The ocular shape and retinal structure in children with a history of treated retinopathy of prematurity

Abstract

This study compared the retinal structure between the eyes of children with a history of treated retinopathy of prematurity (ROP), spontaneously regressed ROP and full-term children using widefield optical coherence tomography angiography (OCTA) and swept-source OCT (SSOCT). We identified patients who were diagnosed with ROP and followed up. Age-matched full-term controls were also recruited. We reviewed the medical records of 34 ROP eyes (34 patients) treated with laser photocoagulation, 5 spontaneously regressed ROP eyes (5 patients), and 15 age-matched full-term controls (15 control eyes). The thickness of the foveal retina and choroid, the foveal avascular zone (FAZ), and ocular shape were assessed using OCTA and SSOCT (Xephilio OCT-S1) images. The mean foveal thickness of the treated ROP eyes was significantly thicker than that of the control eyes. The FAZ of the treated ROP eyes was significantly smaller than that of the control eyes. SSOCT images revealed that 5 of 34 eyes in the treated ROP group had a dome-shaped macula, while 18 had an elongated barrel shape. Understanding these structural changes is crucial for developing targeted treatments and optimizing visual function in this population.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-15271-7.

Introduction

Advancements in perinatal care have enhanced the management of preterm and low-birth-weight infants¹. Consequently, the number of retinopathies of prematurity (ROP) cases requiring treatment has also risen. While there have been reports on retinal thickness following ROP treatment, there is a lack of studies on ocular shape and morphology.

Previous studies have identified the retinal structures of preterm babies via spectral-domain optical coherence tomography (SDOCT). Specifically, the inner retinal layer structure is preserved in preterm children, and their central foveal depression is shallower than that in full-term children^2–5. However, there are no reports on the macular shape in treated ROP using swept-source OCT (SSOCT) or widefield optical coherence tomography angiography (OCTA).

OCTA has also been used to investigate vascular features in infants with ROP. From previous studies, the foveal avascular zone (FAZ) area was significantly smaller in ROP patients than in preterm children^6,7. Mataftsi et al. reported that the foveal vessel density in the superficial capillary plexus and deep capillary plexus was significantly greater in patients with regressed ROP than in preterm patients⁷. The present study evaluates the ocular shape and FAZ of treated ROP using widefield OCTA and SSOCT.

Results

Demographics of the patients

We reviewed the medical records of 34 ROP eyes (34 patients) treated with laser PC, 5 spontaneously regressed ROP eyes (5 patients), and 15 age-matched full-term controls (15 control eyes; Table 1). No significant differences were observed between the groups in terms of age or sex. The mean BCVA in the PC-treated, regressed ROP, and control groups was 0.07 ± 0.16 logMAR units, 0 ± 0.12 logMAR, and − 0.08 ± 0.04 logMAR units. The mean spherical equivalent in the PC-treated, regressed ROP, and control groups was − 1.2 ± 3.0 D, −0.25 ± 0.98 D, and − 0.7 ± 4.3 D. The mean axial length in the PC-treated, regressed ROP, and control groups was 22.8 ± 2.0 mm, 22.4 ± 0.2 mm, and 22.8 ± 0.2 mm, respectively. No significant differences were found in the spherical equivalent and axial length.

Table 1.

Demographic characteristics of the patients and control.

	ROP-treated eyes (n = 34)	ROP-spontaneously regressed eyes (n = 5)	Control eyes (n = 15)	P value
Assessment age (years)	7.3 ± 1.9	7.3 ± 1.9	8.0 ± 1.2	0.632*
Sex (male/female)	16/18	1/4	10/5	0.155†
Gestational age (week)	25.1 ± 1.8	29.4 ± 2.3	39.7 ± 0.8	0.001*
Birth weight (g)	702.8 ± 201.5	1001.4 ± 545.1	2603.3 ± 357.3	0.001*
Best-corrected visual acuity (logMAR)	0.07 ± 0.16	0 ± 0.12	−0.08 ± 0.04	0.001*
Spherical equivalent (D)	−1.00 ± 3.12	−0.25 ± 0.98	−0.73 ± 4.27	0.66*
Axial length (mm)	22.8 ± 2.0	22.4 ± 0.2	22.8 ± 0.2	0.80*
ROP stage 1/2/3	0/0/34	5/0/0	0/0/0	0.001†

Data are expressed as the mean (standard deviation).

P value: ^*t test (ROP treated eyes vs. Control eyes)test; ^†χ² test(ROP treated eyes vs. Control eyes).

logMAR: logarithm of the minimum angle of resolution.

