Characterization of corneal structure in keratoconus David P. Pi ~ nero, PhD, Juan C. Nieto, MSc, Alberto Lopez-Miguel, MSc The increasing volume of patients interested in refractive surgery and the new treatment options available for keratoconus have generated a higher interest in achieving a better characterization of this pathology. The ophthalmic devices for corneal analysis and diagnosis have experienced a rapid development during the past decade with the implementation of technologies such as the Placido- disk corneal topography and the introduction of others such as scanning-slit topography, Scheimpflug photography, and optical coherence tomography, which are able to accurately describe not only the geometry of the anterior corneal surface but also that of the posterior surface, as well as pachymetry and corneal volume. Specifically, anterior and posterior corneal elevation, corneal power, pachymetry maps, and corneal coma-like aberrometry data provide sufficient information for an accurate characterization of the cornea to avoid misleading diagnoses of patients and provide appropriate counseling of refractive surgery candidates. Financial Disclosure: No author has a financial or proprietary interest in any material or method mentioned. J Cataract Refract Surg 2012; 38:2167–2183 Q 2012 ASCRS and ESCRS Keratoconus is an ectatic corneal disorder character- ized by a usually progressive corneal thinning that results in corneal protrusion, irregular astigmatism, and decreased vision. 1 Specifically, the cornea as- sumes a conical shape as a result of the degeneration of the corneal stromal tissue and the subsequent bio- mechanical alteration. 1,2 The incidence of this condi- tion varies depending on factors such as the ethnic group analyzed or the criteria used to establish the di- agnosis; most estimates are between 50 and 230 per 100 000 in the general population. 1 Several risk factors, such as constant eye rubbing, the presence of systemic diseases (eg, sleep apnea), floppy eyelid syndrome, allergies, and eczema, as well as family history, have been defined for the development of this corneal disease. 1 Corneal topography is a valuable tool for confirm- ing the keratoconus diagnosis. 1 Several indices, algo- rithms, and even neural network approaches to the geometry and optical properties of the anterior corneal surface have been developed for keratoconus diagno- sis and detection. 3–13 Therefore, moderate and ad- vanced keratoconus detection is not a difficult issue using corneal topography in combination with biomi- croscopic, retinoscopic, and pachymetric evaluation. 1 Detection problems arise with very early or preclinical stages of the ectatic disorder. The term keratoconus suspect was coined for designating corneas not showing biomicroscopic keratoconic signs but subtle topographic features similar to early stages of kerato- conus. 14–19 In such conditions, other diagnostic tests and examinations may be helpful to confirm the potential presence of an incipient ectatic condition. The biomicroscopic findings in keratoconus, such as stromal thinning and posterior stress line, suggest that posterior corneal geometry may be altered indepen- dently of the anterior surface alterations. There is evidence that both anterior and posterior curvatures are affected in keratoconic eyes and also in keratoco- nus suspect eyes. 20 Likewise, significantly larger values of best-fit sphere (BFS) 14,16,21 and posterior Submitted: October 23, 2011. Final revision submitted: February 8, 2012. Accepted: May 28, 2012. From the Department of Ophthalmology, Oftalmar (Pi ~ nero), Hos- pital Internacional Medimar, the Foundation for the Visual Quality (Pi ~ nero), Fundaci on para la Calidad Visual, the Optics, Pharma- cology and Anatomy Department (Pi ~ nero), University of Alicante, Alicante, the Optometry and Vision Sciences Unit (Nieto), Optics Department, University of Valencia, the Avanza Vision Ophthal- mology Clinic (Nieto), Valencia, and the Institute of Applied Ophthalmobiology (Lopez-Miguel), University of Valladolid, Valladolid, Spain. Corresponding author: David P. Pi ~ nero, PhD, Oftalmar, Department of Ophthalmology, Medimar International Hospital, Avenida Denia, 78, 03016 Alicante, Spain. E-mail: [email protected]. Q 2012 ASCRS and ESCRS Published by Elsevier Inc. 0886-3350/$ - see front matter 2167 http://dx.doi.org/10.1016/j.jcrs.2012.10.022 REVIEW/UPDATE
17
Embed
Characterization of corneal structure in keratoconus · Keratoconus is an ectatic corneal disorder character-ized by a usually progressive corneal thinning that results in corneal
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
REVIEW/UPDATE
Characterizat
ion of cornealstructure in keratoconus
David P. Pi~nero, PhD, Juan C. Nieto, MSc, Alberto Lopez-Miguel, MSc
SubmittFinal revAccepte
From thpital Int(Pi~nero)cology aAlicanteDepartmmologyOphthalValladol
Correspof Ophth78, 030
Q 2012 A
Published
ed: Octision sd: May
e Depaernacio, Fundnd An, the Oent, UClinicmobiolid, Spa
ondingalmolo16 Alic
SCRS an
by Elsev
The increasing volume of patients interested in refractive surgery and the new treatment optionsavailable for keratoconus have generated a higher interest in achieving a better characterization ofthis pathology. The ophthalmic devices for corneal analysis and diagnosis have experienced a rapiddevelopment during the past decade with the implementation of technologies such as the Placido-disk corneal topography and the introduction of others such as scanning-slit topography,Scheimpflug photography, and optical coherence tomography, which are able to accuratelydescribe not only the geometry of the anterior corneal surface but also that of the posteriorsurface, as well as pachymetry and corneal volume. Specifically, anterior and posterior cornealelevation, corneal power, pachymetry maps, and corneal coma-like aberrometry data providesufficient information for an accurate characterization of the cornea to avoid misleading diagnosesof patients and provide appropriate counseling of refractive surgery candidates.
Financial Disclosure: No author has a financial or proprietary interest in any material or methodmentioned.
