Determining the size of retinal features inprematurely born children by fundus photography

Laura Knaapi,1,2 Tuomo Lehtonen,2 Eija Vesti2* and Markku T. Leinonen2*

1Department of Ophthalmology, Satakunta Central Hospital, Pori, Finland2Department of Ophthalmology, Turku University Hospital, Turku, Finland

ABSTRACT.

Purpose: The purpose was to study the effect of prematurity on the macula–disccentre distance and whether it could be used as a reference tool for determining

the size of retinal features in prematurely born children by fundus photography.

Methods: The macula–disc centre distance of the left eye was measured in pixels

from digital fundus photographs taken from 27 prematurely born children aged

10–11 years with Topcon fundus camera. A conversion factor for Topcon fundus

camera (194.98 pixel/mm for a 50° lens) was used to convert the results in pixels

into metric units.

Results: The macula–disc centre distance was 4.74 mm, SD 0.29. No correlation

between ametropia and the macula–disc centre distance was found (r = �0.07,

p > 0.05). One child (subject 20) had high myopia and retinopathy of

prematurity (ROP), and the macula–disc centre distance was longer than

average (6.35 mm).

Discussion: The macula–disc centre distance in prematurely born children at the

age of 10–11 years provides an easy-to-use reference tool for evaluating the size

of retinal features on fundus photographs. However, if complications of ROP,

for example temporal macular dragging or high ametropia, are present, the

macula–disc centre distance is potentially altered and a personal macula–disccentre distance should be determined and used as a refined reference tool.

Key words: blind spot – fundus photography – macula–disc centre distance – prematurely born

children – retinal feature

*Equal contribution.

Acta Ophthalmol.ª 2014 Acta Ophthalmologica Scandinavica Foundation. Published by John Wiley & Sons Ltd

doi: 10.1111/aos.12554

Introduction

In 1994, Bennett et al. published theirwidely used formula of determining thesize of retinal features by fundus pho-tography. In the formula:

t ¼ 0:01306�ðx� 1:82Þ�1:37�s

t stands for the true size of a retinalfeature, x is axial length of the giveneye, constant 1.37 applies to Zeissfundus camera used and s equals the

size of a retinal image (mm) on camerafilm (Bennett et al. 1994).

The formula of Bennett et al. (1994)applies to fundus photographs takenwith a telecentric fundus camera. How-ever, many fundus cameras are nottelecentric by design, thus preventingthe direct use of the formula (Rudnickaet al. 1998). Hence, different methodsfor determining the size of retinalfeatures have been developed (Quigley2003; Bartling et al. 2008).

The macula–disc centre distance hasbeen found to show little variation inadults (Mok & Lee 2002; Bartling et al.2008; Hong et al. 2010). Therefore, themacula–disc centre distance could pre-sumably be used as an internal refer-ence tool when calculating the size ofretinal features on fundus photographstaken with a non-telecentric funduscamera when the optical principals ofthe camera are not known (Bartlinget al. 2008).

In this study, our purpose was toevaluate whether the macula–disc cen-tre distance could provide a referencetool for determining the size of retinalfeatures by fundus photography inprematurely born children.

Subjects and Methods

Fundus photographs taken with Top-con fundus camera (TRC-50DX; Top-con,Tokyo, Japan) from27prematurelyborn children aged from 10 to 11 years,who were recruited originally foranother study, were analysed. Theleft eye of each subject was chosen forthe analyses. Themedian gestational ageat birth was 30 weeks (range 23–36 weeks), and the median birth weightwas 1260 g (range 525–1905 g). Theametropia range was +4.5 to �20.5 D(spherical equivalent) and astigmatismup to 1 D. One child (subject 20) hadhad retinopathy of prematurity (ROP;stage 3+ in both eyes) that had requiredlaser treatment. Three other childrenhad been diagnosed with strabismus,and one with anterior uveitis.

The parents or legal guardians of allparticipants gave informed writtenconsent after explanation of the nature

1

Acta Ophthalmologica 2014

and possible consequences of the study,according to the Finnish Ethics com-mittee of Turku University Hospital(ETMK 2/180/2012). All researchadhered to the tenets of the Declara-tion of Helsinki.

All 27 fundus photographs were infocus and showed both the macula andthe optic disc. They were digitallyanalysed using ImageJ program (ver-sion 1.4.3.67, publisher url: http://imagej.nih.gov/ij/). The centre of themacula was located at the centre of theannular ring reflex visible in the pic-tures. If it was not evident in the colourpicture (two cases), its location wasdetermined with the aid of the blackand white picture.

The macula–disc centre distance wasmeasured in pixels based on a methoddescribed by Hong et al. (2010; Fig. 1).The results were converted into metricunits using a conversion factor of Top-con fundus camera (194.98 pixel/mmfor a 50° lens; Knaapi et al. 2014). Inaddition, the width and the height ofthe optic discs were measured. Allmeasurements were carried out by thefirst author.

Ametropia was determined usingstreak retinoscopy. Fundus examina-tion was performed with slit lampbiomicroscopy (BQ-900; Haag-Streit,Bern, Switzerland) and a + 90 dioptrelens. Axial length was measured usingIOL Master (Carl Zeiss Meditec, Jena,Germany).