SSOCT and widefield OCTA parameters

The mean foveal thickness in the PC-treated, regressed ROP, and control groups was 242.2 ± 38.6 μm, 207.8 ± 34.5 μm and 203.2 ± 22.6 μm, respectively (Table 2). The mean foveal thickness in the PC-treated patients was significantly thicker compared to that in the control children. The FAZ in the PC-treated, regressed ROP and control groups was 0.16 ± 0.06 mm², 0.19 ± 0.11 mm², and 0.39 ± 0.14 mm², respectively. The FAZ in the PC-treated patients was significantly smaller compared to that in the control children (Table 2). SSOCT images indicated that 7 of 34 eyes in the treated ROP group had a dome-shaped macula, while 16 had an elongated barrel shape. The OSI was significantly lower in the PC-treated patients than in the control children (Table 2). A smaller OSI shows a more dome-shaped macula in PC-treated ROP eyes.

Table 2.

OCT findings.

	ROP-treated eyes (n = 34)	ROP-spontaneously regressed eyes (n = 5)	Control eyes (n = 15)	P value
Foveal thickness (µm)	242.2 ± 38.6	207.8 ± 34.5	203.2 ± 22.6	< 0.0011
FAZ(mm2)	0.16 ± 0.06	0.19 ± 0.11	0.39 ± 0.14	< 0.0011
Dome-shaped macula	7/34	0/5	0/13	0.0732
Elongated with a barrel shape	16/34	0/5	0/13	< 0.0012
The ocular shape index	1.2 × 10−5 ± 5.4 × 10−4	1.9 × 10−5 ± 6.5 × 10−4	7.1 × 10−4 ± 10.6 × 10−4	0.0091

Data are expressed as the mean (standard deviation).

P value: ¹t test (ROP treated eyes vs. Control eyes)test; ²χ² test (ROP treated eyes vs. Control eyes).

FAZ: Foveal avascular zone.

We analyzed the relationship between the FAZ, OSI, and OCT parameters. Analysis of covariance (ANCOVA) adjusted for gestational age and birth weight revealed that the FAZ in the PC-treated group was smaller than that in the control group (Table 3). In the Pearson analysis, the FAZ correlated with gestational age and foveal thickness (Table 4). In the Pearson analysis, the better visual acuity was correlated with later gestational age, larger birth weight and larger FAZ (Table 5). In the multiple regression analysis, the smaller OSI was related to the smaller gestational age (Table 6).

Table 3.

Effect of laser treatment on the retinal structure.

Adjiusted mean (thickness, µm) (95% CI)	ROP-treated eyes (n = 34)	ROP-spontaneously regressed eyes (n = 5)	Control eyes (n = 15)	p-value
FAZ	0.16 (0.13 to 0.19)	0.19 (0.13 to 0.26)	0.40 (0.36 to 0.44)	0.004
The ocular shape index	1.3 × 10−5 (−1.5 × 10−4 to 4.1 × 10−4)	1.9 × 10−5 (−6.1 × 10−4 to 9.8 × 10−4)	7.1 × 10−4 (2.8 × 10−4 to11.1 × 10−4)	0.07

Data are expressed as the adjusted mean based on the analysis of covariance (ANCOVA). CI: confidence interval.

p-value: ANCOVA adjusted for gestational age, birth weight.

ROP: retinopathy of prematurity.

FAZ: Foveal avascular zone.

Table 4.

Pearson correlation of patients and controls.

FAZ (µm2)
Patients and controls (n = 54)	r	P value
Gestational age (week)	0.717	< 0.001
The ocular shape index	0.2	0.164
Foveal thickness (µm)	−0.456	0.001

FAZ: Foveal avascular zone.

Table 5.

Pearson correlation of patients and controls.

Best-corrected visual acuity (logMAR)
Patients and controls (n = 54)	r	P value
Gestational age (week)	−0.459	0.001
Birth weight (g)	−0.405	0.004
Foveal thickness (µm)	0.286	0.053
FAZ (mm2)	−0.352	0.015

FAZ: Foveal avascular zone.

Table 6.

Multiple linear regression analysis between OSI and independent variables.

Independent variable
	Standardized β	P value
Gestational age	0.636	0.032
Retinal thickness FAZ	0.328 −0.133	0.085 0.503
Laser treatment	−0.042	0.857

FAZ: Foveal avascular zone.

Representative ROP patient

Case 1: Dome-shaped macula (Fig. 1A).

Fig. 1.

