Content area
To quantify the fundus color on digital color fundus photographs as a surrogate of myopic chorioretinal degeneration in myopic subjects using a novel computer software program. In this cross-sectional study, the following details were recorded in 152 eyes of 78 myopic subjects: Snellen visual acuity, refractive status, ocular biometric parameters, intraocular pressure, color fundus photography, and myopic chorioretinal degeneration index (MCDI) using an automated computer program. The mean age of the subjects was 32.58 ± 11.12 years (range, 21-62 years). Subjects with aged ≥40 years had a significantly higher MCDI when compared with subjects with aged <40 years (MCDI: age ≥ 40 = 0.519; age < 40 = 0.462, P = 0.000). The MCDI demonstrated a statistically significant and positive relationship with axial length (r = 0.47, P = 0.000) and lens thickness (r = 0.26, P = 0.023). A significant positive relationship was observed between the MCDI and visual acuity (r = 0.30, P = 0.007). The MCDI represents an objective technique for quantitative assessment of chorioretinal degenerative changes in myopic subjects. Older myopic subjects with a relatively higher MCDI and longer axial length may represent a high-risk population who have an increased likelihood of developing visually consequential complications of myopia.[PUBLICATION ABSTRACT]
Int Ophthalmol (2012) 32:203209 DOI 10.1007/s10792-012-9542-4
ORIGINAL PAPER
Quantitative analysis of myopic chorioretinal degeneration using a novel computer software program
Kumari Neelam Rebecca Y. K. Chew
Martin H. K. Kwan Chee Chew Yip
Kah-Guan Au Eong
Received: 13 August 2011 / Accepted: 23 February 2012 / Published online: 6 April 2012 Springer Science+Business Media B.V. 2012
Abstract To quantify the fundus color on digital color fundus photographs as a surrogate of myopic chorioretinal degeneration in myopic subjects using a novel computer software program. In this cross-sectional study, the following details were recorded in 152 eyes of 78 myopic subjects: Snellen visual acuity, refractive status, ocular biometric parameters, intraocular pressure, color fundus photography, and myopic chorioretinal degeneration index (MCDI) using an automated computer program. The mean age of the subjects was 32.58 11.12 years (range, 2162 years). Subjects with aged C40 years had a signicantly higher MCDI when compared with subjects with aged\40 years (MCDI: age C 40 = 0.519;
age \ 40 = 0.462, P = 0.000). The MCDI demonstrated a statistically signicant and positive relationship with axial length (r = 0.47, P = 0.000) and lens thickness (r = 0.26, P = 0.023). A signicant positive relationship was observed between the MCDI and
visual acuity (r = 0.30, P = 0.007). The MCDI represents an objective technique for quantitative assessment of chorioretinal degenerative changes in myopic subjects. Older myopic subjects with a relatively higher MCDI and longer axial length may represent a high-risk population who have an increased likelihood of developing visually consequential complications of myopia.
Keywords Myopia Myopic chorioretinal
degeneration Ocular biometric parameters
RGB values
Introduction
High or pathological myopia is a clinically distinct entity characterized by excessive and progressive elongation of the axial length with concomitant degenerative changes in the posterior segment of the globe. It is one of the major causes of legal blindness in the 40- to 50-years age group [1, 2] and, in fact, is the second commonest cause of blindness in the Chinese population. [3].
The degenerative changes at the posterior pole in highly myopic eyes include chorioretinal degeneration, choroidal neovascularization, posterior staphyloma and lacquer cracks. Early stages of chorioretinal degeneration are characterized by thinning of the retinal pigment epithelium (RPE) and choriocapillaris, which results in increased prominence of the larger
K. Neelam R. Y. K. Chew (&) C. C. Yip
Department of Ophthalmology and Visual Sciences, Khoo Teck Puat Hospital, 90 Yishun Central, Singapore 768828, Singaporee-mail: [email protected]
R. Y. K. Chew M. H. K. Kwan
Center for International Studies, Pennsylvania College of Optometry, Philadelphia, PA, USA
K.-G. A. EongSingapore International Eye Cataract Retina Center, Mount Elizabeth Medical Center, Singapore, Singapore
123
204 Int Ophthalmol (2012) 32:203209
choroidal vessels and altered fundus color (the so-called tessellated fundus). [4]. Currently, there is no objective way of quantifying these early chorioretinal degenerative changes in subjects with high myopia.
