Spectral optical coherence tomography: a newimaging technique in contact lens practice

Bartłomiej J. Kału _zny1, Jakub J. Kaluzny1, Anna Szkulmowska2, IwonaGorczynska2, Maciej Szkulmowski2, Tomasz Bajraszewski2, PiotrTargowski2 and Andrzej Kowalczyk2

1Department of Ophthalmology, Collegium Medicum, Nicolaus Copernicus University, Curie-

Skłodowskiej 9, 85-094 Bydgoszcz and 2Institute of Physics, Nicolaus Copernicus University,

Grudziadzka 5, 87-100 Torun, Poland

Abstract

Purpose: Spectral optical coherence tomography (SOCT) is a new non-invasive, non-contact, high-

resolution technique, which provides cross-sectional images of objects that weakly absorb and

scatter light. The aim of this article is to demonstrate the application of SOCT to imaging of eyes fitted

with contact lenses.

Methods: Nine eyes of six different subjects fitted with various contact lenses have been examined

with a slit-lamp and a prototype SOCT instrument.

Results: Our SOCT system provides high-resolution (4–6 lm longitudinal, 10 lm transversal)

tomograms composed of 3000–5000 A-scans with acquisition time of 100–250 ms. The quality of the

images is adequate for detailed evaluation of contact lens fit. Design, shape and lens edge position

were assessed, and complications of contact lens wear could be visualized. Thickness of the lens,

corneal epithelium and stroma as well as the space between the lens and the eye surface have been

measured.

Conclusions: SOCT allows high-resolution, cross-sectional visualization of the eye fitted with a

contact lens. The ability to carry out a detailed evaluation of the fitting relationship between the lens

and the ocular surface might be useful in research and optometric practice. SOCT can also be helpful

in diagnosis, evaluation and documentation of contact lens complications.

Keywords: contact lenses, contact lens complications, cornea, lens fit, spectral optical coherence

tomography

Introduction

Discontinuation from contact lens wear is a majorproblem in contact lens practice, which could result inthe lack of growth of the contact lens market (Pritchardet al., 1999; Morgan, 2003). A majority of lapsedcontact lens wearers cite discomfort as the main reasonfor discontinuation (Young et al., 2002). The patients�experience of inadequate comfort may have arisen from

a variety of causes. Among them, one of the mostimportant is inappropriate lens fit (Young, 2004).

Although slit-lamp examination is irreplaceable in theassessment of contact lens fit, by itself it gives limitedinformation on the space between the surface of the eyeand the posterior surface of the contact lens. For thisreason, fluorescein pattern analysis was introduced andbecame routine in contact lens practice. However, arecent study shows that slit-lamp examination withfluorescein may not always be sufficiently sensitive(Mountford et al., 2005). Better accuracy of assessmentof the fitting relationship between a contact lens and theeye surface could be helpful to achieve an optimal fit andthus to improve success rates (Young et al., 1993),especially in eyes with astigmatism (Reddy et al., 2000),in corneal pathology (Rubinstein, 2003) or in ortho-keratology (Mountford et al., 2005).

Received: 28 April 2005

Revised form: 21 August 2005

Accepted: 29 August 2005

Correspondence and reprint requests to: Bartłomiej Kału_zny.Tel.: +48 52 5854033; Fax: +48 52 5854033.

E-mail address: bartka@by.onet.pl

Ophthal. Physiol. Opt. 2006 26: 127–132

ª 2006 The College of Optometrists 127

Optical coherence tomography (OCT) is a non-invasive, non-contact, high-resolution technique, whichprovides cross-sectional images of objects that weaklyabsorb and scatter light. Thus, it could potentially beused for imaging of the eye fitted with a contact lens.Conventional (time domain) OCT has already proved itsusefulness in visualization of different ocular tissuesincluding the cornea (Drexler, 2004). However, its maindisadvantage is a long acquisition time. For example,the lately introduced prototype of ultra-high-resolutionOCT achieves 3 lm axial and 15–20 lm transverseresolution, but needs approximately 4 s to perform asingle image consisting of 600 A-scans (Wollstein et al.,2005).

