 140 scans in eyes with DME were evaluated (mean: 2.8 scans per eye), from which the 25-line raster confirmed the presence of intraretinal/subretinal fluid in all 140. The 6-line radial was able to identify intraretinal/subretinal fluid in all but 4 scans from 3 patients (2.9%; 95% CI 0.8%–7.2%), resulting in a sensitivity of 97.1% ( Figure 2 ). The observed difference in fluid detection between the 2 scan patterns for cases of DME was not found to be statistically significant ( P = .13).
Retinal Vein Occlusion
A total of 123 scans for CRVO and 126 scans for BRVO with concurrent macular edema were interpreted (mean: 2.46 and 2.52 scans per patient, respectively). The 25-line raster scan pattern detected the presence of intraretinal/subretinal fluid in all 123 CRVO scans and 126 BRVO scans. However, the 6-line radial did not detect fluid in 2 CRVO scans from 2 patients (1.6%; 95% CI 0.2%–5.8%) and 4 BRVO scans from 4 patients (3.2%; 95% CI 0.9%–7.9%) that had been confirmed on the 25-line raster, resulting in a sensitivity of 98.4% and 96.8%, respectively ( Figure 3 ). This difference in fluid detection between the 2 scanning patterns was not statistically significant for both CRVO ( P = .25) and BRVO ( P = .13).
Central Serous Chorioretinopathy
For eyes with central serous chorioretinopathy, 91 SD OCT scans were reviewed (mean: 1.82 scans per eye). While the 25-line raster scan pattern detected the presence of subretinal fluid in all scans, the 6-line radial scan pattern failed to detect subretinal fluid in only 1 scan (1.1%; 95% CI 0.03%–6.0%) that was detected by the 25-line raster scan pattern. The sensitivity of the 6-line radial in detecting subretinal fluid in central serous chorioretinopathy was 98.9%. The difference in subretinal fluid detection between the 2 scanning patterns was not found to be statistically significant ( P = .32).
Vitreomacular Traction and Macular Hole
In addition to evaluating for intraretinal/subretinal fluid, the 6-line radial and 25-line raster scan patterns were compared for their ability to register vitreomacular interface abnormalities. From a total of 80 SD OCT scans for eyes with vitreomacular traction (mean: 1.6 scans per eye), the 6-line radial scan pattern detected vitreomacular traction on all 80 SD OCT scans. Conversely, from a total of 70 scans with focal vitreomacular traction (<1500 μm), 5 25-line raster scans (7.1%; 95% CI 2.5%–17.0%) missed focal traction that was successfully identified using the 6-line radial pattern ( Figure 4 ). The average measured vitreoretinal adhesion in the 5 cases of missed focal vitreomacular traction was 453 μm. The sensitivity of the 6-line radial in detecting focal vitreomacular traction was 100%, as compared to 92.9% for the 25-line raster. The difference between the 2 groups did not achieve statistical significance ( P = .07).
