Ophthalmic & Physiological Optics ISSN 0275-5408
A methodological approach for evaluation of foveal immaturity after extremely preterm birth n1, Johan Sjo € strand2, Maria Nilsson1,* and Kerstin Hellgren1,3,* Rebecka Rose 1
Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden, 2Department of Ophthalmology, University of Gothenburg, €lndal, Sweden, and 3Astrid Lindgren Children’s Hospital, Karolinska University Hospital, Stockholm, Sweden Mo
€strand J, Nilsson M & Hellgren K. A methodological approach for evaluation of foveal immaturity after extremely Citation information: Ros en R, Sjo preterm birth. Ophthalmic Physiol Opt 2015; 35: 433–441. doi: 10.1111/opo.12221
Keywords: foveal development, foveal microanatomy, gestational age, optical coherence tomography, retinopathy of prematurity Correspondence: Maria Nilsson E-mail address: [email protected] *These authors contributed equally to this study and should be considered as co-principal authors. Received: 26 February 2015; Accepted: 18 May 2015
Abstract Purpose: To characterize typical microanatomical alterations of immaturity in the fovea, that remain into childhood, after extremely preterm birth before 27 weeks gestational age (GA) and to suggest a clinical methodological evaluation tool. Methods: Subjects were consecutively recruited at age 6.5 years and organized in four groups (10 subjects in each): Group A (full-term), Group B (GA 25–27 weeks) without retinopathy of prematurity (ROP), Group B* (GA 25–27 weeks) and Group C (GA 23–24 weeks) both with ROP stage 3. Foveal microanatomy was studied using a grading system of OCT-scans. Results: Prematurely born children (including Group B, B* and C) had significantly reduced foveal depth (mean difference 53 lm, p < 0.001), thicker inner retinal layer (mean difference 21.6, p = 0.005) and thicker outer nuclear layer (mean difference 23.5, p < 0.001) than controls. Foveal depth and inner retinal layer thickness were significantly correlated to GA (p = 0.003 and p = 0.017 respectively) within the preterm group. Foveal depth increased with 14.1 lm per week between GA 23 and 27. The three hyper reflective bands of the outer segments as well as a central protuberance of inner and outer segment layers were present in all children. Conclusion: Previous studies have revealed signs of immaturity affecting most retinal layers at time of birth in prematurely born children. The present study adds information to which extent these signs of underdevelopment remains to later in life. The applied method showed that premature birth before GA 27 weeks commonly leads to characteristic anatomical alterations of the foveal anatomy expressed as reduced foveal depth and incomplete extrusion of the inner retinal layers. Although deviations of the outer nuclear layers can be seen in the most extremely preterm born children, the outer part of the fovea generally develops well, independent of prematurity. The single most important determinant for the degree of foveal maturation seems to be GA.
Introduction As shown in histopathological studies, the human fovea develops rapidly during early life with a maturation process starting at foetal week 22 continuing until 15–45 month after birth.1 The foveal pit formation is intense between foetal weeks 25 and 37 as a result of displacement of inner
retinal layers (IRL) in the foveal centre (FC) towards the periphery.2 By the use of optical coherence tomography (OCT) foveal maturation has been studied in preterm infants from 32 weeks gestational age (GA) and the foveal pit was present in all infants, appeared deeper and more distinct with increasing age and seemed to reach an adultlike stage at 17 month of age.3 In a similar study histologic
© 2015 The Authors Ophthalmic & Physiological Optics © 2015 The College of Optometrists Ophthalmic & Physiological Optics 35 (2015) 433–441
433
R Rosen et al.
