Optical Coherence Tomography for Corneal Diseases [605312]

REVIEW ARTICLE
Optical Coherence Tomography for Corneal Diseases
Naoyuki Maeda, M.D.
Abstract: Anterior segment optical coherence tomography (OCT) is
currently used for investigating the distribution of the corneal thickness,shape of the stromal interface after lamellar corneal surgery, associationbetween host and corneal graft in keratoplasty, dimension of theanterior chamber, and lesions of the corneal diseases. In addition, theadvances of OCT technology has enabled three-dimensional imaging,tissue imaging, cell imaging, and topographic analysis. In this review,examples of tissue imaging with 840-nm spectral-domain OCT, cellimaging with full-field OCT, and corneal topographic analysis with1,310-nm swept-source OCT were introduced.
Key Words: Optical coherence tomography—Cornea—Biometry—In
vivo biopsy—Corneal topography
(Eye & Contact Lens 2010;5: 254 /H11002259)
Because optical coherence tomography (OCT) was intro-
duced as a noncontact imaging modality that provides
detailed cross-sectional images of internal structures of thetissues,
1OCT has been attracting a great deal of attention in the
field of Ophthalmology. As a result, a paradigm shift hasoccurred in diagnosing and treating retinal diseases.
The first report of OCT imaging of cornea and anterior
segment (AS) was published in 1994.
2The time-domain (TD)
OCT at 1,310 nm has become commercially available forcross-sectional images of the cornea, anatomic structures ofangle, and anterior chamber biometry. Anterior segment of theeye can be visualized and quantified with noninvasively in theclinic. Anterior segment OCT is now frequently used not onlyfor diagnosing the pathologic condition of AS but also forplanning and evaluating AS surgeries.
3– 6
SPECTRAL-DOMAIN OCT
The resolution in vertical direction is 18 /H9262m, and the speed
of A-scan is 2,000 scan/sec with the 1,310-nm TD-OCT. On theother hand, higher resolution (5
/H9262m) and faster speed of A-scan
(26,000 scan/sec) can be obtained with the 840-nm spectral-domain (SD) OCT (RTVue-100; Optovue, Inc., Fremont, CA).Table 1 shows the currently available OCTs for AS.Figure 1 shows the differences of the cross-sectional images
of a normal cornea between TD-OCT and SD-OCT. Bowmanlayer can be identified as the parallel lines, and the thicknessprofile of epithelial layer and stromal layer of the cornea can beanalyzed separately with SD-OCT. However, it is difficult toobtain images clear enough to separate the epithelial layer fromthe stroma in normal cornea with TD-OCT. The wound-healingprocess of the cornea after iron foreign body examined bySD-OCT is shown in Figure 2. The stromal scar and epithelialthickening were observed. In addition, low-intensity shadowappears to be the remnant iron body that could not be detectedwith slitlamp examination. Epithelial ingrowth under the laserin situ keratomileusis (LASIK) flap is seen in Figure 3. Suchhigh-resolution images with high magnification are helpful tounderstand the wound-healing process or the effects of surgicalprocedures on the ocular surface.
Because of the higher resolution, SD-OCT may be better than
TD-OCT for the detection of keratoconus based on the distri-bution of corneal thickness
7or for the evaluation of corneal
deposits and scars.8On the other hand, SD-OCT instruments
have shorter wavelength and smaller scan width than TD-OCT.TD-OCT still has an advantage over SD-OCT in observing deeptissue or for the anterior chamber biometry.
SWEPT-SOURCE OCT
The introduction of the 1,310-nm swept-source (SS) OCT,
which is categorized as one type of Fourier-domain OCT, madeit possible to reconstruct the three-dimensional images of theAS of the eye more precisely.
9For example, SS-1000 (Tomey
Corporation, Nagoya, Japan) has a vertical resolution of 10 /H9262m
with 30,000 A-Scan/sec for 16-mm diameter. Three-dimen-sional analysis has been performed with the SS-OCT for dis-
From the Department of Ophthalmology, Osaka University Graduate
School of Medicine, Osaka, Japan.