Keratoconus is an ectatic corneal disorder character-ized by a usually progressive corneal thinning thatresults in corneal protrusion, irregular astigmatism,and decreased vision.1 Specifically, the cornea as-sumes a conical shape as a result of the degenerationof the corneal stromal tissue and the subsequent bio-mechanical alteration.1,2 The incidence of this condi-tion varies depending on factors such as the ethnicgroup analyzed or the criteria used to establish the di-agnosis; most estimates are between 50 and 230 per100 000 in the general population.1 Several risk factors,such as constant eye rubbing, the presence of systemic
ober 23, 2011.ubmitted: February 8, 2012.28, 2012.
rtment of Ophthalmology, Oftalmar (Pi~nero), Hos-nal Medimar, the Foundation for the Visual Qualityaci�on para la Calidad Visual, the Optics, Pharma-atomy Department (Pi~nero), University of Alicante,ptometry and Vision Sciences Unit (Nieto), Opticsniversity of Valencia, the Avanza Vision Ophthal-(Nieto), Valencia, and the Institute of Appliedogy (Lopez-Miguel), University of Valladolid,in.
author: David P. Pi~nero, PhD, Oftalmar, Departmentgy, Medimar International Hospital, Avenida Denia,ante, Spain. E-mail: [email protected].
d ESCRS
ier Inc.
diseases (eg, sleep apnea), floppy eyelid syndrome,allergies, and eczema, as well as family history, havebeen defined for the development of this cornealdisease.1
Corneal topography is a valuable tool for confirm-ing the keratoconus diagnosis.1 Several indices, algo-rithms, and even neural network approaches to thegeometry and optical properties of the anterior cornealsurface have been developed for keratoconus diagno-sis and detection.3–13 Therefore, moderate and ad-vanced keratoconus detection is not a difficult issueusing corneal topography in combination with biomi-croscopic, retinoscopic, and pachymetric evaluation.1
Detection problems arise with very early or preclinicalstages of the ectatic disorder. The term keratoconussuspect was coined for designating corneas notshowing biomicroscopic keratoconic signs but subtletopographic features similar to early stages of kerato-conus.14–19 In such conditions, other diagnostic testsand examinations may be helpful to confirm thepotential presence of an incipient ectatic condition.
The biomicroscopic findings in keratoconus, such asstromal thinning and posterior stress line, suggest thatposterior corneal geometry may be altered indepen-dently of the anterior surface alterations. There isevidence that both anterior and posterior curvaturesare affected in keratoconic eyes and also in keratoco-nus suspect eyes.20 Likewise, significantly largervalues of best-fit sphere (BFS)14,16,21 and posterior
0886-3350/$ - see front matter 2167http://dx.doi.org/10.1016/j.jcrs.2012.10.022
2168 REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
elevation15,16,22 have been reported in clinical kerato-conic eyes and keratoconus suspect eyes. Therefore,manifestations of the ectasia also occur at the posteriorcorneal surface, even in early stages of the disease.These manifestations can be currently monitored inthe clinical practice by means of advanced imagingtechniques that are commercially available, such asoptical coherence tomographers, systems combiningthe scanning-slit and Placido disk technologies, andsystems based on Scheimpflug photography.23 Thesesystems also allow the clinician to characterize the pa-chymetric distribution as well as the volume of theevaluated cornea, which have been shown to be usefulfor keratoconus diagnosis and characterization.24
The aim of the current review is to summarize thediagnostic criteria defined for keratoconus with themore advanced topography systems, including analy-sis of the posterior corneal surface and pachymetricdistribution. These criteria are especially useful forscreening corneal refractive surgery candidates toavoid performing the surgical procedure in a patholog-ical or potentially pathological cornea.
ANALYSIS OF THE ANTERIOR CORNEAL SURFACEIN KERATOCONUS
Analysis of the Topographic Map
Analysis of the topographic pattern of the ante-rior corneal surface has been shown to help iden-tify ectatic corneas, especially during the earlystates because the topographic pattern in such con-ditions is different than the normal pattern.1,25–28
Many studies have been conducted to characterizethe keratoconus “topographical phenotype” by
J CATARACT REFRACT SURG - V
means of computerized corneal videokeratogra-phy.1,12,29,30 In most patients affected by keratoco-nus, the topographic map of the anterior cornealsurface is characterized by focal steepening overthe peripheral or midperipheral zone, which coin-cides with the location of the conical protrusionthat commonly expands over fewer than 2 quad-rants.1 Therefore, the typical appearance of thekeratoconic topographic map is the presence ofa well-delimited zone with a high dioptric value,surrounded by progressively decreasing curvaturezones (Figure 1). The cone vertex is typically dis-placed toward the lower midperipheral region ineither the nasal or temporal quadrant, althoughin rare occasions the ectatic alterations could ap-pear in the central portion of the cornea. In con-trast to a with-the-rule topographical astigmatismin a healthy cornea, the topographical keratoconicloop is usually asymmetric with an inferior hemi-meridian showing steeper radii (Figure 1). Like-wise, there is usually a vertical asymmetry witha certain diagonal angling for any of the 2 hemi-meridians, providing a characteristic topographicimage pattern. This topographic pattern is com-monly similar in both eyes, although one of themmay show a more advanced state.
As a summary, the topographic pattern of kerato-conus usually has several particular features(Figure 1): (1) Focal steepening located in the coneprotrusion zone surrounded by concentric decreas-ing power zones. Focal areas with dioptric valuesgreater than 46.0 to 47.0 diopters (D) should alertthe examiner and other diagnostic complementarytests must be performed to confirm or rule out the
Figure 1. Axial curvature map of a kerato-conic cornea obtained by a rotatingScheimpflug camera device.
OL 38, DECEMBER 2012
2169REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
diagnosis of keratoconus. (2) Inferior–superior (I–S)asymmetry within the midperipheral cornea. Specialattention should be taken when obtaining values ofthe I–S index greater than 1.4 D. (3) Angling of thehemimeridians in the bowtie pattern. Keratoconusmust be suspected, especially when this angling ex-ceeds 20 or 30 degrees in relation to the verticalmeridian.