Foveal immaturity after extremely preterm birth
findings were correlated with findings visualised by OCT and a high correlation was found between the methods at all stages of development. The preterm infants had a shallow pit indenting IRL and short, undeveloped foveal photoreceptors. At term, centripetal migration of IRL resulted in a deeper pit, and photoreceptor subcellular structures began to appear.4 The maturation process of the outer retina then continues by displacement of the photoreceptors inwards to increase foveal cone density. The cones also undergo a transformation and their outer segment (OS) decreases in diameter and increases in length (cone specialization). These events start at a few weeks of age and reach near-full to full maturity at 1–4 years of age.2, 4 The maturation of the OS also includes increased thickness and formation of the outer nuclear layer (ONL) at FC. Sj€ ostrand & Popovic5 recently presented a time-line model of developmental events within the human fovea. By merging data based on OCT imaging and histological samples2–4, 6 they presented a model showing how the inner and outer part of the fovea matures within different time frames, the inner retina during the late part of gestation and the outer retina during the first postnatal years. By comparing the model showing normal foveal development with the results from OCT examination of ex-premature eyes (imaged at an age of 5 years or above) it was evident that signs of immaturity of the inner retinal structures were still present in the prematurely born subjects whereas the outer retina seemed to have continued to develop during postnatal years and reached normal or close to normal maturation.7–10 Several factors besides prematurity itself affect the foveal development such as; birth weight, presence or absence of retinopathy of prematurity (ROP) and ROP treatment. It is difficult to investigate the exact contribution of these factors separately since they are closely linked. This problem was therefore addressed in our study by comparing foveal microanatomy in cases born at different GA (23–40 weeks) or at the same GA (25–27 weeks) with and without ROP. Our hypothesis was that the degree of prematurity is the most important determinant for foveal maturation. The first aim of the present study was to characterize typical microanatomical alterations of retinal layers in the fovea, in relation to GA, that remain into childhood after
preterm birth before 27 weeks GA. The second aim was to suggest a methodological clinical approach for identification of foveal immaturity. Materials and methods The study followed the guidelines of the Declaration of Helsinki and was approved by the local ethics committee. Informed consents were obtained from all the children’s parents. Subjects Forty subjects were selected from a national populationbased follow-up study including children born before GA 27 (extremely preterm born; EPT) and age matched fullterm control children.11 Subjects were consecutively recruited and grouped according to GA at birth, i.e., GA 37–42 (full-term), GA 25–27 or GA 23–24. Thereby, 10 typically developing term-born children were recruited to serve as controls (Group A), as well as 10 children born at GA 25–27 (Group B) without ROP and 10 children born at GA 23–24 (Group C). Since all the consecutively recruited children in group C had ROP stage 3, one group including 10 children born at GA 25–27 with ROP stage 3 (Group B*) was added, see Table 1. In Group B* two subjects had received treatment for ROP and in Group C eight subjects. At the time the children (22 boys and 18 girls) underwent the OCT examination they were 6.5 years 2 months. Only OCT examinations resulting in high quality images (detailed description further down) were accepted. All subjects had a visual acuity better than 0.40 logMAR (Snellen 6/15 or 20/50, decimal 0.40) in the examined eye. Retinal imaging and image analysis All subjects (both eyes) underwent macular scanning using a spectral domain OCT (Cirrus HD-OCT version 4.0; Carl Zeiss Meditec, www.zeiss.com/meditec) with the macular cube scan 512 9 128, covering 6 9 6 mm of the retina with the fovea centred. OCT scans with highest signal strength and best image quality were selected resulting in scans from 22 right and 18 left eyes. The signal strength,
Table 1. Demographic characteristics: gestational age, birth weight, Retinopathy of prematurity (ROP) stage and treatment for all four study groups.
Group
Gestational age (weeks days days)
A (n = 10) B (n = 10) B* (n = 10) C (n = 10)
39.6 26.3 26.1 24.2
10 4 5 3
Birth weight (gram)
ROP stage
Any laser treatment
All cases 0 All cases 0 All cases 3 All cases 3
0 0 2 cases 8 cases
3557 859 825 657
486 158 173 94
Group A – controls, Group B – GA 25–27 without ROP, Group B* – GA 25–27 with ROP stage 3, Group C – GA 23–24 with ROP stage 3.
434
© 2015 The Authors Ophthalmic & Physiological Optics © 2015 The College of Optometrists Ophthalmic & Physiological Optics 35 (2015) 433–441
R Ros en et al.
