August 2016
Volume 57, Issue 10
Open Access
Visual Psychophysics and Physiological Optics  |   August 2016
Microperimetry as an Outcome Measure in Choroideremia Trials: Reproducibility and Beyond
Author Affiliations & Notes
  • Ioannis S. Dimopoulos
    Department of Ophthalmology and Visual Sciences University of Alberta, Edmonton, Alberta, Canada
  • Calvin Tseng
    Department of Ophthalmology and Visual Sciences University of Alberta, Edmonton, Alberta, Canada
  • Ian M. MacDonald
    Department of Ophthalmology and Visual Sciences University of Alberta, Edmonton, Alberta, Canada
  • Correspondence: Ioannis S. Dimopoulos, Clinical Research Fellow, Department of Ophthalmology and Visual Sciences, 7-030 Katz Building, University of Alberta, Edmonton, Alberta, Canada, T6G 2E1; dimopoul@ualberta.ca
Investigative Ophthalmology & Visual Science August 2016, Vol.57, 4151-4161. doi:10.1167/iovs.16-19338
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Ioannis S. Dimopoulos, Calvin Tseng, Ian M. MacDonald; Microperimetry as an Outcome Measure in Choroideremia Trials: Reproducibility and Beyond. Invest. Ophthalmol. Vis. Sci. 2016;57(10):4151-4161. doi: 10.1167/iovs.16-19338.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose: To determine test-retest repeatability of microperimetry testing (MP) in choroideremia (CHM) subjects using standard and personalized stimulus grids.

Methods: Fifteen CHM subjects (28 eyes) underwent consecutive repeat examinations with the Macular Integrity Assessment (MAIA) microperimeter using a standard (10°) and a customized macular grid adapted to individual macular pathology. Repeatability of standard-grid mean (MS) and point-wise (PWS) sensitivity was determined and compared with age-matched controls (seven eyes), with PWS separately analyzed for loci within and outside the border of degeneration. Interpolated volumetric indices were used to estimate repeatability of customized grids and compare their performance to standard grids.

Results: Test-retest measures of standard-grid MS yielded higher coefficients of variation (CV) in CHM subjects compared with controls (0.09 vs. 0.02). Volumetric indices from customized grids improved repeatability by driving CV values to 0.05 and close to 0.02 for region-of-interest (ROI) analysis. Variability of PWS was significantly higher in CHM, especially at the border of degeneration (10.68 vs. 4.74 dB at the central retina, P < 0.001).

Conclusions: Microperimetry testing in CHM shows high test-retest variation at the border of degeneration, which influences repeatability of MS measures. Volumetric measures from customized grids can improve reliability of both global and regional sensitivity assessment. Nevertheless, inherent test-retest variation of individual points needs to be taken into account when assessing potential functional decline and/or disease progression.

Choroideremia (CHM) is an X-linked disorder defined by progressive degeneration of the neuroretina, the retinal pigment epithelium (RPE), and the choroid. Choroideremia is caused by loss-of-function mutations in the CHM gene,1 which encodes Rab escort protein 1 (REP1), a protein involved in prenylation of Rabs. Clinically, affected males experience night blindness in early adulthood followed by progressive peripheral visual field loss, with central vision preserved until later in life. Current investigational therapeutic approaches aim to modify the natural history of the disease through viral-mediated gene transfer.2 Visual acuity remains the most widely used outcome measure in these trials, providing functional assessment of foveal integrity; however, psychophysical tests have revealed significant deficits in macular function of CHM patients, including those with normal visual acuity.3 Therefore, more sensitive clinical outcome measures are required to determine early efficacy of therapeutic interventions in CHM. 
In recent years, fundus-driven perimetry, also known as microperimetry, has emerged as a robust method for assessing visual function in patients with macular disease.4 Its comparative advantage to conventional perimetry stems from real-time fundus imaging combined with eye-tracking technology. Fundus viewing allows “locking” of stimuli at predefined retinal locations, which enables structural-functional correlations to be explored while adapting perimetry grids to individual retinal morphology. Eye-tracking provides high-accuracy functional measures, even in cases of unstable or extrafoveal fixation,5 such as ABCA4-associated retinopathies6 and age-related macular degeneration (AMD).7 
However, recent studies examining the repeatability of perimetric sensitivity measures have suggested a high degree of test-retest variability at the border of pathologic changes8,9 and the edge of deep scotomas.10 For CHM and other conditions with centripetally advancing degeneration, these findings may hamper reliability of microperimetry testing especially during later disease stages, when macular involvement is noted. At these stages, which have been shown to provide the best discriminatory power to determine a potential treatment benefit,1113 mapping macular sensitivity becomes particularly challenging due to the highly variable retinal morphology. Most standard perimetry grids fail to sufficiently sample all areas of surviving retina within the macular region. The use of personalized grids could potentially improve sensitivity mapping, but their adoption in clinical trials is currently limited due to the difficulty in comparing examinations acquired with different grid arrangements. Recently though, interpolation methods have enabled the generation of indices that surpass these limitations.14 
In studies described herein, we sought to investigate test-retest repeatability of microperimetry testing in CHM, especially for disease stages targeted by current investigational trials. Our focus expanded beyond conventional measures from standard grids to include repeatability of indices extracted from interpolation of personalized grids. Determining test-retest limits of such indices will allow the adoption of a single estimate for all microperimetry examinations used in current and future CHM clinical trials. 
Methods
Subjects
The study population consisted of 28 eyes of 15 CHM subjects (age 39.1 ± 11.7 years [mean ± SD]; range, 21–65 years). All subjects had genetic or molecular confirmation of their diagnosis (Table 1). Inclusion criteria consisted of best-corrected visual acuity (BCVA) better than or equal to 20/62 (0.50 logMAR), stability of fixation determined with microperimetry (see the Microperimetry Examination section), and presence of active degeneration within the clinical macula determined with spectal-domain optical coherence tomography (SD-OCT) (Heidelberg Engineering, Heidelberg, Germany). None of the subjects had concurrent ocular disease that could affect visual performance. Normal data for microperimetry were collected from seven eyes of seven healthy subjects (age 35.3 ± 5.1 years [mean ± SD]; range, 25–40 years). All procedures conformed to the Code of Ethics of the World Medical Association (Declaration of Helsinki) and were done with the understanding and written consent of each participant. 
Table 1
 