Recently, a new variant of OCT has been developed –spectral optical coherence tomography (SOCT). Becauseof its improved sensitivity (de Boer et al., 2003; Chomaet al., 2003; Leitgeb et al., 2003) and short acquisitiontime, it provides tomograms of outstanding quality andenables real-time and three-dimensional (3D) imaging(Nassif et al., 2004; Wojtkowski et al., 2004; Jiao et al.,2005). The high speed of imaging vastly reduces motionartefacts (Yun et al., 2004); thus, SOCT providestomograms free from distortions caused by the motionof the object, which could be misinterpreted as struc-tural imperfections. Therefore, SOCT might be a valu-able imaging tool not only in research but also inoptometric practice.

SOCT-based instruments, in contrast to conventionalOCT, are not available commercially and only proto-types have been constructed to date. The first low-resolution SOCT image of the eye fitted with a contactlens was published in 2002 (Kału_zny et al., 2002). Atthat time, the authors suggested that it might be a usefultechnique in contactology but since then no furtherwork in this field has been presented. In this article wedemonstrate, for the first time to our knowledge, theapplication of high-resolution SOCT to cross-sectionalimaging of eyes fitted with contact lenses.

Materials and methods

Altogether, nine eyes of six different subjects fitted withvarious contact lenses have been examined with a slit-lamp and a prototype SOCT instrument, constructed inthe Institute of Physics, Nicolaus Copernicus Univer-sity, Torun, Poland.

The basic principle of the SOCT instrument operationis given elsewhere (Wojtkowski et al., 2002). Briefly, thelight emitted by a broadband (Dk ¼ 70 nm, centralwavelength 830 nm) super-luminescent diode islaunched into a fibre-optic Michelson interferometerand is split into reference and object beams. Therespective beams reflected from the reference mirrorand from the structures within the object are brought

together to interference and registered by a line scanCCD camera (AVIIVA M2-Atmel). The resulting signal(spectral fringes) carries information on the locations ofthe object structural interfaces along the path of thepenetrating beam. From this signal one in-depth line (A-scan) of a cross-sectional image is constructed. Then, theobject beam is moved laterally to the adjacent position,and eventually the 2D slice of the object is formed. Animage is composed of 3000–5000 A-scans having 690pixels each – such high density ensures the high qualityof the image. The colours in the images represent theintensity of reflected light. The intensity values areplotted on a logarithmic scale. The exposure time(during which the eye is illuminated) is 43 ls per oneA-scan. Total examination time is 100–250 ms. Thesensitivity of the system is 96 dB. Longitudinal resolu-tion of the instrument is 6 lm in air (4 lm in tissue);transversal resolution is better than 10 lm. The meas-urement head has been specially designed for theanterior segment imaging of an eye and mounted on aclassical slit-lamp stand in order to perform ophthalmicimaging in a convenient way.

The optical power of the beam incident at the corneais 750 lW, and is consistent with the American NationalStandards Institute (ANSI) recommended exposurelimit for continuous beam viewing (ANSI, 2000). Thestudy was approved by the Ethics Committee of theCollegium Medicum, Nicolaus Copernicus University inBydgoszcz, Poland, and examinations were conductedafter written informed consent was obtained.

Results

An example of a SOCT tomogram of the eye fitted withsoft contact lens (Acuvue Advance, Johnson and John-son Visioncare, Jacksonville, USA) is presented inFigure 1. The lens adheres properly to the surfaces ofthe cornea and conjunctiva. The peripheral part of thelens (Figure 1a) is much thicker than the central one(Figure 1b). The enhanced mid-peripheral zone is theelement of design that makes the lens hold its shape onthe fingertip and allows it to be easily applied onto theeye. The thicknesses of the contact lens, corneal epithe-lium and stroma measured at the centre were 66, 45 and405 lm, respectively. In order to calculate these distan-ces the refractive indices of the contact lens (1.41) andthe cornea (1.376) were taken into account. The preci-sion of geometrical thickness determination is 3 lm. Asexpected, the scattering inside the homogeneous mater-ial of the lens is much less than the scattering from thecorneal internal structure.