To test the correlation between themacula–disc centre distance and

ametropia, Pearson correlation coeffi-cient was calculated using IBM SPSSStatistics v. 21 software package. Thenormality of the parameters was testedusing the Kolmogorov–Smirnov test.

Results

Little variation was observed in themacula–disc centre distance (mean4.74 mm, SD 0.29). The dimensionsof the optic disc were mean width1.72 mm, SD 0.11 and mean height1.93 mm, SD 0.13. The mean axiallength of the left eye was 23.01 mm,SD 0.78. In subject 20 who had hadROP, the macula–disc centre distancewas significantly longer than average(6.35 mm). The refractive error of theleft eye of subject 20 was �20.5 (spher-ical equivalent), the axial length was25.64 mm, and the disc was verticallyoval: the width of the disc was1.66 mm, and the height, 2.25 mm.

No correlation was found betweenametropia and the macula–disc centredistance (r = �0.07, p > 0.05). Theoutlier, subject 20, was removed fromall of the analyses and calculations ofthe mean values.

Discussion

In our study, there was little variationin the macula–disc centre distanceof prematurely born children aged10–11 years (mean 4.74 mm, SD0.29). The optic-disc-to-fovea distance(ODF; equivalent to the macula–disccentre distance) in preterm and full-term infants had previously been stud-ied by De Silva et al. (2006). In theirstudy, the postmenstrual age rangedfrom 32 to 50 weeks. The mean ODFwas 4.4 mm, SD 0.4. They discussedthat the growth of the highly organizedarea of the posterior pole appearslimited (11%) from birth to adulthood,which is supported by our study. Inadults, the macula–disc centre distancehas been studied by Mok & Lee (2002),Bartling et al. (2008) and Hong et al.(2010). The results stated 4.69 mm, SD0.08; 4.62 mm, SD 0.33; and 4.503, SD0.373 mm in right eye and 4.458,SD 0.376 mm in left eye, respectively,and are smaller than our measurementson the 10- to 11-year-old subjects.

Ocular pathology can alter physio-logical fundus landmarks, and it maythus potentially affect the measure-ments carried out on fundus photo-

graphs. Hellstr€om et al. (1997) havepreviously studied ocular fundus mor-phology in preterm children whosemean postmenstrual gestational agewas 29.1 weeks. They found no differ-ence in optic disc morphology betweenpreterm children and control subjects.In another study, Hellstr€om et al.(2000) further described ocular fundusabnormalities in children born before29 weeks of gestation. They found thatpreterm birth was associated with sub-normal optic disc and rim areas.

In prematurity, a cicatricial retinop-athy may develop. It can lead tomyopia and temporal vitreoretinalfibrosis followed by dragging of themacula, vitreoretinal folds, retrolentalfibrovascular tissue and partial or totalretinal detachment. Hence, cicatricialROP can alter physiological funduslandmarks and affect the results of themeasurements carried out on fundusphotographs.

In our study, the macula–disc centredistance of subject 20 diagnosed withROP was significantly longer thanaverage (6.35 mm). This was causedby either ROP itself or high myopiaassociated with it (the spherical equiv-alent of the left eye was �20.5 D). Theeffect of possible macular dragging dueto ROP could not be ruled outalthough no retinal folds were detected.No staphyloma was present. Previ-ously, no correlation between the mac-ula–disc centre distance and ametropiaraging from �9 to +6 D has been found(Bartling et al. 2008).

Another difference in subject 20 wasthe shape of the disc; it was verticallyoval. In adult eyes, the average dimen-sions of the optic disc are as follows:width 1.77, SD 0.19 mm and height1.88, SD 0.19 mm (Quigley et al.1990). Previously, in adult eyes, alonger axial length has been shown toassociate with a longer distancebetween the disc and foveola, a largerindex of ovalness and a larger disc(p < 0.01) (Chihara & Chihara 1994).

If ROP or any other posterior polepathology alters physiological funduslandmarks, a personal macula–disc cen-tre distance should be determined andused as a refined reference tool for deter-mining the size of other retinal featuresby fundus photography instead of rely-ing on average dimensions (Knaapiet al. 2014). This personal macula–disccentre distance can be calculated fromfundus photographs taken with a tele-

Fig. 1. Measuring the macula–disc centre dis-

tance in subject 20. A rectangle was positioned

around the border of the optic disc. The centre

of the rectangle was determined by drawing

lines from the corners of the rectangle. The

macula–disc centre distance was calculated

from the centre of the macula to the centre

of the rectangle.