Several topographicmaps of the anterior corneal sur-face can be useful in detecting and characterizing thekeratoconus, such as the tangential or elevation maps.As tangential maps provide more detailed informationabout local changes in curvature than axial maps,31 thetangential algorithm has been shown to be more validfor the clinical identification of topographical patternscompatible with keratoconus.32–37 Tangential mapscombined with elevation (front and posterior) and pa-chymetricmapsareanexcellent tool forkeratoconusde-tection, even in the most incipient forms.
The color scale used to graphically show the topo-graphic datamust be selectedwith care. Several studiesof the ideal dioptric step for topographical scaling havebeen published, with the aim of defining a standard forobtaining themaximum information from the diagnos-tic examination.38–40 TheAmericanNational StandardsInstitute39 suggest the use of 0.5, 1.0, and 1.5 D steps asadequate intervals. Smolek et al.38 report that the stan-dard deviation (SD) for curvature in the central cornealportion of a normal group of corneas was 1.59 D, stat-ing that the use of 1.5 D steps was optimum for colortopographic scaling. Similarly, in their clinical studyusing a standard videokeratograph,Wilson et al.40 con-clude that no topographical discoveries were unno-ticed using a 1.5 D step instead of 1.0 D.
Corneal Asphericity
Some quantitative indexes and descriptors havebeen defined or developed for characterizing theshape of the anterior corneal surface (Table 1). Oneof the most important is the corneal asphericity. Itshould be considered that the geometry of the anterior
Table 1. Values of asphericity and the related parameters for theconic surfaces that may adjust to the anterior corneal contour.
e2 Z eccentricity; p Z shape factor; Q Z asphericity
J CATARACT REFRACT SURG - V
corneal surface of the human cornea can be adjusted toa conical section that is characterized by an asphericity(Q) and a specific value for the radius of the apical cur-vature (r0).41,42 The Q characterizes the gradual curva-ture change in the corneal surface from center toperiphery.42–44 The most commonly accepted valuefor Q in young adult patients is �0.23 G 0.08 (prolateellipsoid).41,45 Positive values of Q would denote thepresence of an oblate surface (ie, corneas after myopicablation with the excimer laser or orthokeratologytreatments), and a null value (QZ 0) would representa completely spherical surface. In an ectatic cornea(Figure 2), the anterior protrusion generates an in-crease of the corneal prolatismwith an associated neg-ativization of Q (Table 1). Indeed, a more markednegative asphericity (significant prolateness) for theanterior corneal surface was reported in a study by Pi-~nero et al.46 comparing some topographic features ofa sample of normal eyes with those found in kerato-conic corneas grade I and II according to the Amsler-Krumeich grading system.4 Specifically, mean Qvalues of �0.29 G 0.09, �0.65 G 0.27, and �1.18G 0.32 were found in the normal, keratoconus gradeI, and keratoconus grade II groups, respectively.46 Itshould be mentioned that other related parametersare used by some authors for representing the cornealcurvature change from center to periphery: eccentricity(e), QZ�e2; form factor (p), pZ 1CQ; shape factor,shape factor Z �Q.
Other Indexes and Descriptors
Several topographic indexes and descriptorshave been developed for characterizing cornealshape that can be especially useful for the detec-tion and diagnosis of keratoconus. Some of themost common indices are the corneal irregularitymeasurement (CIM), the mean toric keratometry(MTK), the surface regularity index (SRI), the pre-dicted corneal acuity (PCA), and the surfaceasymmetry index (SAI). The CIM is a numericvalue representing the degree of irregularity thatis present in the corneal surface and obtained bymeans of a proprietary algorithm.47 Higher proba-bility of a pathological cornea is associated withlarge values of this index. The CIM is gradedwith the following criteria47: healthy from 0.03to 0.68 mm, borderline from 0.69 to 1.00 mm, andanomalous from 1.10 to 5.00 mm. The MTK is de-rived from corneal elevation data. It compares theanalyzed cornea with the elevation values obtainedby means of the best-fit toric surface.47 The largerthe MTK value, the more significant is the cornealtoricity and the higher is the likelihood of a cornealectatic disorder. The MTK distribution in the
OL 38, DECEMBER 2012
Figure 2.A: Front elevation map of a keratoconus. The descriptive parameters (QZ�1.29, eZ 1.13) show high prolate values. These values area result of the cone protrusion. B: High progressive keratoconus showing localized paracentral protrusion, pachymetry thinning, and high pos-terior steepening.
2170 REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
population resembles the Gaussian bell curve, withan average value of 44.5 D and values for this in-dex ranging from 41.25 to 47.25 D for 96% of thepopulation. The MTK is graded with the followingcriteria47: healthy from 43.1 to 45.9 D, borderlinefrom 41.8 to 43.0 D and from 46.0 to 47.2 D, andanomalous from 36.0 to 41.7 and from 47.3 to60.0 D.
The SRI, described by Wilson and Klyce in 1991,48 isa corneal quantitativedescriptor. It attempts to representlocalized fluctuations in the corneal surface along a me-ridianwithin the central corneal area. This descriptor hasbeen shown to behighly correlatedwith the correctedvi-sual acuity (rZ 0.80, P! .001). A cornea considered asnormal and healthy presents values of SRI below 1.55.
The PCA is a quantitative index that provides infor-mation about the corneal optical quality within the3.0 mm central area.9 It is one of the 15 corneal param-eters provided by the Holladay diagnosis system inEyesys videokeratographs (EyeSys Vision).9 Finally,the SAI is a centrally weighted summation of differ-ences in corneal power between corresponding points180 degrees apart on 128 equally spaced meridians.13
This index approaches zero for a radially symmetricalsurface and increases as the corneal shape becomesmore asymmetrical.13
Screening Systems
With the aim of providing faster and better clas-sification of topographic patterns of the anterior
J CATARACT REFRACT SURG - V
corneal surface, several diagnostic or early detec-tion software and algorithms have been developedand incorporated into corneal topography sys-tems.9,48–52 The most commonly used are de-scribed below.
