MACULAR RETINAL LAYER THICKNESS IN CHILDHOOD SCOTT A. READ, PHD, MICHAEL J. COLLINS, PHD, STEPHEN J. VINCENT, PHD, DAVID ALONSO-CANEIRO, PHD Purpose: To examine macular retinal thickness and retinal layer thickness with spectral domain optical coherence tomography in a population of children with normal ocular health and minimal refractive errors. Methods: High-resolution macular optical coherence tomography scans from 196 children aged 4 years to 12 years (mean age: 8 ± 2 years), were analyzed to determine total retinal thickness and thickness of 6 different retinal layers across the central 5 mm of the posterior pole. Automated segmentation with manual correction was used to derive retinal thickness values. Results: The mean total retinal thickness in the central 1-mm foveal zone was 255 ± 16 mm, and this increased significantly with age (mean increase of 1.8 mm per year) in childhood (P , 0.001). Age-related increases in thickness of some retinal layers were also observed, with changes of the highest statistical significance found in the outer retinal layers in the central foveal region (P , 0.01). Significant topographical variations in thickness of each of the retinal layers were also observed (P , 0.001). Conclusion: Small magnitude, statistically significant increases in total retinal thickness and retinal layer thickness occur from early childhood to adolescence. The most prominent changes seem to occur in the outer retinal layers of the central fovea. RETINA 35:1223–1233, 2015

H

earlier than the foveal retina.3 Although pronounced changes in many retinal characteristics take place in the first few years of life, it has been suggested that aspects of human foveal development may not reach adult levels until the early teenage years.4 Optical coherence tomography (OCT) technology5,6 provides high-resolution, quantitative in vivo retinal morphologic information, allowing the examination of normal developmental retinal changes of human children in vivo. A number of previous studies have examined the total retinal thickness of the macular region in healthy pediatric populations using lowerresolution time domain OCT.7–11 These studies have demonstrated significant topographic variations in total retinal thickness across the posterior pole,7–11 alterations in retinal morphology associated with refractive error,8,9 and some evidence of small increases in total retinal thickness with age in childhood.10,11 More recently, a few studies have used higher-resolution spectral domain OCT (SD-OCT) methods to assess macular retinal thickness in healthy children, primarily to establish normative values of total retinal thickness in pediatric populations.12–14 Two of these studies have documented a significant

istologic studies indicate that the development of the normal human retina is not complete at birth, and that a number of changes in retinal morphology occur postnatally.1–4 Substantial postnatal changes in morphology of the foveal retina have been documented, including a central migration of cone photoreceptors, a thinning and elongation of photoreceptor inner and outer segments and axons, and substantial increases in foveal cone packing density.1,2,4 Developmental changes in aspects of retinal morphology in parafoveal and midperipheral retinal regions have also been documented; however, the maturation of midperipheral regions is thought to occur earlier than that of parafoveal regions, which in turn is thought to occur From the Contact Lens and Visual Optics Laboratory, School of Optometry and Vision Science, Queensland University of Technology, Brisbane, Queensland, Australia. Supported by an Australian Research Council “Discovery Early Career Researcher Award” to S. A. Read (DE 120101434). None of the authors have any financial/conflicting interests to disclose. Reprint requests: Scott A. Read, PhD, Contact Lens and Visual Optics Laboratory, School of Optometry and Vision Science, Queensland University of Technology, Room D517, O Block, Victoria Park Road, Kelvin Grove 4059, Brisbane, Queensland, Australia; e-mail: [email protected]

1223

1224 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES

positive association between foveal retinal thickness and age in childhood.13,14 Although the assessment of individual retinal layers is possible with SD-OCT, only one of these recent studies has examined the individual macular retinal layer thicknesses in children.14 Yanni et al14 reported normative ranges for a variety of retinal layer thickness parameters in 83 children aged 5 years to 15 years, based on a horizontal line scan through the fovea; however, the association between age and retinal layer thicknesses in childhood was not examined in detail. A few recent studies have used hand-held SD-OCT imaging to document in vivo developmental changes in the foveal region of premature infants.15–17 These studies have documented quantitative15,17 and qualitative16 changes in the retinal layers (particularly the outer retina at the foveal region) in infants, confirming histologic evidence of the maturation of the foveal retina postnatally. These studies have included only small numbers of children over the age of 4 years (i.e., ,20 children), therefore it is not clear if continued development and maturation of the in vivo retinal structure continues throughout childhood. Given that numerous ocular disorders influencing the retina can begin in childhood,18 and the increasing clinical use of SD-OCT, a comprehensive understanding of the normal in vivo retinal characteristics (including intraretinal layer thickness) and their developmental changes in childhood with SD-OCT is important for the clinical diagnosis and monitoring of the retina in pediatric patients. This study therefore aimed to examine the macular total retinal thickness and retinal layer thickness in a substantial population of children with normal ocular health, using highresolution SD-OCT, and to document any changes in retinal morphology associated with age. Methods Subjects and Procedures All children (n = 735) who were enrolled at an elementary school situated 5 miles from the central business district of Brisbane, Australia, were invited to participate in this cross-sectional study. Two hundred and twenty-six of these children aged 4 years to 12 years participated. Approval from the Queensland University of Technology Human Research Ethics Committee was obtained before the study, and written informed consent was provided by all participating children and their parents. Any children who were too young to give written consent provided verbal assent to participate. All participants were treated in accordance with the tenets of the declaration of Helsinki.