Characteristics of CHM Subjects
Table 1
 
Characteristics of CHM Subjects
Microperimetry Examination
Subjects were assessed with the MAIA microperimeter (Macular Integrity Assessment; CenterVue, Padova, Italy). This instrument integrates scanning laser ophthalmoscopy (SLO) and real-time eye-tracking with computerized perimetry testing. A near-infrared (NIR) superluminescent diode (850 nm, 1024 × 1024 pixel resolution, 36° field of view) is used to visualize the fundus, with eye-tracking performed at a rate of 25 frames per second using the entire fundus as a reference. To obtain sensitivity thresholds, Goldmann-type size III stimuli (duration: 200 ms) were presented against a background 1.27 cd/m2, using a 4-2 staircase strategy. Minimum and maximum stimulus luminance achieved was 0 and 318 cd/m2, respectively, covering a dynamic range of 36 dB. Reliability was evaluated by the frequency of responses to 10-dB stimuli at the physiological blind spot (false positives). Any examination with greater than 25% false-positive responses was discarded and repeated. Fixation stability was assessed using the MAIA P1 fixation stability index, which measures the proportion of fixation points located within a 2° diameter circle centered on the fovea. Stable fixation was defined by P1 values greater 75%. 
Retinal Imaging
Before study enrollment, all CHM subjects had undergone central 30° × 30° blue laser fundus autofluorescence (λ = 488 nm) imaging (FAF) and SD-OCT line scans using the Spectralis SD-OCT unit (Heidelberg Engineering). For FAF, automatic real time was set to at least 30 frames. Spectral-domain OCT volume scans were acquired using a setting of 37 B-scans covering a 30° × 15° area, with high-resolution mode set on and 15 frames averaged per B-scan. Fundus autofluorescence imaging delineated the area of remaining central RPE tissue, which appears hyperfluorescent, from the atrophic, nonfluorescent and degenerated retina (Fig. 1C). The margin of RPE atrophy was usually well defined and easy to identify. For purposes of this study, the border of degeneration was defined as the area within 1° from the margin of RPE atrophy in all directions (Fig. 1C; inset). Fundus autofluorescence images were manually aligned to the NIR-SLO image acquired with the MAIA microperimeter (Fig. 1D), using a free image-editing software (GNU Image Manipulation Program, GIMP version 2.8.14; available in the public domain: http://www.gimp.org/). Optic disc and retinal vessels were used as anatomic landmarks. 
Figure 1
 
Microperimetry examination and FAF image alignment. (A) Fundus photograph in a CHM subject (P2) showing diffuse chorioretinal atrophy at the posterior pole with preservation of the central macula. (B) Microperimetry examination in subject P2 using the MAIA standard macular grid. The grid consists of 37 stimuli covering the central 10 degrees of vision. Infrared (IR) imaging allows real-time grid projection onto the fundus. A color decibel scale is provided showing the stimulus intensity range (0–36 dB). Black dots correspond to points not seen at 0 dB. (C) Corresponding 30° blue laser FAF image. White arrows highlight two separate zones: (i) the hyperfluorescent area of remaining central RPE tissue; and (ii) the atrophic, nonfluorescent, degenerated retina. The border of degeneration between those two zones is outlined with a 2°-wide segmented line (yellow) congruous to the RPE margin, with 1° spanning into zone (i) and 1° into zone (ii) (inset). (D) Manual alignment of the IR and FAF images using commercial image-editing software.
Figure 1
 
Microperimetry examination and FAF image alignment. (A) Fundus photograph in a CHM subject (P2) showing diffuse chorioretinal atrophy at the posterior pole with preservation of the central macula. (B) Microperimetry examination in subject P2 using the MAIA standard macular grid. The grid consists of 37 stimuli covering the central 10 degrees of vision. Infrared (IR) imaging allows real-time grid projection onto the fundus. A color decibel scale is provided showing the stimulus intensity range (0–36 dB). Black dots correspond to points not seen at 0 dB. (C) Corresponding 30° blue laser FAF image. White arrows highlight two separate zones: (i) the hyperfluorescent area of remaining central RPE tissue; and (ii) the atrophic, nonfluorescent, degenerated retina. The border of degeneration between those two zones is outlined with a 2°-wide segmented line (yellow) congruous to the RPE margin, with 1° spanning into zone (i) and 1° into zone (ii) (inset). (D) Manual alignment of the IR and FAF images using commercial image-editing software.
Testing Protocol
Testing was performed by a single experienced examiner (ISD) in a dimly lit room with nondilated pupils and according to the manufacturer's instructions. A small, red circle of 1° diameter was used as a fixation target. Subjects were seen at two separate visits, within a 1-week interval. One eye was tested at each visit, patching the other eye. Subjects were instructed and given a training test to ensure they were confident and reliable in the use of the microperimeter. 
At each visit, subjects underwent a sequence of four microperimetry tests. First, a standard 37-stimuli grid pattern was used, subtending 10 degrees of central visual field. The grid consisted of a single foveal response and three concentric rings of retinal loci distanced 1°, 3°, and 5° from the fovea (Figs. 2A–D). Following a short resting period (mean ± SD; 5 ± 1 minutes), the “follow-up” protocol was used to obtain a repeat of the first standard grid examination. The “follow-up” protocol ensured automatic alignment of the infrared fundus images between the two examinations. After completion of the second standard examination, CHM subjects were given the opportunity to rest (mean ± SD; 10 ± 2 minutes) and customized/personalized grids were created following a two-step process. First, an NIR-SLO image of the fundus was captured and used as a reference for positioning a square grid of 121 equidistant stimuli at the fovea or the central area of surviving retina. Depending on the area of the latter, a sampling density of 1° or 2° was chosen (covering 10 or 20 degrees of retina, respectively). Next, test points within degenerated areas of the retina were removed using FAF images as guidance. Test points one or two rows (equivalent to 1° to 2°) from the border of the surviving retina were preserved. Examples of customized grids generated with this method are provided in Figures 2E through 2H, with corresponding FAF images shown in insets. In some cases, additional test points were manually added at nonsampled areas of the surviving retina confirmed by FAF. For instance, FAF imaging in subject P11 demonstrated an extrafoveal RPE island in the superior outer macular region that remained unmapped with the standard 37-stimuli grid (Figs. 2D, 2H). The sampling density for additional tests points was either 1° or 2°, depending on the area sampled. 
Figure 2
 
Generation of customized/personalized grids in CHM. (AD) Representative examples of MAIA 37-stimuli grid examinations in four CHM subjects. The grid consists of a single foveal response and three concentric rings of retinal loci distanced 1°, 3°, and 5° from the fovea. (EH) Customized/personalized grids corresponding to the same CHM subjects. A square grid of 121 equidistant stimuli is centered at the fovea and then modified by removing excess test points within the degenerated retina, using FAF images as guidance (insets). Additional test points are added at non- or undersampled areas of the surviving retina, as shown in (H) for subject P11.
Figure 2
 
