- Email: [email protected]
Subclinical Keratoconus
Sensitivity and Specificity of Posterior Corneal Elevation Measured by Pentacam in Discriminating Keratoconus/Subclinical Keratoconus Ugo de Sanctis, MD, PhD,1 Carlotta Loiacono, MD,1 Lorenzo Richiardi, MD, PhD,2 Davide Turco,1 Bernardo Mutani, MD,1 Federico M. Grignolo, MD, PhD1 Purpose: To estimate the sensitivity and specificity of posterior elevation in discriminating keratoconus and subclinical keratoconus from normal corneas. Design: Prospective case-control study. Participants: Seventy-five patients with keratoconus, 25 with subclinical keratoconus, and 64 refractive surgery candidates with normal corneas. Methods: In one eye of each patient, posterior corneal elevation was measured in the central 5 mm using the Pentacam rotating Scheimpflug camera (Oculus, Wetzlar, Germany). Posterior corneal elevation in keratoconus and subclinical keratoconus were compared with that in normal corneas in separate analyses. Receiver operating characteristic (ROC) curves were used to determine the test’s overall predictive accuracy (area under the curve) and to identify optimal posterior corneal elevation cutoff points to maximize sensitivity and specificity in discriminating keratoconus and subclinical keratoconus from normal corneas. Logistic regression was used to support cutoff points identified through ROC curve analysis, and to check for model validity; model goodnessof-fit was estimated using r2, and its internal validation was by bootstrapping analysis. Main Outcome Measures: Posterior corneal elevation in keratoconus, subclinical keratoconus, and normal corneas. Results: Mean posterior corneal elevation was statistically higher in keratoconus (100.7⫾49.2 m; P⬍0.001), and subclinical keratoconus (39.9⫾15.0 m; P ⫽ 0.01) versus normal corneas (19.8⫾6.37 m). ROC curve analyses showed high overall predictive accuracy of posterior elevation for both keratoconus and subclinical keratoconus (area under the curve 0.99 and 0.93, respectively). Optimal cutoff points were 35 m for keratoconus and 29 m for subclinical keratoconus. These values were associated with sensitivity and specificity of 97.3% and 96.9%, respectively, for keratoconus, and 68% and 90.8% for subclinical keratoconus. Similar cutoff points were obtained with logistic regression analysis (38 m for keratoconus and 32 m for subclinical keratoconus). The models showed good fit to the data, including after internal validation. Conclusions: Posterior corneal elevation very effectively discriminates keratoconus from normal corneas. Its efficacy is lower for subclinical keratoconus, and thus data concerning posterior elevation should not be used alone to stratify patients with this condition. Financial Disclosure(s): The authors have no proprietary or commercial interest in any materials discussed in this article. Ophthalmology 2008;115:1534 –1539 © 2008 by the American Academy of Ophthalmology.
Keratoconus is a noninflammatory disorder, characterized by corneal thinning and anterior protrusion.1,2 Detection of this disease is particularly important among patients considering refractive surgery, in whom the prevalence of keratoconus and subclinical keratoconus has been reported to be greater than in the normal population3,4; keratorefractive procedures may have unsatisfactory results and cause postoperative complications in these patients.5–7 Moreover, apart from excessive ablation of corneal tissue, unidentified subclinical keratoconus is considered to be the main cause of ectasia after LASIK.8,9 Clinical diagnosis of keratoconus in eyes with biomicroscopic, keratometric, and retinoscopic signs is not difficult.
1534
© 2008 by the American Academy of Ophthalmology Published by Elsevier Inc.