Suzuki has quantied the fundus color of patients with VogtKoyanagiHarada (VKH) disease by estimating the sunset glow index using digital photographs and an image analysis software program. [5]. He showed that the altered fundus color in VKH-affected eyes is associated with a signicant increase in the sunset glow index, and this index continued to increase with progressive de-pigmentation of the fundus during the course of the disease. He suggested the sunset glow index as a useful technique for quantitative assessment of fundus color in healthy and diseased eyes.
Since progression of myopic chorioretinal degeneration is associated with a change in the color of the fundus clinically, an index similar to the sunset glow index could potentially be useful in quantifying this color change as a surrogate of myopic chorioretinal degeneration. This study aimed to quantify the fundus color in myopic eyes using the myopic chorioretinal degeneration index (MCDI), using a novel computer software program.
Methods
In this cross-sectional study, 78 myopic subjects were recruited from responders to poster advertisements and by word of mouth over a 6-month period. The study was conducted in the Department of Ophthalmology and Visual Sciences in Alexandra Hospital, Singapore, after ethics approval was obtained from the institutional review board. Its protocol adhered to the tenets of the Declaration of Helsinki.
Subjects were eligible to participate if they were C21 years or older, ethnic Chinese, and have spherical
equivalent of -0.5 diopter (D) in at least one eye. Each subject underwent comprehensive eye examination using slit-lamp biomicroscopy by a single trained ophthalmologist for verication of eligibility. Subjects with only clear lenses (nuclear colour/opalescence 01, lens opacities classication system III) were included into the study. Furthermore, subjects with a history of ocular co-morbidity, such as glaucoma and diabetic retinopathy, or previous ocular surgery were excluded from the study.
After informed written consent was obtained, the following data were collected in each study participant: best-corrected visual acuity using Snellen chart, refraction with autorefractor, ocular biometric parameters, intraocular pressure (IOP), and 50-degree color fundus photograph of the posterior pole following pupil dilation using Zeiss fundus cameraTM (Carl Zeiss Meditec AG, Germany). The fundus photographs in all study subjects were captured by a single study optometrist using the same digital fundus camera with standardized settings, which were strictly adhered to during the entire study period.
The fundus color was quantied on the fundus photograph using the MCDI with the aid of image analysis software that displays red, green, and blue (RGB) pixels under full color conditions. To estimate the MCDI, a rectangular area of 1,000 9 500 pixels superotemporal to the optic disc, was rst selected (Fig 1). This area was chosen within the posterior pole but outside the fovea by measuring 50 pixels superior and temporal to the optic disc margin. Next, a Retinal Vessel Extraction Program was then used to selectively identify the RGB pixels corresponding to the blood vessels within the selected area. These RGB pixels were then subtracted from the RGB pixels of the entire rectangular area. This was followed by construction of the red, green, and blue histograms of the remaining pixels in the rectangular area.
The MCDI (mean red intensity) was calculated as follows:
MCDI
Mean red intensityMean red intensity + Mean green intensity + Mean blue intensity
123
Int Ophthalmol (2012) 32:203209 205
Data analysis
The data were analyzed using the SPSS statistical software package (SPSS Inc., Chicago, IL, USA). Since there was a strong correlation between the ocular biometric parameters (axial length: r = 0.85; anterior chamber depth: r = 0.88; lens thickness: r = 0.89) of the two eyes in each subject, for the purpose of reporting the relationships between the MCDI and ocular biometric parameters, we used data from the right eyes only. Correlations were evaluated by linear regression analysis and multiple linear regression analysis were was used to assess cross-sectional relationships. All P values were considered statistically signicant when these values were less than 0.05.
Results
In this study, weincluded 152eyes of 78 myopic subjects (74 subjects = both study eyes; 4 subjects = one study eye because fellow eyes were having advanced myopic macular degeneration). All study participants had early stagesofmyopic chorioretinaldegeneration,i.e.,varying degrees of choroidal pallor and tessellation; however,
there were no geographic areas of chorioretinal atrophy, choroidal neovascularization, and macular retinoschisis, at the fovea in any of the study eyes.
The mean age of the study population was32.58 11.12 years (range, 2162 years). Of the 78 subjects, 42 (53.84 %) were males and 36 (46.16 %) were females. High myopia, dened as refractive error of -6.0 D and above, was present in 38 subjects(48.71 %). Moderate (-3.00 to -5.75 D) and low (-0.50 to -2.75 D) myopia were present in 22(28.20 %) and 18 (23.07 %) subjects, respectively.