Generation of customized/personalized grids in CHM. (AD) Representative examples of MAIA 37-stimuli grid examinations in four CHM subjects. The grid consists of a single foveal response and three concentric rings of retinal loci distanced 1°, 3°, and 5° from the fovea. (EH) Customized/personalized grids corresponding to the same CHM subjects. A square grid of 121 equidistant stimuli is centered at the fovea and then modified by removing excess test points within the degenerated retina, using FAF images as guidance (insets). Additional test points are added at non- or undersampled areas of the surviving retina, as shown in (H) for subject P11.
To limit patient fatigue, caution was given not to exceed a total number of 68 stimuli at any customized grid, which is the number of test points used in the 10-2 program of the Humphrey automated perimeter. Customized grid examinations were repeated after a final short resting period (mean ± SD; 5 ± 2 minutes) using the “follow-up” option. Control subjects underwent the same testing protocol, but with a 10-2 grid pattern of 68 stimuli for examinations three and four (similar to the Humphrey 10-2 test). 
Interpolated Surface Analysis of Microperimetry Data
Microperimetry data were first exported as xyz triplets, with z representing the sensitivity of each tested point and x, y its Cartesian coordinates with respect to fixation. An interpolated two-dimensional matrix surface was then generated from xyz microperimetry data of customized examinations using the thin plate spline (TPS) function. The TPS algorithm appends a thin, elastic “spline” (or surface) through all data points, which is then deformed to find the best fit that minimizes the bending energy function.15 Two-dimensional integration can then be applied to calculate the volume between any given z-plane and the matrix surface using a numerical integral method. In our case, the central macular region was considered a relatively “flat,” continuous surface z = f(x,y),(x,y) ∈ σ, with σ being the domain of interpolated points. Therefore, the volume underneath the macular surface was computed as:  where m, n are the number of rows and columns of the interpolated matrix. Volumetric units are reported in dB•deg2.  
All interpolation and volumetric analyses were performed using OriginPro 2015 (OriginLab Corp., Northampton, MA, USA). Color-scaled contour plots and three-dimensional rendering of the interpolated surface were used for visualization. Custom-sized region-of-interest (ROI) rectangles were created to perform integration within a specific ROI of the contour plot. The two-dimensional integration gadget (OriginPro 2015) was used to define the position and size of the ROI rectangle by entering the desired visual field XY coordinates for the left/right and top/bottom lines, respectively. Region-of-interest coordinates corresponded to the opposite of retinal XY coordinates in NIR-SLO images. The ROI rectangle was prevented from moving in volume integration of repeat examinations. 
Statistical Analysis
Dense scotomas were assigned a value of −1 dB so as to be included in the calculation of mean sensitivity (MS). The distribution of sensitivity values was substantially positively skewed by the large number of loci with dense scotomas. For that reason, the Tukey's trimean sensitivity (TmS)16 was also computed as a more appropriate measure of central tendency; TmS is a weighted average of the median sensitivity (Smedian) and top and bottom quartile values (Q1 = 25th percentile and Q3 = 75th percentile), defined by the following equation:    
For point-wise analysis (PWS), all points recorded as not seen (−1 dB) or having a sensitivity value of 0 dB in both microperimetry tests were excluded to minimize floor effects. No spatial averaging was applied to the raw data. Changes in MS, TmS and volumetric measures between repeated examinations were explored using the Wilcoxon signed-rank test. For PWS, a linear mixed effects model was constructed (SPSS; IBM, Armonk, New York, NY, USA), with test number used as a fixed effect and stimulus points nested within subjects as a random effect to account for multiple measurements from same subjects and intereye correlations.7 Bland-Altman plots were used to evaluate test-retest agreement of all parameters. For those parameters that did not exhibit a significant test-retest change, the coefficient of repeatability (CoR) was computed using the following equation17:  where Sw was the within-subject SD.  
Differences in estimated CoR between controls and CHM patients were investigated using the Mann-Whitney U test and among different regions using the Kruskal-Wallis test. The relationship of Sw with patient-specific characteristics was explored using Spearman correlations. The root mean square coefficient of variation (CV)18 was used to compare reliability of volumetric against conventional sensitivity measures. Significance was set at P < 0.05. All statistical analyses were performed using GraphPad Prism 6 (GraphPad Software, La Jolla, CA, USA). 
Results
Repeatability of Global Sensitivity Measures
We first examined trend changes in test-retest measures of MS from standard grid examinations to exclude a learning effect. No significant test-retest differences were noted for either control (P = 0.24) or choroideremia eyes (P = 0.81). Bland-Altman plots were used to illustrate the agreement between the tests (Figs. 3A, 3B). The CoR for MS was estimated to be ±1.51 dB and ±1.33 dB for control and choroideremia eyes, respectively. There was no significant difference between the two groups of eyes (Fig. 3C; 2-tailed Mann-Whitney U test; P = 0.85). However, the CV was 5-fold greater in CHM (0.09) compared with control eyes (0.02) (Fig. 3D). Similar results were obtained for the TmS, with CV values of 0.08 and 0.02, respectively (Supplementary Fig. S1B). 
Figure 3
 
Repeatability of MS and relationship with patient-specific characteristics. (A, B) Bland-Altman plots illustrating test-retest agreement for controls (A) and CHM subjects (B). (C) Comparison of MS CoR between CHM and control eyes. No significant difference was observed (±1.33 dB and ±1.51 dB, respectively). (D) Root mean square CV for MS is significantly higher in CHM compared with control eyes (0.09 vs. 0.02, P = 0.003). (E) Correlation of within-subject SD (Sw) with MS. A positive correlation can be noted (r2 = 0.32, P = 0.007). Choroideremia Sw values do not exceed at any point those in controls (gray-shaded area). (F) Correlation of within-subject CV (CVw) and FAF area. Although no correlation can be noted for the range of FAF areas tested, CVw values were significantly higher than expected from controls (gray-shaded area) for FAF areas <5 mm2. Asterisk denotes significance at P < 0.05. Error bars represent 95% confidence intervals.
Figure 3
 
Repeatability of MS and relationship with patient-specific characteristics. (A, B) Bland-Altman plots illustrating test-retest agreement for controls (A) and CHM subjects (B). (C) Comparison of MS CoR between CHM and control eyes. No significant difference was observed (±1.33 dB and ±1.51 dB, respectively). (D) Root mean square CV for MS is significantly higher in CHM compared with control eyes (0.09 vs. 0.02, P = 0.003). (E) Correlation of within-subject SD (Sw) with MS. A positive correlation can be noted (r2 = 0.32, P = 0.007). Choroideremia Sw values do not exceed at any point those in controls (gray-shaded area). (F) Correlation of within-subject CV (CVw) and FAF area. Although no correlation can be noted for the range of FAF areas tested, CVw values were significantly higher than expected from controls (gray-shaded area) for FAF areas <5 mm2. Asterisk denotes significance at P < 0.05. Error bars represent 95% confidence intervals.
Relationship With Patient-Specific Characteristics
To explore any dependence of test-retest variability of global measures on patient-specific characteristics, we correlated individual test-retest SD (Sw) with the magnitude of sensitivity and the area of residual RPE on FAF. As shown in Figure 3E, Sw was proportional to the magnitude of MS, without ever exceeding control values (r2 = 0.32, P = 0.007). Because of that observation, we converted Sw to within-subject CV (CVw) and investigated its dependence to FAF area (Fig. 3F). For most CHM patients with areas less than 5 mm2, higher CVw values (0.05–0.10) were noted compared with controls (0.02); CVw values approximated those from controls in patients with larger areas. A similar trend was observed for TmS (Supplementary Fig. S1C). 
Changes in PWS
Next, we sought to determine CoR for PWS. Using a linear mixed effects model, we found that PWS from all subjects remained unchanged between the two tests, excluding a learning effect. Bland-Altman plots were constructed for control and choroideremia eyes (Figs. 4A, 4B) and PWS CoR calculated to be ±4.5 and ±8.7 dB, respectively. These CoRs were significantly different between the two groups of eyes (P < 0.001; Fig. 4C). 
Figure 4
 
