August 2015
Volume 56, Issue 9
Free
Retina  |   August 2015
Polarization-Sensitive Optical Coherence Tomography and Conventional Retinal Imaging Strategies in Assessing Foveal Integrity in Geographic Atrophy
Author Affiliations & Notes
  • Ramzi G. Sayegh
    Department of Ophthalmology Medical University of Vienna, Vienna, Austria
  • Stefan Zotter
    Center for Medical Physics and Biomedical Engineering, Medical University of Vienna, Vienna, Austria
  • Philip K. Roberts
    Department of Ophthalmology Medical University of Vienna, Vienna, Austria
  • Maciej M. Kandula
    Chair of Bioinformatics Research Group, Department of Biotechnology, BOKU University Vienna, Austria
  • Stefan Sacu
    Department of Ophthalmology Medical University of Vienna, Vienna, Austria
  • David P. Kreil
    Chair of Bioinformatics Research Group, Department of Biotechnology, BOKU University Vienna, Austria
  • Bernhard Baumann
    Center for Medical Physics and Biomedical Engineering, Medical University of Vienna, Vienna, Austria
  • Michael Pircher
    Center for Medical Physics and Biomedical Engineering, Medical University of Vienna, Vienna, Austria
  • Christoph K. Hitzenberger
    Center for Medical Physics and Biomedical Engineering, Medical University of Vienna, Vienna, Austria
  • Ursula Schmidt-Erfurth
    Department of Ophthalmology Medical University of Vienna, Vienna, Austria
Investigative Ophthalmology & Visual Science August 2015, Vol.56, 5246-5255. doi:10.1167/iovs.14-15114
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Ramzi G. Sayegh, Stefan Zotter, Philip K. Roberts, Maciej M. Kandula, Stefan Sacu, David P. Kreil, Bernhard Baumann, Michael Pircher, Christoph K. Hitzenberger, Ursula Schmidt-Erfurth; Polarization-Sensitive Optical Coherence Tomography and Conventional Retinal Imaging Strategies in Assessing Foveal Integrity in Geographic Atrophy. Invest. Ophthalmol. Vis. Sci. 2015;56(9):5246-5255. doi: 10.1167/iovs.14-15114.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: To compare current imaging methods with respect to their ability to detect the condition of the fovea in patients with geographic atrophy (GA).

Methods: The retinas of 176 eyes with GA were imaged using two spectral-domain optical coherence tomography (SD-OCT) systems, Cirrus HD-OCT and Spectralis HRA+OCT, and fundus autofluorescence (FAF) and infrared imaging (IR) was used in the scanning laser ophthalmoscope (SLO) mode. Polarization-sensitive OCT (PS-OCT), which selectively visualizes the RPE in addition to SD-OCT features, was used to image 95 eyes. Geographic atrophy lesions were categorized as fovea spared, involved, or not quantifiable (grades 0, 1, and 2). Morphologic gradings were subsequently correlated with best-corrected visual acuity (BCVA) measurements to independently identify the corresponding functional condition of the fovea. Cohen's κ statistics with a bootstrap method was applied to compare retinal imaging methods.

Results: In PS-OCT, 84% of eyes with BCVA greater than or equal to 20/40 were detected, whereas in conventional retinal imaging the rate ranged from 27% in FAF to 45% in the SD-OCT segment. Cohen's κ statistics revealed significant differences between the gradings of PS-OCT and conventional imaging with κ = 0.488 and a global Hotelling's T2 statistic of 17.9 with a P value of P = 0.003. Statistical tests revealed no statistically significant differences between the conventional retinal imaging modalities.

Conclusions: Polarization-sensitive OCT can better allow correct grading of the fovea in relation to BCVA and identify foveal sparing than other imaging modalities. The differences in imaging precision should be considered in diagnostic and therapeutic evaluations.