MCDI and characteristics of study participants
The mean MCDI for the entire study population was0.473 0.03 and 0.474 0.03 in the right and left eye, respectively. There was a good degree of interocular agreement in the MCDI with a maximum right left eye difference of 0.106 (r = 0.76, P = 0.000). A statistically signicant and positive relationship was observed between the MCDI and age (r = 0.62, P = 0.000, Fig. 2). Subjects aged C40 years had a signicantly higher MCDI when compared with those aged \40 years old [0.519 (age C 40 years) vs0.462 (age \ 40 years), P = 0.000]. No signicant
Fig. 1 A colour fundus photograph showing red, green and blue pixels using image analysis software. The selected rectangular area superotemporal to optic disc is enclosed within the dotted line
123
206 Int Ophthalmol (2012) 32:203209
difference in the mean MCDI was observed between males and females [0.477 (males) vs 0.466 (females), P = 0.18].
MCDI and spherical equivalent
The mean spherical equivalent of the study population was -5.80 4.19 and -5.16 4.22 D, with a maximum of -19.00 and -19.25 D, in the right and left eye, respectively. A statistically signicant inverse relationship was observed between the MCDI and spherical equivalent (r = -0.43, P = 0.000). Subjects with high myopia demonstrated a signicantly higher MCDI when compared with subjects with low and moderate myopia [0.489 (high mopia) vs 0.459 (moderate and low myopia), P = 0.000].
MCDI and ocular biometric parameters
The mean axial length was 25.89 1.78 and25.76 1.74 mm in the right and left eye, respectively. A statistically signicant and positive relationship was observed between the MCDI and axial length (r = 0.47, P = 0.000, Fig. 3), and this relationship remained statistically signicant after controlling for age (P = 0.002).
The mean anterior chamber depth was 3.67 0.28 and 3.66 0.28 mm in the right and left eye, respectively. There was no association between the MCDI and anterior chamber depth (P = 0.70).
The mean lens thickness was 3.75 0.29 and3.74 0.33 mm in the right and left eye, respectively. A statistically signicant and positive relationship was observed between the MCDI and lens thickness (r = 0.26, P = 0.023), and this relationship remained statistically signicant after controlling for age (P = 0.033).
MCDI and visual acuity
The mean best-corrected visual acuity of the study population was 6/7.5 and 6/6 in the right and left eye, respectively. The MCDI demonstrated a statistically signicant positive relationship with visual acuity (r = 0.30, P = 0.007).
MCDI and intraocular pressure
The mean IOP was 15.21 3.82 and 15.08 3.49 mmHg in the right and left eye, respectively. No signicant relationship was observed between the MCDI and IOP (r = 0.19, P = 0.12).
Discussion
In this study, we quantied the fundus color as a surrogate of myopic chorioretinal degeneration in 152 eyes of 78 myopic subjects by estimating the MCDI with the aid of a novel computer software program and digital fundus photographs. The results demonstrate that the MCDI was signicantly higher in subjects who were older and in subjects with longer axial length, thicker lens and worse visual acuity.
The color of the ocular fundus depends on the architecture of its layers and the optical properties and quantities of the pigment (mainly melanin) residing in the RPE-choroidal complex. The standard color
Fig. 2 Scatterplot showing the relationship between the myopic chorioretinal degeneration index (MCDI) and age
Fig. 3 Scatterplot showing the positive relationship between the myopic chorioretinal degeneration index (MCDI) and axial length
123
Int Ophthalmol (2012) 32:203209 207
fundus photograph provides the RGB image of the fundus, which is a M 9 N 9 3 array of color pixels, where each color pixel is a triplet corresponding to red, green and blue components of the RGB image at a specied location. [6]. Since RGB values of each pixel in the fundus image provide the index to the model of coloration, the fundus color of a digital image can be predicted using indices based on RGB values.
We chose a rectangular area of 1,000 9 500 pixels within the posterior pole but outside the fovea for estimation of the MCDI in our study. Since the model of coloration is parameterized by the concentrations of the main absorbers in the fundus (melanin in the RPE, choroidal melanin and hemoglobin in the choroid) and assume that all other properties of the tissues are consistent, it is believed that RGB values estimated at the perifoveal region may accurately reproduce the reectance spectra. Furthermore, the foveal region was avoided due to the presence of macular pigment at the human macula, which will attenuate the reectance in the blue spectral regions with negligible effect in the red region.