Repeatability of PWS. (A, B) Bland-Altman plots illustrating test-retest agreement for controls (A) and CHM subjects (B). (C) Comparison of PWS CoR between CHM and control eyes. Point-wise sensitivity CoRs were significantly higher in CHM eyes (±8.7 vs. ±4.5 dB; P < 0.001). Asterisk denotes significance at P < 0.05. Error bars represent 95% confidence intervals.
Figure 4
 
Repeatability of PWS. (A, B) Bland-Altman plots illustrating test-retest agreement for controls (A) and CHM subjects (B). (C) Comparison of PWS CoR between CHM and control eyes. Point-wise sensitivity CoRs were significantly higher in CHM eyes (±8.7 vs. ±4.5 dB; P < 0.001). Asterisk denotes significance at P < 0.05. Error bars represent 95% confidence intervals.
Sensitivity Changes of Loci at the Border and Central Retina
To further investigate the origin of the high PWS variability in choroideremia eyes, we aligned and registered microperimetry SLO images to FAF images and determined the precise location of each tested sensitivity locus relative to the margins of the residual RPE island. A total of 441 points were analyzed and subdivided into three distinct groups: n = 192 points at the border of degeneration or within 1° from the margin of RPE atrophy (in all directions), n = 160 points within the central “healthier” retina, and n = 89 points within the degenerated retina (Fig. 5A). After excluding a systematic test-retest change in average PWS for all of the three groups of points, the PWS CoRs were computed to be, on average, ±10.68 dB, ±4.74 dB, and ±4.78 dB for points at the border, central, and degenerated retina, respectively. Points close to the border of degeneration had significantly greater CoR compared with points at the central or degenerated retina (P < 0.001; Fig. 5E). 
Figure 5
 
Point-wise sensitivity repeatability at different regions. (A) Representative segregation of tested points into three groups based on their relationship with the border of degeneration (highlighted in dark green): blue loci (with the blue square symbol) correspond to test points within the degenerated retina; green loci (with the green square symbol) correspond to test points within the central, ``healthier'' retina; and red loci (with the red square symbol) correspond to test points at the border of atrophy. (BD) Bland-Altman plots illustrating PWS test-retest agreement for border loci (B), central loci (C), and degenerated loci (D). (E) One-way ANOVA comparing CoR from the three regions. Coefficient of repeatability for loci at the border were significantly higher compared with the other two regions (10.68 vs. 4.74 dB, P < 0.001). Asterisk denotes significance at P < 0.05. Error bars represent 95% confidence intervals.
Figure 5
 
Point-wise sensitivity repeatability at different regions. (A) Representative segregation of tested points into three groups based on their relationship with the border of degeneration (highlighted in dark green): blue loci (with the blue square symbol) correspond to test points within the degenerated retina; green loci (with the green square symbol) correspond to test points within the central, ``healthier'' retina; and red loci (with the red square symbol) correspond to test points at the border of atrophy. (BD) Bland-Altman plots illustrating PWS test-retest agreement for border loci (B), central loci (C), and degenerated loci (D). (E) One-way ANOVA comparing CoR from the three regions. Coefficient of repeatability for loci at the border were significantly higher compared with the other two regions (10.68 vs. 4.74 dB, P < 0.001). Asterisk denotes significance at P < 0.05. Error bars represent 95% confidence intervals.
Repeatability of Interpolated Surface Sensitivity Indices
Can the reproducibility of global sensitivity measures be improved with the use of customized microperimetry grids in CHM? To investigate this, one should first create visual function indices that are independent of the sampling density of test points. Such indices can be derived from interpolated surface maps of perimetry data as illustrated in Figure 6. Subject P6 underwent consecutive examinations using four different stimulus arrangements: a standard 37-stimuli grid, a 10-2 grid of 68 stimuli, an equally spaced (1°) grid of 121 stimuli, and a customized grid of 40 stimuli. Although MS values were noncomparable (4.4, 1.2, 0.2, and 5.6 dB), even when including seeing points only (14.16, 12.14, 7.75, and 12.89 dB), measurement of the volume underneath the interpolated surface yielded consistent measures across the three grids, with values of 184.18, 177.15, 182.65, and 186.23 dB•deg2, respectively. 
Figure 6
 
Repeatability of volumetric measures from different stimulus grids. Representative examples of microperimetry examinations and corresponding interpolated contour plots generated from the same subject (P6) using a standard 37-stimuli macular grid (A); a 10-2 grid of 68 stimuli (B); an equally spaced (1°) custom grid of 121 stimuli (C); and a customized grid of 40 stimuli (D). Sensitivity levels have been color-coded at 4-dB intervals. Black dots correspond to points not seen at 0 dB. Although MS values are noncomparable, volumetric measures yield consistent measurements across the three grids.
Figure 6
 
Repeatability of volumetric measures from different stimulus grids. Representative examples of microperimetry examinations and corresponding interpolated contour plots generated from the same subject (P6) using a standard 37-stimuli macular grid (A); a 10-2 grid of 68 stimuli (B); an equally spaced (1°) custom grid of 121 stimuli (C); and a customized grid of 40 stimuli (D). Sensitivity levels have been color-coded at 4-dB intervals. Black dots correspond to points not seen at 0 dB. Although MS values are noncomparable, volumetric measures yield consistent measurements across the three grids.
For that reason, we relied on volumetric measures of interpolated data to extract a single estimate of test-retest repeatability for customized MP examinations. Table 2 summarizes the parameters used for each patient's customized grid. The number of test points was on average 57.0 ± 9.1 (mean ± SD), with the area of retina covered ranging from 6° to 20°. Time to complete a customized examination was not statistically different from the time required for control subjects to undertake a 10-2 68-loci examination (8:12 minutes ± 58 seconds vs. 8:52 minutes ± 32 seconds, P = 0.11). After excluding a significant test-retest change in volumetric measures (P = 0.67), the CoR for customized grids was computed to be ±62.15 dB•deg2; this translates into an overall improved CV value (0.05) than that obtained from standard-grid MS measures (0.09), as shown in Figure 7D. Nevertheless, volumetric measures from interpolated 10-2 grid examinations in control subjects continued to demonstrate lower overall test-retest error (CoR = ±249.50 dB•deg2 and CV = 0.02). 
Table 2
 