Supported in part by Scientific Research from the Japanese Ministry of
the Education, Culture, Sports, Science and Technology grant 21592257(to N.M.).
Address correspondence and reprint requests to Naoyuki Maeda,
M.D., Department of Ophthalmology, Osaka University MedicalSchool, Room J7, Yamadaoka 2-2, Suita 565-0871, Japan; e-mail:nmaeda@ophthal.med.osaka-u.ac.jp
Accepted June 27, 2010.
DOI: 10.1097/ICL.0b013e3181ef0dbb
FIG. 1. Cross-sectional images of the normal cornea. The image by
840-nm SD-OCT ( top) has higher resolution than the 1310-nm
TD-OCT ( bottom ).
254 Eye & Contact Lens • Volume 36, Number 5, September 2010

tribution of corneal depositions in granular corneal dystrophy,10
filtering bleb,11or anterior biometry.12
The OCT-based corneal topographer with SS-OCT as the
next generation was developed to solve problems of currentlyavailable corneal topographers. Although corneal power hasbeen measured previously, there was a limitation in the repeat-ability compared with standard keratometry because of the useof TD-OCT.
13Parameters for corneal topographic analysis were
set as follows: scan mode; radial, scan number; 512 /H1100316/H11005
8,192 points, scan range; 10 mm, measurement time; 0.34 sec.
Figures 4 and 5 are examples comparing topographic findings
from a Placido-based corneal topographer (TMS-4 Advance;Tomey), Scheimpflug-based (Pentacam HR; OCULUS Op-tikgeräte GmbH, Wetzlar, Germany), and OCT-based instru-ment (SS-1000; Tomey). In an eye with moderate keratoconus(Fig. 4), all the anterior axial power maps revealed similarinferior steepening. In addition, elevation maps of the anteriorand posterior corneal surfaces and pachymetry map measuredwith the Pentacam and SS-1000 were similar. For an eye afterpenetrating keratoplasty (Fig. 5), Pentacam and SS-1000 alsoshowed similar pattern. Although it is not easy for TMS-4Advance to measure outside the host-graft junction, the othertwo topographers could cover a larger area of cornea.
This is because the Placid-based topographer uses mire
images produced by the reflex in precorneal tear film, and it isdifficult for the videokeratoscope to digitize heavily distortedmire images in eyes with severe irregular astigmatism. On theother hand, the Scheimpflug- and the OCT-based topographersare resistant to the severe-distorted corneas because of thesequential acquisition of cross-sectional images. There seems to
be an obvious tradeoff between acquisition time and toleranceof severe corneal distortion.
With the OCT corneal topographer, topographic analysis can
be performed even in the area of a severe scar or even when theedema exists in the corneal stroma. This may be the potentialadvantage of the OCT corneal topographer over conventionalinstruments for evaluating various types of corneal diseases orcorneal surgeries.
Recent advances of corneal lamellar surgeries such as De-
scemet stripping automated endothelial keratoplasty (DSAEK)and deep anterior lamellar keratoplasty have their advantagesand disadvantages over penetrating keratoplasty. The OCT-based corneal topographer with SS-OCT has potential to ana-lyze anterior and posterior parts of lamellar surgeries sepa-rately, as OCT can detect the interface after DSAEK, deepanterior lamellar keratoplasty, femtosecond laser-enabled ker-atoplasty, and LASIK.
14 –19
Figure 6 is an example of a new output using SS-1000. In this
display, not only conventional maps including axial powermaps, elevation maps, and total pachymetry maps but alsoelevation of the interface and pachymetric maps of anterior andposterior part can be shown. This is a case after DSAEK, andlocalized steepening of the posterior axial power map, asym-metric elevation of the posterior elevation map, and asymmetricthickening of the pachymetry map for posterior part (graft)showed that there was an asymmetric cut of the donor thatyielded irregular astigmatism of the posterior surface althoughthe donor was well centered in the eye.