Age-related macular degeneration is a chronic progressive disease with strong phenotype variability.1 Focal hyperpigmentation and drusen are among the earliest signs of AMD, eventually progressing to either choroidal neovascularization or geographic atrophy (GA).2 Geographic atrophy is a distinct entity characterized by slow progression of neurosensory layer atrophy with loss of photoreceptors (PR) and RPE.35 Areas with PR and RPE loss show functional deficiency associated with relative or absolute scotomas, which affect the fovea in late-stage disease and lead to severe and irreversible loss of central visual acuity.6 Although advanced GA disease affects one third of patients with late AMD,7,8 it may not be detected clinically until late in the course of the disease as GA typically starts extra-fovealy forming an atrophic island and/or an atrophic ring surrounding the fovea, which preserves central visual acuity.6,9 This process, termed foveal sparing, is seen by fundus photography and can be delineated using blue light-enhanced fundus autofluorescence (FAF) and recently also by optical coherence tomography (OCT).6,913 The pathophysiological background of this phenotypic foveal sparing is not understood but is speculated to relate to the high density of cones in the central retina, a protective effect of macular pigment and/or a unique choroidal blood supply at the fovea.1418 Many strategies are used for imaging GA, each visualizing different pathognomonic features. In fundus photography, enhanced visibility of choroidal vessels is the most obvious sign of a severely atrophic RPE and PR layer.19 Fundus autofluorescence imaging is based on the autofluorescent features of lipofuscin, which is composed of a mixture of metabolites originating from the visual cycle such as N-retinylidene-N-retinylethanolamine (A2E). Autofluorescence from lipofuscin is a diagnostic marker for retinal aging and an impaired lipid degradation process resulting from incomplete phagocytosis of PR outer segments by the RPE.20,21 Fundus autofluorescence is used to measure the extent of atrophy, which it identifies as a dark area of hypofluorescence, reflecting a complete absence of fluorophores. Detection of this dark area has been accepted as an endpoint in recent GA progression studies.18,22,23 By contrast, spectral-domain OCT (SD-OCT), which is predominantly used in therapeutic monitoring of patients with neovascular AMD, serves to determine morphologic features of the neurosensory retina and RPE including subretinal fluid, intraretinal cysts, RPE detachment, and retinal layer integrity based on tissue-specific reflectivity.2426 The value of OCT in dry AMD imaging is still being investigated.2729 Polarization-sensitive SD-OCT (PS-OCT) is a technological extension of conventional intensity-based OCT and combines the advantages of high-resolution structural imaging and selective identification of the RPE layer.30 The sensitivity of PS-OCT for RPE is related to the characteristic intracellular pattern of melanosomes in physiological RPE cells modifying the polarization state of backscattered light in a characteristic fashion.31 The RPE depolarizes the backscattered light (i.e., it scrambles the polarization state of the light) an effect that is measured and quantified by the PS-OCT device in a depth-resolved manner. A reduction of this polarization scrambling effect indicates a loss of RPE cells and/or an alteration related to the progression from early to advanced AMD.3235 To date, no prospective clinical study has compared all the systems available for imaging the fovea in GA in a comprehensive way using conventional technologies (infrared imaging [IR], FAF, SD-OCT) and new technologies (PS-OCT). Therefore, the purpose of this study was to determine which imaging modality is most reliable for evaluating foveal integrity or alteration correlated with best-corrected visual acuity (BCVA) measurements. A precise assessment of the fovea, a tiny area of only 1.5 mm,2,36 which solely determines visual outcome in AMD6 is of utmost value for monitoring patients in therapeutic studies and future clinical practice, as patients with foveal sparing have the highest risk of vision loss and will benefit most from upcoming treatment options. For this reason, introduction of a retinal imaging strategy that reliably identifies foveal sparing is of great importance. 
Methods
Inclusion and Exclusion Criteria
This prospective noninterventional study was conducted at the Department of Ophthalmology of the Medical University of Vienna (Vienna, Austria). Enrolled in the study were 176 eyes from 96 consecutive patients with GA secondary to AMD and each patient had given informed consent. Patients were followed according to a standardized protocol that adhered to the tenets of the Declaration of Helsinki and was approved by the local ethics committee (EK569/2011). Patients were excluded if they had clinical signs or a history of choroidal neovascularization (CNV) in biomicroscopy or OCT in either eye. Clear ocular media were required to provide good imaging quality. At each study visit, patients underwent standardized BCVA testing with Early Treatment Diabetic Retinopathy Study charts (ETDRS), slit-lamp examination, biomicroscopy, SD-OCT imaging, FAF, and IR imaging. When a new PS-OCT system became available a subgroup of 95 eyes from 54 consecutive patients were additionally imaged with this novel system. 
Conventional Retinal Imaging
The imaging modalities analyzed in this study included two SD-OCT systems, the Cirrus HD-OCT (Carl Zeiss Meditec, Dublin, CA, USA) and the Spectralis HRA+OCT (Heidelberg Engineering, Heidelberg, Germany). Spectralis OCT combines an SD-OCT and a confocal scanning laser ophthalmolscope (cSLO), visualizing FAF, and IR images. 
The Cirrus HD-OCT Macular Cube 512 × 128 scans analyzed covered a 6 × 6 × 2 mm area of the macula. Scanning patterns with a density of 512 × 128 × 1024 voxels were acquired for each patient. 
The Spectralis HRA+OCT volume scans were acquired using the high speed mode (scan width: 768 A scans) with a standard of 49 B scans per volume set with a width and height of 20° × 20°. A scans covered a depth of 1.9 mm and consisted of 496 voxels. Each B-scan was averaged using the automatic real-time mode of the Spectralis device with 30 frames per B scan. The resulting density was 768 × 49 × 496 voxels. 
For this specific study, nonnormalized FAF data acquisition was used in a resolution of 768 × 768 pixels in the high speed mode and brightness and contrast were individually adjusted during data acquisition for optimal visualization of the FAF intensity distribution at the posterior pole and the foveal location. Fundus autofluorescence images were obtained by recording a 7-second video using a 30° × 30° frame size in the FAF mode of the Spectralis HRA+OCT. One image comprising the mean FAF intensity was calculated out of 15 frames. 
The Spectralis HRA+OCT cSLO device for IR imaging uses a wavelength of 820 nm. Infrared images were acquired with a frame size of 30° × 30° and a resolution of 768 × 768 pixels. The technical properties of FAF, IR, and SD-OCT and procedures are described in detail elsewhere.27,37 
No general comparison of the absolute dimensions of the scanned areas in the different OCT models was possible because the Cirrus HD-OCT device does not provide angular field-of-view information about the scan patterns applied and the Spectralis HRA+OCT device calculates the individual scaling based on the focus setting. Furthermore, the physical dimension of the area scanned by the Spectralis HRA+OCT device varies with the setting of the scanner focus. 
Polarization-Sensitive (PS)-OCT Imaging
A new wide-field, high-speed PS-OCT system developed by the Center for Medical Physics and Biomedical Engineering, Medical University of Vienna was used in this study.38 Compared with earlier PS-OCT systems,32,33 it acquires a denser volume scan (up to 1024 × 250 A scans) over a larger scan field (up to 40° × 40°). We used the highest density pattern with a scan field of 30° × 30° for this study. The technical principles of this device have been described previously.32,33,39,40 The system enables measurement of four variables simultaneously: the intensity of the backscattered light (as in conventional SD-OCT imaging), the birefringence retardation, optic axis orientation, and the degree of polarization uniformity (DOPU). Retinal layers can then be classified in polarization preserving (e.g., photoreceptor layers), depolarizing (e.g., RPE cells) and birefringent layers (e.g., retinal nerve fiber layer). As only melanin-containing structures in the RPE distinctly depolarize backscattered light,31 PS-OCT is able to specifically identify the RPE layer under physiological and pathological conditions. Degree of polarization uniformity values will be close to 1 in polarization-preserving layers but lower in depolarizing (RPE) structures. Retinal structures are consistent with polarization-preserving and birefringent layers, while the RPE is displayed as a depolarizing layer. Areas of DOPU less than 0.75 were delineated to generate a precise RPE segmentation and displayed in red on top of intensity B scans, generating a point-to-point overlay of depolarizing tissue within the regular intensity-based SD-OCT image (Figs. 115523). Thus, a complete axial dataset displaying distinct retinal and RPE layers was obtained. 
Figure 1
 
Example of a patient with foveal sparing in polarization-sensitive retinal imaging. Example of a patient with BCVA of 20/22 Snellen equivalents imaged by PS-OCT. (A) The acquired en face intensity projection image during the PS-OCT volume scan acquisition. (B) En face map of depolarizing tissue thickness and visualizing the multifocal areas of atrophy and the foveal sparing in this patient. (C) The DOPU image used to identify depolarizing tissue in the position of the yellow line in image (A). (D) The intensity B-scan image in the position of the yellow line in image (A). (E) The same B scan as in (C) segmented by the GA software and showing the depolarizing tissue overlay (RPE) in red.
Figure 1
 
Example of a patient with foveal sparing in polarization-sensitive retinal imaging. Example of a patient with BCVA of 20/22 Snellen equivalents imaged by PS-OCT. (A) The acquired en face intensity projection image during the PS-OCT volume scan acquisition. (B) En face map of depolarizing tissue thickness and visualizing the multifocal areas of atrophy and the foveal sparing in this patient. (C) The DOPU image used to identify depolarizing tissue in the position of the yellow line in image (A). (D) The intensity B-scan image in the position of the yellow line in image (A). (E) The same B scan as in (C) segmented by the GA software and showing the depolarizing tissue overlay (RPE) in red.
The extent of areas with disturbed RPE is delineated by summing up the number of depolarizing pixels (DOPU < 0.75) along each A scan throughout the entire three-dimensional (3D) PS-OCT dataset. A 2D en face map visualizing the pattern of depolarizing material at the RPE level is formed (Figs. 1B, 2E1, 2E2). In addition to detection of areas with a complete absence of RPE, another map is obtained that allows detection of any thinning, thickening or distortion of RPE. This map is consistent with the en face maps obtained by FAF and IR imaging, as proven by our group.27,29,33 The algorithm used to generate RPE thickness maps has been described.33,34,41 In brief, the algorithm first detects the most intensively depolarizing pixel along each A scan. This pixel is typically located in the physiological RPE cells. Segmented pixels located more than 10 pixels away from the average of the segmented locations in the ±2 neighboring A scans (e.g., outliers in the choroid) are detected and interpolated. The resulting RPE profile is then smoothed using a Savitzky-Golay filter (polynominal order 3, filter length 200 pixels). The results approximately fit the RPE position in high resolution. Subsequently, the anterior border of the RPE is located by searching for the first depolarizing pixel in each A scan, starting from the inner retinal layers. Outliers are again deleted from the resulting data points and fitted with the Savitzky-Golay filter (polynominal order 3, filter length 6).33 Finally, a B-scan image (used for the analyses in this study) and a RPE thickness map are reconstructed by calculating the difference between the estimated physiological RPE fit and the anterior border of the RPE. 
Figure 2
 
