Semi-Automated Analysis of Dome-Shaped Macula in Preterm and Full-Term Infants Using Handheld Swept-Source Optical Coherence Tomography

Abstract

Purpose: Dome-shaped macula (DSM) is known to occur in highly myopic adults and, recently, preterm infants. This study uses investigational handheld swept-source optical coherence tomography (SS-OCT) to further characterize infantile DSM.

Methods: In this prospective, observational study, preterm infants undergoing retinopathy of prematurity screening and full-term infants within 72 hours of birth were imaged. Two trained graders assessed macular features, including DSM, subretinal fluid, and macular edema. A semi-automated program measured foveal immaturity, dome height, and diameter.

Results: Two hundred seventeen imaging sessions from 50 full-term and 30 preterm infants were included (46% female, preterm birth weight 1038 ± 335 g, and gestational age 28.7 ± 3.1 weeks). DSM occurred in 40% preterm versus 14% full-term infants (P = 0.01). Mean postmenstrual age at first DSM diagnosis was 38.4 ± 0.0 weeks among preterm and 40.4 ± 1.1 weeks among full-term infants (P < 0.001). Dome height and diameter measured 55.67 ± 44.22 µm and 3583.15 ± 1090.35 µm for preterm versus 88.37 ± 44.73 µm and 3581.97 ± 355.07 µm for full-term infants (P = 0.24 and P = 0.96, respectively). All 27 images (11 preterm and 7 full-term infants) with 3-dimensional analysis had round dome configuration. No other associations were seen, including macular fluid (P = 0.17).

Conclusions: Infants frequently exhibit DSM without an association with macular fluid. Preterm infants were more likely than full-term infants to have DSM. Unlike DSM in children and adults, infantile DSM configuration is mostly round rather than ridge-shaped.

Dome-shaped macula (DSM) is a convex elevation of the chorioretinal contour at the macula, appreciated on optical coherence tomography (OCT).^1,2 The prevalence of DSM has been estimated at between 30% and 80% in adults with myopia,^3–5 most notably in eyes that exhibit high myopia.² DSM has been documented across all ethnicities, with a marginally higher prevalence observed in the Asian population.^6,7 Until recently, its prevalence in children was felt to be rare (0.77%).¹ 

The presence and height of DSM in adults is associated with serous macular detachment^7,8 and thickened choroid.^9,10 DSM is categorized into the following three configurations, listed in order from most common to least common: horizontal ridge-shaped (sometimes called “horizontal oval-shaped”) dome detected on vertical OCT scan, round dome detected on both horizontal and vertical OCT scans, and vertical ridge-shaped (sometimes called “vertical oval-shaped”) dome detected on horizontal OCT scan.^11–13 

Handheld OCT may have great potential as a noninvasive clinical screening device for retinopathy of prematurity (ROP) as well as other ocular conditions, reducing the need for more aggressive techniques, such as indirect ophthalmoscopy with scleral depression and an eyelid speculum. Among infants, handheld OCT has identified a number of frequent macular findings of unclear significance, including cystoid macular edema (CME)^14,15 in preterm infants and subretinal fluid (SRF) in full-term infants.^16,17 Foveal immaturity, characterized by shallow foveal contour with persistent inner retinal layers, has been noted in both full-term and preterm infants.^18,19 Surprisingly, DSM was identified in 65% of preterm infants screened for ROP using handheld spectral domain OCT (SD-OCT), and was found to be associated with low birth weight, ROP, and plus disease.²⁰ OCT quality in that study was not adequate to determine the dome configuration (vertical ridge-shaped, horizontal ridge-shaped, or round dome). 

The present study aims to better characterize and compare DSM between preterm and full-term infants using handheld swept-source OCT (SS-OCT) technology from an investigational device developed at the University of Washington,^21–23 correlating DSM findings to chorioretinal measurements. 