The Pathfinder Corneal Analysis module, which isincluded in the Atlas videokeratograph (anterior cor-neal topography analyzer) software (Carl Zeiss Medi-tec AG), allows the detection of irregular topographicconditions compatible with keratoconus.47,49 It isbased on the combined analysis of the quantitative de-scriptors CIM, shape factor, and MTK. The algorithmused classifies the result of the analysis as follows:
Normal: The cornea shows a typically aspheric pro-file with CIM, shape factor, and MTK values withinnormal limits.
Corneal distortion, also called pseudokeratoconus:This is corneal molding or warpage usually derivedfrom contact lens wear. It is characterized by a CIMvalue outside the normal range but with normal shapefactor and MTK values.
Subclinical keratoconus (keratoconus suspect bydefinition): The topographic pattern usually showsan I–S asymmetry with curvature values lower than50.0 D at the cone apex. The CIM and MTK valuesare usually outside the normal range, whereas theshape factor is normal or slightly outside normality.
Keratoconus: The topographic map alteration isevident, with CIM, MTK, and shape factor outsidenormal limits.
OL 38, DECEMBER 2012
2171REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
Another keratoconus screening algorithm is de-scribed by Rabinowitz51 and based on the use of 2descriptive parameters: the central keratometry (K)value and the I–S index. The I–S index characterizesthe dioptric asymmetry between the superior and infe-rior corneal hemispheres. According to these detectioncriteria, the cornea can be classified as normal, kerato-
conus suspect, or clinical keratoconus. Specifically, thecornea is considered as keratoconus suspect when ei-ther the central K value or the I–S index value is abovethe magnitude obtained by adding 2 times the SD tothe mean value; it is considered as keratoconus whenat least 1 of the 2 indexes is above 3 times the SD addedto the mean value.51 The modified Rabinowitz-McDonnell criteria12 were derived from these initialdetection criteria, using the same descriptive parame-ters, central K value and I–S index. These criteria estab-lished that a cornea is susceptible to be catalogued askeratoconus when the central K value is equal to orabove 47.2 D, the value of the I–S index is equal to orabove 1.4 D, and the keratometric difference betweenboth eyes is above 1.0 D (sensitivity and specificity of96% and 85%, respectively).51,53–55
TheMaeda et al.52 criteria allow keratoconus screen-ing by analysis of the keratoconus index (KCI). A KCIvalue above 0% is codified as “similarity to keratoco-nus.” This method has been shown to have an associ-ated sensitivity and specificity for the detection ofkeratoconus of 98% and 99%, respectively. On theother hand, the keratoconus percentage index (KISA%)is a detectionmethod derived from the topographic in-dexes that were originally created for analysis of thecorneal surface with the TMS videokeratograph(Tomey Corp.).55 It uses a combination of 4 topo-graphic parameters: K value (defined as the weightedaverage of the paracentral dioptric value), I–S index,corneal toricity, and nonorthogonal corneal toricityvalue (SRAX).51,55,56 The KISA% index is calculatedas follows:
KISA%ZK � I � S� Cylinder � SRAX � 100
300(1)
A value of KISA% index higher than 100 is classifiedas a susceptible pattern of keratoconus.53,55
The keratoconus prediction index (KPI) is anotherdescriptor used as a clinical tool for keratoconusdetection.57 It is calculated as the combination of 7
J CATARACT REFRACT SURG - V
topographic parameters characterizing the morphol-ogy of the anterior corneal surface: simulated keratom-etry (SimK), differential sector index (DSI), oppositesector index (OSI), center/surround index (CSI), SAI,irregular astigmatism index (IAI), and percent ana-lyzed area (AA). The following formula is used forsuch calculation:
Values of this index above 0.38 can be considered aspathological, with a sensitivity of 86% and a specificityof 100%.57
Other specific descriptors of some corneal topogra-phy systems, such as those provided by theScheimpflug-based Pentacam system (Oculus, Inc.)(anterior and posterior corneal topography analyzer,corneal aberrometer, pachymeter, and anteriorsegment biometer), are calculated according toproprietary algorithms:58 index of surface variation(deviation of the corneal radius respect to the meanvalue), index of vertical asymmetry (characterizationof the level of symmetry of curvature data relative tothe horizontal meridian), keratoconus index, centerkeratoconus index, the smallest radius of curvatureof the corneal area analyzed, index of height asymme-try (characterization of the level of symmetry ofelevation data respect to the horizontal meridian),index of height decentration (estimation obtainedfrom Fourier analysis), and aberration coefficient (esti-mation obtained from Zernike polynomials-basedanalysis). The abnormality ranges for all theseparameters are shown in Table 2.
In addition to these approaches, numerous innova-tions have been developed to obtain more sensitivedetection criteria, such as neural network approachesand specific mathematical models for reconstructionof the geometry of the anterior corneal surface.7,8,59–63
The digital analysis of the image of the Placido disksprojectedonthecorneahasalsobeenshowntobeavalidtool for keratoconus diagnosis, avoiding the use of esti-mated data or parameters obtained indirectly.64 Specif-ically, the use of primary corneal indices characterizingthe asymmetry of the mires seems to be especially use-ful for keratoconus detection.64
ANALYSIS OF THE POSTERIOR CORNEAL SURFACEIN KERATOCONUS
The elevation and curvature of the posterior cornealsurface have been shown to be screening factors for
OL 38, DECEMBER 2012
Table 2. Abnormality ranges for the topographic parametersprovided by the Pentacam system.