2015  VOLUME 35  NUMBER 6

All participating children underwent a series of vision screening tests to determine their visual acuity, ocular refraction, biometry, and ocular health status. These tests included measures of unaided and bestcorrected monocular visual acuity, noncycloplegic distance retinoscopy and subjective refraction, binocular vision measures, color vision screening, Lenstar LS 900 ocular biometry (Haag Streit AG, Koeniz, Switzerland), and slit-lamp biomicroscopy. One hundred and ninety-six of the 226 participating children met our inclusion criteria of having spherical equivalent refraction between +1.25 diopters (D) and −1.25 D, visual acuity of 0.1 logMAR or better, no evidence of strabismus, no history of ocular disease, surgery, or injury, and no history of systemic disease or medication with known ocular effects, and successfully underwent detailed SD-OCT imaging to assess their macular retinal thickness. After each subject’s vision screening, retinal OCT imaging was performed using a high-resolution SDOCT instrument (Copernicus SOCT-HR; Optopol Technology SA, Zawiercie, Poland). This instrument uses a super-luminescent diode light source with a peak wavelength of 850 nm (bandwidth, 100 nm), providing OCT images with an axial resolution of 3 mm, a transverse resolution of 12 mm to 18 mm, and a scanning speed of 52,000 A-scans per second. This instrument has been shown to exhibit excellent reproducibility for retinal thickness and retinal layer thickness measures in healthy subjects.19 Initially for each subject, 3-dimensional OCT scans across a 6- · 6-mm grid (consisting of 128 B-scans each with 500 A-scans), were captured of both eyes to detect any retinal abnormalities and to confirm fixation reliability in the subsequent high-resolution scans. A series of high-resolution scans were then collected on 1 randomly selected eye of each subject. Figure 1 illustrates this scanning protocol along with representative images from 1 subject. For these images, the instrument’s “cross” scanning protocol was used, which captures multiple sets of two perpendicular OCT line scans centered on the fovea. For each measurement, a series of scans were collected including horizontal/vertical and 45°/135° “cross” scans, resulting in the acquisition of 3 series of 4 radial line scans each separated by 45°, all centered on the fovea (Figure 1A). Each of the line scans consisted of 30 B-scans, with 999 A-scans per B-scan. The two series of four radial line scans exhibiting the best image quality for each subject (as defined by the instruments’ quality index and evidence of the most central fixation [i.e., deepest foveal pit]), were saved for additional analysis. Only OCT scans with an average image quality index of .4 were included, as per the manufacturer instructions.

RETINAL THICKNESS IN CHILDHOOD  READ ET AL

1225

Fig. 1. An overview of the OCT scanning and image analysis protocol used in the study, with an example B-scan image (vertical scan) from a representative subject. A series of 4 radial OCT images centered on the fovea separated by 45° (A), each consisting of 30 individual B-scans (B) was captured for each subject. The individual B-scans from each scan line were then registered and averaged to create averaged B-scan images (C) for each of the four scan lines. An automated method (with manual correction) was then used to segment eight retinal boundaries in each of the averaged B-scans (D), including the RPE, ISe, external limiting membrane (ELM), boundary between the OPL/INL, boundary between the INL and IPL, boundary between the GCL and NFL, and inner limiting membrane (ILM). The segmentation was used to derive the mean of seven retinal thickness parameters (total retinal thickness, RPE to ISe thickness, IS thickness, the ONL + OPL thickness, the INL thickness, the IPL + GCL thickness, and NFL thickness), at the foveal point (red line), at the foveal zone (1), parafoveal zone (2), and perifoveal zone (3) in each averaged B-scan.

Data Analysis Custom written software was used to analyze the OCT data from each subject. Initially, each of the 30 individual raw B-scan images along each scan line was registered and aligned, to create averaged OCT images with reduced speckle noise and enhanced visibility of the individual retinal layers, using a method that has been described in detail previously.20 These averaged images were then analyzed to segment the boundaries of eight different retinal layers, using a method based on graph theory described by Chiu et al21 (with the preprocessing values modified based on the Copernicus SOCT-HR resolution). The boundaries that were segmented are illustrated in Figure 1D and included the outer boundary of the retinal pigment epithelium (RPE), the inner boundary of the inner segment ellipsoid band (ISe), the inner boundary of the external limiting membrane, the boundary between the outer plexiform layer and inner nuclear layer (OPL/INL), the boundary between the INL and the inner plexiform layer (INL/IPL), the boundary between the ganglion cell layer and the nerve fiber layer (GCL/NFL), and the inner boundary of the inner limiting membrane.