Moreover, advances in Placido disk-based computerized videokeratoscopy have lead to a variety of quantitative indices that have been found to be highly sensitive and specific in aiding with the diagnosis of keratoconus.10 –13 Diagnosis of forme fruste keratoconus or subclinical keratoconus is more challenging. The terms “forme fruste keratoconus” and “subclinical keratoconus” were introduced to indicate a very early preclinical stage of the disease, in eyes that do not show the classical keratometric, retinoscopic, or biomicroscopic signs, but show subtle topographic features similar to clinical keratoconus on videokeratoscopy.14 –18 However, an exact diagnosis of subclinical keratoconus is more difficult, because threshold criteria remain to be defined. ISSN 0161-6420/08/$–see front matter doi:10.1016/j.ophtha.2008.02.020
de Sanctis et al 䡠 Keratoconus Detection by Posterior Corneal Elevation with Pentacam Placido disk-based computerized videokeratoscopy only examines the anterior corneal surface. The advent of Orbscan slit-scanning topography (Orbtek, Inc., Salt Lake City, UT) meant that the posterior corneal curvature could also be studied in keratoconic eyes.19,20 In particular, several studies suggest that posterior corneal elevation may be used to detect this disease.21–24 However, the accuracy of the Orbscan system in the measurement of the posterior corneal surface has been criticized by some authors.25,26 The Pentacam (Oculus, Wetzlar, Germany) is a relatively new instrument that images the anterior and posterior corneal surfaces employing a rotating Scheimpflug camera. Measurements of corneal thickness and posterior elevation with the Pentacam rotating Scheimpflug camera have been reported to be highly reproducible and repeatable,27,28 but unlike those obtained with Orbscan topography,29 little is known about what constitutes normal or abnormal posterior corneal elevation measured with this method. The purpose of this study was to measure posterior corneal elevation in keratoconus, subclinical keratoconus, and normal corneas using the rotating Scheimpflug camera and, with this method, to estimate sensitivity and specificity of posterior elevation in discriminating keratoconus and subclinical keratoconus from normal corneas.
Methods Patient Inclusion This prospective, case-control study included patients with keratoconus, subclinical keratoconus, and candidates for refractive surgery with normal corneas. They were enrolled among consecutive patients examined at the Cornea Service of Turin University Ophthalmology Institute between April 2006 and June 2007 by a single cornea specialist. An eye was diagnosed as having keratoconus if it met the criteria of the Collaborative Longitudinal Evaluation of Keratoconus study.30 The ocular findings that defined keratoconus were (1) an irregular cornea determined by distorted keratometry mires, distortion of the retinoscopic or ophthalmoscopic red reflex (or a combination of the two) and (2) at least 1 of the following biomicroscopic signs: Vogt’s striae; Fleischer’s ring of ⬎2 mm arc; corneal scarring consistent with keratoconus. Keratoconus cases with a history of corneal surgery or with extensive corneal scarring were excluded from the study. An eye was diagnosed as having subclinical keratoconus if it was the fellow eye of a patient with keratoconus and showed the following features: (1) normal-appearing cornea at slit-lamp biomicroscopy, keratometry, retinoscopy, and ophthalmoscopy; (2) inferior–superior asymmetry and/or bow-tie pattern with skewed radial axes, detected on tangential Placido disk-based videokeratographs (CSO EyeTop, Florence, Italy), which were displayed using an absolute scale (1.5 dioptric steps); and (3) no history of contact lens wear, ocular surgery, or trauma. Candidates for refractive surgery included patients who underwent preoperative screening for myopia (⬍7 diopters) and/or astigmatism (⬍4 diopters) with normal corneal examination findings. Patients with history of ocular surgery and trauma or with any suspicious for subclinical keratoconus, keratoconus, or pellucid marginal degeneration at Placido disk-based videokeratography were excluded. For contact lens wearing patients, they were asked to stop wearing contact lenses for ⱖ3 weeks for rigid contact lenses and 1 week for soft contact lenses before assessment. After stopping contact lens
use for the recommended period, patients who still showed apparent corneal warping were also excluded. Informed consent was obtained from all patients after the purpose and characteristics of the study had been explained. When the study was begun, review by the local ethics committee was not required. The tenets of the Declaration of Helsinki were followed for all study procedures.
Pentacam Topography and Data Calculation For the purpose of the study, only one eye from each patient was examined. In patients with bilateral keratoconus and in candidates for refractive surgery, the eye to be examined was chosen using a random numbers table. In patients with subclinical keratoconus, the fellow eye with keratoconus was excluded from the analysis. All topographic examinations were performed using the Pentacam rotating Scheimpflug camera, software version 1.15 (Oculus). A single expert examiner, who was blind to the clinical condition of the patient, acquired Pentacam images in the manner described elsewhere.31 Briefly, the patient’s chin was placed on the chin rest and the forehead against the forehead strap. The patient was asked to open both eyes and stare at the fixation target. The examiner adjusted the joystick until alignment was perfect; then, using the automatic release mode, 25 single images were captured by the rotating Scheimpflug camera within 2 seconds for each eye. Images quality was checked, and for each eye only 1 examination with high-quality factor was recorded. For posterior corneal elevation measurement, a best-fit sphere (BFS) was used as reference surface. The sphere that best fit the posterior corneal surface was automatically generated by the software, with the float option over a 9-mm fit. The float option means that the BFS fits any direction of the corneal surface, rather than being set to fit only one direction, such as apex or axial alignment. A 9-mm fit was chosen because a larger fit may lead to a nonuniform BFS in eyes with few data points in the corneal periphery. The posterior elevation maps, which represent the radial distance between the sphere and the posterior corneal surface, were then displayed with 2.5-m colorcoded scales. On these maps, posterior elevation was measured as the maximum value above the BFS in the central 5 mm of the posterior cornea.