Customized Grid Parameters
Table 2
 
Customized Grid Parameters
Figure 7
 
Interpolated surface analysis of customized grids. (A) Representative example of a CHM microperimetry examination using a customized grid. (B) Corresponding three-dimensional rendering of the interpolated surface generated from the customized examination in (A). The x- and y-axes of the Cartesian coordinate system represent degrees of visual field (deg) and the z-axis the measured sensitivity threshold in decibels. Numerical integration of the volume underneath the interpolated surface was used to compute volumetric indices. (C) Bland-Altman plots illustrating test-retest agreement for total volume calculation. (D) Comparison of repeatability between MS from 37-stimuli grid examinations and total volume from customized grids. Total volume yields significantly lower CV values than MS (0.05 vs. 0.09, P = 0.03).
Figure 7
 
Interpolated surface analysis of customized grids. (A) Representative example of a CHM microperimetry examination using a customized grid. (B) Corresponding three-dimensional rendering of the interpolated surface generated from the customized examination in (A). The x- and y-axes of the Cartesian coordinate system represent degrees of visual field (deg) and the z-axis the measured sensitivity threshold in decibels. Numerical integration of the volume underneath the interpolated surface was used to compute volumetric indices. (C) Bland-Altman plots illustrating test-retest agreement for total volume calculation. (D) Comparison of repeatability between MS from 37-stimuli grid examinations and total volume from customized grids. Total volume yields significantly lower CV values than MS (0.05 vs. 0.09, P = 0.03).
Slightly lower reproducibility was encountered in the construction of isopter contours from customized grids, which are lines defining borders of regions with equal retinal sensitivity (Fig. 8A). Isopter area calculation at three threshold levels (20, 16, and 12 dB) yielded CoR of ±2.70 deg2, irrespective of isopter analyzed (Figs. 8B, 8C). This translates into a CV value of 0.07 for the 16-dB and 12-dB isopters, but in a considerably low reliability index for the 20-dB isopter (0.27). These findings highlight that certain measures in customized grids continue to remain susceptible to border variation. 
Figure 8
 
Repeatability in calculation of isopter areas from customized grids. (A) Representative color-coded contour plot generated from the customized examination in Figure 7A. Isopter threshold levels have been color-coded at 4-dB intervals. (B) Bland-Altman plots illustrating test-retest agreement for isopter area calculation at three threshold levels: 20 dB, 16 dB, 12 dB. (C) Isopter area calculation yields similar CoR, irrespective of threshold level.
Figure 8
 
Repeatability in calculation of isopter areas from customized grids. (A) Representative color-coded contour plot generated from the customized examination in Figure 7A. Isopter threshold levels have been color-coded at 4-dB intervals. (B) Bland-Altman plots illustrating test-retest agreement for isopter area calculation at three threshold levels: 20 dB, 16 dB, 12 dB. (C) Isopter area calculation yields similar CoR, irrespective of threshold level.
A comparative advantage of volumetric indices over traditional global measures is the ability to perform ROI-based analyses of the interpolated surfaces. Based on our previous regional PWS analysis, we hypothesized that limiting sensitivity assessment in areas within the central “healthier” retina would minimize any potential variation from the transitional zone and increase reproducibility. We thus explored repeatability of ROI volume integration for areas well within the residual island of RPE. To achieve this, we first positioned an ROI rectangle within the central hyperfluorescent area of the FAF image, as outlined in Figure 9A (area within the red rectangle), and estimated its xy coordinates by adopting the scale of the coregistrered microperimetry SLO image (microperimetry data on the NIR-SLO image have known xy retinal coordinates). Next, the opposite values were applied as xy constraints on the two-dimensional interpolated contour plot to perform volume integration within the specified ROI. Central retina ROI analysis was able to yield significantly lower CoR (± 27.34 dB•deg2) and CV values that were comparable to controls (0.02). Table 3 summarizes repeatability for all volumetric indices used in this study. 
Figure 9
 
Repeatability in central retina ROI volume integration. (A) Fundus autofluorescence–guided positioning of a rectangle in a region well within the hyperfluorescent RPE residual tissue. The coordinates of the rectangle are estimated by adopting the scale of the coregistrered microperimetry SLO image. (B) Region-of-interest volume integration performed on the two-dimensional contour plot of an interpolated customized examination using the xy constraints estimated in (A). (C) Bland-Altman plots illustrating test-retest agreement in central retina ROI volume integration.
Figure 9
 
Repeatability in central retina ROI volume integration. (A) Fundus autofluorescence–guided positioning of a rectangle in a region well within the hyperfluorescent RPE residual tissue. The coordinates of the rectangle are estimated by adopting the scale of the coregistrered microperimetry SLO image. (B) Region-of-interest volume integration performed on the two-dimensional contour plot of an interpolated customized examination using the xy constraints estimated in (A). (C) Bland-Altman plots illustrating test-retest agreement in central retina ROI volume integration.
Table 3
 