Although additional retinal layers (e.g., the boundary of the cone outer segment tips, and the outer nuclear layer [ONL] and the GCL) are visible in the SD-OCT scans, the 8 boundaries used in this study were chosen because they could be reliably/unambiguously detected across the entire central 5 mm in all of the scans. The automated segmentation of each of the average B-scan images from each subject was then checked by an experienced masked observer, who manually corrected any segmentation errors. Because the subjects exhibited a range of different axial lengths (range: 20.45–25.15 mm), the transverse magnification of each scan was also adjusted to account for ocular magnification associated with different eye size, based on each subject’s individual axial length measures, using a previously described method.22 The morphologic characteristics of the foveal region mean that relatively small shifts in fixation away from foveal center can result in variations in retinal thickness.23 To ensure that scans exhibiting evidence of noncentral fixation were not included in the analysis, each subject’s three-dimensional OCT scans were analyzed to determine the total retinal

1226 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES

thickness at the center of the foveal pit. Any individual radial line scans exhibiting a foveal point thickness difference of $12 mm (i.e., $4 pixels difference in the OCT image) compared with the central foveal pit thickness from the 3-dimensional scan, were assumed to have evidence of noncentral fixation and were excluded from further analysis. The segmentation data from each of the OCT images were then used to derive the total retinal thickness (the distance from the RPE to the inner limiting membrane), along with three thickness parameters describing the outer retinal layers, including the RPE to ISe thickness (the distance from the RPE to the ISe), the inner segment (IS) thickness (the distance from the ISe to the external limiting membrane), and the ONL + OPL thickness (the distance from the external limiting membrane to the OPL/INL); and three thickness values describing the inner retinal layers, including INL thickness (the distance from the OPL/ INL to the INL/IPL), IPL + GCL thickness (the distance from the INL/IPL to the GCL/NFL), and NFL thickness (the distance from the GCL/NFL to the inner limiting membrane) (Figure 1D). The thickness for each of the layers at foveal center was determined in each of the averaged scans, and referred to as the “foveal point thickness.” The average thickness from foveal center to 0.5 mm away from the fovea was also calculated, on either side of foveal center, and referred to as the “foveal zone thickness.” The average thickness from 0.5 mm away from foveal center out to 1.5 mm was calculated and referred to as the “parafoveal zone thickness” and also from 1.5 mm from foveal center out to 2.5 mm from foveal center, which was referred to as the “perifoveal zone thickness.” Analysis of the images along each of the four radial scan lines allowed thickness values to be determined in the foveal, parafoveal, and perifoveal zones in eight different locations: temporal, superiortemporal, superior, superior-nasal, nasal, inferiornasal, inferior, and inferior-temporal. For the inner retinal layers, only the parafoveal and perifoveal zones were examined, because these layers were not present at the center of the fovea in the majority of subjects. Because each of the thickness measures within a subject was determined from the mean thickness from two repeated scan lines, the reproducibility of these data was assessed by determining the intraclass correlation coefficient (ICC), along with the mean (and standard deviation) of the difference between the two repeated thickness measures in each subject for each of the considered retinal layers in all of the retinal zones. All statistical analysis was carried out using IBM SPSS Statistics version 21 (IBM Corp, Armonk, NY). Parametric statistical tests were used, because the data



2015  VOLUME 35  NUMBER 6

did not depart significantly from a normal distribution (Kolmogorov–Smirnov test, P . 0.05 for all parameters). For the foveal point thickness values, a univariate analysis of variance was carried out, examining the influence of age and gender as covariates. A repeated-measures analysis of variance was also carried out to examine the variations in each of the retinal thickness values over the central 5 mm of the posterior pole, with two within-subjects effects (retinal zone [foveal, parafoveal, and perifoveal zone] and retinal location), also including age and gender as covariates. Bonferroni-adjusted pairwise comparisons were used to examine any significant main effects. To further examine the influence of age on each of the retinal parameters examined, simple linear regression analysis was performed for each of the variables in each of the retinal zones. To reduce the chance of Type 1 statistical errors associated with multiple statistical testing, a P , 0.01 was considered significant across all analyses.

Results All results presented in this article are the mean ± standard deviation. All 196 children whose data were included in the analysis had a mean age of 8.2 ± 1.9 years (range: 4–12 years) and a mean spherical equivalent refraction of +0.06 ± 0.21 D (range: +1.25 to −0.50 D). One hundred of the participants were female. The majority of participating children were of white ethnic origin (n = 177), with the remaining children having East Asian (n = 7), Middle Eastern (n = 5), South American (n = 3), Melanesian (n = 2), or Indian (n = 2) ethnic origins. The mean ± standard deviation quality index of the OCT measurements from all subjects was 6.7 ± 1.1 The retinal thickness at the center of the foveal pit from the three-dimensional scans was generally highly correlated (r2 = 0.95) and in close agreement with the foveal point thickness values from the high-resolution line scans (mean difference, 2.7 ± 3.2 mm), indicating central foveal fixation during the high-resolution scan acquisition for the majority of subjects. Fifty-two (3.3% of 1,568) individual high-resolution B-scans exhibited evidence of noncentral fixation, and these scans were excluded from the final analysis. Any subject without a complete set of at least 4 valid highresolution radial line scans was excluded from the analysis of retinal thickness across the posterior pole, which meant this aspect of the analysis included 188 of 196 subjects. Considering each of the retinal layer boundaries, on average, the masked observer performed some manual correction of the automated segmentation in 6 ± 6% of