Topographic Parameter Suspicious Pathological
Index of surface variation R 0.37 R 41Index of vertical asymmetry R 0.28 R 0.32Keratoconus index O 1.07 O 1.07Center keratoconus index R 1.03 R 1.03Smallest radius of curvature ! 6.71 ! 6.71Index of height asymmetry R 19 O 21Index of height decentration R 0.014 O 0.016Aberration coefficient R 1 R 1
2172 REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
keratoconus.14–16,20–22 Specifically, in keratoconus andalso in keratoconus suspect, significantly larger valuesof BFS14,16,21 and posterior elevation15,16,22 have beenreported (Figure 3). Schlegel et al.16 report a meanaconic astigmatism for the posterior corneal surfaceof 2.29 D G 1.71 (SD) in a sample of 48 keratoconus-suspect patients using scanning-slit technology. Inthe same sample, they also found a mean maximumposterior elevation of 0.0288 G 0.0102 mm at 1.0 mmof radius from the center.16 De Sanctis et al.14 evalu-ated the diagnostic ability of maximum posterior cor-neal elevation measured in the 5.0 mm central zonewith the Scheimpflug photography-based Pentacamsystem (anterior and posterior corneal topography an-alyzer, corneal aberrometer, pachymeter, and anteriorsegment biometer) in 75 patients with keratoconusand 25 eyes with subclinical keratoconus (keratoconussuspect by definition). They found the cutoff values of35 mm (sensitivity 97.3%, specificity 96.9%) and 29 mm(sensitivity 68%, specificity 90.8%) for clinical and sub-clinical keratoconus, respectively. Fam and Lim65 de-fined the posterior elevation ratio as the ratio of themaximum posterior elevation in the central 5.0 mmcorneal zone to the BFS for the posterior cornea.
J CATARACT REFRACT SURG - V
They obtained a mean value for this ratio of 1.874 G0.532 in a sample of 43 keratoconus patients and of1.103 G 0.462 in a sample of 23 keratoconus suspectsusing scanning-slit technology.65 In addition, theyfound excellent keratoconus detection ability of thesame ratio obtained using the anterior elevation data(cutoff: 0.5122, sensitivity: 99%, specificity: 95.2%).65
Another tool used for keratoconus detection bysome clinicians is the dioptric value of the BFS forthe posterior corneal surface provided by the Orbscansystem (scanning-slit technology) (anterior and poste-rior corneal topography analyzer, pachymeter, andanterior segment biometer). Posterior BFS valuesabove 54.0 D should theoretically alert the examinerif the anterior BFS is within the normal range (41.0 to46.0 D).66 However, this dioptric value for the poste-rior corneal surface is not a real estimation because itis calculated assuming a change in refractive index atthe posterior surface identical to that occurring at theanterior surface. Indeed, the posterior corneal curva-ture is always negative due to the refractive indexchange at this surface.3
CORNEAL ABERROMETRIC ANALYSIS IN KERATOCONUS
Anterior corneal aberration analysis has been provento be an effective tool to detect and grade keratoconus(Figure 4).2–6,17,67,68 Higher amounts of vertical comaand larger values of coma-like root mean square(RMS) have been reported in patients with keratoco-nus or keratoconus suspect.2–6,17,67 In 2006, the firstkeratoconus severity grading system that used themagnitude of anterior corneal coma-like aberrationsas the main classifying criterion was developed(Table 3).4 In 23 eyes with keratoconus, 10 eyes withsubclinical keratoconus (keratoconus suspect by defi-nition), and 127 controls, B€uhren et al.17 evaluatedthe keratoconus detection ability of anterior corneal
Figure 3. Tangential map of the posteriorcorneal surface obtained by the PentacamScheimpflug photography. In this kerato-conic cornea, there is an I–S asymmetryin the posterior corneal surface that is sim-ilar to that normally found in the anteriorcorneal surface in keratoconus.
OL 38, DECEMBER 2012
2173REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
higher-order aberrations (HOAs). The found the cutoffpoints of 0.555 (sensitivity 100%, specificity 98.4%) and0.248 mm (sensitivity 100%, specificity 73.6%) for clin-ical and subclinical keratoconus, respectively.17
In addition to anterior corneal aberrations, posteriorcorneal aberrometric data are useful for keratoconusdiagnosis.69,70 Nakagawa et al.69 found large amountsof primary coma in the anterior and posterior cornealsurfaces (primary coma RMS anterior/posterior:3.57/0.87 mm) in 28 keratoconic eyes, although comafrom the posterior surface compensated partly forthat from the anterior surface. The Scheimpflugphotography-based Pentacam system, (anterior andposterior corneal topography analyzer, corneal aberr-ometer, pachymeter, and anterior segment biometer)was the first commercially available topographer toprovide an estimation of the wavefront aberrationsof the posterior corneal surface, although with somelimitations and inaccuracies.3 Currently, accurateaberrometric analysis of the posterior cornea is
Figure 4. Aberrometric analysis (VOL-CT, version 6.20) of the anterior corntopography system. A: Placido-based axial corneal topography. B: Corneasame keratoconic eye. D: Individual representation of each Zernike coeffic
J CATARACT REFRACT SURG - V
provided by almost all Scheimpflug photography-based topography systems. It may be especially usefulfor the detection of keratoconus in its incipient stages,although some evidence demonstrates that posterioraberrations as well as thickness of the spatial profiledo not markedly improve discriminative ability overthat of anterior wavefront alone.71
It is preferable to use corneal wavefront aberrome-try for keratoconus detection rather than total ocularaberration analysis. Total ocular aberrations (globalaberrometry) are difficult if not impossible to mea-sure accurately in highly aberrated eyes. Wavefrontsensors that are available nowadays can analyze upto 1452 points maximum.72 This sampling is limitedfor characterizing a very complex surface that induceslarge amounts of lower-order aberrations and HOAs.In addition, crowding or superimposing light spots orthe assumption of flat slope of each analyzed portionof the wavefront, as happens with Hartmann-Shackdevices, are features that limit the potential of
eal surface in a keratoconic eye obtained by means of a Placido diskl higher-order wavefront map. C: The point spread function of theient.
OL 38, DECEMBER 2012
Table 3. Keratoconus grading system described by Ali�o andShabayek4 and based on anterior corneal coma-like aberrations(coma-like RMS) as the main classifying criterion.