Upper image: SSOCT image. Middle image: OCT-A image. Lower image: The ocular shape index was defined as the coefficient of determination of the quadratic curve equation. (a) Optical coherence tomography (OCT) images of the treated dome-shaped macular retinopathy of prematurity (ROP). Case 1 was a 9-year-old girl born at 24 weeks of gestation (weighing 784 g). FAZ was 0.20 mm². The ocular shape showed a dome-shaped macula, with an OSI of −2.0 × 10⁻⁴. (b) OCT image of a representative eye of a treated elongated barrel-shaped ROP. Case 2 was a 9-year-old boy born at 26 weeks of gestation (weighing 900 g). FAZ was 0.11 mm². The ocular shape showed an elongated barrel-shaped macula, with an OSI of −6.8 × 10⁻⁵. (c) OCT image of the control eye. Case 3 was an 8-year-old girl. FAZ was 0.67 mm². The OSI was 3.6 × 10⁻⁴.

Case 1 was a 9-year-old girl whose gestational age was 24 weeks (weighing 784 g). She had Zone II Stage III ROP without plus disease; she had no additional complications, such as severe retinal vascular dilation or tortuosity. PC therapy was performed to treat the ROP. Her BCVA was 0 logMAR in both eyes. Her axial length was 23.39 mm, SE was − 6 D, and FAZ was 0.20 mm². In terms of ocular shape, she had a dome-shaped macula with an OSI of −2.0 × 10⁻⁴.

Case 2: Barrel shaped elongated macula(Fig. 1B).

Case 2 was 9-year-old boy whose gestational age was 26 weeks (weighing 900 g). He had Zone II Stage III ROP without plus disease. PC therapy was performed to treat the ROP. His BCVA was 0 logMAR in both eyes. His axial length was 24.06 mm, SE was − 5.25 D, and FAZ was 0.11 mm². In terms of ocular shape, he had an elongated barrel-shaped macula with an OSI of −6.8 × 10⁻⁵.

Case 3: Control eye (Fig. 1C).

Case 3 was an 8-year-old girl. Her BCVA was 0 logMAR in both eyes. Her axial length was 22.70 mm, SE was − 5 D, and FAZ was 0.67 mm². Her OSI was 3.6 × 10⁻⁴.

Discussion

We revealed that PC-treated ROP eyes had a thicker foveal retinal thickness, a smaller FAZ, a smaller OSI, a dome-shaped and an elongated barrel shape compared to control eyes. And there was no significantly difference in the spherical equivalent between the ROP and control eyes.

The present study revealed that the shape of eyes treated for ROP differed from that of regressed ROP eyes and normal eyes and often had a barrel-shaped or dome-shaped macula. Barrel-shaped ocular morphology is commonly observed in highly myopic eyes. Myopic eyes with treated ROP typically have shorter ocular axis lengths than normal myopic eyes; differences in the shape of the anterior segment of the eye are thought to be the cause. Previous longitudinal studies revealed that PC-treated eyes showed a significantly shallower anterior chamber depth, thicker lens, and shorter axial length than age-matched control eyes⁸. In severely elongated barrel-shaped eyes, PVD progresses with irregular vitreoretinal adhesion at the sites of the retinal vessels⁹. Several studies have shown that ROP patients who underwent laser treatment exhibited higher levels of myopia than those exhibited by ROP patients without such a history, premature births without ROP, and full-term controls; however, these differences were not statistically significant^10,11. The eyes treated for ROP demonstrated progressive myopia, likely due to the effects of prematurity.

The Gutenberg Prematurity Eye Study examined a total of 755 eyes and determined that central foveal thickness and foveal hypoplasia were both independently associated with gestational age¹². Furthermore, previous studies have demonstrated that this foveal configuration, along with prematurity, is linked to a reduced FAZ size^13–15. Additionally, prematurity has been associated with a small FAZ, and our reports indicate that a smaller FAZ is correlated with shorter gestational periods.

We analyzed ocular shape using the OSI and discovered that treated ROP eyes exhibited a smaller OSI compared to control eyes. Numerous reports have utilized SSOCT to study treated ROP¹⁶; however, there have been no reports where ocular shape was objectively measured. Eyes treated for ROP exhibited either a dome-shaped or barrel-shaped macula, whereas control eyes typically exhibited a conical shape. We objectively measured ocular shape and found that it differed between treated ROP eyes, regressed ROP eyes, and control eyes. Previous reports measured the macular shape of the normal school-aged children and found that the macular irregularities develop before the age of school-age¹⁷. We also determined that a smaller OSI was associated with a smaller gestational age rather than laser treatment. In addition, the better visual acuity was correlated with later gestational age, larger birth weight and larger FAZ. Immature baby showed severe ROP and must be treated by laser treatment. Fetal immaturity than laser treatment impacts foveal shape, resulting in a smaller OSI.

This study has some limitations. The sample size was small. A comparison between treated ROP eyes and those with spontaneous resolution of ROP is planned for the future.

In conclusion, we compared the retinal structures of the eyes of children with treated ROP and control children using OCTA and SSOCT. We found that treated ROP eyes had thicker foveal retinal thickness, a smaller FAZ, and an elongated barrel shape compared to the control eyes. Understanding these structural changes is crucial for developing targeted treatments and optimizing visual function in this population.