RETINAL THICKNESS IN CHILDHOOD  READ ET AL

scans (ranging from 0.5% of scans for the external limiting membrane boundary to 17% of scans for the OPL/INL boundary). The mean absolute difference between the automatically detected boundary position and the corrected boundary position in all valid scans across the 5-mm analysis zone was 0.3 ± 0.2 mm (ranging from 0.2 ± 0.3 mm for the ISe boundary to 0.6 ± 1.4 for the NFL boundary), indicating that generally only small adjustments to the automated segmentation were required. Analysis of reproducibility of the thickness data for each of the considered retinal layers revealed a high level of reproducibility, with no significant bias between the two repeated measures, and the standard deviations of the differences being similar to the instrument’s axial resolution: total retinal thickness (ICC: 0.997, mean difference: 0.03 ± 3.4 mm), RPE to ISe thickness (ICC: 0.971, mean difference: 0.01 ± 1.8 mm), IS thickness (ICC: 0.942, mean difference: 0.01 ± 1.4 mm), OPL + ONL thickness (ICC: 0.972, mean difference: 0.1 ± 3.4 mm), INL thickness (ICC: 0.756, mean difference: −0.01 ± 3.1 mm), IPL + GCL thickness (ICC: 0.983, mean difference: −0.03 ± 2.7 mm), and NFL thickness (ICC: 0.987, mean difference: 0.05 ± 2.4 mm). To examine for the presence of inner retinal layers at foveal center, we compared the thickness of the total retina, with the thickness of the outer retina (distance from the RPE to the OPL/INL boundary). This revealed an average difference between total retina thickness and outer retina thickness of 3.4 mm at foveal center, indicating for the majority of children, there was no evidence of inner retinal layers present at foveal center. Eight subjects exhibited more than 12 mm difference between total retina and outer retina thickness at foveal center, suggesting some remnants of inner retinal layers at foveal center for 4.3% (8 of 188) of the children examined. Total Retina Thickness The average foveal point total retina thickness in the population was 218 ± 15 mm, and total retina thickness increased significantly with age (P = 0.001). Regression analysis indicated an average increase in foveal retina thickness of 1.8 mm per year (Table 1). There was no significant effect of gender on the total retina thickness values (P . 0.05). Table 2 and Figure 2A demonstrate the mean total retina thickness in the children across the central posterior pole. Total retina thickness exhibited a significant variation with age (P , 0.01), retinal zone (P , 0.001) and retinal location (P , 0.001), and a significant zone by location interaction (P , 0.001). The total retina was significantly thinner in the foveal zone (mean thickness

1227

across all locations, 255 ± 16 mm), compared with the perifoveal zone (mean: 301 ± 13 mm), which was significantly thinner than the parafoveal zone (mean: 336 ± 13 mm) (P , 0.001 for all comparisons). The temporal retina (mean thickness across all zones, 285 ± 12 mm) was significantly thinner than the nasal retina (mean: 305 ± 13 mm), and the inferior retina (mean: 294 ± 12 mm) was significantly thinner than the superior retina (mean: 303 ± 12 mm), and the magnitude of the retinal location–related differences in thickness was smaller in the foveal zone compared with the parafoveal and perifoveal zones. Regression analysis revealed a significant positive association between age and total retinal thickness for foveal and parafoveal zone measures (both P , 0.01, Figure 2B), but no significant association between age and total retinal thickness was observed in the perifoveal zone (P . 0.1). Outer Retina Layer Thickness Table 1 presents the average foveal point thickness for each of the outer retina layer thickness values. At foveal center, only the RPE to ISe thickness exhibited a significant influence of age (P , 0.001), with regression analysis indicating an increase in the RPE to ISe thickness of 0.4 mm per year. The foveal point IS thickness and the ONL + OPL thickness did not exhibit significant changes with age (all P . 0.05). Table 2 and Figure 3 illustrate the variations in each of the 3 outer retinal layer thickness values across the central 5 mm of the posterior pole. Each of the three outer retinal layers exhibited significant variations in thickness with retinal zone and location, and significant zone by location interactions (all P , 0.001). In terms of retinal zone, a similar pattern of change was observed in all three outer retina layer thicknesses, with peak thickness in the central foveal zone being significantly thicker than that in the parafoveal zone, and the perifoveal zone being the thinnest. Any meridional variations in the outer retinal layers (i.e., differences in thickness between the eight considered locations) were only of small magnitude. However, some small (on average 1 mm) but statistically significant differences were found across the various locations, for each of the layers. For the RPE to ISe, the inferior-nasal and inferior locations were typically the thinnest and were each significantly thinner than the temporal, superior, and nasal locations (P , 0.01). For the IS thickness, the inferior, inferiortemporal and inferior-nasal locations were typically thinner than the other locations, and the difference was significant compared with the temporal, superior, superior-nasal, and nasal locations (P , 0.01). For the ONL + OPL thickness, the inferior, inferior-temporal,

