+ +

Screening indices

General, definitions

The vast majority of testing indexes of the Keratoconus Derive from the information collection of curvature topography (Placido topography). The keratoconus is clinically characterized by the occurrence of a corneal deformation ()Directly and exclusively linked to repeated eye rubbing According to the author of this site). The corneal topography is an exploration that allows to apprehend and represent the corneal geometry. Thus, topographic explorations are based on the detection and quantification of the irregular deformation of the corneal dome.

During the introduction of the corneal videokeratoscopie computer-assisted (now known under the term of "") corneal topography"), the first warning signs that can be mentioned the presence of a Keratoconus were." qualitative : they were based on the simple aspect of the inspection of the topographic maps present pattern, while the "numbers" of curvature are considered to quantify the characteristics of previous analyzed corneal curvature.

Screening indices quantitative were naturally then designed from the analysis and the collection of encrypted data from the analysis of the specular reflection. They will be detailed later, and stress that they result from data restricted to the only front of the cornea.


It is useful to understand that every screening test quantity is associated with a 'threshold '. This threshold is a number, which separates into two categories the people tested for the chosen test: normal vs. abnormal - positive vs negative, etc. In medicine, no test has an absolute reliability; This is not necessarily related to test him - even, but to the fact that there is almost never a real border value, but rather a interval boundary values between individuals or criteria, normal and abnormal. This is for example the case for the measurement of blood sugar levels with the risk of diabetes, or blood pressure towards the risk of HTA (high blood pressure); There is no absolute threshold of blood sugar or blood pressure beyond which we are absolutely certain that the subject is diabetic or hypertensive (or threshold below which it is certain that all tested individuals are healthy).

For practical reasons, however he must choose a threshold, and it is determined according to statistical and clinical criteria. "" You can choose a threshold that minimizes number of misclassified comments: detected as positive observations are labelled "false positives", and classified as negative comments are designated as case of "false negatives. You can for example decide to adjust the threshold of the test will not pass too many 'false negatives '. For example, the ringtones of the airport security gates, "" often sound for which turns after control be false alarms, because it is important to minimize the risk of letting a weapon, etc.) : this test is very sensitive, but not specific: it gives a lot of false positives for (hope - the) no false negative.

screening test: false negative false positives

Due to the distribution of Gaussian ("bell-shaped") of biological variables, it is difficult to separate two groups without error (healthy vs pathological). Some suspicious forms can be normal corneas (false negatives) and vice versa (false positives). The sensitivity and the specificity of a test is calculated from the proportion of observations well classified (true positives, true negatives) and misclassified (false positives, false negatives). The threshold is arbitrary: it is chosen from preliminary studies on known samples. The threshold is "low", the test is sensitive, but loses in specificity, since a growing number of observations will be falsely classified as positive.

Let's go back to the field of early forms of Keratoconus screening: in the context of the' preoperative assessment in LASIKIt makes sense to focus on the specificity sensitivity because it is certainly less annoying not to operate a cornea ranked by excess as positive for Keratoconus infra clinical ('false positive'), to operate a cornea "innocent" by mistake but also reached rough form of Keratoconus ("false negative").

The screening of Keratoconus in its advanced forms is now relatively easy because ophthalmologists are accustomed to the topographical appearance of Frank forms of the disease.

Many clues derived from the videotopographie of curvature (technology Placido) have been developed for Keratoconus screening: all were based on the topographic examination of the anterior face of the cornea. Some corneas yet considered to be 'normal' in the light of these indices of screening have nevertheless developed a form of ectasia post LASIK. Force is to recognize the sensitivity of these indices is sometimes insufficient to current issues. Fortunately, since the availability of topographers of elevation (Orbscan, Galilei, Pentacam), it is possible to have topographical maps of the posterior face of the cornea and corneal thickness analysis "point by point". These data, added to those provided by bending cards, allow to increase the sensitivity and specificity of early diagnosis of Keratoconus infra clinical.


The real challenge of the contemporary techniques of topographical screening is very beginner forms of Keratoconus screening, because they represent an absolute contraindication to performing a LASIK, at the risk of occurrence of a corneal ectasia type complication. However, the fledgling infra clinical forms (which are grouped together under the term of fruste Keratoconus) are often have a minor topographic expression. Screening indices are 'markers' that help the clinician to judge of the risk of being in the presence of a fledgling form of Keratoconus.