FIG. 2. Cross-sectional image with 840-nm SD-OCT in a patient of
iron foreign body. Stromal thinning with fibrosis and epithelialthickening is remarkable. There is a shadow caused by residualforeign body.
FIG. 3. Cross-sectional image with 840-nm SD-OCT in a patient of
epithelial ingrowth after LASIK. The interface between flap andstromal bed was recognized clearly, and epithelial ingrowth wasshown as high-intensity area.TABLE 1. Currently Available Anterior Segment OCTs
Instrument/manufacture Type YearWave
length (nm)Resolution horizontal
/H11003vertical ( /H9262m)Range horizontal /H11003
vertical (mm) Speed A mode/sec
SL-OCT/Heidelberg Engineering TD 2006 1,310 20 /H1100325 15 /H110037 200
Visante OCT/Carl Zeiss Meditec TD 2007 1,310 60 /H1100318 16 /H110036 2,000
SS-1000 CASIA/Tomey
Corporation SS 2008 1,310 30 /H1100310 16 /H110036 30,000
3D OCT-1000 MARK II/Topcon SD 2006 840 20 /H1100356 /H110032.3 27,000
Cirrus HD-OCT/Carl Zeiss Meditec SD 2007 840 10 /H1100353 /H110032 27,000
RTVue-100/Optovue SD 2007 840 15 /H1100356 /H110032.3 26,000
iVue-100/Optovue SD 2009 840 15 /H1100356 /H110032.3 25,000
3D OCT-2000/Topcon SD 2010 840 20 /H1100356 /H110032.3 50,000
TD, time domain; SD, spectral domain, SS, swept source.Eye & Contact Lens • Volume 36, Number 5, September 2010 Optical Coherence Tomography
© 2010 Lippincott Williams & Wilkins 255

FIG. 4. Comparison of topographic appearance in a patient with moderate keratoconus among three
corneal topographers. Note that Placido-based ( left), Scheimpflug-based ( center ), and OCT-based
(right ) topographers revealed similar pattern.
FIG. 5. Comparison of topographic appearance in a patient after penetrating keratoplasty among
three corneal topographers. Scheimpflug-based ( center ) and OCT-based ( right ) topographers can cover
wider area than Placido-based ( left) topographer.
FIG. 6. Output of OCT-corneal topographer in a patient after DSAEK. The top row shows axial power
maps of anterior surface ( left) and posterior surface ( right ). The center is digitized cross-sectional image.
The second row indicates the elevation maps (best-fit-sphere) of anterior surface ( left), interface
(center ), and posterior surface ( right ). The bottom row is for the pachymetry maps of anterior part ( left),
total ( center ), and posterior part ( right ). Please note, there is an obvious steepening and elevation of the
posterior surface caused by asymmetric cut of the implanted lamellar graft.N. Maeda Eye & Contact Lens • Volume 36, Number 5, September 2010
256 Eye & Contact Lens • Volume 36, Number 5, 2010

Similarly, comparisons of topographic appearance at the flap
interface during LASIK between femtosecond laser and micro-keratome were possible with OCT corneal topographer asshown in Figure 7. These topographic maps were taken imme-diately after creating flaps in porcine eyes, and flaps werereturned without excimer laser ablation. The differences between theplanar flap created by a femtosecond laser (left in Fig. 7) and themeniscus flap by a microkeratome (right in Fig. 7) were obvious in thepachymetric maps of the anterior part (flap) and posterior part (bed).It appears that the OCT-based corneal topographer may be useful forevaluating and improving lamellar corneal surgeries.