The data from our study demonstrate that myopic subjects with relatively longer axial length have a signicantly higher MCDI. Since excessive globe elongation is one of the important factors for the development of pathological changes in myopic subjects, eyes with longer axial length will exhibit greater severity of chorioretinal degenerative changes, and consequently, greater alteration of fundus color and a higher MCDI. Indeed, histopathologic studies have demonstrated that chorioretinal degeneration in early stages is associated with atrophy of the RPE layer and loss of choroidal melanocytes, both of which are capable of altering the RGB values of a digital image [7].
The ndings of this study are similar to Suzukis report that altered fundus color due to progressive hypopigmentation in patients with VKH disease was associated with a higher sunset glow index [5]. The investigator proposed that an increase in the sunset glow index might have occurred due to altered RGB values that resulted from de-pigmentary changes in the choroidal melanocytes appearing 26 months after the disease onset. In addition, this index was capable of identifying subtle de-pigmentation of fundus that may remain undetected during routine fundus examination.
Our study also demonstrated a signicantly higher MCDI with increasing age in myopic subjects. The
exact reason for this observation is not known but one possible explanation is that age-related structural changes in the RPE-choroid complex may alter the RGB values, and consequently, resulted in a higher MCDI. These changes include reduction of melanin granules and increase in lipofuscin within the RPE cell layer, decreased thickness of the choroidal layer, and slow lling of the choroidal vasculature [4]. Alternatively, a signicantly higher MCDI with increasing age may reect a higher prevalence of chorioretinal degenerative changes in older myopic subjects, a nding that is in full agreement with past studies where an age-related trend in the prevalence of myopic retinopathy was observed [8, 9].
Myopia progression is mainly derived by an increase in axial length and vitreous chamber depth, with minimal changes in the anterior segment of the eye [10, 11]. Although lens thickness does not appear to be related to the degree of myopia, the lens may thin during a period of rapid progression of myopia and then thicken as myopic progression begins to slow, with variation by age [12]. We observed a signicant and positive relationship between the MCDI and lens thickness. This observation suggests that degenerative myopic changes in the retina are associated with secondary changes in the lens, in the form of an increase in lens thickness. Given that there is a linear increase in lens thickness with aging [13], disease processes associated with myopia may have exaggerated the age-related changes in the crystalline lens.
Furthermore, a signicantly higher MCDI was observed in subjects with worse visual acuity when compared with subjects with relatively better visual acuity. Since a higher MCDI suggests greater alteration of fundus color due to probably more severe chorioretinal degenerative changes, it is not surprising to have a positive relationship betweeen the MCDI and visual acuity as observed in our study. And by extension, the MCDI may represent an alternative indicator of visual function in subjects with high myopia by acting as a surrogate marker for underlying myopic chorioretinal degeneration.
We also conducted the testretest exercise in 6 subjects where fundus photographs were captured during two sessions separated in time by at least 24 h, but no more than 2 weeks, for estimating the MCDI. The results demonstrated a high degree of correlation between the MCDI at two sessions within an individual, with a maximum difference of 0.013 (mean
123
208 Int Ophthalmol (2012) 32:203209
MCDI: Session 1 = 0.420; Session 2 = 0.415, r2 =
0.958). These observations suggest that the intra-individual variability is reassuringly small.
It is possible that morphological alterations in the architecture of the layers and/or pigment in the RPE-choroidal complex will lead to alteration of fundus color and RGB values at an early stage. The reason is that RGB values when plotted on an orthogonal axis occupy a well-dened volume bounded by physiologically meaningful axes, and these axes includes the concentration of melanin in the RPE, blood in the choroid, and melanin in the choroidal melanocytes. [14]. Therefore, degenerative changes in the RPE-choroidal complex associated with myopia may be quantied using the MCDI that is based on RGB values.
It is important to note that we included subjects with early stages of myopic chorioretinal degeneration, i.e., varying degrees of choroidal pallor and tessellation, the reason being that it is difcult to quantify any progression in choroidal pallor and tessellation on routine fundus examination and/or fundus photograph. The authors believe that worsening of early myopic chorioretinal degeneration will perhaps be pregured clinically by a change in the proposed index.