Qualitative screening criteria


The topographic criteria for the screening of Keratoconus were initially proposed by Rabinowitz for the rapid identification of a suspicious form of keratoconus were qualitative.  It is profitable to recall them because they have been the source of certain quantitative criteria.  These aspects are presented in the following figure:

qualitative aspects for the screening of Keratoconus

Patterns (patterns) schematic and examples correspondents (with intermediate forms) of corneas 'at risk' of sub clinical form of Keratoconus according to Rabinowitz et al. These aspects have been obtained in axial curvature (previous specular topography), through an absolute scale, and arbitrarily designated with a letter of the alphabet between A and J. here are represented as particularly suspicious aspects to the beginner Keratoconus:
D: appearance of pronounced camber isolated lower
G: excessive angulation of the arched meridians: 'throw Radial Axes"(SRAX)
H: "bow tie" asymmetric: "Asymmetry Bowtie" (AB)
J: combination of G and H: forms this pattern is referred to by the acronym AB/SRAX: this is a certain risk factor to a proven form.
It should be noted that the islets of warm color do not correspond to 'bumps', but of the regions where the anterior corneal surface has a pronounced camber.


These qualitative topographic aspects were described from the data gathered with relatives of patients with Keratoconus, or eyes known (sometimes retrospectively, after previous topography study conducted before the appearance of more forthright signs) with new forms of Keratoconus. Used topographer was the TMS - 1 (Tomey) with an absolute scale covering the range 35-50 D in steps of 1.5 D (either 10 colors), the curvatures are outside these values were represented by 5 d No.  This is important, because it is essential to emphasize that evocative topographical aspects of a suspect form for Keratoconus can be removed by a too large scale. The reliability of the topographic examination (adequate fixation) is also essential to correctly interpret these patterns.


The existence of a pattern of AB/SRAX, combining an asymmetrical bow (asymmetry Bowtie: AB) and an excessive angulation of the arched Meridian hemi (throw Radial Axes: SRAX) is frequently found (80%) on the image of topography in curvature of the 'healthy' eye in patients of Keratoconus "unilateral." This aspect represents a risk of progressing towards a form proven to be higher than 50%.


These topographic aspects, a central keratometry value greater than 47.2D, and/or a central keratometry difference greater than 1 d between the two eyes were other clues that might evoke the presence of fruste Keratoconus. In order to increase their specificity (a significant proportion of uninjured Keratoconus corneas may show one or more of these characteristics), other quantitative elements were then proposed.


 Quantitative indices

 Index I-S

This index uses data from the specular topography by Placido disc.  It is to average the IOL values obtained at different points of the lower hemicornee (I) and subtract them the average correspondent for the corresponding points in the hemicornee lower (S). The calculation of I is equal to the average of 5 measurements of keratometry on the ring of the 3 mm Central, at the intersection of the Meridian lines 210, 240, 270, 300 and 330 °. The calculation of S is made according to the same principle with the meridians 30, 60, 90, 120 and 150 °.

index I-S for the screening of Keratoconus

Index I-S: the diameter of the circle centered on the vertex is 3 mm. The average of the values in the upper hemicornee is S = 40.214D. The average of the values measured in the lower hemicornee is I 41.614 = D. The index (I-S) is equal to 1. 4 d. Despite the slightly below average keratometry figures, index (I-S) is positive for the topographic diagnosis of "suspicious cornea in Keratoconus.

A threshold value of 1. 4 d for the index I-S has been chosen as allowing to decide between "physiological" asymmetry and proven presence of Keratoconus infra clinic. A value greater than 1.9D is commonly associated with a clinical form of Keratoconus (visible thinning at the slit, Fleischer ring lamp, etc.). In order to increase the sensitivity of this test, it is recommended to place one of the points used for the collection of the IOL on the hemi-Meridian the arched lower values.

Index KISA%

It was developed by Rabinowitz and Rasheed (15). It is based on the use of the value expressed in degrees of the angulation of the two most bulging hémiméridiens (SRAX), as well as the values of the maximal keratometry removed of D, of theastigmatism k, and subtraction (I-S). THE formula for the calculation of the composite index KISA% is:


Kisa% = K x (I-S) x 0.3 x SRAX

With: K = 1 if the maximum keratometry is less than 47 d, (I-S) = 1 if (I-S) < 1 and Ast = 1 if Ast < 1.


indices SRAX and KISA

KISA index is based on the use of a combination of qualitative and quantitative indices, including the calculation of the SRAX (for "Throw Radial Axes", or radial deviation of the axes). The axis of the hemimeridien of maximum curvature is identified within each hemicornee (upper and lower). The SRAX angle is the most acute angle formed by the directions of each hemimeridien. This index is often determined "manually", but it was also provided by the software equipping the topographer "Eyemap" (Alcon), which was equipped with a disc of 22 rings Placido. The reflection indirectly SRAX ' one-sidedness ' of measures of axial curvatures. This pattern disappears from the representation of the average curvature or Gaussian.

According to Rabinowitz and Rasheed:

-KISA% < 60: normal cornea

– 60 < KISA% < 100 : kératocône fruste

-KISA% > 100: Keratoconus.