Subjects and Methods

Preterm infants undergoing routine ROP screening and full-term newborns were included in this prospective, observational study. This study was approved by and conducted in compliance with the institutional review board at the University of Washington and Seattle Children's Hospital. Informed consent was obtained from all guardians after an explanation of the nature and possible consequences of the study. Participants were recruited between March 2018 and June 2019. Eligible preterm infants met the standard ROP screening criteria of less than 30 weeks’ gestation and/or under 1500 g birth weight. Eligible full-term infants were born at minimum 37 weeks’ gestation and were at least 12 hours old at the time of imaging. Full-term infants were excluded for systemic disease, uncertain guardianship, or complex social situations. Both preterm and full-term patients who were deemed too medically unstable for SS-OCT imaging were also excluded. Two drops of phenylephrine hydrochloride 1% and cyclopentolate ophthalmic 0.2% solution (Cyclomydril; Alcon, Fort Worth, TX, USA) were administered to dilate the pupils of each participant prior to imaging and the ROP screening examination. The order of OCT imaging and clinical examinations with scleral depression alternated for each visit. 

SS-OCT Imaging

Awake, unsedated full-term and preterm infants were imaged using an investigational handheld SS-OCT device powered by a 1060 nm swept laser source with an imaging speed of 200,000 A-scans per second (200 kHz).^21,23 The spatial resolution of the device was 5.5 µm axially and 15 µm laterally. This device was built with a pupil-finding feature to align with the optical axis more efficiently. Full-term infants were imaged once between 12 and 72 hours of birth, and preterm infants were imaged each time they had a routine ROP examination (typically starting at 31 weeks postmenstrual age or 1 month after birth and extending up to 45 weeks postmenstrual age, depending on clinical findings for the routine ROP examination). Neither an eyelid speculum nor ocular contact were used during OCT imaging. The eyelids were gently retracted by the imagers’ fingers. High-quality SS-OCT images of the posterior pole were obtained during a maximum timeframe of 15 minutes for both eyes. 

Clinical Examinations

Pediatric ophthalmologists (authors M.T.C. or K.T.H.) in the neonatal intensive care unit conducted ROP examinations on the same day as SS-OCT imaging. The ophthalmologists were unaware of the SS-OCT imaging findings during their examinations. These examinations involved indirect ophthalmoscopy with scleral depression. Decisions regarding further monitoring or treatment were in accordance with published standard guidelines.^24,25 Subsequent evaluations occurred every 1 to 3 weeks for typically 2 to 7 visits. The medical records of each patient were reviewed for patient data including sex, race, gestational age, birth weight, and ROP findings based on clinical indirect ophthalmoscopic results. 

Image Analysis

The OCT volumetric scans for each eye from each imaging session were evaluated by two trained, independent, masked graders (authors S.E.L. and T.R.M.) for the presence of DSM, CME, and SRF. The presence of DSM was defined as any chorioretinal convexity of the macula without minimum required measurements, following previously published definitions.^{1–3,26–28} This was determined by subjective evaluation for elevation of the retinal pigment epithelium above a straight line between the parafoveal retinal pigment epithelium (Fig. 1 demonstrates this line). Images with motion artifact or other image distortion impairing the ability to accurately detect DSM were excluded. CME was defined as hyporeflective intraretinal spaces visible in three or more adjacent B-frames, resulting in distortion of the retinal layers. SRF was defined as hyporeflective or sometimes turbid fluid beneath the neuroretina but above the retinal pigment epithelium at the fovea. A third trained, independent, masked grader acted as a tiebreaker for all disagreements between the two graders. 

Figure 1.

 

Example of dome-shaped macula measurements using the semi-automated MATLAB program. The grader selected a high-quality B scan including the foveal center for analysis. The dome width was measured at the base from inflection point to inflection point of the retinal pigment epithelium, selected by the grader. The dome height was measured from the base to the peak of the retinal pigment epithelium, whose location was selected by the grader.

For best quality volumes of those imaging sessions graded as having DSM, 3-dimensional renderings were created and viewed using a Java plugin on Image J (National Institute of Health, Bethesda, MD, USA). A single grader (author J.N.D.) graded each 3-dimensional volume as having round dome (the dome appears circular without a vertical or horizontal elongation), vertical ridge-shaped (dome appears vertically elongated with a shorter horizontal length), or horizontal ridge-shaped (dome appears horizontally elongated with a shorter vertical length) configuration. 