Examples of patients with fovea spared and involved by the GA process in conventional and polarization-sensitive retinal imaging. Images (A1F1): Example of a patient with BCVA of 20/20 Snellen equivalent graded as grade 0 (fovea spared), by near IR (A1), SD-OCT (Spectralis HRA+OCT [B1]), FAF (C1), SD-OCT (Cirrus HD-OCT [D1]), and PS-OCT (E1, F1). Image (E1) represents the calculated en face map of depolarizing tissue based on PS-OCT B scans (F1). The fovea contour and neurosensory layers in the fovea are well defined in the PS-OCT and SD-OCT images (yellow arrows). The choroidal signal enhancement is detectable at the borders of the foveal depression (red arrows). The fovea is spared and encircled in yellow in the conventional en face images (A1, C1). In PS-OCT (F1), the depolarizing tissue is segmented in red by the GA software (white arrow indicates intact RPE in the foveal area), and the fovea contour and neurosensory layers are preserved. Images (A2F2): Example of a patient with best corrected visual acuity of 20/160 Snellen equivalent graded as grade 1 (fovea involved) in all retinal imaging modalities. (A2) Infrared imaging, (C2) Fundus autofluorescence imaging, (B2, D2) SD-OCT, and (E2, F2) PS-OCT. The fovea contour and the neurosensory layers are altered in the PS-OCT and SD-OCT images (red arrow SD-OCT; green arrow PS-OCT). The fovea has a typical GA fluorescence pattern in the en face images (red circle).
Figure 2
 
Examples of patients with fovea spared and involved by the GA process in conventional and polarization-sensitive retinal imaging. Images (A1F1): Example of a patient with BCVA of 20/20 Snellen equivalent graded as grade 0 (fovea spared), by near IR (A1), SD-OCT (Spectralis HRA+OCT [B1]), FAF (C1), SD-OCT (Cirrus HD-OCT [D1]), and PS-OCT (E1, F1). Image (E1) represents the calculated en face map of depolarizing tissue based on PS-OCT B scans (F1). The fovea contour and neurosensory layers in the fovea are well defined in the PS-OCT and SD-OCT images (yellow arrows). The choroidal signal enhancement is detectable at the borders of the foveal depression (red arrows). The fovea is spared and encircled in yellow in the conventional en face images (A1, C1). In PS-OCT (F1), the depolarizing tissue is segmented in red by the GA software (white arrow indicates intact RPE in the foveal area), and the fovea contour and neurosensory layers are preserved. Images (A2F2): Example of a patient with best corrected visual acuity of 20/160 Snellen equivalent graded as grade 1 (fovea involved) in all retinal imaging modalities. (A2) Infrared imaging, (C2) Fundus autofluorescence imaging, (B2, D2) SD-OCT, and (E2, F2) PS-OCT. The fovea contour and the neurosensory layers are altered in the PS-OCT and SD-OCT images (red arrow SD-OCT; green arrow PS-OCT). The fovea has a typical GA fluorescence pattern in the en face images (red circle).
Figure 3
 
Examples of patients demonstrating the grading system in conventional and PS-OCT imaging. Images (A1A5) show five different eyes imaged by all the retinal imaging devices used in this study and are examples for grade 0 (fovea spared). (A1) Infrared image of a patient with BCVA of 20/20 Snellen equivalents. (A2) Fundus autofluorescence image of a patient with BCVA 20/22. (A3) Spectral-domain OCT B scan (Spectralis) of a patient with BCVA of 20/32. (A4) Spectral-domain OCT B scan (Cirrus HD-OCT) of a patient with BCVA of 20/32. (A5) Segmented PS-OCT B scan of a patient with BCVA 20/40. The images (B1B5) represent five different eyes of patients where the fovea was graded as fovea involved (grade 1) in all retinal imaging devices. Respective BCVA results were (B1) IR image, BCVA = 20/80; (B2) FAF image, BCVA = 20/80; (B3) Spectralis SD-OCT, BCVA = 20/200; (B4) Cirrus SD-OCT, BCVA = 20/160; (B5) PS-OCT B scan, BCVA = 20/100. Images (C1C5) represent five examples of eyes where the fovea was graded as not quantifiable (grade 2). Respective BCVA results corresponding to the images were (C1) IR image, BCVA = 20/50; (C2) FAF image, BCVA = 20/160; (C3) Spectralis SD-OCT, BCVA = 20/200; (C4) Cirrus SD-OCT, BCVA = 20/100; (C5) PS-OCT B scan, BCVA = 20/25.
Figure 3
 
Examples of patients demonstrating the grading system in conventional and PS-OCT imaging. Images (A1A5) show five different eyes imaged by all the retinal imaging devices used in this study and are examples for grade 0 (fovea spared). (A1) Infrared image of a patient with BCVA of 20/20 Snellen equivalents. (A2) Fundus autofluorescence image of a patient with BCVA 20/22. (A3) Spectral-domain OCT B scan (Spectralis) of a patient with BCVA of 20/32. (A4) Spectral-domain OCT B scan (Cirrus HD-OCT) of a patient with BCVA of 20/32. (A5) Segmented PS-OCT B scan of a patient with BCVA 20/40. The images (B1B5) represent five different eyes of patients where the fovea was graded as fovea involved (grade 1) in all retinal imaging devices. Respective BCVA results were (B1) IR image, BCVA = 20/80; (B2) FAF image, BCVA = 20/80; (B3) Spectralis SD-OCT, BCVA = 20/200; (B4) Cirrus SD-OCT, BCVA = 20/160; (B5) PS-OCT B scan, BCVA = 20/100. Images (C1C5) represent five examples of eyes where the fovea was graded as not quantifiable (grade 2). Respective BCVA results corresponding to the images were (C1) IR image, BCVA = 20/50; (C2) FAF image, BCVA = 20/160; (C3) Spectralis SD-OCT, BCVA = 20/200; (C4) Cirrus SD-OCT, BCVA = 20/100; (C5) PS-OCT B scan, BCVA = 20/25.
Identification of Foveal Sparing
A systematic grading scale was established and SD-OCT and PS-OCT scans, FAF and IR images were graded (RS) according to a predetermined grading protocol in an anonymized and masked manner (i.e., in a random sequence for each eye). 
The grading system was designed according to disease progression with grade 0 fovea spared, grade 1 fovea involved and grade 2 fovea not quantifiable. By SD-OCT, the fovea was graded as involved when the RPE in the central portion of the fovea was continuously absent, associated with an enhanced light transmission into the choroid, and/or the RPE band was absent compared with the normal RPE band. The fovea was graded as spared when the central portion of the foveal depression showed a regular morphology with normal inner and outer retinal layers and could be identified in five consecutive B scans in Cirrus HD-OCT and three B scans in Spectralis HRA+OCT and PS-OCT (Figs. 1, 2B1, 2D1, 2F1, 3A3–A5). The number of B scans analyzed were different because the OCT systems used acquired different scan densities. Polarization-sensitive OCT identified foveal sparing using the algorithm and color-coding described above. Foveal sparing was graded whenever a continuous depolarizing RPE band extended to the central foveal area. By FAF, the fovea was defined as involved when the gray-scale image in FAF had identical intensity values in the region of the supposed fovea as in the central region of the GA lesion. The images were analyzed simultaneously when the combination of FAF and IR images was used. When none of the criteria described above were met, the fovea was graded as fovea not gradable. Examples of typical gradings are given in Figure 3
Correlation of BCVA and Morphologic Sparing or Involvement
After grading the morphology in the different imaging modalities, the results of the grading were related with functional values from BCVA testing. The diameter of the fovea was assumed to be 1.5 mm.36 Therefore, the radius of the physiological fovea is 0.75 mm or 2.65° (assuming a standard eye length of 24 mm).42 From the center of the focal BCVA distribution of the posterior pole (Fig. 4; see also fig. 4 in Ref. 43), a BCVA greater than 20/40 can be expected within a distance of 2.65° from the foveal center. Therefore, for our study we arbitrarily concluded that a BCVA greater than 20/40 can be associated with a morphologic foveal sparing and a BCVA less than 20/40 with a fovea, which is affected by disease. This functional information was used in the comparative analysis of the retinal imaging modalities' ability to be graded correctly according to the BCVA threshold. 
Figure 4
 