1228 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES



2015  VOLUME 35  NUMBER 6

Table 1. Mean Foveal Point Thickness for the Total Retina, and for the Outer Retinal Layers (RPE to ISe, IS, and ONL + OPL Thickness) and Results From Regression Analysis Examining the Relationship Between Age and Foveal Center Point Thickness Mean ± SD Thickness (mm)

Regression Analysis (Retinal Thickness and Age)

All Children (n = 196) Total retina RPE to ISe IS ONL + OPL

218 74 34 108

± ± ± ±

15 3 2 11

r

B (mm/year)

P

0.23 0.27 0.06 0.09

1.8 0.4 0.06 0.5

0.001 ,0.001 0.4 0.2

SD, standard deviation.

magnitude of effect observed in the central foveal zone (0.5 mm increase in RPE to ISe thickness per year). There was no significant influence of age on the IS thickness values (P . 0.05), and for the ONL + OPL thickness, only the central foveal zone exhibited a significant positive association between age and thickness (r = 0.20, slope, 0.7 mm/year, P , 0.01). There were

and inferior-nasal locations were also significantly thinner than each of the other locations (P , 0.01). A significant effect of age was also found for the RPE to ISe thickness across the posterior pole (P , 0.001). Regression analysis revealed a significant positive association with age in each of the three considered zones (all P , 0.001), with the largest

Table 2. Mean Thickness Across the Central 5 mm of the Posterior Pole for the Total Retina Thickness and for the Outer Retinal Layers (RPE to ISe, IS, and ONL + OPL Thickness) for All 188 Children in the Study With Complete Data for All 4 Radial Scan Lines Mean ± SD Thickness (mm) Foveal Zone Temporal

Superior-temporal

Superior

Superior-nasal

Nasal

Inferior-nasal

Inferior

Inferior-temporal

SD, standard deviation.

Total retina RPE to ISe IS ONL + OPL Total Retina RPE to ISe IS ONL + OPL Total retina RPE to ISe IS ONL + OPL Total retina RPE to ISe IS ONL + OPL Total retina RPE to ISe IS ONL + OPL Total retina RPE to ISe IS ONL + OPL Total retina RPE to ISe IS ONL + OPL Total retina RPE to ISe IS ONL + OPL

247 65 33 109 254 64 33 107 259 65 33 107 258 65 33 108 256 65 34 110 258 65 33 108 257 65 33 107 252 65 33 106

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

17 2 1 8 16 3 1 7 16 2 1 7 16 3 1 7 16 3 1 7 16 3 1 7 16 2 1 7 16 2 1 7

Parafoveal Zone 321 57 29 95 332 56 27 91 343 57 28 92 343 57 28 93 340 57 30 96 340 56 28 91 337 56 27 88 332 56 28 88

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

14 2 1 7 13 2 1 7 14 2 1 7 14 2 1 7 14 2 1 7 13 2 1 7 14 2 1 7 13 2 1 7

Perifoveal Zone 285 54 25 79 289 55 25 80 306 55 25 82 320 54 25 82 320 54 26 81 310 53 24 76 290 54 24 73 288 54 24 76

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

14 2 1 7 12 2 1 6 13 2 1 6 15 2 1 7 15 2 1 7 15 2 1 7 14 2 1 6 13 2 1 6

RETINAL THICKNESS IN CHILDHOOD  READ ET AL

1229

thinnest in the temporal (mean thickness across all zones, 42 ± 3 mm) and the thickest in nasal (mean: 44 ± 3 mm) and superior-nasal (mean: 44 ± 3 mm) locations. The IPL + GCL exhibited its thinnest location inferiorly (mean: 75 ± 5 mm) and was the thickest in temporal (mean: 89 ± 6 mm) and nasal (mean: 84 ± 5 mm) locations. The NFL was the thinnest in temporal locations (mean: 2 ± 1 mm) and the thickest in inferior-nasal (mean: 40 ± 4 mm) and superior-nasal locations (mean: 38 ± 4 mm). No significant age or gender differences were observed for the INL thickness or the IPL + GCL thickness (both, P . 0.05). Significant variations with age were observed in the NFL thickness data (P , 0.01), and regression analysis revealed a significant positive association between age and NFL thickness, with an increase in NFL thickness of 0.3 mm per year observed in both parafoveal and perifoveal zones. There was no significant effect of gender on NFL thickness (P . 0.01). Discussion

Fig. 2. Average thickness profile of the total retina across the central 5 mm of the posterior pole for 188 children in the study with complete data for each of the 4 radial scan lines (error bars represent ±1 standard deviation) (A), and the influence of age on the foveal zone total retinal thickness (the solid line is best-fit regression line, and the dashed line is 95% confidence intervals) (B). Regression analysis revealed the association between total retinal thickness and age to be significant (r = 0.22, P = 0.003).