Figure 2 displays an example of the eye fitted with arigid gas permeable (RGP) contact lens (Boston, LensSerwis, Warszawa, Poland). The lens adheres properlyto the surface of the cornea. The essential features of the

128 Ophthal. Physiol. Opt. 2006 26: No. 2

ª 2006 The College of Optometrists

lens such as the position of the edge, its thickness(194 lm, assuming n ¼ 1.46) and design can beassessed. The edge of the lens is slightly elevated, whichallows better tear circulation (Figure 2a). One may notethat the central thickness of the RGP lens is almost threetimes that of the soft lens (Figure 2b).The SOCT technique also allows visualization of

contact lens associated complications. Figure 3a showssuperficially localized keratitis (arrows) caused by pro-longed contact lens wear. It is possible to assess thedepth (137 lm) and the width (230 lm) of the lesiontogether with the extent of the epithelial damage. Thelatter additionally introduces an imperfection of thecorneal surface in the vicinity of the lesion. An increaseof corneal thickness (678 lm) due to the corneal oedemais observed. The refractive index used in the geometricaldistance calculations was the same (1.376) for the corneaand for the lesion. However, unpredictable changes inthe refractive indices of pathological tissues may intro-duce minor measurement error. The changed tissueapparently absorbs more light of 830 nm than a healthy

cornea, which results in a reduction of the light intensityreflected from the structures behind it. The shadowingeffect does not influence the precision of geometricaldimensions determination. The slit-lamp examinationprovides information on size and localization of thelesion, but the depth can only be estimated approxi-mately.

Figure 3b presents the mucin ball (arrow) developedunder a silicone hydrogel lens. This common compli-cation for this lens type causes a hollow in theepithelium.

Figure 4 shows SOCT cross-sectional images ofsteep, acceptable and flat contact lens fit. In all casesthe eyes were fitted with Focus Night and Day contactlenses (Ciba Vision, Atlanta, USA). Figure 4a showstoo steep fit of a subject eye with 8.4 mm back opticzone radius (BZOR) lens. The result of too strongpressure of the lens edge on the eye surface results in adeformation (arrow) of the surrounding conjunctiva.Figure 4b, c presents the same eye of another subjectoptimally and flatly fitted with 8.4 and 8.6 mm BZORlenses, respectively. In the latter a gap of 25 lm(arrows) between the lens and the eye surface is clearly

Figure 2. The eye optimally fitted with rigid gas permeable contact

lens (Boston, Lens Serwis, Warsaw, Poland; n lens ¼ 1.46): (a)

peripheral part and (b) central part. Central thicknesses of the

contact lens, corneal epithelium and stroma are 194, 45 and 474 lm,

respectively. The bars correspond to 500 lm.

Figure 1. The eye optimally fitted with soft contact lens (Acuvue

Advance, Johnson and Johnson Visioncare, Jacksonville, USA;

n lens ¼ 1.41): (a) peripheral part and (b) central part. Central

thicknesses of the contact lens, corneal epithelium and stroma are

66, 45 and 405 lm, respectively. The geometrical distances are

taken along the vertical line connecting two white vertical segments.

The bars correspond to 500 lm.

Spectral optical coherence tomography: B. J. Kał u _zny et al. 129

ª 2006 The College of Optometrists

distinguished, while in the former the lens properlyaligns to the eye surface. Information on the inappro-priate soft lens fit from the slit-lamp examination isobtained indirectly and is usually based on the obser-vation of lens movement.

Discussion

Slit-lamp examination is a �gold standard� in evaluationof the anterior segment of the eye and contact lens fit.Although in the majority of cases this simple techniqueenables proper assessment, there is evidence that 7% ofcontact lens discontinuations are caused by inappropri-ate lens fit (Young, 2004).

The ability to accurately judge rigid contact lens fit isimproved by a fluorescein pattern analysis. Experiencedpractitioners using slit-lamp examination with fluoresc-ein can detect a variation of 0.05 mm in back optic zoneradius, which corresponds to a 10 lm difference in tearlayer thickness (Brunagert, 1961). However, a recentstudy shows that only 61–67% of experienced practi-tioners can correctly identify the static fluoresceinpattern of an ideally fitted conventional RGP lens and

reveals that fluorescein pattern analysis may not besufficiently sensitive for assessing ortho-k (reversegeometry) lens fitting (Mountford et al., 2005).

The clinical fitting performance of soft contact lens isgenerally considered to relate to its geometrical design.Research studies have shown that the back surfacedesign especially has a significant effect on lens fit andpatients� comfort (Young et al., 1993). On the marketthere are a number of contact lenses with different back

Figure 4. Assessment of contact lens fit (Focus Night and Day, Ciba

Vision, Atlanta, USA; n lens ¼ 1.43): (a) steep, too strong pressure of

the lens edge on the eye surface (arrow); (b) acceptable; (c) flat, a

gap of 25 lm (arrows) between the lens and the cornea. The bars

correspond to 500 lm.