Repeatability of Conventional and Volumetric Measures
Table 3
 
Repeatability of Conventional and Volumetric Measures
Discussion
Our study investigated the reproducibility of microperimetry testing in CHM and its potential to serve as a reliable outcome measure, in light of gene therapy clinical trials under way.19 The degenerative process in CHM resembles that of other inherited retinal disorders, such as retinitis pigmentosa, in that the chorioretinal atrophy is progressing in a centripetal fashion with foveal involvement occurring only at the very late stages of the disease.3 During the course of the disease, a transitional zone can be delineated between degenerated and relatively healthy retina, which is indicative of disease activity.20 Monitoring sensitivity changes at this zone could constitute an attractive functional marker of CHM progression. Our study demonstrated that reproducible microperimetry measures cannot be obtained for that region in CHM; points close to the border of degeneration had significantly greater CoR, compared with points at the central retina. Although our reported CoR for central retinal points (±4.74 dB) are comparable to that obtained in patients with ABCA4-associated retinopathies (±4.21 dB),6 juvenile retinoschisis (±5.4 dB),21 and other macular disorders (±5.6 dB),5 the issue of higher variability at the border of degeneration has been inconsistently reported by other studies. 
Wu et al.10 investigated the variability of microperimetric sensitivity measures at the border of the optic nerve head, in an effort to model deep scotomas, and compared it with other areas of normal retina. Coefficient of repeatability for points at the border of the optic nerve head were significantly higher than points at the macular region, in agreement with the higher CoR estimated in our study at the border of atrophy in CHM. Similar observations have been made for automated perimetry measures at the border of glaucomatous defects in glaucoma patients and the border of the blind spot in healthy subjects.8,9 However, a microperimetry study by Cideciyan et al.6 in patients with ABCA4-associated maculopathy suggested that a single estimate of test-retest repeatability can be adopted for all points, independent of their relative distance from a deep scotoma. Apart from the smaller dynamic stimulus range (0-20 dB) of the microperimeter (MP1; Nidek, Inc., Fremont, CA, USA) used in this study, reasons explaining the observed discrepancy pertain mainly to the noninclusion of retinal loci at the immediate boundaries of deep scotomas and the three-point spatial averaging applied to the microperimetry data. Wu et al. elegantly showed that by applying the same spatial averaging to their data, similar PWS CoR could be obtained across all regions. 
What could explain the higher test-retest variability observed at the border of degeneration? A plausible origin should be sought at the microsaccades occurring during the short stimulus presentation, which cannot be fully compensated with the current fundus tracking frequency (25 Hz). Therefore, loci at the retinal border may shift into the degenerated retina demonstrating greater intrasession variability. This interpretation is supported by the work of Wyatt et al.,9 who showed a substantial contribution of small fixational eye movements to test-retest variability by correlating the latter with the gradient, the rate at which sensitivity changes with location. Higher variability at the margin of RPE atrophy also can be attributed to the presence of degenerating photoreceptors, which have been shown to form outer retinal tubulations in CHM and other disorders.22 These structures may contain highly dysfunctional photoreceptors that exhibit inconsistency between responses. Although less likely in cases of photoreceptor loss, variability may also originate from dysfunctional ganglion cells subserving the transitional zone, as previously shown for glaucoma.9 
In later stages of CHM, the high variability of the residual retina's morphology warrants the use of individualized perimetric grids to sufficiently map visual sensitivity in every case. However, MS measures from these grids become dependent on the parameters of the applied grid pattern, such as test point number, spacing, and local condensation.14 Several methods have been used to interpolate perimetric test grids and generate indices that are not grid specific. These include neural network algorithms,23 TPS,14 and nearest neighbor interpolations.14 Our study adapted the TPS algorithm, previously shown by Weleber et al.14 to have good performance and accuracy for modeling full-field perimetry data. Volumetric indices were chosen to quantify the visual sensitivity of the customized interpolated grid.14 In our case, volume integration was performed in Cartesian rather than spherical coordinates, because of the small extension of the central macular region, which was considered a relatively “flat” surface. Customized grids were able to improve not only the efficiency but also the reliability of sensitivity mapping in CHM. Similar observations have been made for the use of individually condensed test grids in detection of glaucomatous visual field defects.24 Regional analysis of those locally condensed areas of interest can be performed with high levels of reproducibility in CHM, a capability particularly useful for evaluating the outcome of localized therapeutic interventions, such as subretinal gene delivery.2 
Nevertheless, there are certain limitations in our study that need to be considered. Apart from the small sample size, it is important to keep in mind that most CHM subjects were evaluated at late disease stages. Low reliability in global and PWS might not be encountered at early or intermediate stages of CHM. In addition, the estimation of test-retest variability was based on two repeat tests (within the same session) rather than three or four examinations across a short period of time. Further studies are needed to determine the short-term intersession variability in microperimetry testing that should be expected in CHM. Another aspect to consider is the “fixed” sequence of testing in our study's protocol, which may have partially influenced the reduced variability observed with the customized grid. Even though fatigue was expected to increase test-retest measurement error in customized examinations, the effect of a learning curve cannot be fully excluded. Finally, it is noteworthy to mention that the volumetric parameter, although providing complementary information not obtainable through traditional measures, is not easily applicable at this point in routine clinical settings, as there is no commercial software available to calculate this parameter. 
In summary, microperimetry testing in CHM is limited by the high test-retest variation at the border of chorioretinal atrophy. As a result, assessing functional decline and disease progression may become less reliable during later stages of the disease. Our study showed that customized grids could improve both reliability and efficiency of sensitivity mapping at these stages through the use of volumetric measures, which yield similar values regardless of stimulus array. All approaches discussed herein can be expanded to other forms of inherited retinal disorders, in which perimetric sensitivity assessment is used as an outcome measure. 
Acknowledgments
The authors thank Rafael C. Caruso, MD, at Princeton University for his insightful comments on this manuscript. 
Supported by Canadian Institutes of Health Research Grant 119190, Alberta Innovation Health Services Grant 201201139, and Canadian Foundation for Innovation Grant 28916. 
Disclosure: I.S. Dimopoulos, None; C. Tseng, None; I.M. MacDonald, None 
References
McTaggart KE, Tran M, Mah DY, Lai SW, Nesslinger NJ, MacDonald IM. Mutational analysis of patients with the diagnosis of choroideremia. Hum Mutat. 2002; 20: 189–196.
MacLaren RE, Groppe M, Barnard AR, et al. Retinal gene therapy in patients with choroideremia: initial findings from a phase 1/2 clinical trial. Lancet. 2014; 383: 1129–1137.
Jacobson SG, Cideciyan AV, Sumaroka A, et al. Remodeling of the human retina in choroideremia: rab escort protein 1 (REP-1) mutations. Invest Ophthalmol Vis Sci. 2006; 47: 4113–4120.
Acton JH, Greenstein VC. Fundus-driven perimetry (microperimetry) compared to conventional static automated perimetry: similarities, differences, and clinical applications. Can J Ophthalmol. 2013; 48: 358–363.
Chen FK, Patel PJ, Xing W, et al. Test-retest variability of microperimetry using the Nidek MP1 in patients with macular disease. Invest Ophthalmol Vis Sci. 2009; 50: 3464–3472.
Cideciyan AV, Swider M, Aleman TS, et al. Macular function in macular degenerations: repeatability of microperimetry as a potential outcome measure for ABCA4-associated retinopathy trials. Invest Ophthalmol Vis Sci. 2012; 53: 841–852.
Wu Z, Ayton LN, Guymer RH, Luu CD. Intrasession test-retest variability of microperimetry in age-related macular degeneration. Invest Ophthalmol Vis Sci. 2013; 54: 7378–7385.
Haefliger IO, Flammer J. Fluctuation of the differential light threshold at the border of absolute scotomas. Comparison between glaucomatous visual field defects and blind spots. Ophthalmology. 1991; 98: 1529–1532.
Wyatt HJ, Dul MW, Swanson WH. Variability of visual field measurements is correlated with the gradient of visual sensitivity. Vision Res. 2007; 47: 925–936.
Wu Z, Jung CJ, Ayton LN, Luu CD, Guymer RH. Test–retest repeatability of microperimetry at the border of deep scotomas. Invest Ophthalmol Vis Sci. 2015; 56: 2606.
Seitz IP, Zhour A, Kohl S, et al. Multimodal assessment of choroideremia patients defines pre-treatment characteristics. Graefe's Arch Clin Exp Ophthalmol. 2015; 253: 2143–2150.
Coussa RG, Kim J, Traboulsi EI. Choroideremia: effect of age on visual acuity in patients and female carriers. Ophthalmic Genet. 2012; 33: 66–73.
Freund PR, Sergeev YV, MacDonald IM. Analysis of a large choroideremia dataset does not suggest a preference for inclusion of certain genotypes in future trials of gene therapy. Molecular Genetics & Genomic Medicine. 2016; 4: 344–358.
Weleber RG, Smith TB, Peters D, et al. VFMA: topographic analysis of sensitivity data from full-field static perimetry. Transl Vis Sci Technol. 2015; 4: 14.
Wahba G. Spline Models for Observational Data. Philadelphia, PA: Society for Industrial and Applied Mathematics; 1990.
Tukey JW. Exploratory Data Analysis. Reading MA: Pearson City; 1977.
Bartlett JW, Frost C. Reliability repeatability and reproducibility: analysis of measurement errors in continuous variables. Ultrasound Obstet Gynecol. 2008; 31: 466–475.
Glüer C-C, Blake G, Lu Y, Blunt BA, Jergas M, Genant HK. Accurate assessment of precision errors: how to measure the reproducibility of bone densitometry techniques. Osteoporos Int. 1995; 5: 262–270.
Dimopoulos IS, Chan S, MacLaren RE, MacDonald IM. Pathogenic mechanisms and the prospect of gene therapy for choroideremia. Expert Opin Orphan Drugs. 2015; 3: 787–798.
Lazow MA, Hood DC, Ramachandran R, et al. Transition zones between healthy and diseased retina in choroideremia (CHM) and Stargardt disease (STGD) as compared to retinitis pigmentosa (RP). Invest Ophthalmol Vis Sci. 2011; 52: 9581–9590.
Jeffrey BG, Cukras CA, Vitale S, Turriff A, Bowles K, Sieving PA. Test-retest intervisit variability of functional and structural parameters in X-linked retinoschisis. Trans Vis Sci Tech. 2014; 3 (5): 5.
Zweifel SA, Engelbert M, Laud K, Margolis R, Spaide RF, Freund KB. Outer retinal tubulation: a novel optical coherence tomography finding. Arch Ophthalmol. 2009; 127: 1596–1602.
Henson DB. Perimetry Update 2002/2003. Kugler Publications. Amsterdam NY: Kugler City; 2004.
Schiefer U, Flad M, Stumpp F, et al. Increased detection rate of glaucomatous visual field damage with locally condensed grids: a comparison between fundus-oriented perimetry and conventional visual field examination. Arch Ophthalmol. 2003; 121: 458–465.
Figure 1
 