Illustration of potential visual acuity in an SD-OCT B scan. Overlay of the distribution graph of visual acuity across the retina and delineation of 2.65° from the center of the fovea and an SD-OCT B scan to illustrate the potential visual acuity when the fovea is affected or spared by the GA disease process. The yellow line delineate the center of the fovea with maximum BCVA of 20/20 and red lines represent the margins (20/50–20/33) of possible BCVA when the fovea is altered (modified with permission from Coren S, Ward LM, Enns JT. Sensation and Perception. 5th ed. Fort Worth, TX: Harcourt Brace College Publishers; 1999).
Figure 4
 
Illustration of potential visual acuity in an SD-OCT B scan. Overlay of the distribution graph of visual acuity across the retina and delineation of 2.65° from the center of the fovea and an SD-OCT B scan to illustrate the potential visual acuity when the fovea is affected or spared by the GA disease process. The yellow line delineate the center of the fovea with maximum BCVA of 20/20 and red lines represent the margins (20/50–20/33) of possible BCVA when the fovea is altered (modified with permission from Coren S, Ward LM, Enns JT. Sensation and Perception. 5th ed. Fort Worth, TX: Harcourt Brace College Publishers; 1999).
Statistical Analysis
An ANOVA was used to analyze the relation between the BCVA functional values and the morphologic grading results from all the imaging systems (the BCVA results in letters were analyzed and then converted into Snellen VA for descriptive purposes). To compare the ability of the retinal imaging methods to reliably grade the fovea, we constructed a reference grading (“known truth”) based on the BCVA of each eye. We then computed a weighted and an unweighted Cohen's κ statistic4446 to obtain a κ statistic for each method compared with the reference grading. For a weighted κ, where weight was set to equal, we assumed that an unclear grading (grade 2) was equally better than a wrong grading as it was worse than a right grading. For an unweighted κ, any grading different from the reference grading, including an unclear grading, was considered equally bad. To identify significant differences between κ coefficients, κ statistics for different methods were obtained for the same cohort and were thus expected to be correlated. We tested the hypothesis that κ1 = … = κn, where κ1 is the agreement between FAF and the reference BCVA and κi for i = 2…n are the agreements of the alternative methods with the BCVA reference. We tested the significance of differences relative to randomly expected variance using an established bootstrap method47 with 2001 iterations. 
Results
Patient Characteristics and Balance Between Groups
One hundred seventy-six eyes from 96 patients (60 women, mean age 78 years) were included and imaged using a Cirrus HD-OCT and a Spectralis HRA+OCT device. Thus, two different SD-OCT volume scans, FAF and IR images were obtained. As soon as the new PS-OCT system became available, a subset of 95 eyes from 54 consecutive patients (34 women, mean age 77 years) were also imaged with PS-OCT to selectively identify the condition of the RPE layer. The PS-OCT system had not been available to scan the other eyes. Mean BCVA was 20/51 Snellen equivalents for the first 176 eyes and 20/50 for the 95 eyes in the PS-OCT group, showing a balanced functional composition. In total, 72 eyes from 57 patients (41% of all eyes included) had a BCVA greater than or equal to 20/40, suggesting functional preservation of the central neurosensory fovea. In the PS-OCT group, 38 eyes from 31 patients (40% of the eyes included in this group) had a BCVA greater than or equal to 20/40, again demonstrating equivalent conditions. 
Analysis of Foveal Grading in All Imaging Modalities and Relation to BCVA
All grading results and comparisons between the imaging modalities and comparisons ofer the grading results with the threshold of BCVA greater than or equal to 20/40 or BCVA less than 20/40 are listed in Tables 1 through 4 and Figure 5. Furthermore, the ANOVA relating the gradings in the different imaging systems to the corresponding BCVA values showed that the differences between the groups (grades 0, 1, and 2) were highly significant with P less than 0.0001 for all systems. 
Table 1
 
Morphologic Grading and BCVA Results in All Imaging Modalities
Table 1
 
Morphologic Grading and BCVA Results in All Imaging Modalities
Figure 5
 
Relative grading results distribution in axial and en face imaging modalities in relation to functional values. Proportion of gradings indicating the morphologic condition of the fovea for each imaging modality in each grading category for patients with BCVA ≥ 20/40 (top) and BCVA less than 20/40 (bottom). Top: Polarization-sensitive OCT is the most sensitive method for the detection of foveal sparing. Bottom: There is an equal distribution among all retinal imaging modalities in finding an involved fovea in the disease process. Furthermore, PS-OCT achieves the highest values for identification of a spared foveal RPE integrity and lowest values for failed gradability.
Figure 5
 