Other statistical indicators (specular topography)

The topography specular to calculate the local k power in many ways. From these values, various indexes can be calculated, sometimes weighted by the area of the corneal surface that they are interested in. These indices was initially provided by the topographer TMS-1 (21). Here is a list of clues, whose value is likely to be affected by the presence of a Keratoconus (table  According to bring intervals of normal values and thresholds of suspicious values, the following Figure represents a graphic interpretation of indications)


 Clues related to the degree of toricite of the corneal surface:


SimK1/SimK2: Purpose keratometry / simulated Keratometry

Corneal powers of the arched and flat areas in the area of 3 mm.

CYL: Purpose keratometric cylinder / cylinder k simulates

Difference between the values of SimK1 and SimK2.


 Clues related to the degree of fluctuation of power k:

CVP: Coefficient of Variation of corneal Power / Coefficient of variation of corneal power

This index corresponds to the scope of the distribution of the IOL powers present on the surface analyzed around the average: CVP = 1000 x (standard deviation of total corneal power (DSP) / average total corneal power).


SDP: Standard Deviation of corneal Power / Standard Deviation of the corneal power

Value of the standard deviation of the corneal power total


AA: Analyzed Area / analyzed Surface

The value of the corneal surface covered by the targets could be analyzed. It can be reduced in Keratoconus advanced case (excessive deformation of rods).

 Clues related to the degree of asphericity of the corneal surface:

CIS: Corneal Eccentricity Index / Index of eccentricite of the cornea

It corresponds to the calculated value of the eccentricite of the cornea. This eccentricity is becoming increasingly prolate in case of keratoconique deformation.

 Clues related to the degree of asymmetry and irregularity of the corneal surface:

DSI: Differential Sector Index / Index of sectoral difference

The corneal surface is divided into 8 equal pieces (sectors), including the internal angle sweeps 45 °. The average axial k power is calculated for each sector. DSI index is equal to the maximum difference between any of these sectors. Its value increases in case of asymmetric camber.


SRI: Surface Regularity Index / Index of regularity of surface

The SRI is correlated to the ACP index because it matches local irregularities of the corneal surface. It is calculated as the summation of the fluctuation of the local power along 256 meridians hemi regularly spaced, on 10 central targets. Affection responsible for a reduction of the regularity of corneal surface induces an increase in the value of this index.


CBC: Area compensated Surface Regularity Index / index of regularity compensated for the surface

This index is calculated as a SRI standardized according to the area of the analyzed surface. It increases for the same reasons as the SRI.


SAI: Surface Asymmetry Index / Index of asymmetry of surface

The SAI is a measure that incorporates the power difference k between pairs of two opposite points located on 128 meridians.


IAI: Irregular Astigmatism Index / Index of irregular astigmatism

The IAI is a function of the power variation k of ring ring along a given Meridian.


OSI: Opposite Sector Index / Index of sector opposite

The corneal surface is divided into 8 equal sectors, including internal angle is equal to 45 °. The average power of each sector is calculated, and OSI is equal to the maximum difference measured between two opposite sectors. Its value increases in case of irregular astigmatism, characterized by the presence of a significant degree of asymmetry


CSI: Center-Surrounded Index / Index Center - periphery

The CSI is the difference between the average axial coubure in the Central 3 mm, and that calculated in the ring (from 3 to 6 mm).


DSI OSI CSI screening indices

Calculation of indices DSI, OSI and CSI. The corneal surface is divided into 8 arbitrary sectors (45 ° each). The average axial curvature is calculated within each sector. The calculation is repeated by changing the orientation of each sector to sweep all of the possible cutouts (for example, if the keratometry is analyzed on points distributed according to 256 hemimeridiens, there will be 256/8 = 32 possible cutouts). The DSI (Differential Sector Index) index is the maximum difference between any sectors. OSI (Opposite Sector Index) is the maximum difference between two opposite sectors. The CSI (Center Surround Index) is the average difference k between the Central 3 mm and including ring between 3 and 6 mm.


To allow the use of these indexes in a statistical manner for Keratoconus screening, some have been grouped in a function called KPI (Keratoconus Prediction Index).


KPI: Keratoconus Prediction Index / Keratoconus prediction Index

The KPI is a discriminant function calculated from the values obtained for the following indexes: SimKl, SimK2, SAI, DSI, OSI, CSI, IAI, AA. It is described in the article: Maeda N, Klyce SD, Smolek MK, Thompson HW: Automated keratoconus screening with corneal topography analysis. Invest invest live Sci, 1994; 35:2749 - 2757



Automated detection of suspicious forms of Keratoconus

Calculated indices are as much quantitative data that can be used to build statistical models designed to separate into diagnostic groups tested corneas. The topographer TMS-1 (Tomey) software has the criteria of Rabinowitz (KISA%) and Klyce & Maeda; the following figure provides an example where the criteria of Rabinowitz conclude to the presence of a suspect form of Keratoconus, unlike the criteria Klyce and Maeda.