Foveal immaturity after extremely preterm birth
defined automatically by the machine, ranges between 1 and 10, higher value indicates higher signal strength. All scans in this study had signal strength of seven or above. High image quality was defined as images with small or negligible influence by eye movements and/or blinking. One single B-scan was selected for further analysis. This horizontal scan was identified by using anatomical landmarks to ensure that the scan represented the FC. Those landmarks were the foveal reflex when visible and/or where the maximum foveal depression was found as well as the highest peak of the foveal photoreceptor segment layer. The parameters/signs graded in the image analysis were based on normal foveal developing events, i.e., formation of the foveal pit, extrusion of IRL, ONL thickening and OS lengthening as well as separation of the photoreceptor layers.1, 12, 13 The qualitative and quantitative analyses were performed as follows: Qualitative analyses/descriptions: The inner part of the retina was evaluated by estimating the foveal pit formation and extrusion of IRL in the FC. The outer part of the retina was graded by observing the presence or absence of the three hyperreflective bands; external limiting membrane, inner segment/outer segment boundary14, and the inner border of the retinal pigment epithelium (RPE) complex3 together with a peak formation, i.e., protuberance of the ellipsoid zone of the photoreceptor layer (ISe) at the FC. In a few cases the true Henle fibre layer (HFL) and true ONL could be visualized by studying tilted OCT scans.15, 16 Quantitative measurements: All quantitative measurements were performed in a customized MATLAB program and in Image J. Thickness values were expressed in micrometres (lm) (Table 2) and positions were identi-
fied using two landmarks: FC and rim of the foveal pit, i.e., foveal wall maximum (FWM) (see Figure 1a). The different measurements are presented under separate headings. Inner part of the retina The measure describing FD was defined as the difference between the retinal thickness at the FC and the mean retinal thickness of the temporal and nasal rim or FWM, i.e., the difference between retinal thickness at the pit and at the rim in lm. In order to facilitate the qualitative estimation of the FD as above, this difference was also expressed as a ratio. Retinal thickness was measured from internal limiting membrane to the outer border of the RPE (see Figure 1a). Extrusion of the sublayers of the inner retina was graded by measuring the thickness of IRL equal to the distance between inner border of the internal limiting membrane and outer border of IRL at the FC (Figure 1a and b). Outer part of the retina The outer retinal layers (ORL), extending from the border between the IRL and the highly reflective inner border of the outer plexiform layer (iOPL) and the outer border of RPE (Figure 1a), contain the photoreceptors and the RPE. In order to evaluate the microstructural changes during development, the ORL was divided into two sublayers by a segmentation line along the inner border of the ISe layer14, previously defined as the inner and outer segment junction. The thickness of these sublayers, defined as ONL+ and OS+, respectively (Figure 1b) was quantified and described under the headings ONL thickening and OS lengthening (Figure 1).
Table 2. Retinal layer thicknesses in the foveal centre of the four study groups, mean SD Group
FD (lm)
A (B+B*+C) B B* C p-Values A vs (B+B*+C) A vs B B vs B* B* vs C
133.7 80.7 99.1 81.3 61.6
IRL thickness (lm) 27.5 27.8 23.0 14.4 31.2
A methodological approach for evaluation of foveal immaturity after extremely preterm birth n1, Johan Sjo € strand2, Maria Nilsson1,* and Kerstin Hellgren1,3,* Rebecka Rose 1
Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden, 2Department of Ophthalmology, University of Gothenburg, €lndal, Sweden, and 3Astrid Lindgren Children’s Hospital, Karolinska University Hospital, Stockholm, Sweden Mo
€strand J, Nilsson M & Hellgren K. A methodological approach for evaluation of foveal immaturity after extremely Citation information: Ros en R, Sjo preterm birth. Ophthalmic Physiol Opt 2015; 35: 433–441. doi: 10.1111/opo.12221
Keywords: foveal development, foveal microanatomy, gestational age, optical coherence tomography, retinopathy of prematurity Correspondence: Maria Nilsson E-mail address: [email protected] *These authors contributed equally to this study and should be considered as co-principal authors. Received: 26 February 2015; Accepted: 18 May 2015
Abstract Purpose: To characterize typical microanatomical alterations of immaturity in the fovea, that remain into childhood, after extremely preterm birth before 27 weeks gestational age (GA) and to suggest a clinical methodological evaluation tool. Methods: Subjects were consecutively recruited at age 6.5 years and organized in four groups (10 subjects in each): Group A (full-term), Group B (GA 25–27 weeks) without retinopathy of prematurity (ROP), Group B* (GA 25–27 weeks) and Group C (GA 23–24 weeks) both with ROP stage 3. Foveal microanatomy was studied using a grading system of OCT-scans. Results: Prematurely born children (including Group B, B* and C) had significantly reduced foveal depth (mean difference 53 lm, p < 0.001), thicker inner retinal layer (mean difference 21.6, p = 0.005) and thicker outer nuclear layer (mean difference 23.5, p < 0.001) than controls. Foveal depth and inner retinal layer thickness were significantly correlated to GA (p = 0.003 and p = 0.017 respectively) within the preterm group. Foveal depth increased with 14.1 lm per week between GA 23 and 27. The three hyper reflective bands of the outer segments as well as a central protuberance of inner and outer segment layers were present in all children. Conclusion: Previous studies have revealed signs of immaturity affecting most retinal layers at time of birth in prematurely born children. The present study adds information to which extent these signs of underdevelopment remains to later in life. The applied method showed that premature birth before GA 27 weeks commonly leads to characteristic anatomical alterations of the foveal anatomy expressed as reduced foveal depth and incomplete extrusion of the inner retinal layers. Although deviations of the outer nuclear layers can be seen in the most extremely preterm born children, the outer part of the fovea generally develops well, independent of prematurity. The single most important determinant for the degree of foveal maturation seems to be GA.