KeratoconusGrade
Coma-likeRMS (mm) Other Features
1 1.50–2.50 No corneal scarringMean central K ! 48 D
2 O 2.50% 3.50
No corneal scarringMean central K ! 53 D
Minimum corneal thicknessO 400 mm
3 O 3.50% 4.50
No corneal scarringMean central K O 53 D
Minimum corneal thickness200–400 mm
4 O 4.50 Corneal scarringMean central K O 55 D
Minimum corneal thickness 200 mm
K Z keratometry; RMS Z root mean square
2174 REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
analysis of these devices in keratoconus. With cornealtopography systems, more than 6000 points can bestudied and then a very exhaustive analysis can beperformed without limitation to the pupillary area.3
For all these reasons, corneal aberrometry has beenestablished as a potential diagnostic tool forkeratoconus.
PACHYMETRIC ANALYSIS IN KERATOCONUS
The pachymetric study, as well as the analysis of thecorneal topographic pattern, is considered an indis-pensable tool in the preoperative screening of refrac-tive surgery candidates and in diagnosing cornealectasia.73 Significant differences in central and mini-mum corneal thickness have been shown betweensubclinical, early, moderate, and advanced keratoco-nus, 67,74 which is useful to characterize the severityof keratoconus. The advances in imaging technologyallow the clinician to obtain pachymetric maps pro-viding point-by-point information of the entire cor-nea (Figure 5). A more detailed and accuratecontrol of the progression of an ectatic disordercan be performed with these imaging advancedsystems, such as those based on Scheimpflug pho-tography, optical coherence tomographers, or veryhigh-frequency ultrasonography systems (Figure 6).
In 46 eyes diagnosed with mild to moderate kerato-conus and364normal eyes,Ambr�osio et al.24 evaluatedthe corneal thickness at the thinnest point and the aver-ages of the points on 22 imaginary circles centered onthe thinnest point with increased diameters at 0.4 mmsteps using the Pentacam system and the Orbscan sys-tem (scanning-slit technology) (anterior and posterior
J CATARACT REFRACT SURG - V
corneal topography analyzer, pachymeter, and ante-rior segment biometer). They found statistically signif-icant differences between groups in all positions ofcorneal-thickness spatial profile as well as in the per-centage increase in thickness between 3.5 mm and7.0 mm diameters.24 These authors conclude that thispachymetric analysis could diagnose keratoconusand screen refractive candidates. In a study using thescanning-slit technology, Saad and Gatinel75 foundthat indices generated from corneal thickness (percent-age of thickness increase from the thinnest point to theperiphery) and curvature measurements (percentageof variation of anterior and posterior curvatures) overthe entire cornea centered on the thinnest point canidentify very mild forms of ectasia undetected bya Placido-based neural network program. In a recentstudybyAmbr�osioetal.76usingthePentacamtechnology,the tomographic-derived pachymetric parameters(pachymetric progression indices) were better factorsto differentiate between normal and keratoconic cor-neas than single-point pachymetricmeasurements. Us-ing a specific model of an optical coherencetomographer, Li et al.77 found that pachymetric mapsof 27 keratoconic eyes were more asymmetric (I–S,�44.8 G 28.7 mm versus �9.9 G 9.3 mm; cutoff,�31.3 mm; inferotemporal–superonasal, �63G 35.7 mmversus �22 G 11.4 mm; cutoff, �48.2 mm) than mapscorresponding to 34 controls evaluated. The cutoffpoints in this study had an associated sensitivity andspecificity similar to KISA% index.77
Analysis of the epithelial thickness profile has alsobeen shown to be useful for keratoconus detectionand characterization.78,79 Specifically, the presenceof an underlying stromal cone seems to be associ-ated with an epithelial doughnut pachymetricpattern. Therefore, this characteristic epithelial thick-ness doughnut may be considered a diagnostic testfor keratoconus. However, it should be consideredthat epithelial compensation can mask the presenceof an underlying cone in very early keratoconus.78
CORNEAL VOLUME ANALYSIS IN KERATOCONUS
Corneal volume was recently identified as anadditional screening factor for keratoco-nus.24,46,74,80,81 Significant differences in corneal vol-ume have been reported between normal andmoderate keratoconic eyes (Pentacam system: 60.83G 3.27 mm3 controls versus 57.98 G 2.65 mm3 mod-erate keratoconus),46 suggesting the potential rolefor corneal volume as a diagnostic factor for cornealectatic disorders. However, there is not enough sci-entific evidence of the potential usefulness of cornealvolume as a screening factor for keratoconussuspect.
OL 38, DECEMBER 2012
Figure 5. Pachymetric map of a keratoconiceye obtained by means of the Scheimpflugphotography-based system.
2175REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
CORNEAL BIOMECHANICS IN KERATOCONUS
As all the topographic and aberrometric alterationsin keratoconic eyes are the consequence of the bio-mechanical changes in the corneal structure,2 thestudy of corneal biomechanics has been proposedas an additional test for keratoconus diagnosis.82,83
However, the in vivo evaluation of corneal biome-chanical properties is not easy. To date, only oneclinical device has been developed for characterizingthe viscoplasticity of the cornea, the Ocular Re-sponse Analyzer (ORA) (Reichert, Inc.) (corneal bio-mechanics analyzer and air tonometer), which is
Figure 6.Module of pachymetric analysis from the Pentacam system. Normperiphery. A: Normal cornea. B: Keratoconic cornea.