Screening criteria Rabinowicz Klyce and Maeda Keratoconus indices

: Topographic map in axial mode and Indices of screening of Keratoconus provided by the topographer TMS-1. There is an increase of curvature in lower answers. The index (I-S) is positive for Keratoconus suspect (1.51 > 1.4): this value is suspect according to the values proposed by Rabinowitz. Despite the values higher some clues (ACP: average corneal power, AA), the composite index KPI is negative for Keratoconus. Despite these warnings, that eye was operated of LASIK, and an ectasia occurs in the postoperative period, confirming a posteriori the presence of a suspect form of Keratoconus.

This example highlights the lack of sensitivity and absolute specificity of screening tests. In addition to the topographer TMS software, defined above indices are used in the computer program "Corneal Navigator", implemented in software OPDstation (topographer OPD scan®, Magellan®, Nidek, Japan topographer). From the calculated values of all the index mentioned in this paragraph c), presented a percentage of similarity with one of the following clinical conditions: normal (NRM), astigmatism (AST), Keratoone Suspect (KCS), Keratoconus (KC), pellucid marginal degeneration (PMD), surgery corneal myopia (SRM), corneal of hyperopia (HRS) surgery, indeterminate form (OTH).

Suspicious keratoconus klyce maeda topography specular indices

: Example of the report using the clues topograhiques of Klyce and Maeda. The clues so the values are suspicious are yellow, those who are frankly pathological red (here the IAS). After analysis of these clues by a neural network, a percentage of similarity with a particular condition is provided (26.4% for a suspect form of Keratoconus in this example).


Alio et al. have studied the analysis of aberrations (HOA) high degree of the anterior corneal surface as a tool to assess the severity of a Keratoconus by using the videokeratoscopie with analysis of the previous corneal maps (3). They found that the earlier corneal HOA, including type coma aberrations, were significantly more important in bearing eyes of Keratoconus than in normal eyes. Type coma aberrations could be good indicators for the early detection and evaluation of Keratoconus. Choosing a threshold and the corresponding values of sensitivity and specificity no were not reported, limiting the value of this approach for the screening of suspicious forms help for now.


Composite index including elevation anterior, posterior and thickness data: SCORE Analyzer

This index consists of a single number, the calculation requires the integration of 12 clues in a discriminant function. A page is dedicated to this strategy, which has been available since January, 2014 for users of the Orhscan IIz topographer: SCORE Analyzer



(1) Rabinowitz YS, McDonnell PJ. Computer-assisted corneal topography in keratoconus. Refract Corneal Surg, 1989; 5 (6): 400-8

(2) Rabinowitz YS, Garbus J, McDonnell PJ. Computer-assisted corneal topography in family members of patients with keratoconus. Arch invest. 1990; 108 (3): 365-71.

(3) Rabinowitz YS. Keratoconus. Sud invest. 1998; 42 (4): 297-31

(4) Rabinowitz YS. Tangential vs sagittal videokeratographs in the 'early' detection of keratoconus. Am J invest. 1996; 122 (6): 887-9.

(5) MK, Klyce SD Smolek. Current detection methods compared with keratoconus is neural network approach. Invest invest Vis Sci. 1997; 38 (11): 2290-9.

(13) Saad, Gatinel D. Topography and Tomography Properties of form Fruste Keratoconus Corneas. Invest invest Vis Sci. 2010, June 16 online

(15) Rabinowitz YS, Rasheed K. KISA% index: a quantitative algorithm embodying videokeratography minimum topographic criteria for diagnosing keratoconus. J Cataract Refract Surg. 1999; 25 (10): 1327-35.

(16) li X, Rabinowitz YS, Rasheed K, Yang H. Longitudinal study of the normal eyes in unilateral keratoconus patients. Ophthalmology. 2004; 111 (3): 440-6.

(17) Maeda N, SD, MK, Thompson HW Smolek Klyce. Automated keratoconus screening with corneal topography analysis. Invest invest Vis Sci. 1994; 35 (6): 2749-57.

(18) Alio JL, Shabayek MH. Corneal higher order aberrations: a method to rank keratoconus. J Refract Surg. 2006; 22 (6): 539-45


Leave a comment

You can ask questions or comment on this content: for this, use the "comments" form located below. The questions and comments of a general interest will be processed and published, and the information provided on the relevant pages should be clarified or supplemented.

Your e-mail address will not be published. Required fields are indicated with *