Figure 3. Contact lens complications: (a) localized keratitis (ar-

rows), depth of the lesion 137 lm, width of the lesion 230 lm,

corneal thickness 678 lm; (b) mucin ball (arrow) under a silicone

hydrogel lens (Focus Night and Day, Ciba Vision, Atlanta, USA;

n lens ¼ 1.43). The bars correspond to 500 lm.

130 Ophthal. Physiol. Opt. 2006 26: No. 2

ª 2006 The College of Optometrists

surface designs (monocurve, bicurve, multicurve andaspheric), different diameters and back optic zone radii.Slit-lamp examination may not always answer thequestion of which contact lens is the most suitable forthe topography of a particular eye.We conclude that in some cases high-resolution cross-

sectional visualization of the ocular surface fitted with acontact lens may be useful. Currently, there are threetechnologies that could potentially be used for thispurpose. The first one is high-frequency ultrasoundimaging. The main disadvantage of this technique is thenecessity of use of the immersion bath, which canchange examination conditions. High-frequency(50 MHz) ultrasound can achieve only 20–30 lm axialresolution (Reinstein et al., 1994).The second possibility is an application of the optical

method – Scheimpflug camera. The most advancedcommercially available instrument based on this tech-nology is Pentacam (Oculus, Wetzlar, Germany). It usesa rotating Scheimpflug digital camera for visualizationand analysis of the anterior segment of the eye (Holl-aday et al., 2005). Our own observations show thatPentacam is not capable of visualizing a contact lens onthe eye surface, probably due to the very low scatteringof lens material.The third technique that might be considered is OCT.

It is analogous to conventional ultrasonic imaging,except that OCT does not require direct contact with theinvestigated tissue. Conventional (time domain) OCT isa well-established technique but long acquisition timecauses a decrease of image quality and limits its clinicalapplications. On the contrary, SOCT has a very shortacquisition time and most of the motion artefacts areeliminated.Our prototype SOCT system provides images of

quality adequate for detailed evaluation of the eye fittedwith a contact lens (Figures 1 and 2). Design, shape andlens edge position can be visualized. Thickness of thelens, corneal epithelium and stroma can be measured.However, the most important, from the practical pointof view, seems to be the ability of cross-sectionalvisualization of the contact lens–eye interface. SOCTcan also be useful in diagnosis and documentation ofcontact-lens-associated complications (Figure 3). Incontrast to slit-lamp examination the depth and thearea of corneal lesions can be precisely assessed as longas the refractive index is known. The time evolution ofthe depth and the area as well as the light scatteringproperties of the changed tissue can be evaluated. Thismay be useful in the monitoring of treatment resultsbecause progression or regression of the lesion can beobjectively assessed.The unique feature of this technique is the ability to

cross-sectionally assess the fitting relationship betweenocular surface and contact lens. It allows evaluation of

contact lens fit (Figure 4). In particular, the spacebetween the surface of the eye and the contact lens canbe visualized and measured (Figure 4c). The slit-lampexamination provides information on inadequate lens fitbut it cannot be objectively determined.

SOCT is a rapidly developing technique. Potentially,3D SOCT could provide data to calculate the volume ofthe lesions and to determine the whole space between acontact lens and the eye surface. Further work on fastSOCT will probably lead to the introduction of a newcorneal topography instrument based on this technol-ogy. It might be used for precise analysis of bothanterior and posterior surfaces of the cornea and toperform full corneal pachymetry. An instrument thatwould combine SOCT imaging and SOCT cornealtopography might be a valuable tool for every contactlens practitioner.

The results of this study confirm that SOCT allowshigh-resolution, cross-sectional visualization of the eyefitted with a contact lens. The ability to preciselyevaluate the fitting relationship between the lens andthe ocular surface might be useful not only in researchbut also in optometric practice. Moreover, SOCT couldalso be helpful in diagnosis, evaluation and documen-tation of contact lens complications.

References

ANSI (2000) American National Standard for Safe Use ofLasers. ANSI Z 136.1, Orlando, FL.

de Boer, J. F., Cense, B., Park, B. H., Pierce, M. C., Tearney,

G. J. and Bouma, B. E. (2003) Improved signal-to-noiseratio in spectral-domain compared with time-domain opticalcoherence tomography. Opt. Lett. 28, 2067–2069.

Brunagert, T. F. (1961) Fluorescein patterns: they are accurateand they can be mastered. J. Am. Optom. Assoc. 32, 973–974.