J CATARACT REFRACT SURG - V
based on analysis of the corneal reaction during a bi-directional applanation process.84 Specifically, thisdevice delivers an air pulse to the eye that causesthe cornea to move inward, achieving a specific ap-planation state or flattening (Pressure 1). Millisec-onds after the first applanation, the pressuredecreases and the cornea passes through a second ap-planated state (Pressure 2) while returning from con-cavity to its normal convex curvature (Figure 7).The initial version of this instrument provided only2 biomechanical parameters associated with thisprocess: corneal hysteresis (CH) and the corneal
ality ranges are provided for the increase in thickness from center to
OL 38, DECEMBER 2012
Figure 7. Signal diagram obtained with theORA in a normal eye. The red line representsthe applanation signal and the green line thepressure changes. As shown, the device de-livers an air pulse to the eye, which causesthe cornea tomove inward achieving a specificapplanation state or flattening (Pressure 1).Milliseconds after the first applanation, thepressure decreases and the cornea passesthrough a second applanated state (Pressure 2)while returning from concavity to the normalconvex curvature.
2176 REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
resistance factor (CRF). Corneal hysteresis is definedas the difference between the 2 pressures (Pressures1 and 2) recorded during the measurement processdescribed. The CRF is calculated using a proprietaryalgorithm and is said to be predominantly related tothe elastic properties of the cornea.84 These parame-ters, CH and CRF, have been shown to be reproduciblein nonoperated healthy eyes.85
Some authors have demonstrated that CH andCRF are significantly lower in keratoconic eyesthan in normal eyes.82,83 Ortiz et al.83 comparedthe ORA outcomes in 165 normal eyes and 21eyes with keratoconus (47% grade I according tothe Amsler-Krumeich grading system), obtainingthe mean values of 10.8 G 1.5 mm Hg and 7.7 G1.3 mm Hg, respectively. The between-group differ-ence was statistically significant. However, CH andCRF have been demonstrated as poor parametersfor discriminating between mild keratoconus andnormal corneas.86 Specifically, Fontes et al.86 per-formed a receiver operating characteristic curveanalysis, obtaining a poor overall predictive accu-racy of CH (cutoff, 9.64 mm Hg; sensitivity, 87%;specificity, 65%; test accuracy, 74.83%) and CRF(cutoff, 9.60 mm Hg; sensitivity, 90.5%; specificity,66%; test accuracy, 76.97%) for detecting mild kera-toconus. In addition, the CRF has been better thanCH in detecting keratoconic corneas once the effectof central corneal thickness on ORA measurementsis considered.87 It should be noted that althoughthe manufacturer states that the CH may primarilyreflect corneal viscosity and the CRF may predom-inantly relate to the elastic properties of the cornea,the exact physical meaning of these parameters isnot well understood. The parameters are said torepresent the viscoelastic properties of the cornea,but no study proves whether these parameters arerelated to the standard mechanical properties used
J CATARACT REFRACT SURG - V
for describing the elastic materials (Youngmodulus).
To overcome the limited diagnostic performance ofCH and CRF, new parameters based on the analysis ofthe response signal curve have been developed.88 Firstresults with this type of analysis are promising,88 butmore scientific evidence of its diagnostic value isrequired.
INTEGRAL ANALYSIS OF CORNEAL STRUCTURE
No clinical parameter provides 100% sensitivity andspecificity for the detection of keratoconus, especiallythe subclinical forms. The combination of several clin-ical tests is extremely useful for detection and com-plete characterization of keratoconus.67 It should bementioned that anamnesis is crucial because some an-tecedents, such as the diagnosis of corneal ectasia inother members of the family, can alert the clinician.Some studies reveal the presence of topographic pat-terns compatible with keratoconus in family membersof patients diagnosed with keratoconus, as well as inthe fellow eyes of patients in whom keratoconus hasbeen clinically diagnosed.89–91 Therefore, new diag-nostic criteria based on parameters obtained from dif-ferent clinical examinations, not only analysis of theanterior corneal surface, are necessary, especiallywhen the most incipient cases are to be detected. Re-cent research on keratoconus is focused on this aim, at-tempting to find multiple models for predicting thepresence of the ectasia. Using logistic regression anal-ysis, Ucakhan et al.92 found that the combined analysisof anterior and posterior corneal power, elevation, andthickness data provided by a Scheimpflug device ef-fectively discriminated between ectatic corneas andnormal corneas. Kovacs et al.93 observed that therewas a threshold level of posterior corneal elevation(40 mm) and corneal thickness (450 mm) beyond which
OL 38, DECEMBER 2012
Figure 8.Keratoconus suspect topographical findings.A andD: Right eye (RE) and left eye (LE) sagittal maps show inferior steepening. B and E:Posterior elevation maps of both eyes do not show high elevation values. C and F: Pachymetry map values from both eyes are within normallimits.
2177REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
the level of corneal protrusion in keratoconus acceler-ated. Ali�o et al.67 found that the vectorial difference be-tween refractive (calculated to the corneal plane) and
Figure 9. Appropriate selection of the reference surface is mandatory for anerence surface is a sphere, themap does not showvalues outside normal limshows suspicious values. C: Similarly, when a toric ellipsoid is chosen as rated with ectasia.
J CATARACT REFRACT SURG - V
corneal astigmatism was a relevant factor to includein predictive models for keratoconus detection (ocularresidual astigmatism). Specifically, these authors
accurate diagnosis of an incipient ectasia. A: When the selected ref-its. B: If an ellipsoid is selected as reference surface, the elevationmapeference surface, posterior elevation map again shows values associ-
OL 38, DECEMBER 2012
2178 REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
found a significantly highermagnitude of ocular resid-ual astigmatism in keratoconic eyes showing morevisual deterioration.67
KERATOCONUS SUSPECT
In the keratoconus suspect condition, the cornea doesnot show any typical alteration of a clinical keratoco-nus such as Munson sign, scissors-like retinoscopic re-flex, asymmetric corneal thinning, Fleischer ring, orVogt striae. Therefore, the keratoconus suspect corneais apparently normal, except for the suspicious topo-graphic signs that can be detected (Figure 8). As notedpreviously, anterior corneal topography is crucial forthe diagnosis of keratoconus suspect: I–S index of 1.4 Dor higher, hemimeridian angling higher than 20 de-grees in relation to the vertical meridian, and inferiorsteepening.94 Studying the posterior corneal surfaceusing the elevation map also provides valuable infor-mation for detecting incipient states of ectatic disor-ders (Figure 9), as noted in a previous section. Thistype of examinationmay be especially useful in detect-ing very incipient keratoconus because hyperplasic ac-tivity of the corneal epithelium is able to mask subtleectatic changes.