Introduction As shown in histopathological studies, the human fovea develops rapidly during early life with a maturation process starting at foetal week 22 continuing until 15–45 month after birth.1 The foveal pit formation is intense between foetal weeks 25 and 37 as a result of displacement of inner
retinal layers (IRL) in the foveal centre (FC) towards the periphery.2 By the use of optical coherence tomography (OCT) foveal maturation has been studied in preterm infants from 32 weeks gestational age (GA) and the foveal pit was present in all infants, appeared deeper and more distinct with increasing age and seemed to reach an adultlike stage at 17 month of age.3 In a similar study histologic
© 2015 The Authors Ophthalmic & Physiological Optics © 2015 The College of Optometrists Ophthalmic & Physiological Optics 35 (2015) 433–441
433
R Rosen et al.
Foveal immaturity after extremely preterm birth
findings were correlated with findings visualised by OCT and a high correlation was found between the methods at all stages of development. The preterm infants had a shallow pit indenting IRL and short, undeveloped foveal photoreceptors. At term, centripetal migration of IRL resulted in a deeper pit, and photoreceptor subcellular structures began to appear.4 The maturation process of the outer retina then continues by displacement of the photoreceptors inwards to increase foveal cone density. The cones also undergo a transformation and their outer segment (OS) decreases in diameter and increases in length (cone specialization). These events start at a few weeks of age and reach near-full to full maturity at 1–4 years of age.2, 4 The maturation of the OS also includes increased thickness and formation of the outer nuclear layer (ONL) at FC. Sj€ ostrand & Popovic5 recently presented a time-line model of developmental events within the human fovea. By merging data based on OCT imaging and histological samples2–4, 6 they presented a model showing how the inner and outer part of the fovea matures within different time frames, the inner retina during the late part of gestation and the outer retina during the first postnatal years. By comparing the model showing normal foveal development with the results from OCT examination of ex-premature eyes (imaged at an age of 5 years or above) it was evident that signs of immaturity of the inner retinal structures were still present in the prematurely born subjects whereas the outer retina seemed to have continued to develop during postnatal years and reached normal or close to normal maturation.7–10 Several factors besides prematurity itself affect the foveal development such as; birth weight, presence or absence of retinopathy of prematurity (ROP) and ROP treatment. It is difficult to investigate the exact contribution of these factors separately since they are closely linked. This problem was therefore addressed in our study by comparing foveal microanatomy in cases born at different GA (23–40 weeks) or at the same GA (25–27 weeks) with and without ROP. Our hypothesis was that the degree of prematurity is the most important determinant for foveal maturation. The first aim of the present study was to characterize typical microanatomical alterations of retinal layers in the fovea, in relation to GA, that remain into childhood after
preterm birth before 27 weeks GA. The second aim was to suggest a methodological clinical approach for identification of foveal immaturity. Materials and methods The study followed the guidelines of the Declaration of Helsinki and was approved by the local ethics committee. Informed consents were obtained from all the children’s parents. Subjects Forty subjects were selected from a national populationbased follow-up study including children born before GA 27 (extremely preterm born; EPT) and age matched fullterm control children.11 Subjects were consecutively recruited and grouped according to GA at birth, i.e., GA 37–42 (full-term), GA 25–27 or GA 23–24. Thereby, 10 typically developing term-born children were recruited to serve as controls (Group A), as well as 10 children born at GA 25–27 (Group B) without ROP and 10 children born at GA 23–24 (Group C). Since all the consecutively recruited children in group C had ROP stage 3, one group including 10 children born at GA 25–27 with ROP stage 3 (Group B*) was added, see Table 1. In Group B* two subjects had received treatment for ROP and in Group C eight subjects. At the time the children (22 boys and 18 girls) underwent the OCT examination they were 6.5 years 2 months. Only OCT examinations resulting in high quality images (detailed description further down) were accepted. All subjects had a visual acuity better than 0.40 logMAR (Snellen 6/15 or 20/50, decimal 0.40) in the examined eye. Retinal imaging and image analysis All subjects (both eyes) underwent macular scanning using a spectral domain OCT (Cirrus HD-OCT version 4.0; Carl Zeiss Meditec, www.zeiss.com/meditec) with the macular cube scan 512 9 128, covering 6 9 6 mm of the retina with the fovea centred. OCT scans with highest signal strength and best image quality were selected resulting in scans from 22 right and 18 left eyes. The signal strength,
Table 1. Demographic characteristics: gestational age, birth weight, Retinopathy of prematurity (ROP) stage and treatment for all four study groups.