Microperimetry examination and FAF image alignment. (A) Fundus photograph in a CHM subject (P2) showing diffuse chorioretinal atrophy at the posterior pole with preservation of the central macula. (B) Microperimetry examination in subject P2 using the MAIA standard macular grid. The grid consists of 37 stimuli covering the central 10 degrees of vision. Infrared (IR) imaging allows real-time grid projection onto the fundus. A color decibel scale is provided showing the stimulus intensity range (0–36 dB). Black dots correspond to points not seen at 0 dB. (C) Corresponding 30° blue laser FAF image. White arrows highlight two separate zones: (i) the hyperfluorescent area of remaining central RPE tissue; and (ii) the atrophic, nonfluorescent, degenerated retina. The border of degeneration between those two zones is outlined with a 2°-wide segmented line (yellow) congruous to the RPE margin, with 1° spanning into zone (i) and 1° into zone (ii) (inset). (D) Manual alignment of the IR and FAF images using commercial image-editing software.
Figure 1
 
Microperimetry examination and FAF image alignment. (A) Fundus photograph in a CHM subject (P2) showing diffuse chorioretinal atrophy at the posterior pole with preservation of the central macula. (B) Microperimetry examination in subject P2 using the MAIA standard macular grid. The grid consists of 37 stimuli covering the central 10 degrees of vision. Infrared (IR) imaging allows real-time grid projection onto the fundus. A color decibel scale is provided showing the stimulus intensity range (0–36 dB). Black dots correspond to points not seen at 0 dB. (C) Corresponding 30° blue laser FAF image. White arrows highlight two separate zones: (i) the hyperfluorescent area of remaining central RPE tissue; and (ii) the atrophic, nonfluorescent, degenerated retina. The border of degeneration between those two zones is outlined with a 2°-wide segmented line (yellow) congruous to the RPE margin, with 1° spanning into zone (i) and 1° into zone (ii) (inset). (D) Manual alignment of the IR and FAF images using commercial image-editing software.
Figure 2
 
Generation of customized/personalized grids in CHM. (AD) Representative examples of MAIA 37-stimuli grid examinations in four CHM subjects. The grid consists of a single foveal response and three concentric rings of retinal loci distanced 1°, 3°, and 5° from the fovea. (EH) Customized/personalized grids corresponding to the same CHM subjects. A square grid of 121 equidistant stimuli is centered at the fovea and then modified by removing excess test points within the degenerated retina, using FAF images as guidance (insets). Additional test points are added at non- or undersampled areas of the surviving retina, as shown in (H) for subject P11.
Figure 2
 
Generation of customized/personalized grids in CHM. (AD) Representative examples of MAIA 37-stimuli grid examinations in four CHM subjects. The grid consists of a single foveal response and three concentric rings of retinal loci distanced 1°, 3°, and 5° from the fovea. (EH) Customized/personalized grids corresponding to the same CHM subjects. A square grid of 121 equidistant stimuli is centered at the fovea and then modified by removing excess test points within the degenerated retina, using FAF images as guidance (insets). Additional test points are added at non- or undersampled areas of the surviving retina, as shown in (H) for subject P11.
Figure 3
 
Repeatability of MS and relationship with patient-specific characteristics. (A, B) Bland-Altman plots illustrating test-retest agreement for controls (A) and CHM subjects (B). (C) Comparison of MS CoR between CHM and control eyes. No significant difference was observed (±1.33 dB and ±1.51 dB, respectively). (D) Root mean square CV for MS is significantly higher in CHM compared with control eyes (0.09 vs. 0.02, P = 0.003). (E) Correlation of within-subject SD (Sw) with MS. A positive correlation can be noted (r2 = 0.32, P = 0.007). Choroideremia Sw values do not exceed at any point those in controls (gray-shaded area). (F) Correlation of within-subject CV (CVw) and FAF area. Although no correlation can be noted for the range of FAF areas tested, CVw values were significantly higher than expected from controls (gray-shaded area) for FAF areas <5 mm2. Asterisk denotes significance at P < 0.05. Error bars represent 95% confidence intervals.
Figure 3
 
Repeatability of MS and relationship with patient-specific characteristics. (A, B) Bland-Altman plots illustrating test-retest agreement for controls (A) and CHM subjects (B). (C) Comparison of MS CoR between CHM and control eyes. No significant difference was observed (±1.33 dB and ±1.51 dB, respectively). (D) Root mean square CV for MS is significantly higher in CHM compared with control eyes (0.09 vs. 0.02, P = 0.003). (E) Correlation of within-subject SD (Sw) with MS. A positive correlation can be noted (r2 = 0.32, P = 0.007). Choroideremia Sw values do not exceed at any point those in controls (gray-shaded area). (F) Correlation of within-subject CV (CVw) and FAF area. Although no correlation can be noted for the range of FAF areas tested, CVw values were significantly higher than expected from controls (gray-shaded area) for FAF areas <5 mm2. Asterisk denotes significance at P < 0.05. Error bars represent 95% confidence intervals.
Figure 4
 