Relative grading results distribution in axial and en face imaging modalities in relation to functional values. Proportion of gradings indicating the morphologic condition of the fovea for each imaging modality in each grading category for patients with BCVA ≥ 20/40 (top) and BCVA less than 20/40 (bottom). Top: Polarization-sensitive OCT is the most sensitive method for the detection of foveal sparing. Bottom: There is an equal distribution among all retinal imaging modalities in finding an involved fovea in the disease process. Furthermore, PS-OCT achieves the highest values for identification of a spared foveal RPE integrity and lowest values for failed gradability.
Cohen's κ Statistics for the Comparison of Conventional Retinal Imaging
We considered the three possible grading results from the 176 eyes imaged by all retinal imaging systems except for PS-OCT and compared them with our reference BCVA threshold (BCVA ≥ 20/40 and BCVA < 20/40). We did this using a weighted and unweighted approach (see Methods section for details) of Cohen's κ statistics. The weighted κ ranged between κ = 0.255 with P less than 0.01 for Spectralis-OCT and κ = 0.364 with P less than 0.01 for the combination of FAF and IR imaging for all the conventional retinal imaging methods. In an unweighted κ statistic, κ ranged between κ = 0.229 with P less than 0.01 for Cirrus HD-OCT and κ = 0.295 with P less than 0.01 for FAF. 
Then, we assessed significant deviations by means of a bootstrap procedure for comparing correlated κ coefficients for all conventional retinal imaging methods. We obtained a global Hotelling's T2 of 2.04 with a P value of 0.73 in a weighted approach and of 1.17 with a P value of 0.88 in an unweighted approach for the comparison of conventional retinal imaging modalities with each other. 
Cohen's κ Statistics for the Comparison of Conventional Retinal Imaging and PS-OCT
We used a weighted approach of Cohen's κ statistics for the 95 eyes imaged by all retinal imaging systems including PS-OCT for comparing the conventional retinal imaging modalities with PS-OCT. κ ranged between κ = 0.221 for Cirrus HD-OCT with P less than 0.01 and κ = 0.493 with P less than 0.01 for PS-OCT. The bootstrap test yielded a global Hotelling's T2 statistic of 5.73 with P = 0.017. 
We also used an unweighted approach with which κ ranged between κ = 0.187 for Cirrus HD-OCT with P less than 0.01 and κ = 0.488 with P less than 0.01 for PS-OCT. The bootstrap test then yielded a global Hotelling's T2 statistic of 17.9 with a P value of P = 0.003. 
Discussion
Identifying the foveal status could be relevant for the visual prognosis of patients during disease progression as well as in the evaluation of therapeutic efficacy in upcoming clinical trials. Therefore, the aim of this study was to analyze if PS-OCT imaging is superior to conventional imaging and if conventional retinal imaging modalities differ in their assessments of the condition of the fovea, particularly foveal sparing. Best-corrected visual acuity was used as an external factor in our analysis to differentiate between foveal sparing and foveal involvement. We compared the functional threshold of BCVA greater than or equal to or less than 20/40 with the grading results of the fovea anatomically denoted as fovea spared (grade 0), fovea involved (grade 1), or fovea not quantifiable (grade 2) and found statistically significant differences between BCVA and the grading groups with all imaging methods, with P < 0.0001 (ANOVA). 
The statistical analyses presented in Table 1 show that the BCVA results of the gradings in the 95% confidence interval of every system consistently adhered to the BCVA thresholds. For example, BCVA in grade 0 ranged from 20/44 to 20/31 Snellen for all imaging systems. 
A comparison of SD-OCT (Spectralis only) with FAF and a combination of near IR plus FAF imaging found a BCVA ranging from 0.30 to 0.12 logarithm of minimal angle resolution in eyes with foveal sparing.11 This result accords with our results of a BCVA of 20/40 or higher for grade 0, if the fovea is to be graded as spared. 
The mean BCVA in all retinal imaging methods for grade 1 in our study was between 20/125 and 20/74. Therefore, we can conclude that our grading 1 for an involved fovea was reliable. 
The mean BCVA in en face imaging for grade 2 was between 20/53 and 20/34. A possible explanation is that the fovea in this group of eyes graded with grade 2 was in the process of ongoing destruction and therefore assessing if the fovea was spared or involved was difficult. The BCVA for grade 2 had a wider range for the OCT systems (i.e., between 20/154 in Spectralis and 20/37 in Cirrus imaging and in one case in PS-OCT; BCVA = 20/25). The explanation for the wider range in the OCT systems could be that residual RPE-like hyperreflective structures observed in OCT scans are functional or that they are not functional and are over interpreted. 
We found no statistically significant differences between the conventional retinal imaging modalities with the conservative and robust Cohen's κ statistical measure. However, although not statistically significant, the SD-OCT segment did show more positive results in discerning foveal sparing in the descriptive statistics. When comparing the modalities the drawbacks of the FAF and IR systems should also be considered. Making decisions upon foveal involvement is challenging with FAF imaging as the macular pigment in the neurosensory foveal retina blocks the blue excitation light.9,11,27,48 In both FAF and IR en face imaging distinguishing the exact position of the foveal center in the absence of a structural foveal depression is harder than with SD-OCT. 
Although we found no statistically significant differences between the conventional retinal imaging modalities, when analyzing our grading results in combination with BCVA, the subgroup of eyes scanned with PS-OCT showed statistically significant differences in grading the fovea relative to BCVA compared with conventional retinal imaging. 
The descriptive grading results showed PS-OCT correlated better than en face imaging with BCVA (e.g., from Table 2, in 22 eyes PS-OCT grade 0/FAF grade 1, mean BCVA of 20/37; or in 12 eyes, PS-OCT grade 0/IR grade 1, mean BCVA of 20/38). 
Table 2
 
Comparison of Gradings and BCVA in Conventional Retinal Imaging Systems (176 Eyes) and PS-OCT (95 Eyes)
Table 2
 
Comparison of Gradings and BCVA in Conventional Retinal Imaging Systems (176 Eyes) and PS-OCT (95 Eyes)
Fundus autofluorescence has the advantage over PS-OCT of being able to predict lesion progression toward hyperfluorescence at the edges of the lesion.18,22,49,50 However, PS-OCT provides more detailed information about the condition of the RPE layer in terms of thinning, thickening, detachment, and porosity at the GA rim. This information is more useful for advanced analysis of GA pathophysiology than the hyperfluorescent feature surrounding the GA lesion in FAF alone without any corresponding morphologic details. Moreover, 3D RPE mapping by PS-OCT segmentation is an automatic function and may lead to a GA classification based on distinct morphologic features in the near future. 
Polarization-sensitive OCT detected 84% of eyes with BCVA greater than or equal to 20/40, whereas SD-OCT only detected 47% (Table 4). Furthermore, 15 eyes were graded 1 with PS-OCT and either 0 or 2 with Cirrus HD-OCT with a mean BCVA of 20/77 and 20/154, respectively. In 14 eyes, the grading differed between PS-OCT grade 1 and Spectralis HRA+OCT either grade 0 or 2 with mean BCVA 20/71 and 20/100 (Table 3). The explanation for the different gradings could be that in SD-OCT pathologic residual structures in the RPE no longer contribute to visual function. Cirrus HD-OCT and Spectralis HRA+OCT unlike PS-OCT have no tissue-specific differentiation function and could therefore mislabel and overinterpret hyperreflective unspecific residual structures as viable. 
Table 3
 
Comparison of All OCT Grading Results and BCVA Values
Table 3
 
Comparison of All OCT Grading Results and BCVA Values
Table 4
 
Grading and BCVA Results in Conventional Retinal Imaging Systems (176 Eyes) and in PS-OCT (95 Eyes) in Eyes Where BCVA Was Over and Under 20/40 Snellen Equivalents
Table 4
 