Statistical Analyses Eyes with keratoconus and those with subclinical keratoconus were compared with normal corneas in separate series of analyses. One-way analysis of variance with the least significant difference procedure was used to compare mean posterior corneal elevation values. Receiver operating characteristic (ROC) curves were used to determine the overall predictive accuracy of the test as described by the area under the curve. These curves are obtained by plotting sensitivity against 1⫺specificity, calculated for each value observed. An area of 100% implies that the test perfectly discriminates between groups. We also used this approach to calculate specificity, sensitivity, and positive [sensitivity/(1⫺specificity)] and negative [(1⫺sensitivity)/specificity] likelihood ratios (LR) for cutoff points of posterior corneal elevation selected a priori, and to identify posterior corneal elevation cutoff points that maximized sensitivity and specificity in discriminating keratoconus and subclinical keratoconus from normal corneas.32 Logistic regression was used to support the cutoff points identified with the ROC curves analysis and to check for model validity.33 For this purpose, the goodness-of-fit of the model, with the posterior elevation included as a continuous variable, was estimated using r2, and its internal validation was evaluated by bootstrapping analysis.34 Analyses were carried out using the software STATA9 and R (StataCorp, College Station, TX).
1535
Ophthalmology Volume 115, Number 9, September 2008
Figure 1. Distribution of posterior elevation (PE) in normal corneas and eyes with subclinical keratoconus and keratoconus.
Results One eye of 75 patients (45 men, 30 women) with keratoconus, 25 patients (18 men, 7 women) with subclinical keratoconus, and 64 candidates for refractive surgery (33 men/31 women) with normal corneas were analyzed. The mean ages were 40⫾15, 35⫾14, and 43⫾14 years, respectively, in patients with keratoconus, subclinical keratoconus, and normal corneas. Mean posterior corneal elevation was 100.7⫾49.2 in keratoconus, 39.9⫾15.0 in subclinical keratoconus, and 19.8⫾6.37 in normal eyes. The distribution of posterior corneal elevation in the three groups is summarized in Figure 1. Differences versus normal eyes were statistically significant for both keratoconus (P⬍0.001) and subclinical keratoconus (P ⫽ 0.01). Figure 2 reports the distribution density of posterior corneal elevation in the three groups, to illustrate the overlap between normal corneas, and eyes with subclinical keratoconus and keratoconus. Figure 3 compares results of the ROC curve analyses for keratoconus and subclinical keratoconus with those relating to normal corneas. The ROC graph showed that posterior corneal elevation discriminated keratoconus from normal corneas almost
perfectly. The area under the curve was close to 1.0 (0.99) for keratoconus and 0.93 (95% confidence interval [CI]: 0.87– 0.98) for subclinical keratoconus. Table 1 reports the specificity, sensitivity, and positive and negative LRs, identified by cutoff points of posterior corneal elevation selected a priori, in discriminating between eyes with keratoconus, subclinical keratoconus, and normal corneas. Based on the ROC curves, the optimal cutoff point to identify eyes with keratoconus was estimated to be 35 m. This cutoff point was associated with a sensitivity of 97.3% (95% CI, 90.7–99.7) and a specificity of 96.9% (95% CI, 89.3–99.6). To identify eyes with subclinical keratoconus, the optimal cutoff point was estimated to be 29 m. This cutoff point was associated with a sensitivity of 68% (95% CI, 46.5– 85), a specificity of 90.8% (95% CI, 81–96.5), and a LR⫹ of 7.39 and a LR⫺ of 0.35. With logistic regression analysis, cutoff points were similar to those identified by ROC analysis (38 m for keratoconus and 32 m for subclinical keratoconus). In comparing keratoconus and normal eyes, the model had a good fit to the data (r2 ⫽ 0.93), and r2 changed only marginally after validation through bootstrapping analysis. Similarly, we found an r2 of 0.63 for the comparison between subclinical keratoconus and normal eyes; r2 remained unchanged after validation.