A single foveal B-scan image of the highest quality from the best OCT volumetric scan for each eye of each visit was selected for additional semi-automated foveal maturity analysis by two additional trained, masked graders (authors E.K.T. and K.L.). Images were excluded from foveal maturity analysis if they were of poor quality or lacked a visible fovea. These methods were previously validated.²⁹ Briefly, a MATLAB (MathWorks, Inc., Natick, MA, USA) program was utilized to segment and measure the inner retinal thickness from the internal limiting membrane to the outer plexiform layer. The choroidal thickness was measured from the retinal pigment epithelium to the choroid/sclera border. Both measurements were performed at the foveal center and parafovea (2.5 mm from the center of the fovea on each side, averaged). The fovea/parafoveal (F/P) ratio was calculated for the inner retinal thickness. Outer retinal thickness was not assessed in this study because it did not previously correlate well with other parameters of foveal development or demographics.²⁹ The foveal angle was measured in degrees from the lowest point of the foveal internal limiting membrane to the parafoveal points on both sides where the retinal contour started to flatten (as specified by the grader). 

For OCT volumes containing DSM, the B-scan image where the dome displayed its peak was chosen for measurement of the dome's height and diameter. A novel semi-automated program in MATLAB (MathWorks, Inc., Natick, MA, USA) measured the dome diameter by tangentially connecting the two points of the retinal pigment epithelium at the base of the dome, selected by the grader at the inflection point on each side (see Fig. 1). The height was calculated by measuring the perpendicular distance from the base to the peak of the dome, whose location was selected by the grader. Three repeated measurements performed in this fashion were averaged to give the final result. This analysis was performed by a single grader (author J.N.D.). 

Statistical Analysis

Regarding the primary outcome, preterm infants were compared to full-term infants for the rate at which DSM was observed and for the height and diameter of the observed domes. 

For secondary outcomes, infants with and without DSM were compared for demographic features, including sex, race, gestational age, and birth weight. Eyes with and without DSM were compared for ROP severity (based on the clinical standard ROP examination) and other SS-OCT findings (CME in preterm infants, SRF in full-term infants, inner retinal F/P ratio, foveal angle, and choroidal thickness at the fovea and parafovea). 

Generalized linear mixed models were used to account for two eyes and multiple visits from the same infant. Demographic comparisons between infants were carried out using the nonparametric Mann-Whitney U test for continuous variables. Categorical data were analyzed by either the chi-squared test or the Fisher's exact test. Intergrader reliability for DSM diagnosis was evaluated using a kappa statistic. A repeatability analysis was performed using the coefficient of variation (CV) to evaluate the precision of the DSM height and diameter measurements. Sampling one high-quality B-scan image previously analyzed from each of five imaging sessions containing DSM (4 eyes from 3 infants on 4 dates), a single grader repeated the semi-automated analysis on the same images as described above, to compare results to the original measurements. Statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). Statistical significance was defined by an alpha level of 0.05, or a P value < 0.05 in a 2-tailed test. 

Results

Out of 278 total SS-OCT imaging sessions, 15 were excluded because the fovea was not captured, 14 were excluded due to poor image quality, and 32 were excluded because the presence of DSM could not be determined. The remaining 217 imaging sessions from 155 eyes of 80 infants (30 preterm and 50 full-term infants) were included in this study. Preterm and full-term infants were similar in terms of sex, race, and ethnicity (Table 1). Mean birth weight was 1038 ± 335 g among preterm infants and 3415 ± 502 g among full-term infants (P < 0.001). Mean gestational age was 28.7 ± 3.1 weeks among preterm infants and 39.3 ± 1.3 weeks among full-term infants (P < 0.001). On clinical examination, 23 of 59 eyes (39%) from 30 preterm infants were diagnosed with ROP, 8 (14%) had stage 3 ROP, 10 (17%) had pre-plus disease, and 0 (0%) had plus disease. On SS-OCT, 13 of 59 eyes (22%) from 30 preterm infants had CME, and 14 of 96 eyes (15%) from 50 full-term infants had SRF. There were no disagreements between the 2 graders of > 50 µm or 15 degrees in retinal thickness or foveal angle, respectively. 

Table 1.