Repeatability of PWS. (A, B) Bland-Altman plots illustrating test-retest agreement for controls (A) and CHM subjects (B). (C) Comparison of PWS CoR between CHM and control eyes. Point-wise sensitivity CoRs were significantly higher in CHM eyes (±8.7 vs. ±4.5 dB; P < 0.001). Asterisk denotes significance at P < 0.05. Error bars represent 95% confidence intervals.
Figure 4
 
Repeatability of PWS. (A, B) Bland-Altman plots illustrating test-retest agreement for controls (A) and CHM subjects (B). (C) Comparison of PWS CoR between CHM and control eyes. Point-wise sensitivity CoRs were significantly higher in CHM eyes (±8.7 vs. ±4.5 dB; P < 0.001). Asterisk denotes significance at P < 0.05. Error bars represent 95% confidence intervals.
Figure 5
 
Point-wise sensitivity repeatability at different regions. (A) Representative segregation of tested points into three groups based on their relationship with the border of degeneration (highlighted in dark green): blue loci (with the blue square symbol) correspond to test points within the degenerated retina; green loci (with the green square symbol) correspond to test points within the central, ``healthier'' retina; and red loci (with the red square symbol) correspond to test points at the border of atrophy. (BD) Bland-Altman plots illustrating PWS test-retest agreement for border loci (B), central loci (C), and degenerated loci (D). (E) One-way ANOVA comparing CoR from the three regions. Coefficient of repeatability for loci at the border were significantly higher compared with the other two regions (10.68 vs. 4.74 dB, P < 0.001). Asterisk denotes significance at P < 0.05. Error bars represent 95% confidence intervals.
Figure 5
 
Point-wise sensitivity repeatability at different regions. (A) Representative segregation of tested points into three groups based on their relationship with the border of degeneration (highlighted in dark green): blue loci (with the blue square symbol) correspond to test points within the degenerated retina; green loci (with the green square symbol) correspond to test points within the central, ``healthier'' retina; and red loci (with the red square symbol) correspond to test points at the border of atrophy. (BD) Bland-Altman plots illustrating PWS test-retest agreement for border loci (B), central loci (C), and degenerated loci (D). (E) One-way ANOVA comparing CoR from the three regions. Coefficient of repeatability for loci at the border were significantly higher compared with the other two regions (10.68 vs. 4.74 dB, P < 0.001). Asterisk denotes significance at P < 0.05. Error bars represent 95% confidence intervals.
Figure 6
 
Repeatability of volumetric measures from different stimulus grids. Representative examples of microperimetry examinations and corresponding interpolated contour plots generated from the same subject (P6) using a standard 37-stimuli macular grid (A); a 10-2 grid of 68 stimuli (B); an equally spaced (1°) custom grid of 121 stimuli (C); and a customized grid of 40 stimuli (D). Sensitivity levels have been color-coded at 4-dB intervals. Black dots correspond to points not seen at 0 dB. Although MS values are noncomparable, volumetric measures yield consistent measurements across the three grids.
Figure 6
 
Repeatability of volumetric measures from different stimulus grids. Representative examples of microperimetry examinations and corresponding interpolated contour plots generated from the same subject (P6) using a standard 37-stimuli macular grid (A); a 10-2 grid of 68 stimuli (B); an equally spaced (1°) custom grid of 121 stimuli (C); and a customized grid of 40 stimuli (D). Sensitivity levels have been color-coded at 4-dB intervals. Black dots correspond to points not seen at 0 dB. Although MS values are noncomparable, volumetric measures yield consistent measurements across the three grids.
Figure 7
 
Interpolated surface analysis of customized grids. (A) Representative example of a CHM microperimetry examination using a customized grid. (B) Corresponding three-dimensional rendering of the interpolated surface generated from the customized examination in (A). The x- and y-axes of the Cartesian coordinate system represent degrees of visual field (deg) and the z-axis the measured sensitivity threshold in decibels. Numerical integration of the volume underneath the interpolated surface was used to compute volumetric indices. (C) Bland-Altman plots illustrating test-retest agreement for total volume calculation. (D) Comparison of repeatability between MS from 37-stimuli grid examinations and total volume from customized grids. Total volume yields significantly lower CV values than MS (0.05 vs. 0.09, P = 0.03).
Figure 7
 
Interpolated surface analysis of customized grids. (A) Representative example of a CHM microperimetry examination using a customized grid. (B) Corresponding three-dimensional rendering of the interpolated surface generated from the customized examination in (A). The x- and y-axes of the Cartesian coordinate system represent degrees of visual field (deg) and the z-axis the measured sensitivity threshold in decibels. Numerical integration of the volume underneath the interpolated surface was used to compute volumetric indices. (C) Bland-Altman plots illustrating test-retest agreement for total volume calculation. (D) Comparison of repeatability between MS from 37-stimuli grid examinations and total volume from customized grids. Total volume yields significantly lower CV values than MS (0.05 vs. 0.09, P = 0.03).
Figure 8
 
Repeatability in calculation of isopter areas from customized grids. (A) Representative color-coded contour plot generated from the customized examination in Figure 7A. Isopter threshold levels have been color-coded at 4-dB intervals. (B) Bland-Altman plots illustrating test-retest agreement for isopter area calculation at three threshold levels: 20 dB, 16 dB, 12 dB. (C) Isopter area calculation yields similar CoR, irrespective of threshold level.
Figure 8
 
Repeatability in calculation of isopter areas from customized grids. (A) Representative color-coded contour plot generated from the customized examination in Figure 7A. Isopter threshold levels have been color-coded at 4-dB intervals. (B) Bland-Altman plots illustrating test-retest agreement for isopter area calculation at three threshold levels: 20 dB, 16 dB, 12 dB. (C) Isopter area calculation yields similar CoR, irrespective of threshold level.
Figure 9
 
Repeatability in central retina ROI volume integration. (A) Fundus autofluorescence–guided positioning of a rectangle in a region well within the hyperfluorescent RPE residual tissue. The coordinates of the rectangle are estimated by adopting the scale of the coregistrered microperimetry SLO image. (B) Region-of-interest volume integration performed on the two-dimensional contour plot of an interpolated customized examination using the xy constraints estimated in (A). (C) Bland-Altman plots illustrating test-retest agreement in central retina ROI volume integration.
Figure 9
 
Repeatability in central retina ROI volume integration. (A) Fundus autofluorescence–guided positioning of a rectangle in a region well within the hyperfluorescent RPE residual tissue. The coordinates of the rectangle are estimated by adopting the scale of the coregistrered microperimetry SLO image. (B) Region-of-interest volume integration performed on the two-dimensional contour plot of an interpolated customized examination using the xy constraints estimated in (A). (C) Bland-Altman plots illustrating test-retest agreement in central retina ROI volume integration.
Table 1
 
Characteristics of CHM Subjects
Table 1
 
Characteristics of CHM Subjects
Table 2
 
Customized Grid Parameters
Table 2
 
Customized Grid Parameters
Table 3
 
Repeatability of Conventional and Volumetric Measures
Table 3
 
Repeatability of Conventional and Volumetric Measures
×
×

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.

×