no significant gender effects observed in any of the outer retinal layers in any of the considered zones (P . 0.01 for all layers). Inner Retina Layer Thickness The variation in the thickness of the examined inner retinal layers across the central 5 mm of the posterior pole (within the parafoveal and perifoveal zones) is illustrated in Figure 4 and Table 3. Significant variations were evident in each of these layers with retinal zone, and location, and there was also a significant zone by location interactions (all P , 0.001). The INL and IPL + GCL thickness were both significantly thicker in the parafoveal zone compared with the perifoveal zone (P , 0.001), whereas the NFL on average was thinner in the parafovea compared with the perifovea (P , 0.001). The INL exhibited only relatively small meridional variations but was generally the

We have examined the average macular total retinal thickness, and retinal layer thickness with highresolution SD-OCT, in a population of healthy pediatric subjects, with minimal refractive errors, and found small magnitude but statistically significant increases in the total retinal thickness, and of some of the individual retinal layers with age in this population. To our knowledge, this study provides the first in vivo evidence of statistically significant increases in retinal layer thickness from early childhood to adolescence. These findings are consistent with previous histologic evidence that suggest that aspects of foveal retinal morphology such as the foveal cone outer segment continue to develop postnatally and well into the childhood years.4 The topographic and age-related changes in total retinal thickness that we observed are generally consistent with other recent studies reporting on macular total retinal thickness in pediatric populations (ages examined, 4–17 years) with SD-OCT.12–14 These studies have reported mean total retinal thickness in the central 1-mm foveal zone ranging from 253 mm to 271 mm, and 2 of these recent articles have also reported very similar magnitude increases in foveal total thickness with age in childhood (of 1.7–1.8 mm per year).13,14 The changes we observed in total retinal thickness with age were statistically significant in the foveal and parafoveal zones, but not in the perifoveal zones, which suggests that the development/maturation of the more peripheral retinal

1230 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES



2015  VOLUME 35  NUMBER 6

Fig. 3. Average thickness profile (averaged across the 4 measured meridians) for each of the outer retinal layers examined, across the central 5 mm of the posterior pole for 188 children in the study with complete data for each of the 4 radial scan lines (error bars represent ±1 standard deviation) (A). The influence of age on the mean foveal zone RPE to ISe thickness (B), the IS thickness (C), and the ONL + OPL thickness (D) is also illustrated (the solid line is best-fit regression line, and the dashed line is 95% confidence interval). Regression analysis revealed that the relationship between age and foveal zone thickness was only statistically significant for the RPE to ISe thickness (P , 0.001), and the ONL + OPL thickness (P , 0.001).

regions occurs earlier than that of central foveal regions in childhood. Previous histologic studies have also suggested that the midperipheral retinal regions develop sooner than the central foveal regions do.3 We also examined the changes in a range of different retinal layers with age and found small but statistically significant increases from early childhood to adolescence in some of the retinal layers examined. The most prominent changes with age in childhood seemed to occur in the outer retinal layers close to the fovea. The changes observed in the RPE to ISe thickness exhibited the strongest association with age. Taking into account the mean magnitude of increase in the outer retinal layers across the range of ages examined in this study, our findings suggest that the foveal outer retinal morphology is approaching close to adult levels24 by adolescence. Statistically significant age-related changes were observed in only one of the examined inner retinal layers (the NFL), and the magnitude of these changes was small and unlikely to be of clinical significance. This also agrees with previous histologic evidence of changes in the foveal photoreceptors occurring later into childhood compared with inner retinal structures.3

The total retinal thickness exhibited topographic variations in thickness across the posterior pole, with temporal regions being thinner compared with nasal regions, and inferior regions thinner compared with superior regions. These results are similar to those reported in other studies with SD-OCT in children12–14 and adults.25,26 We also found significant topographic variations occurring in the outer and inner retinal layers across the posterior pole. The outer retinal layers all exhibited their peak in thickness at the center of the fovea, which is consistent with previous observations from OCT studies in children14 and adults24 and the known spatial density of photoreceptors across the posterior pole.27 The topographic variations observed in the inner retinal layers were also generally qualitatively similar to previous studies of healthy adults28 and to anatomical studies of the density of inner retinal cells across the retina.29 The similarities in thickness distribution of the total retina and retinal layers across the macular region with previous studies of adults25,26,28 suggest that the distribution of retinal tissue and cells is morphologically similar between the adult and pediatric retinae across the posterior pole.

RETINAL THICKNESS IN CHILDHOOD  READ ET AL

1231

Fig. 4. Average thickness profile (left) across the central 5 mm of the posterior pole and the influence of age on the parafoveal zone thickness (error bars represent ±1 standard deviation) (right) for the INL thickness (A), IPL + GCL thickness (B), and NFL thickness (C) for each of 188 children in the study with complete data for each of the 4 radial scan lines (the solid line is best-fit regression line, and the dashed line is 95% confidence interval). Note there are no measures in the central foveal 1 mm zone because of the absence of the inner retinal layers in the fovea in the majority of subjects.