A higher MCDI, reecting greater degree of chorioretinal degenerative changes, in older myopic subjects with relatively longer axial length may suggest an increased likelihood for developing visually consequential complications, such as geographic areas of chorioretinal atrophy. In research settings, the MCDI may represent an objective technique to quantify and/or monitor the progression of chorioretinal degenerative changes in the myopic population, and therefore, could potentially be invaluable for interventional studies on myopia where prevention or arrestation of the degenerative process at the macula is the primary aim.
This is the rst study to quantify fundus color in subjects as a surrogate of myopic chorioretinal degeneration using the MCDI generated from a novel computer software program. The main limitations of this study are small sample size, lack of control group, and its cross-sectional nature. Although we have restricted subjects to Chinese ethnicity, the results of this study may be confounded by variation in the fundus color among subjects with similar ethnic
origin, due to differences in the melanin content within the RPE-choroidal complex.
In conclusion, our study has demonstrated that the MCDI represents an objective technique for quantitative assessment of underlying chorioretinal degenerative changes in myopic subjects. A higher MCDI may represent an increased likelihood for developing visually consequential complications of myopia. However, future studies with larger sample size and of a longitudinal nature are warranted to conrm these preliminary observations.
Acknowledgments The authors would like to acknowledge Dr Mong-Li Lee, PhD and Dr Hsu Wynne, PhD from the School of Computing, National University of Singapore for providing us the Retinal Vessel Extraction Program.
References
1. Curtin BJ (1976) Visual acuity in myopia. In: Faye EE (ed) Clinical low vision, Chap. 26. Little Brown, Boston, pp 295300
2. Duke-Elder S, Abrams D (1970) Pathological myopia. In: Duke-Elder S (ed) System of ophthalmology, vol 5. C.V. Mosby, St. Louis, pp 300362
3. Xu L, Wang Y, Li Y et al (2006) Causes of blindness and visual impairment in urban and rural areas in Beijing. The Beijing Eye Study. Ophthalmology 113:11341141
4. Akyol N, Kukner SA, Ozdemir T et al (1996) Choroidal and retinal blood ow changes in degenerative myopia. Can J Ophthalmol 31(3):113119
5. Suzuki S (1999) Quantitative evaluation of sunset glow fundus in VogtKoyanagiHarda disease. Jpn J Ophthalmol 43(4):327333
6. Orihuela-Espina F, Claridge E, Preece JS (2003) Histological parametric maps of the human ocular fundus: preliminary result. http://www.cs.bham.ac.uk/~exc/Research/Papers/OrihuelaMIUA2003_3.pdf
Web End =http://www.cs.bham.ac.uk/*exc/Research/
http://www.cs.bham.ac.uk/~exc/Research/Papers/OrihuelaMIUA2003_3.pdf
Web End =Papers/OrihuelaMIUA2003_3.pdf . Accessed 20107. Grossniklaus HE, Green WR (1992) Pathologic ndings in pathologic myopia. Retina 12:127133
8. Vongphanit J, Mitchell P, Wang JJ (2002) Prevalence and progression of myopic retinopathy in an older population. Ophthalmology 109:704711
9. Lai TYY, Fan DSP, Lai WWK et al (2008) Peripheral and posterior pole retinal lesions in association with high myopia: a cross-sectional community-based study in Hong Kong. Eye 22:209213
10. Saw SM, Chua WH, Gazzard G et al (2005) Eye growth changes in myopic children in Singapore. Br J Ophthalmol 89:14891494
11. Logan NS, Gilmartin B, Wildsoet CF et al (2004) Posterior retinal contour in adult human anisomyopia. Invest Ophthalmol Vis Sci 45:21522162
12. Norton TT, Hyman L, Dong L et al. (2008) Longitudinal changes in the lens thickness in myopic children enrolled in
123
Int Ophthalmol (2012) 32:203209 209
the correction of myopia evaluation trial (COMET). Invest Ophthalmol Vis Sci 49: E-Abstract 260313. Cook CA, Koretz JF, Pfahnl A (1994) Aging of the human crystalline lens and anterior segment. Vision Res 34(22): 29452954
14. Preece JS, Claridge E (2003) Physics based medical image understanding of the coloration of the ocular fundus with application to detection of diabetic retinopathy. Department of computer science, University of Birmingham. http://www.cs.bham.ac.uk/~exc/Research/Papers/miua00_eye.pdf
Web End =http://www. http://www.cs.bham.ac.uk/~exc/Research/Papers/miua00_eye.pdf
Web End =cs.bham.ac.uk/*exc/Research/Papers/miua00_eye.pdf
123
Springer Science+Business Media B.V. 2012