Grading and BCVA Results in Conventional Retinal Imaging Systems (176 Eyes) and in PS-OCT (95 Eyes) in Eyes Where BCVA Was Over and Under 20/40 Snellen Equivalents
It is noteworthy, however, that PS-OCT can be used in an en face mode in addition to axial OCT scanning based on the PS-OCT algorithm described earlier.3335,41 Hence, PS-OCT integrates the advantages of axial and en face imaging in a single approach but for now it is more time consuming and unavailable commercially. 
The limitations of our study are mostly related to the subjective grading of each imaging method by the grader and the BCVA threshold chosen. Although other thresholds have been tried and the results were no different, comparing retinal imaging methods using an external factor other than BCVA in GA, for example, fixation stability in microperimetry, could have led to other results. 
In conclusion, our study shows that PS-OCT, an advanced OCT technology, can reliably assess the fovea and delineate the area of atrophy, while simultaneously discerning neurodegenerative processes in the retina. Polarization-sensitive OCT may provide new insights into the pathophysiology of atrophic AMD. Our results should encourage investigators to include PS-OCT in future clinical trials seeking to develop agents and strategies to preserve the RPE and the neurosensory retina in the sensitive fovea. 
Acknowledgments
The authors thank Christoph Mitsch, MD, for his technical advice. 
Disclosure: R.G. Sayegh, None; S. Zotter, Canon, Inc. Tokyo (F); P.K. Roberts, Canon, Inc. Tokyo (F); M.M. Kandula, None; S. Sacu, None; D.P. Kreil, None; B. Baumann, Canon, Inc. Tokyo (F); M. Pircher, Canon, Inc. Tokyo (F); C.K. Hitzenberger, Canon, Inc. Tokyo (F); U. Schmidt-Erfurth, Canon, Inc. Tokyo (F) 
References
Green WR, Key SN. Senile macular degeneration: a histopathologic study. Trans Am Ophthalmol Soc. 1977; 75: 180–254.
Sarks SH, Arnold JJ, Killingsworth MC, Sarks JP. Early drusen formation in the normal and aging eye and their relation to age related maculopathy: a clinicopathological study. Br J Ophthalmol. 1999; 83: 358–368.
Schatz H, McDonald HR. Atrophic macular degeneration. Rate of spread of geographic atrophy and visual loss. Ophthalmology. 1989; 96: 1541–1551.
Maguire P, Vine AK. Geographic atrophy of the retinal pigment epithelium. Am J Ophthalmol. 1986; 102: 621–625.
Sunness JS. The natural history of geographic atrophy the advanced atrophic form of age-related macular degeneration. Mol Vis. 1999; 5: 25.
Sunness JS, Rubin GS, Zuckerbrod A, Applegate CA. Foveal-sparing scotomas in advanced dry age-related macular degeneration. J Vis Impair Blind. 2008; 102: 600–610.
Klein R, Klein BE, Linton KL. Prevalence of age-related maculopathy. The Beaver Dam Eye Study. Ophthalmology. 1992; 99: 933–943.
Klein R, Klein BE, Knudtson MD, Meuer SM, Swift M, Gangnon RE. Fifteen-year cumulative incidence of age-related macular degeneration: the Beaver Dam Eye Study. Ophthalmology. 2007; 114: 253–262.
Schmitz-Valckenberg S, Fleckenstein M, Helb HM, Charbel Issa P, Scholl HP, Holz FG. In vivo imaging of foveal sparing in geographic atrophy secondary to age-related macular degeneration. Invest Ophthalmol Vis Sci. 2009; 50: 3915–3921.
Pilotto E, Vujosevic S, Melis R, et al. Short wavelength fundus autofluorescence versus near-infrared fundus autofluorescence, with microperimetric correspondence, in patients with geographic atrophy due to age-related macular degeneration. Br J Ophthalmol. 2011; 95: 1140–1144.
Forte R, Querques G, Querques L, Leveziel N, Benhamou N, Souied EH. Multimodal evaluation of foveal sparing in patients with geographicatrophy due to age-related macular degeneration. Retina. 2013; 33: 482–489.
Monés J, Biarnés M, Trindade F, Arias L, Alonso J. Optical coherence tomography assessment of apparent foveal swelling in patients with foveal sparing secondary to geographic atrophy. Ophthalmology. 2013; 120: 829–836.
Brader HS, Ying GS, Martin ER, Maguire MG. Characteristics of incident geographic atrophy in the complications of age-related macular degeneration prevention trial. Ophthalmology. 2013; 120: 1871–1879.
Chen C, Wu L, Wu D, et al. The local cone and rod system function in early age-related macular degeneration. Doc Ophthalmol. 2004; 109: 1–8.
Curcio CA. Photoreceptor topography in ageing and age-related maculopathy. Eye (Lond). 2001; 15: 376–383.
Curcio CA, Millican CL, Allen KA, Kalina RE. Aging of the human photoreceptor mosaic: evidence for selective vulnerability of rods in central retina. Invest Ophthalmol Vis Sci. 1993; 34: 3278–3296.
Jackson GR, Owsley C, Curcio CA. Photoreceptor degeneration and dysfunction in aging and age-related maculopathy. Ageing Res Rev. 2002; 1: 381–396.
Mauschitz MM, Fonseca S, Chang P, et al. Topography of geographic atrophy in age-related macular degeneration. Invest Ophthalmol Vis Sci. 2012; 53: 4932–4939.
Sunness JS, Bressler NM, Tian Y, Alexander J, Applegate CA. Measuring geographic atrophy in advanced age-related macular degeneration. Invest Ophthalmol Vis Sci. 1999; 40: 1761–1769.
Delori FC. Spectrophotometer for noninvasive measurement of intrinsic fluorescence and reflectance of the ocular fundus. Appl Opt. 1994; 33: 7439–7452.
Delori FC, Dorey CK, Staurenghi G, Arend O, Goger DG, Weiter JJ. In vivo fluorescence of the ocular fundus exhibits retinal pigment epithelium lipofuscin characteristics. Invest Ophthalmol Vis Sci. 1995; 36: 718–729.
Bindewald A, Bird AC, Dandekar SS, et al. Classification of fundus autofluorescence patterns in early age-related macular disease. Invest Ophthalmol Vis Sci. 2005; 46: 3309–3314.
Bindewald A, Schmitz-Valckenberg S, Jorzik JJ, et al. Classification of abnormal fundus autofluorescence patterns in the junctional zone of geographic atrophy in patients with age related macular degeneration. Br J Ophthalmol. 2005; 89: 874–878.
Witkin AJ, Vuong LN, Srinivasan VJ, et al. High-speed ultrahigh resolution optical coherence tomography before and after ranibizumab for age-related macular degeneration. Ophthalmology. 2009; 116: 956–963.
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.
Kiss CG, Geitzenauer W, Simader C, Gregori G, Schmidt-Erfurth U. Evaluation of ranibizumab-induced changes in high-resolution optical coherence tomographic retinal morphology and their impact on visual function. Invest Ophthalmol Vis Sci. 2009; 50: 2376–2383.
Sayegh RG, Simader C, Scheschy U, et al. A systematic comparison of spectral-domain optical coherence tomography and fundus autofluorescence in patients with geographic atrophy. Ophthalmology. 2011; 118: 1844–1851.
Sayegh RG, Kiss CG, Simader C, et al. A systematic correlation of morphology and function using spectral domain optical coherence tomography and microperimetry in patients with geographic atrophy. Br J Ophthalmol. 2014; 98: 1050–1055.
Simader C, Sayegh RG, Montuoro A, et al. A longitudinal comparison of spectral-domain optical coherence tomography and fundus autofluorescence in geographic atrophy. Am J Ophthalmol. 2014; 158: 557–566.
Pircher M, Hitzenberger CK, Schmidt-Erfurth U. Polarization sensitive optical coherence tomography in the human eye. Prog Retin Eye Res. 2011; 30: 431–451.
Baumann B, Baumann SO, Konegger T, et al. Polarization sensitive optical coherence tomography of melanin provides intrinsic contrast based on depolarization. Biomed Opt Express. 2012; 3: 1670.
Götzinger E, Pircher M, Geitzenauer W, et al. Retinal pigment epithelium segmentation by polarization sensitive optical coherence tomography. Opt Express. 2008; 16: 16410–16422.
Baumann B, Gotzinger E, Pircher M, et al. Segmentation and quantification of retinal lesions in age-related macular degeneration using polarization-sensitive optical coherence tomography. J Biomed Opt. 2010; 15: 61704.
Schlanitz FG, Baumann B, Spalek T, et al. Performance of automated drusen detection by polarization-sensitive optical coherence tomography. Invest Ophthalmol Vis Sci. 2011; 52: 4571–4579.
Schütze C, Bolz M, Sayegh R, et al. Lesion size detection in geographic atrophy by polarization-sensitive optical coherence tomography and correlation to conventional imaging techniques. Invest Ophthalmol Vis Sci. 2013; 54: 739–745.
Gass JDM. Stereoscopic atlas of macular diseases. 1987; 1:2.
Sayegh RG, Georgopoulos M, Geitzenauer W, Simader C, Kiss C, Schmidt-Erfurth U. High-resolution optical coherence tomography after surgery for vitreomacular traction: a 2-year follow-up. Ophthalmology. 2010; 117: 2010–2017.
Zotter S, Pircher M, Torzicky T, et al. Large-field high-speed polarization sensitive spectral domain OCT and its applications in ophthalmology. Biomed Opt Express. 2012; 3: 2720.
Pircher M, Götzinger E, Findl O, et al. Human macula investigated in vivo with polarization-sensitive optical coherence tomography. Invest Ophthalmol Vis Sci. 2006; 47: 5487–5494.
Götzinger E, Pircher M, Hitzenberger CK. High speed spectral domain polarization sensitive optical coherence tomography of the human retina. Opt Express. 2005; 13: 10217–10229.
Ritter M, Zotter S, Schmidt WM, et al. Characterization of Stargardt disease using polarization-sensitive optical coherence tomography and fundus autofluorescence imaging. Invest Ophthalmol Vis Sci. 2013; 54: 6416–6425.
Gullstrand A. The optical system of the eye. Helmholtz's Treatise Physiol Opt. 1910; 1: 350–358.
Coren S, Enns JT, Ward LM, . Sensation and Perception. 5th ed. Boston: Wiley and Sons; 1999.
Cohen J. A coefficient of agreement of nominal scales. Educ Psychol Meas. 1960; 20: 37–46.
Cohen J. Weighted kappa: nominal scale agreement with provision for scaled disagreement or partial credit. Psychol Bull. 1968; 70: 213–220.
Fleiss JL, Cohen J, Everitt BS. Large sample standard errors of kappa and weighted kappa. Psychol Bull. 1969; 72: 323–327.
Vanbelle S, Albert A. A bootstrap method for comparing correlated kappa coefficients. J Stat Comput Simul. 2008; 78: 1009–1015.
Wolf-Schnurrbusch UE, Wittwer VV, Ghanem R, et al. Blue-light versus green-light autofluorescence: lesion size of areas of geographic atrophy. Invest Ophthalmol Vis Sci. 2011; 52: 9497–9502.
Biarnés M, Monés J, Trindade F, Alonso J, Arias L. Intra and interobserver agreement in the classification of fundus autofluorescence patterns in geographic atrophy secondary to age-related macular degeneration. Graefe's Arch Clin Exp Ophthalmol. 2012; 250: 485–490.
Fleckenstein M, Schmitz-Valckenberg S, Adrion C, et al. Progression of age-related geographic atrophy: role of the fellow eye. Invest Ophthalmol Vis Sci. 2011; 52: 6552–6557.
Figure 1
 