2

Acta Ophthalmologica 2014

centric fundus camera using the formulaof Bennett et al. (1994). However, manyfundus cameras are not telecentric bydesign. If a telecentric fundus camera isnot available and the optical principalsbehind the fundus camera used are notknown, a personal macula–disc centredistance in degrees (U0) can be used if thelocation of the centre of the physiolog-ical blind spot in the visual field isdetermined by visual field examination(Knaapi et al. 2014). This location indegrees is divided with the refractiveindex of the final ocular medium (1.336)to calculate the personal macula–disccentre distance in degrees (Knaapi et al.2014). The personal macula–disc centredistance in metric units, t, can then beclosely approximated using the formula:

U0 ¼ ðt�360Þ=ð2�p�ðx� 1:82ÞÞwhere x is axial length subtracted withthe location of the second principalpoint of the eye from the cornea (1.82,Bennett et al. 1994) (Fig. 2) (Knaapiet al. 2014).

In our study, there was little varia-tion in the macula–disc centre distance(4.74 mm, SD 0.29) in prematurelyborn children at the age of 10–11 years,and thus, it provides an easy-to-usereference tool for evaluating the size of

retinal features on fundus photo-graphs. If the macula–disc centre dis-tance is potentially altered by ocularpathology or high ametropia, a per-sonal macula–disc centre distanceshould be determined and used as arefined reference tool.

References

Bartling H, Wanger P & Martin L (2008):

Measurement of optic disc parameters on

digital fundus photographs: algorithm devel-

opment and evaluation. Acta Ophthalmol

86: 837–841.Bennett AG, Rudnicka AR & Edgar DF

(1994): Improvements on Littmann’s

method of determining the size of retinal

features by fundus photography. Graefes

Arch Clin Exp Ophthalmol 232: 361–367.Chihara E & Chihara K (1994): Covariation of

optic disc measurements and ocular param-

eters in the healthy eye. Graefes Arch Clin

Exp Ophthalmol 232: 265–271.De Silva JD, Cocker KD, Lau G, Clay ST,

Fielder AR & Moseley MJ (2006): Optic

disk size and optic disk-to-fovea distance in

preterm and full-term infants. Invest Oph-

thalmol Vis Sci 47: 4683–4686.Hellstr€om A, H�ard AL, Chen Y, Niklasson A

& Albertsson-Wikland K (1997): Ocular

fundus morphology in preterm children.

Influence of gestational age, birth size,

perinatal morbidity, and postnatal growth.

Invest Ophthalmol Vis Sci 38: 1184–1192.Hellstr€om A, H�ard A-L, Svensson E & Nik-

lasson A (2000): Ocular fundus abnormali-

ties in children born before 29 weeks of

gestation: a population-based study. Eye 14:

324–329.Hong SW, Ahn MD, Kang SH & Im SK

(2010): Analysis of peripapillary retinal

nerve fiber distribution in normal young

adults. Invest Ophthalmol Vis Sci 51: 3515–3523.

Knaapi L, Aarnisalo E, Vesti E & Leinonen

MT (2014): Clinical verification of the

formula of Bennett et al. (1994) of determin-

ing the size of retinal features by fundus

photography. Acta Ophthalmol. Epub

ahead of print.

Mok KH & Lee VW (2002): Disc-to-macula

distance to disc-diameter ratio for optic disc

size estimation. J Glaucoma 11: 392.

Quigley HA, Brown AE, Morrison JD &

Drance SM (1990): The size and shape of

the optic disc in normal human eyes. Arch

Ophthalmol 108: 51–57.Quigley MGDP (2003): A new fundus camera

technique to help calculate eye-camera mag-

nification: a rapid means to measure disc

size. Arch Ophthalmol 121: 707–709.Rudnicka AR, Burk RO, Edgar DF & Fitzke

FW (1998): Magnification characteristics of

fundus imaging systems. Ophthalmology

105: 2186–2192.

Received on December 22nd, 2013.

Accepted on August 7th, 2014.

Correspondence:

Laura Knaapi, MD

Department of Ophthalmology

Satakunta Central Hospital

Sairaalantie 3

FIN-28500

Pori, Finland

Tel: +358 2 627 7841

Fax: +358 2 627 7799

Email: laura.knaapi@utu.fi

The authors thank Kari Nummelin for part of

fundus photography. This study was supported by

the Nissi foundation (LK), the Kaukomarkkinat Oy

fund (LK), the Esther and Gustaf Nikula fund

(LK), the Rauno and Anne Puolimatka foundation

(LK), the Finnish Ophthalmological Society (LK),

the J. R. Danielsson-Kalmari fund (LK), the Paavo

Salminen fund (LK), the Combined Research

Foundation of the University of Turku (LK), the

Intermunicipal Hospital District of Southwest Fin-

land (TL), the TYKS foundation (LK, TL) and the

Finnish Fund of Neonatal Research (TL).

Fig. 2. Calculating the personal macula–disc centre distance, t, in metric units. The location of the

centre of the physiological blind spot (B) from the fixation point (F) in the visual field equals U

(degrees). Macula is indicated with M and the centre of the optic disc with (D). U0 is the personalmacula–disc centre distance in degrees. P stands for the first and P’ for the second principal point

of the eye. U0 = U/n, where n = 1.336 and stands for the refractive index of the final ocular

medium. x � A1P0 = x � 1.82, x being axial length and 1.82 the location of the second principal

point (P0) of the eye from the cornea. Adapted from Bennett et al. (1994).

3

Acta Ophthalmologica 2014