 

Characteristics of Preterm and Full-Term Infants

 

Dome-Shaped Macula on Handheld SS-OCT

The two graders agreed on the DSM diagnosis in 190 of 217 (87.6%) of imaging sessions (kappa agreement 0.60 ± 0.07, 95% confidence interval [CI] = 0.46 to 0.72, indicating moderate to substantial agreement). Twelve of 30 (40%) preterm infants compared to 7 of 50 (14%) full-term infants had DSM in at least one eye (P = 0.01; Figs. 2, 3). Most of these were bilateral (8 of 12 or 67% of preterm infants and 3 of 7 or 43% of full-term infants). Mean postmenstrual age at the time of first DSM diagnosis was 38.4 ± 0.0 weeks among preterm infants and 40.4 ± 1.1 weeks among full-term infants (P < 0.001). Based on eyes, 20 of 59 (34%) of preterm eyes versus 10 of 96 (10%) of full-term eyes had DSM (P < 0.001). Of these, 17 of 20 (85%) eyes of 11 preterm infants and 10 of 10 (100%) eyes of 7 full-term infants had adequate image quality in at least one imaging session to perform 3-dimensional DSM analysis. Of these, 17 of 17 (100%) preterm infant eyes and 10 of 10 (100%) full-term infant eyes were graded as having round dome configuration (see Fig. 2B). No images were graded as ridge-shaped. DSM tended to occur at a higher postmenstrual age at imaging (38.8 ± 2.5 weeks with DSM versus 37.3 ± 3.4 weeks without DSM, P = 0.01) and a lower gestational age (33.0 ± 6.1 weeks with DSM versus 36.0 ± 5.3 weeks without DSM, P = 0.04). DSM was not correlated with other demographic features or ROP severity (Table 2). 

Figure 2.

 

DSM on handheld SS-OCT. (A) SS-OCT B-scans of DSM in preterm and full-term infants. Some preterm infants displayed varying severity of CME. Some full-term infants presented with SRF. No significant association between DSM and either SRF or CME were seen. (B) Slices of a handheld SS-OCT volume from a full-term and a preterm infant, both right eyes, were compiled into 3-dimensional representations of DSM, revealing a round rather than ridge-shaped dome configuration. This round configuration was seen in 10 of 10 (100%) full-term eyes and 17 of 17 (100%) preterm eyes with adequate quality for 3-dimensional analysis, and no cases of ridge-shaped configuration were identified. DSM, dome-shaped macula; SS-OCT, swept-source optical coherence tomography; CME, cystoid macular edema; SRF, subretinal fluid.

Figure 3.

 

Bar graph displaying the number of preterm and full-term infants in the study who had DSM identified at least once by handheld SS-OCT in one eye, both eyes, or not at all. Forty percent of preterm infants versus 14% of full-term infants had DSM in at least one eye (P = 0.01). Among those with DSM, 67% of preterm infants and 43% of full-term infants had DSM in both eyes. DSM, dome-shaped macula; SS-OCT, swept-source optical coherence tomography.

Table 2.

 

Dome-Shaped Macula and Associated Characteristics

 

Analyzing the presence of DSM and its relationship to other OCT macular features, DSM coincided with signs of foveal immaturity, including a greater F/P ratio at the inner retina (P = 0.08), and wider foveal angle (P = 0.08), but these findings did not reach statistical significance. DSM did not correlate with choroidal thickness at the fovea (P = 0.56) or parafovea (P = 0.39). This study found no association between DSM and CME in preterm infants (P = 0.17) or between DSM and SRF in full-term infants (P = 0.17; see Table 2). 

A repeatability analysis of dome height and width measurements revealed a maximum CV of 1.82% across all measurements, indicating very high precision.³⁰ The average dome measured 55.67 ± 44.22 µm in height and 3583.15 ± 1090.35 µm in diameter among preterm infants versus 88.37 ± 44.73 µm in height and 3581.97 ± 355.07 µm in diameter among full-term infants (P = 0.25 and P = 0.96, respectively; Fig. 4). There were no associations between DSM measurements and demographics or ROP severity (Table 3). 

Figure 4.

 

Scatterplot of dome-shaped macula height and diameter in µm of preterm (n = 20) and full-term (n = 10) infant eyes with adequate quality images for measurements. The average dome in 20 eyes from 12 preterm infants measured 55.67 ± 44.22 µm in height and 3583.15 ± 1090.35 µm in diameter. The average dome in 10 eyes from 7 full-term infants measured 88.37 ± 44.73 µm in height and 3581.97 ± 355.07 µm in diameter (P = 0.25 and P = 0.96, respectively).

Table 3.