Example of a patient with foveal sparing in polarization-sensitive retinal imaging. Example of a patient with BCVA of 20/22 Snellen equivalents imaged by PS-OCT. (A) The acquired en face intensity projection image during the PS-OCT volume scan acquisition. (B) En face map of depolarizing tissue thickness and visualizing the multifocal areas of atrophy and the foveal sparing in this patient. (C) The DOPU image used to identify depolarizing tissue in the position of the yellow line in image (A). (D) The intensity B-scan image in the position of the yellow line in image (A). (E) The same B scan as in (C) segmented by the GA software and showing the depolarizing tissue overlay (RPE) in red.
Figure 1
 
Example of a patient with foveal sparing in polarization-sensitive retinal imaging. Example of a patient with BCVA of 20/22 Snellen equivalents imaged by PS-OCT. (A) The acquired en face intensity projection image during the PS-OCT volume scan acquisition. (B) En face map of depolarizing tissue thickness and visualizing the multifocal areas of atrophy and the foveal sparing in this patient. (C) The DOPU image used to identify depolarizing tissue in the position of the yellow line in image (A). (D) The intensity B-scan image in the position of the yellow line in image (A). (E) The same B scan as in (C) segmented by the GA software and showing the depolarizing tissue overlay (RPE) in red.
Figure 2
 
Examples of patients with fovea spared and involved by the GA process in conventional and polarization-sensitive retinal imaging. Images (A1F1): Example of a patient with BCVA of 20/20 Snellen equivalent graded as grade 0 (fovea spared), by near IR (A1), SD-OCT (Spectralis HRA+OCT [B1]), FAF (C1), SD-OCT (Cirrus HD-OCT [D1]), and PS-OCT (E1, F1). Image (E1) represents the calculated en face map of depolarizing tissue based on PS-OCT B scans (F1). The fovea contour and neurosensory layers in the fovea are well defined in the PS-OCT and SD-OCT images (yellow arrows). The choroidal signal enhancement is detectable at the borders of the foveal depression (red arrows). The fovea is spared and encircled in yellow in the conventional en face images (A1, C1). In PS-OCT (F1), the depolarizing tissue is segmented in red by the GA software (white arrow indicates intact RPE in the foveal area), and the fovea contour and neurosensory layers are preserved. Images (A2F2): Example of a patient with best corrected visual acuity of 20/160 Snellen equivalent graded as grade 1 (fovea involved) in all retinal imaging modalities. (A2) Infrared imaging, (C2) Fundus autofluorescence imaging, (B2, D2) SD-OCT, and (E2, F2) PS-OCT. The fovea contour and the neurosensory layers are altered in the PS-OCT and SD-OCT images (red arrow SD-OCT; green arrow PS-OCT). The fovea has a typical GA fluorescence pattern in the en face images (red circle).
Figure 2
 