Group
Gestational age (weeks days days)
A (n = 10) B (n = 10) B* (n = 10) C (n = 10)
39.6 26.3 26.1 24.2
10 4 5 3
Birth weight (gram)
ROP stage
Any laser treatment
All cases 0 All cases 0 All cases 3 All cases 3
0 0 2 cases 8 cases
3557 859 825 657
486 158 173 94
Group A – controls, Group B – GA 25–27 without ROP, Group B* – GA 25–27 with ROP stage 3, Group C – GA 23–24 with ROP stage 3.
434
© 2015 The Authors Ophthalmic & Physiological Optics © 2015 The College of Optometrists Ophthalmic & Physiological Optics 35 (2015) 433–441
R Ros en et al.
Foveal immaturity after extremely preterm birth
defined automatically by the machine, ranges between 1 and 10, higher value indicates higher signal strength. All scans in this study had signal strength of seven or above. High image quality was defined as images with small or negligible influence by eye movements and/or blinking. One single B-scan was selected for further analysis. This horizontal scan was identified by using anatomical landmarks to ensure that the scan represented the FC. Those landmarks were the foveal reflex when visible and/or where the maximum foveal depression was found as well as the highest peak of the foveal photoreceptor segment layer. The parameters/signs graded in the image analysis were based on normal foveal developing events, i.e., formation of the foveal pit, extrusion of IRL, ONL thickening and OS lengthening as well as separation of the photoreceptor layers.1, 12, 13 The qualitative and quantitative analyses were performed as follows: Qualitative analyses/descriptions: The inner part of the retina was evaluated by estimating the foveal pit formation and extrusion of IRL in the FC. The outer part of the retina was graded by observing the presence or absence of the three hyperreflective bands; external limiting membrane, inner segment/outer segment boundary14, and the inner border of the retinal pigment epithelium (RPE) complex3 together with a peak formation, i.e., protuberance of the ellipsoid zone of the photoreceptor layer (ISe) at the FC. In a few cases the true Henle fibre layer (HFL) and true ONL could be visualized by studying tilted OCT scans.15, 16 Quantitative measurements: All quantitative measurements were performed in a customized MATLAB program and in Image J. Thickness values were expressed in micrometres (lm) (Table 2) and positions were identi-
fied using two landmarks: FC and rim of the foveal pit, i.e., foveal wall maximum (FWM) (see Figure 1a). The different measurements are presented under separate headings. Inner part of the retina The measure describing FD was defined as the difference between the retinal thickness at the FC and the mean retinal thickness of the temporal and nasal rim or FWM, i.e., the difference between retinal thickness at the pit and at the rim in lm. In order to facilitate the qualitative estimation of the FD as above, this difference was also expressed as a ratio. Retinal thickness was measured from internal limiting membrane to the outer border of the RPE (see Figure 1a). Extrusion of the sublayers of the inner retina was graded by measuring the thickness of IRL equal to the distance between inner border of the internal limiting membrane and outer border of IRL at the FC (Figure 1a and b). Outer part of the retina The outer retinal layers (ORL), extending from the border between the IRL and the highly reflective inner border of the outer plexiform layer (iOPL) and the outer border of RPE (Figure 1a), contain the photoreceptors and the RPE. In order to evaluate the microstructural changes during development, the ORL was divided into two sublayers by a segmentation line along the inner border of the ISe layer14, previously defined as the inner and outer segment junction. The thickness of these sublayers, defined as ONL+ and OS+, respectively (Figure 1b) was quantified and described under the headings ONL thickening and OS lengthening (Figure 1).
Table 2. Retinal layer thicknesses in the foveal centre of the four study groups, mean SD Group
FD (lm)
A (B+B*+C) B B* C p-Values A vs (B+B*+C) A vs B B vs B* B* vs C
133.7 80.7 99.1 81.3 61.6
IRL thickness (lm) 27.5 27.8 23.0 14.4 31.2