This type of ectasia constitutes a contraindication tolaser in situ keratomileusis because the flap creation inconjunction with the excimer laser ablation might pro-duce a biomechanical disorder that can trigger themechanisms toward progressive forms.95 For thisreason, photorefractive keratectomy or laser-assistedsubepithelial keratectomy has been presented as anadequate technique in keratoconus suspect patients,although the potential risk for iatrogenic ectasia isnot completely eliminated.96,97
ATYPICAL FORMS OF KERATOCONUS
Although the most common keratoconus form is thepresence of localized steepening over the inferior mid-peripheral zone, cases of superior keratoconus havebeen reported.98,99 The characteristic topographicpattern is similar to that corresponding to classic kera-toconus but with the localized steepening locatedabove the horizontal meridian.
Posterior keratoconus consists of a central or para-central depression in the posterior corneal surfacewith an intact but abnormal Descemet membrane.100
Two forms of posterior keratoconus can be differen-tiated according to the extent of the area of the pos-terior corneal surface that is affected: the entireposterior corneal curvature (keratoconus posticus to-talis) or a localized portion (keratoconus posticuscircumscriptus).101,102 It is an uncommon, nonin-flammatory, and nonprogressive corneal disease
J CATARACT REFRACT SURG - V
that is normally unilateral. Posterior keratoconus isusually detected during a routine ophthalmologicexamination when moderate amblyopia or irregularretinoscopic reflexes are observed. In these cases, theanterior corneal surface is not altered, althoughMannis et al.102 also observed specific abnormalitiesin this surface.
PSEUDOKERATOCONUS: ARTIFACTS IN THECHARACTERIZATION OF CORNEAL STRUCTURE
Some acquisition artifacts with currently availablecorneal topography systems can lead to errors in in-terpretation of the measurements by the topogra-pher software. Some of these errors consist of thedetection of unreal corneal surfaces susceptible ofbeing interpreted by the clinician as keratoconic orsuspicious of keratoconus (pseudokeratoconus).The following are factors that can induce this typeof artifact:
Incorrect patient positioning during measurement:The correct patient position during the topographic ac-quisition as well as the correct eye alignment with thefixation stimulus is crucial to obtain repeatable andclinically useful topographic data. Eyelids under con-tinuing pressure from different gaze positions havebeen shown to influence anterior corneal shape.103 Inaddition, patients showing normal videokeratogra-phies with a symmetrical bowtie pattern might showaltered videokeratographies after some eye move-ments (Figure 10).
Contact lens–induced corneal warpage: Contactlens wear, especially with rigid gas-permeable lenses,might significantly alter corneal topographic patterns.Rigid gas-permeable lenses should be considered tohave a higher rigidity module, and their bearing onthe cornea can cause some transitory effect. This isthe reason for asking contact lens wearers to discon-tinue use of the lenses at least 3 weeks before refractivesurgery screening to obtain information of the cornealbasal state (Figure 11).104
Tear-film stability and dry eye: Tear-film alter-ations constitute one of the most important factorsfor inducing the presence of artifacts in a topo-graphic examination. A low tear film quality (lowbreakup time) or quantity (low Schirmer testvalues) might generate the presence of alterationsin the anterior corneal surface such as epitheliopa-thies that can contribute to obtaining altered topo-graphic patterns (Figure 12).105 For all thesefactors, it is recommended that topographic acquisi-tion be performed after an eye blink and that theconsistency of measurement in cases of dry eye beconfirmed by taking 3 consecutive measurementsand comparing them.
OL 38, DECEMBER 2012
Figure 10. Appropriate alignment during acquisition is mandatory to avoid misdiagnosis. A: Topography map in supraversion eye position.B: Topography map in infraversion eye position. It looks similar to an inferior steepening that might mislead the diagnosis.
2179REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
In conclusion, the following key points can beextracted from this review:
Combined analysis of anterior and posterior cornealpower, elevation, and pachymetry data provides effi-cient discrimination between normal and ectatic cor-neas and a proper profound analysis for keratoconusfollow-up purposes.
The use of tangential maps having adequatedioptric power steps is necessary for a high-qualitycorneal ectasia detection process when performing to-pographical analysis because erroneous selectionmight mislead decision-making procedures.
Numerous helpful topographic indexes and de-scriptors have been developed to characterize cornealshape. Clinicians should be familiar with the indexes
J CATARACT REFRACT SURG - V
provided by the ophthalmic devices commonly usedduring their daily clinics.
Wavefront analysis has recently been incorporatedinto the common screening test battery in outpatienteye clinics; thus, this technology might also bea valuable tool when diagnosing ectatic corneas.Other new cutting-edge technological analysessuch as corneal volume assessment might be futurescreening factors that can be incorporated into theclinical practice.
Finally, accurate detection of subclinical or forme-fruste keratoconus is mandatory when counselingcorneal refractive surgery candidates. Similarly, elimi-nation of image acquisition artifacts during diagnosticprocedures is obligatory to avoidmisleadingdiagnoses.
OL 38, DECEMBER 2012
Figure 11. Differential keratometric map be-tween the topography taken after 1 week (B)and 6 weeks (A) of RGP contact lens removal.A central cornea flattening has occurred afterthe RGP contact lens removal.
Figure 12. Tear-film breakup must be avoided during acquisition. A: Placido disk image shows irregularity while performing noninvasivebreakup time test. B: Topography corresponding to image A. C: Homogeneous Placido disk image after preservative-free eyedrops instillation.D: Topography corresponding to image C.
2180 REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS
J CATARACT REFRACT SURG - VOL 38, DECEMBER 2012
2181REVIEW/UPDATE: CORNEAL STRUCTURE IN KERATOCONUS