Studies have documented changes in a range of different retinal layer thicknesses (particularly the outer retinal layers) associated with a variety of retinal diseases29–31 that can begin in childhood. Therefore, an improved understanding of the average thickness of various retinal layers and the thickness distribution across the posterior pole in healthy pediatric populations may assist in the clinical diagnosis of retinal disease in childhood. Our finding of an increase in thickness of outer retinal layers from early childhood

to adolescence, coupled with the previous finding of significant thinning (particularly in outer retinal layers), associated with a range of retinal dystrophies30–32 suggests that a clinical observation of a thinning in the outer retinal layers in childhood should raise a high clinical suspicion of a retinal abnormality. In our population of children, we observed reasonable consistency in the pattern of thickness distribution of the retina and retinal layers across the macular region. All children exhibited a minimum in their total

1232 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES



2015  VOLUME 35  NUMBER 6

Table 3. Mean Thickness in the Parafoveal and Perifoveal Retinal Zones for the Inner Retinal Layers Examined (INL, IPL + GCL, and NFL Thickness) for 188 Children in the Study With Complete Data for All 4 Radial Scan Lines Mean ± SD Thickness (mm) Parafoveal Zone Temporal Superior-temporal Superior Superior-nasal Nasal Inferior-nasal Inferior Inferior-temporal

INL IPL + GCL NFL INL IPL + GCL NFL INL IPL + GCL NFL INL IPL + GCL NFL INL IPL + GCL NFL INL IPL + GCL NFL INL IPL + GCL NFL INL IPL + GCL NFL

43 95 2 46 91 21 46 92 28 47 94 25 45 93 18 46 93 26 46 91 28 46 92 21

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

4 6 1 3 5 2 4 6 3 3 6 3 4 6 2 3 5 3 3 6 2 3 6 3

Perifoveal Zone 41 83 2 40 64 26 40 63 42 41 65 52 44 76 40 40 64 53 39 64 42 41 67 27

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

3 6 1 3 5 3 3 6 5 3 6 6 3 6 4 3 6 6 3 6 5 3 6 3

SD, standard deviation.

retina thickness, a peak in the thickness of outer retinal layers, and a substantial thinning and merging together of the inner retinal layers at foveal center. However, the error bars and spread of data in Figures 2–4 indicate the presence of considerable between-subject variation in the absolute retina and retina layer thickness values across the macula (e.g., total retina thickness at foveal center ranged from a maximum of 292 mm to a minimum of 189 mm), suggesting some inherent between-subject variability in macular retinal morphology, that is consistent with previous studies in adults.33 A few children in our study (4%) also exhibited some evidence of remnants of inner retinal layers persistent across the fovea. Persistence of inner retinal layers at foveal center has previously been reported to be associated with preterm birth and retinopathy of prematurity34; however, there is also evidence from OCT studies of healthy adults that some remnants of inner retinal layers at foveal center are a feature of the foveal morphology in a small percentage (6%) of healthy adults with no history of prematurity.33 Although our study has examined retinal layer thickness in a larger number of children and across a larger number of retinal regions than the previous study in pediatric subjects (that examined retinal layer

thickness at three discrete locations in a single horizontal OCT scan),14 a limitation of the data presented in this article is that the topographic thickness data were derived from only four foveal radial line scans. Further work examining the topographic characteristics of retinal layer thickness, using a denser scanning protocol, will provide further insights into the regional variations in retinal morphology in childhood. The cross-sectional nature of this study is another limitation, and future longitudinal studies will provide greater insight into the nature and time-course of retinal changes with age in childhood. In conclusion, this study demonstrates small but statistically significant changes in total retinal thickness and retinal layer thickness that occur with age from early childhood to adolescence. The relatively small magnitude of the observed age-related changes, coupled with similarities between the distribution of retinal layer thickness across the posterior pole in the children in this study and that documented in adults previously, suggests that many aspects of macular retinal morphology are reaching close to adult levels by adolescence. Key words: foveal development, macular thickness, optical coherence tomography, pediatric.