 

Dome-Shaped Macula Height and Diameter and Associated Characteristics

 

Discussion

The key findings from this study are that DSM occurs in both preterm and full-term infants in a round rather than ridge-shaped configuration. This study did not find an association between DSM and macular fluid such as CME or SRF. This is the first study to characterize DSM in both preterm and full-term infants, and the first 3-dimensional analysis of DSM in the infant population. 

Prior studies of pediatric DSM have identified a ridge-shaped configuration (only noted on vertically oriented B-scans)¹² without typical Bruch's membrane defects^12,13,31 or greater degrees of myopia^1,7,12,20,31 seen in adults with DSM. All full-term and preterm infants with DSM and adequate image quality to interpret the 3-dimensional configuration in the present study appeared to have round domes without ridge-shaped configuration, differing from prior pediatric studies of older children and adolescents. These results suggest that infant DSM may have unique pathophysiology compared to DSM of older children and adolescents. 

Legocki et al. first described DSM in preterm infants.²⁰ That study found a relationship between DSM and ROP severity, as well as DSM and lower birth weight. These findings suggest that DSM may be considered a biomarker for ROP severity. These associations did not reach statistical significance in the present study, possibly due to lower n and inadequate power in the preterm group. Nonetheless, by including full-term infants in the present study, we add to our understanding of DSM in the infant population. DSM was more prevalent in healthy full-term newborns (14% in at least one eye) than we would previously have expected based on prior research in the pediatric general population (1.1%).¹ Considering the high prevalence of DSM in newborn infants in the present study compared to children and adults in previous studies, it is likely that DSM in infants is a temporary condition that resolves with age. Legocki et al. proposed that an immature retina and sclera may display mismatched size, resulting in DSM.²⁰ Therefore, the relationship between newborn sclera and the chorioretina in the macular region may change as eyes mature. Supporting this, we saw a trend toward an association between DSM and foveal immaturity (P = 0.08). Such anatomic differences may explain the higher prevalence of DSM among preterm compared to full-term eyes (34% vs. 10%, P < 0.001). Nonetheless, the measurements of DSM height and diameter did not differ greatly between preterm and full-term infants in this study. 

Adults with DSM have been reported to frequently develop macular serous retinal detachments, retinoschisis, choroidal neovascularization, and choroidal thickening.^7–10,27 CME and SRF are known macular OCT findings in preterm and full-term newborns, respectively.^14–17 Nonetheless, this study did not identify an association between DSM and macular fluid in full-term and preterm infants. Nor were differences seen in choroidal thickness measurements for infants with DSM. 

DSM in adults is strongly associated with high myopia.^1,7,12,20,31 Best corrected visual acuity, however, is often preserved.²⁷ Legocki et al. did not identify refractive or visual acuity differences between ex-preterm infants with DSM and those without at 9 months of age, as measured by cycloplegic streak retinoscopy and Teller acuity, respectively.²⁰ Further studies are needed with longer follow up of visual and refractive outcomes to understand the long-term implications of this OCT finding in the infant population. 

Limitations of this study include small sample size and lack of long-term follow-up. A higher number of imaging sessions for preterm infants undergoing longer follow-up due to ROP risk may have introduced bias toward higher risk infants. Furthermore, cohort size differences and differences in timing of imaging between full-term and preterm groups may alter the demographic associations seen in this study. Nonetheless, this study is strengthened by data comparing preterm to full-term infants, 3-dimensional configuration analysis, and the use of newer handheld SS-OCT technology. 

In summary, DSM is more prevalent in preterm infants than full-term infants. Infantile DSM has a round configuration compared to the ridge-shaped configuration previously found in children and adolescents, alluding to a different pathophysiology. Further studies are needed to determine the long-term implications of this finding in the infant population. 

Acknowledgments

Supported by the Alcon Research Institute Young Investigator Award, Violet Sees, Seattle Children's Research Institute Summer Scholars Program, and unrestricted grants from Research to Prevent Blindness and National Institutes of Health (EY00130) to the University of Washington Department of Ophthalmology. 

Disclosure: J.N. Dang, None; J. Wu, None; Y. Zheng, None; J.J. Bunk, None; E.K. Tam, None; K.E. Lee, None; S.E. Lawson, None; T.R. Monger, None; A.T. Legocki, None; J.P. Kelly, None; O. Egeolu, None; L. Ding, None; R.K. Wang, None; K. Tarczy-Hornoch, None; M.T. Cabrera, None