Examples of patients with fovea spared and involved by the GA process in conventional and polarization-sensitive retinal imaging. Images (A1F1): Example of a patient with BCVA of 20/20 Snellen equivalent graded as grade 0 (fovea spared), by near IR (A1), SD-OCT (Spectralis HRA+OCT [B1]), FAF (C1), SD-OCT (Cirrus HD-OCT [D1]), and PS-OCT (E1, F1). Image (E1) represents the calculated en face map of depolarizing tissue based on PS-OCT B scans (F1). The fovea contour and neurosensory layers in the fovea are well defined in the PS-OCT and SD-OCT images (yellow arrows). The choroidal signal enhancement is detectable at the borders of the foveal depression (red arrows). The fovea is spared and encircled in yellow in the conventional en face images (A1, C1). In PS-OCT (F1), the depolarizing tissue is segmented in red by the GA software (white arrow indicates intact RPE in the foveal area), and the fovea contour and neurosensory layers are preserved. Images (A2F2): Example of a patient with best corrected visual acuity of 20/160 Snellen equivalent graded as grade 1 (fovea involved) in all retinal imaging modalities. (A2) Infrared imaging, (C2) Fundus autofluorescence imaging, (B2, D2) SD-OCT, and (E2, F2) PS-OCT. The fovea contour and the neurosensory layers are altered in the PS-OCT and SD-OCT images (red arrow SD-OCT; green arrow PS-OCT). The fovea has a typical GA fluorescence pattern in the en face images (red circle).
Figure 3
 
Examples of patients demonstrating the grading system in conventional and PS-OCT imaging. Images (A1A5) show five different eyes imaged by all the retinal imaging devices used in this study and are examples for grade 0 (fovea spared). (A1) Infrared image of a patient with BCVA of 20/20 Snellen equivalents. (A2) Fundus autofluorescence image of a patient with BCVA 20/22. (A3) Spectral-domain OCT B scan (Spectralis) of a patient with BCVA of 20/32. (A4) Spectral-domain OCT B scan (Cirrus HD-OCT) of a patient with BCVA of 20/32. (A5) Segmented PS-OCT B scan of a patient with BCVA 20/40. The images (B1B5) represent five different eyes of patients where the fovea was graded as fovea involved (grade 1) in all retinal imaging devices. Respective BCVA results were (B1) IR image, BCVA = 20/80; (B2) FAF image, BCVA = 20/80; (B3) Spectralis SD-OCT, BCVA = 20/200; (B4) Cirrus SD-OCT, BCVA = 20/160; (B5) PS-OCT B scan, BCVA = 20/100. Images (C1C5) represent five examples of eyes where the fovea was graded as not quantifiable (grade 2). Respective BCVA results corresponding to the images were (C1) IR image, BCVA = 20/50; (C2) FAF image, BCVA = 20/160; (C3) Spectralis SD-OCT, BCVA = 20/200; (C4) Cirrus SD-OCT, BCVA = 20/100; (C5) PS-OCT B scan, BCVA = 20/25.
Figure 3
 
Examples of patients demonstrating the grading system in conventional and PS-OCT imaging. Images (A1A5) show five different eyes imaged by all the retinal imaging devices used in this study and are examples for grade 0 (fovea spared). (A1) Infrared image of a patient with BCVA of 20/20 Snellen equivalents. (A2) Fundus autofluorescence image of a patient with BCVA 20/22. (A3) Spectral-domain OCT B scan (Spectralis) of a patient with BCVA of 20/32. (A4) Spectral-domain OCT B scan (Cirrus HD-OCT) of a patient with BCVA of 20/32. (A5) Segmented PS-OCT B scan of a patient with BCVA 20/40. The images (B1B5) represent five different eyes of patients where the fovea was graded as fovea involved (grade 1) in all retinal imaging devices. Respective BCVA results were (B1) IR image, BCVA = 20/80; (B2) FAF image, BCVA = 20/80; (B3) Spectralis SD-OCT, BCVA = 20/200; (B4) Cirrus SD-OCT, BCVA = 20/160; (B5) PS-OCT B scan, BCVA = 20/100. Images (C1C5) represent five examples of eyes where the fovea was graded as not quantifiable (grade 2). Respective BCVA results corresponding to the images were (C1) IR image, BCVA = 20/50; (C2) FAF image, BCVA = 20/160; (C3) Spectralis SD-OCT, BCVA = 20/200; (C4) Cirrus SD-OCT, BCVA = 20/100; (C5) PS-OCT B scan, BCVA = 20/25.
Figure 4
 
Illustration of potential visual acuity in an SD-OCT B scan. Overlay of the distribution graph of visual acuity across the retina and delineation of 2.65° from the center of the fovea and an SD-OCT B scan to illustrate the potential visual acuity when the fovea is affected or spared by the GA disease process. The yellow line delineate the center of the fovea with maximum BCVA of 20/20 and red lines represent the margins (20/50–20/33) of possible BCVA when the fovea is altered (modified with permission from Coren S, Ward LM, Enns JT. Sensation and Perception. 5th ed. Fort Worth, TX: Harcourt Brace College Publishers; 1999).
Figure 4
 
Illustration of potential visual acuity in an SD-OCT B scan. Overlay of the distribution graph of visual acuity across the retina and delineation of 2.65° from the center of the fovea and an SD-OCT B scan to illustrate the potential visual acuity when the fovea is affected or spared by the GA disease process. The yellow line delineate the center of the fovea with maximum BCVA of 20/20 and red lines represent the margins (20/50–20/33) of possible BCVA when the fovea is altered (modified with permission from Coren S, Ward LM, Enns JT. Sensation and Perception. 5th ed. Fort Worth, TX: Harcourt Brace College Publishers; 1999).
Figure 5
 
Relative grading results distribution in axial and en face imaging modalities in relation to functional values. Proportion of gradings indicating the morphologic condition of the fovea for each imaging modality in each grading category for patients with BCVA ≥ 20/40 (top) and BCVA less than 20/40 (bottom). Top: Polarization-sensitive OCT is the most sensitive method for the detection of foveal sparing. Bottom: There is an equal distribution among all retinal imaging modalities in finding an involved fovea in the disease process. Furthermore, PS-OCT achieves the highest values for identification of a spared foveal RPE integrity and lowest values for failed gradability.
Figure 5
 
Relative grading results distribution in axial and en face imaging modalities in relation to functional values. Proportion of gradings indicating the morphologic condition of the fovea for each imaging modality in each grading category for patients with BCVA ≥ 20/40 (top) and BCVA less than 20/40 (bottom). Top: Polarization-sensitive OCT is the most sensitive method for the detection of foveal sparing. Bottom: There is an equal distribution among all retinal imaging modalities in finding an involved fovea in the disease process. Furthermore, PS-OCT achieves the highest values for identification of a spared foveal RPE integrity and lowest values for failed gradability.
Table 1
 
Morphologic Grading and BCVA Results in All Imaging Modalities
Table 1
 
Morphologic Grading and BCVA Results in All Imaging Modalities
Table 2
 
Comparison of Gradings and BCVA in Conventional Retinal Imaging Systems (176 Eyes) and PS-OCT (95 Eyes)
Table 2
 
Comparison of Gradings and BCVA in Conventional Retinal Imaging Systems (176 Eyes) and PS-OCT (95 Eyes)
Table 3
 
Comparison of All OCT Grading Results and BCVA Values
Table 3
 
Comparison of All OCT Grading Results and BCVA Values
Table 4
 
Grading and BCVA Results in Conventional Retinal Imaging Systems (176 Eyes) and in PS-OCT (95 Eyes) in Eyes Where BCVA Was Over and Under 20/40 Snellen Equivalents
Table 4
 
Grading and BCVA Results in Conventional Retinal Imaging Systems (176 Eyes) and in PS-OCT (95 Eyes) in Eyes Where BCVA Was Over and Under 20/40 Snellen Equivalents
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×