RETINAL THICKNESS IN CHILDHOOD  READ ET AL

References 1. Hendrickson AE, Yuodelis C. The morphological development of the human fovea. Ophthalmology 1984;91:603–612. 2. Yuodelis C, Hendrickson A. A qualitative and quantitative analysis of the human fovea during development. Vis Res 1986;26:847–855. 3. Hendrickson A, Drucker D. The development of parafoveal and mid-peripheral human retina. Behav Brain Res 1992;49: 21–31. 4. Hendrickson A, Possin D, Vajzovic L, Toth CA. Histologic development of the human fovea from midgestation to maturity. Am J Ophthalmol 2012;154:767–778. 5. Huang D, Swanson EA, Lin CP, et al. Optical coherence tomography. Science 1991;254:1178–1181. 6. Wojtkowski M, Srinivasan V, Fujimoto JG, et al. Three-dimensional retinal imaging with high-speed ultrahigh resolution optical coherence tomography. Ophthalmology 2005;112: 1734–1746. 7. Hess DB, Asrani SG, Bhide MG, et al. Macular and retinal nerve fiber layer analysis of normal and glaucomatous eyes in children using optical coherence tomography. Am J Ophthalmol 2005;139:509–517. 8. Huynh SC, Wang XY, Rochtchina E, Mitchell P. Distribution of macular thickness by optical coherence tomography: findings from a population-based study of 6-year old children. Invest Ophthalmol Vis Sci 2006;47:2351–2357. 9. Luo H-D, Gazzard G, Fong A, et al. Myopia, axial length, and OCT characteristics of the macula in Singaporean children. Invest Ophthalmol Vis Sci 2006;47:2773–2781. 10. Huynh SC, Wang XY, Burlutsky G, et al. Retinal and optic disc findings in adolescence: a population-based OCT study. Invest Ophthalmol Vis Sci 2008;49:4328–4335. 11. El-Dairi MA, Asrani SG, Enyedi LB, Freedman SF. Optical coherence tomography in the eyes of normal children. Arch Ophthalmol 2009;127:50–58. 12. Turk A, Ceylan OM, Arici C, et al. Evaluation of the nerve fiber layer and macula in the eyes of healthy children using spectral domain optical coherence tomography. Am J Ophthalmol 2012;153:552–559. 13. Barrio-Barrio J, Noval S, Galdós M, et al. Multicenter spanish study of spectral-domain optical coherence tomography in normal children. Acta Ophthalmol 2013;91:e56–e63. 14. Yanni SE, Wang J, Cheng CS, et al. Normative reference ranges for the retinal nerve fiber layer, macula, and retinal layer thicknesses in children. Am J Ophthalmol 2013;155:354–360. 15. Maldonado RS, O’Connell RV, Sarin N, et al. Dynamics of human foveal development after premature birth. Ophthalmology 2011;118:2315–2325. 16. Dubis AM, Costakos DM, Subramaniam CD, et al. Evaluation of normal human foveal development using optical coherence tomography and histologic examination. Arch Ophthalmol 2012;130:1291–1300. 17. Vajzovic L, Hendrickson AE, O’Connell RV, et al. Maturation of the human fovea: correlation of spectral domain optical

18. 19.

20.

21.

22.

23.

24.

25.

26.

27. 28.

29. 30.

31.

32.

33.

34.

1233

coherence tomography findings with histology. Am J Ophthalmol 2012;154:779–789. Steinkuller PG, Du L, Gilbert C, et al. Childhood Blindness. J AAPOS 1999;3:26–32. Thomas M, Kumar A, Thompson JR, et al. Is high-resolution spectral domain optical coherence tomography reliable in nystagmus? Br J Ophthalmol 2013;97:534–536. Alonso-Canerio D, Read SA, Collins MJ. Speckle reduction in optical coherence tomography imaging by affine-motion image registration. J Biomed Opt 2011;16:116027. Chiu SJ, Li XT, Nicholas P, et al. Automatic segmentation of seven retinal layers in SDOCT images congruent with expert manual segmentation. Opt Express 2010;18:19413–19428. Vincent SJ, Collins MJ, Read SA, Carney LG. Retinal and choroidal thickness in myopic anisometropia. Invest Ophthalmol Vis Sci 2013;54:2445–2456. Odell D, Dubis AM, Lever JF, et al. Assessing errors inherent in OCT-derived macular thickness maps. J Ophthalmol 2011; 2011: 692574. Srinivasan VJ, Monson BK, Wojtkowski M, et al. Characterization of outer retinal morphology with high-speed ultrahighresolution optical coherence tomography. Invest Ophthalmol Vis Sci 2008;49:1571–1579. Grover S, Murthy RK, Brar VS, Chalam KV. Normative data for macular thickness by high-definition spectral-domain optical coherence tomography (spectralis). Am J Ophthalmol 2009; 148:266–271. Harb E, Hyman L, Fazzari M, et al. Factors associated with macular thickness in the COMET myopic cohort. Optom Vis Sci 2012;89:620–631. Curcio CA, Sloan KR, Kalina RE, Hendrickson AE. Human photoreceptor topography. J Comp Neurol 1990;292:497–523. Ooto S, Hangai M, Tomidokoro A, et al. Effects of age, sex, and axial length on the three-dimensional profile of normal macular layer structures. Invest Ophthalmol Vis Sci 2011;52: 8769–8779. Curcio CA, Allen KA. Topography of ganglion cells in human retina. J Comp Neurol 1990;300:5–25. Witkin AJ, Ho TH, Fujimoto JG, et al. Ultra-high resolution optical coherence tomography assessment of photoreceptors in retinitis pigmentosa and related diseases. Am J Ophthalmol 2006;142:945–952. Lim JI, Fawzi AA, Hopkins J, et al. A pilot study of fourierdomain optical coherence tomography of retinal dystrophy patients. Am J Ophthalmol 2008;146:417–426. Hood DC, Lin CE, Lazow MA, et al. Thickness of receptor and post-receptor retinal layers in patients with retinitis pigmentosa measured with frequency-domain optical coherence tomography. Invest Ophthalmol Vis Sci 2009;50:2328–2336. Tick S, Rossant F, Ghorbel I, et al. Foveal shape and structure in a normal population. Invest Ophthalmol Vis Sci 2011;52: 5105–5110. Yanni SE, Wang J, Chan M, et al. Foveal avascular zone and foveal pit formation after preterm birth. Br J Ophthalmol 2012; 96:961–966.