FULL-FIELD OCT
There is another type of OCT that may be potentially useful
in the field of ophthalmology. The full-field OCT system usedis based on a white light interference microscope, where the
sample is illuminated by a thermal light source and a horizontalen face image is detected using a two-dimensional charge-coupled device camera (Fig. 8). A conventional four-framephase-shift detection technique is used to extract the interfero-metric image from the charge-coupled device output. An axialresolution is 2.0
/H9262m, and a transverse resolution is 2.4 /H9262m.20
Figure 9 shows the en face images of the ex vivo porcine
cornea. The superficial cells, wing cells, basal cells, stromalkeratocyte, Descemet membrane, and endothelial cells are
clearly observed.
In Figure 10, images of the ex vivo human donor cornea with
a tiny stromal scar are shown. As the volume-rendering proce-dure of the en face images were made, cross-sectional images,three-dimensional images, or movie clips of the en face imagescan be shown by the reconstruction of the data. The association
FIG. 7. Comparison of flap shape between femtosecond laser and microkeratome with OCT corneal
topographer in porcine cornea; femtosecond laser ( left), microkeratome ( right ). The elevation map of
the interface (second row, center ), pachymetry maps of anterior part (bottom row, left), and posterior
part (bottom row, right ) show the difference between planar flap with femtosecond laser and meniscus
flap with microkeratome.
FIG. 8. System of the full-field OCT. Conventional OCT acquires cross-sectional image first. In contrast,
full-field OCT has a two-dimensional charge-coupled device camera as the detector and an en face image.Eye & Contact Lens • Volume 36, Number 5, September 2010 Optical Coherence Tomography
© 2010 Lippincott Williams & Wilkins 257

FIG. 9. En face images of ex vivo por-
cine cornea by full-field OCT. The toprow indicates superficial cell layer ( left),
wing cell layer ( center ), and basal cell
layer ( right ). The bottom row indicates
keratocyte in the superficial stroma(left), Descemet membrane ( center ), and
endothelium ( right ).
FIG. 10. The en face images of ex vivo
human donor cornea with faint cornealscar by full-field OCT. The associationbetween superficial stromal scar ( arrow )
and break of Bowman layer was easilyrecognized. Reprinted with permissionfrom Akiba M, Maeda N, Yumikake K, etal. Ultrahigh-resolution imaging of humandonor cornea using full-field optical co-herence tomography. J Biomed Opt 2007;
12:041202. © 2007 Society of Photo-Op-tical Instrumentation Engineers.
FIG. 11. Sequential observation of the
effects of intraocular pressure on cornealepithelium and stroma. Cross-sectionalimage of the epithelial and superficialstroma at hypertension (70 mm Hg, topright ) showed higher signal at the epi-
thelial cell border and stromal kerato-cyte than those at normal tension (18mm Hg, top left). En face image at the
basal cell layer (bottom left:1 8m mH g ,
bottom right : 70 mm Hg) confirmed
similar trend.N. Maeda Eye & Contact Lens • Volume 36, Number 5, September 2010
258 Eye & Contact Lens • Volume 36, Number 5, 2010

between superficial stromal scar and break of Bowman layer or
epithelial edema was easily recognized.
Similar to the in vivo confocal microscope, it is possible to
perform sequential observation of the corneal tissue with full-field OCT. As shown in Figure 11, the effect of intraocularpressure to the corneal epithelium and stroma can be comparedin porcine corneas (ex vivo).
SUMMARY
The advancement of OCT technology will play a more and
more important role as one of the most useful noninvasive andquantitative modalities for examining the anterior pathologiccondition of the eye and planning and evaluating AS surgeriesin the near future.
ACKNOWLEDGMENT
The author thanks Prof. Teruo Nishida for giving extraordi-
nary opportunity to participate in the Yamaguchi InternationalSymposium 2010 “Corneal Trilogy” and also for his kindnessfor many years.
REFERENCES
1. Huang D, Swanson EA, Lin CP, et al. Optical coherence tomography.
Science 1991;254:1178–1181.