Methods

This study was conducted at Nara Medical University Hospital. It was a single-center study and retrospective in nature. The study included patients diagnosed with ROP and treated at Nara Medical University Hospital. The study protocol adhered to the Declaration of Helsinki and was approved by the Nara Medical University Ethics Committee. Due to the retrospective nature of the study, written informed consent was waived by the Nara Medical University Ethics Committee. The data were stored in a deidentified format at Nara Medical University and retrospectively retrieved.

Participants in this study were children aged 4 to 12 years diagnosed with ROP who had received treatment. Only those with available OCT images who cooperated during the examination were included in the study. Patients with eye diseases were excluded. The entire retinal avascular area was subjected to laser photocoagulation (PC) therapy in accordance with the diagnostic criteria of the Early Treatment for Retinopathy of Prematurity (ET-ROP) study¹⁸. ROP was monitored at least once per week during the acute phase. The patients who received PC therapy were treated in all quadrants of the eyes. Some participants presenting with stage ≤ 3 disease had ROP that regressed spontaneously without receiving any ROP treatment. Equivalent spherical power testing was performed using a KR-8100 instrument (TOPCON) following mydriasis induction. Best-corrected visual acuity (BCVA) was measured by applying a Snellen chart and converting to the logarithm of the minimum angle of resolution (logMAR) for statistical analysis. An intraocular lens (IOL) Master (Carl Zeiss Meditec, Jena, Germany) was used to evaluate the axial length. The gestational age, birth weight, and sex of the participants were investigated.

We presented data as means ± standard deviation. We compared the gestational age at birth, age at the evaluation, BCVA, spherical equivalent, axial length, and ocular shape of the eyes between PC-treated ROP patients, and the control children using t test. The existence of dome-shaped macula and barrel-shaped macula was compared using the χ² test. The sample size of the spontaneously regressed ROP group was small, so we present the data descriptively only in t-test and χ² test. In t test, gestational age and birth weight were significantly different between two groups. So we selected the gestational age and birth weight as potential confounders of ANCOVA. The correlative significance between the FAZ and ocular findings in ROP eyes was assessed using univariate and multivariate linear regression analyses. Multivariate linear regression analyses were performed using potential confounders (gestational age, birth weight, BCVA, laser treatment, and ocular shape index (OSI)). The standardized coefficients (ß) were computed for the independent variable. A p-value of < 0.05 was considered statistically significant. All statistical analyses were performed using licensed statistical software (SPSS version 22.0; SPSS Inc., Chicago, IL, USA). The intra-rater agreement for the OSI measurement was an intraclass correlation coefficient of 0.800 (95% confidence interval: 0.641–0.896).

We used a swept-source OCT (Xephilio OCT-S1; Canon, Tokyo, Japan). The OCT-S1 captures cross-sectional images measuring 23 × 20 mm (80 × 70°). We evaluated only the vertical scans (20 mm) through the fovea. The lines for the retinal pigment epithelium were manually drawn using the line tool in ImageJ software (National Institutes of Health, Bethesda, MD). Tomographic images from the 20 mm vertical scan were exported to ImageJ, and lines were manually drawn on the posterior surface of the retinal pigment epithelium and chorioscleral junction. The lines were drawn by two specialists (T.N. and Y.M.). The line of the RPE is plotted at six points within 3000 μm. The ocular shape index was defined as the coefficient of determination of the quadratic curve equation. (Fig. 2). In the present study, we analyzed vertical images across the fovea using a previously reported method¹⁹. From previous report, dome-shaped macula was effectively found with vertical OCT images than with horizontal OCT images¹⁹. The OSI was calculated by two of this report’s authors, and the average was determined. From the method, a larger OSI shows a more conical-shaped macula, whereas a smaller OSI shows a more dome-shaped macula. The elongated barrel-shaped and dome-shaped macula were independently evaluated by two specialists (T. N. and Y. M.). The FAZ segmentation was manually delineated by two specialists (T. N. and Y. M.). The FAZ was measured by two of this report’s authors, and the average was determined.

Fig. 2.

Upper image: SSOCT image. The line of the RPE is plotted at six points within 3000 μm. Lower image: The ocular shape index was defined as the coefficient of determination of the quadratic curve equation. (a) OSI was − 1.1 × 10⁻⁴. The ocular shape was dome shaped. (b) OSI was 1.3 × 10⁻⁴. The ocular shape was conical.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

T.N. and T.U. wrote the manuscript. T.N. and Y.M. collected the data. T.N., H.T., T.S. and N.O. revised the manuscript. All authors have reviewed the manuscript and agree to its publication.

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.