Choma, M. A., Sarunic, M. V., Yang, C. and Izatt, J. A.(2003) Sensitivity advantage of swept source and Fourierdomain optical coherence tomography. Opt. Express 11,

2183–2189.

Drexler, W. (2004) Ultrahigh-resolution optical coherencetomography. J. Biomed. Opt. 9, 47–74.

Holladay, J. T., Belin, M. W., Chayet, A. S., Maus, M. and

Vinciguerra, P. (2005) Next-generation technology for thecataract and refractive surgery. Suppl Cataract Refract.Surg. Today 1, 1–11.

Jiao, S., Knighton, R., Huang, X., Gregori, G. and Puliafito,C. A. (2005) Simultaneous acquisition of sectional andfundus ophthalmic images with spectral-domain optical

coherence tomography. Opt. Express 13, 444–452.Kału_zny, J. J., Wojtkowski, M. and Kowalczyk, A. (2002)Imaging of the anterior segment of the eye by spectraloptical tomography. Opt. Appl. 32, 581–589.

Leitgeb, R. A., Hitzenberger, C. K. and Fercher, A. F. (2003)Performance of Fourier domain vs. time domain opticalcoherence tomography. Opt. Express 11, 889–894.

Spectral optical coherence tomography: B. J. Kał u _zny et al. 131

ª 2006 The College of Optometrists

Morgan, P. (2003) Healthcheck on the contact lens market.

Optician 226, 32–33.Mountford, J., Cho, P. and Chui, W. S. (2005) Is fluoresceinpattern analysis a valid method of assessing the accuracy of

reverse geometry lenses for orthokeratology. Clin. Exp.Optom. 88, 33–38.

Nassif, N., Cense, B., Park, B. H., Yun, S. H., Chen, T. C.,

Bouma, B. E., Tearney, G. J. and de Boer J. F. (2004)In vivo human retinal imaging by ultrahigh-speed spectraldomain optical coherence tomography. Opt. Lett. 29, 480–482.

Pritchard, N., Fonn, D. and Brazeau, D. (1999) Discontinu-ation of contact lens wear: a survey. Int. Contact Lens Clin.26, 157–162.

Reddy, T., Szczotka, L. B. and Roberts, C. (2000) Peripherialcorneal contour measured by topography influences softtoric contact lens fitting success. CLAO J. 26, 180–185.

Reinstein, D. Z., Silverman, R. H., Rondeau, M. J. andColeman, D. J. (1994) Epithelial and corneal thicknessmeasurements by high-frequency ultrasound digital signalprocessing. Ophthalmology 101, 140–146.

Rubinstein, M. P. (2003) Applications of contact lens devicesin the management of corneal disease. Eye 17, 872–876.

Wojtkowski, M., Leitgeb, R., Kowalczyk, A., Bajraszewski, T.

and Fercher, A. F. (2002) In vivo human retinal imaging by

Fourier domain optical coherence tomography. J. Biomed.

Opt. 7, 457–463.Wojtkowski, M., Bajraszewski, T., Gorczynska, I., Targowski,

P., Kowalczyk, A., Wasilewski, W. and Radzewicz, Cz.

(2004) Ophthalmic imaging by spectral optical coherencetomography. Am. J. Ophthalmol. 138, 412–419.

Wollstein, G., Paunescu, L. A., Ko, T. H., Fujimoto, J. G.,

Kowalevicz, A., Hartl, I., Beaton, S., Ishikawa, H., Mattox,C., Singh, O., Duker, J., Drexler, W. and Schuman, J. S.(2005) Ultrahigh-resolution optical coherence tomographyin glaucoma. Ophthalmology 112, 229–237.

Young, G. (2004) Why one million contact lens wearersdropped out. Cont. Lens Anterior Eye 27, 83–85.

Young, G., Holden, B. and Cooke, G. (1993) Influence of soft

contact lens design on clinical performance. Optom. Vis. Sci.70, 394–403.

Young, G., Veys, J., Pritchard, N. and Coleman, S. (2002) A

multi-centre study of lapsed contact lens wearers. Ophthal-mic Physiol. Opt. 22, 516–527.

Yun, H., Tearney, G. J., de Boer, J. F. and Bouma, B. E.(2004) Motion artifacts in optical coherence tomography

with frequency-domain ranging. Opt. Express 12, 2977–2998.

132 Ophthal. Physiol. Opt. 2006 26: No. 2

ª 2006 The College of Optometrists