2. Izatt JA, Hee MR, Swanson EA, et al. Micrometer-scale resolution imaging
of the anterior eye in vivo with optical coherence tomography. Arch
Ophthalmol 1994;112:1584–1589.
3. Konstantopoulos A, Hossain P, Anderson DF. Recent advances in ophthal-
mic anterior segment imaging: A new era for ophthalmic diagnosis? Br J
Ophthalmol 2007;91:551–557.
4. Chen J, Lee L. Clinical applications and new developments of optical
coherence tomography: An evidence-based review. Clin Exp Optom 2007;
90:317–335.
5. Simpson T, Fonn D. Optical coherence tomography of the anterior segment.
Ocul Surf 2008;6:117–127.
6. Ramos JL, Li Y, Huang D. Clinical and research applications of anteriorsegment optical coherence tomography—A review. Clin Exp Ophthalmol
2009;37:81–89.
7. Li Y, Meisler DM, Tang M, et al. Keratoconus diagnosis with optical
coherence tomography pachymetry mapping. Ophthalmology 2008;115:
2159–2166.
8. Khurana RN, Li Y, Tang M, et al. High-speed optical coherence tomogra-
phy of corneal opacities. Ophthalmology 2007;114:1278–1285.
9. Yasuno Y, Madjarova VD, Makita S, et al. Three-dimensional and high-
speed swept-source optical coherence tomography for in vivo investigationof human anterior eye segments. Opt Express 2005;13:10652–10664.
10. Miura M, Mori H, Watanabe Y, et al. Three-dimensional optical coherence
tomography of granular corneal dystrophy. Cornea 2007;26:373–374.
11. Kawana K, Kiuchi T, Yasuno Y, et al. Evaluation of trabeculectomy blebs
using 3-dimensional cornea and anterior segment optical coherence tomog-raphy. Ophthalmology 2009;116:848–855.
12. Fukuda S, Kawana K, Yasuno Y, et al. Anterior ocular biometry using
3-dimensional optical coherence tomography. Ophthalmology 2009;116:
882–889.
13. Tang M, Li Y, Avila M, et al. Measuring total corneal power before and
after laser in situ keratomileusis with high-speed optical coherence tomog-raphy. J Cataract Refract Surg 2006;32:1843–1850.
14. Kymionis GD, Suh LH, Dubovy SR, et al. Diagnosis of residual Descemet’s
membrane after Descemet’s stripping endothelial keratoplasty with anteriorsegment optical coherence tomography. J Cataract Refract Surg 2006;32:
1827–1835.
15. Stahl JE, Durrie DS, Schwendeman FJ, et al. Anterior segment OCT
analysis of thin IntraLase femtosecond flaps. J Refract Surg 2007;23:555–
558.
16. Lim LS, Aung HT, Aung T, et al. Corneal imaging with anterior segment
optical coherence tomography for lamellar keratoplasty procedures. Am J
Ophthalmol 2008;145:81–90.
17. Holz HA, Meyer JJ, Espandar L, et al. Corneal profile analysis after
Descemet stripping endothelial keratoplasty and its relationship to postop-erative hyperopic shift. J Cataract Refract Surg 2008;34:211–214.
18. von Jagow B, Kohnen T. Corneal architecture of femtosecond laser and
microkeratome flaps imaged by anterior segment optical coherence tomog-raphy. J Cataract Refract Surg 2009;35:35–41.
19. Lombardo M, Terry MA, Lombardo G, et al. Analysis of posterior donor
corneal parameters 1 year after Descemet stripping automated endothelialkeratoplasty (DSAEK) triple procedure. Graefes Arch Clin Exp Ophthalmol
2010;248:421–427.
20. Akiba M, Maeda N, Yumikake K, et al. Ultrahigh-resolution imaging of
human donor cornea using full-field optical coherence tomography.J Biomed Opt 2007;12:041202.Eye & Contact Lens • Volume 36, Number 5, September 2010 Optical Coherence Tomography
© 2010 Lippincott Williams & Wilkins 259