Discussion The study showed that posterior corneal elevation measured with the Pentacam rotating Scheimpflug camera is higher in eyes with keratoconus or subclinical keratoconus than in normal corneas, and that posterior elevation is a useful index for discriminating this disease. Past approaches to estimating the performance of diagnostic indices for keratoconus have frequently used the classical method of data splitting, in which the data are arbitrarily or randomly divided and then used to develop a classification method (training set), or to estimate the generalization error of a trained classifier (testing set). So as to use the available data more efficiently, in this study the
Figure 2. Distribution density of posterior corneal elevation (PE) in normal corneas, and in eyes with subclinical keratoconus and keratoconus.
1536
de Sanctis et al 䡠 Keratoconus Detection by Posterior Corneal Elevation with Pentacam
Figure 3. Receiver operator characteristic curves for keratoconus and subclinical keratoconus versus normal corneas.
performance of posterior elevation was estimated using the whole dataset, and model validity was assessed by logistic regression and bootstrapping analysis. The overall predictive accuracy of posterior elevation, as described by the area under the ROC curve, was high for both keratoconus and subclinical keratoconus; values ⬎0.90 indicate outstanding discrimination of a test.32 Posterior elevation was very effective for the discrimination of keratoconus. The cutoff point of 35 m showed high sensitivity and specificity (97.3% and 96.9%, respectively), values that are comparable or above those obtained with other topographic indices derived from Placido disk-based videokeratography.10 –13,35–37 However, unlike posterior elevation, which derives from a single data, most topographic indices derived from Placido disk-based videokeratography include multiple parameters, require integration of the data into a
decision-making process, such as neural network or automated tree classification or are based on a more sophisticated polynomial analysis.11–13,35–37 The posterior elevation was less effective in discriminating subclinical keratoconus than it was in discriminating keratoconus. The cutoff point of 29 m had 68% sensitivity and 90.8% specificity. Although it has been suggested that an increase in posterior elevation may be the earliest sign of subclinical keratoconus,21 indices derived from Placido diskbased videokeratography may be more sensitive in discriminating this condition, which is in large part defined on the basis of topographic patterns produced with this method. However, in eyes with subclinical keratoconus, the Pentacam rotating Scheimpflug camera may add useful information to Placido disk-based videokeratography. Axial curvature analysis, provided by the latter method, suffers from some limitations when
Table 1. Specificity, Sensitivity, Positive and Negative Likelihood Ratios Identified by Cutoff Points of Posterior Corneal Elevation Selected a priori (eyes with keratoconus or subclinical keratoconus vs normal corneas) Keratoconus vs Normal Corneas
Subclinical Keratoconus vs Normal Corneas
Cutoff Point (m)
Sensitivity (%)
Specificity (%)
LR⫹
LR⫺
Sensitivity (%)
Specificity (%)
LR⫹
LR⫺
10 15 20 25 30 35 40 45
100 100 100 100 97.3 97.3 96.0 92.0
0 20.0 55.4 75.4 90.8 96.9 98.5 100
1 1.25 2.24 4.06 10.5 31.6 62.4 —
— 0 0 0 0.03 0.03 0.04 0.08
100 100 100 92.0 68.0 64.0 44.0 32.0
0 20.0 55.3 75.4 90.8 92.3 98.5 100
1 1.25 2.24 3.74 7.37 8.32 28.6 —
— 0 0 0.11 0.35 0.39 0.57 0.68
LR⫺ ⫽ negative likelihood ratio; LR⫹ ⫽ positive likelihood ratio.
1537
Ophthalmology Volume 115, Number 9, September 2008 studying an abnormal shape. Axial curvature analysis is based on the assumption that the reference axis used to generate the maps is the same as the visual axis and the corneal apex. In many normal eyes, the corneal apex and the corneal sighting point do not correspond,38 and Placido disk-based videokeratography may generate incomplete or sometimes misleading pictures.39,40 Because the cornea is analyzed around a point other than its center, a normal aspherical surface may generate an asymmetrical bow-tie pattern or inferior steepening, which are the commonest patterns seen in subclinical keratoconus. In such eyes, the Pentacam rotating Scheimpflug camera may help to distinguish normal from abnormal corneal shapes, because this system directly acquires elevation points from the corneal surfaces, which are independent of both the visual axis and the corneal apex. In this study, the cutoff points of posterior elevation values measured with the Pentacam rotating Scheimpflug camera were lower than those reported using Orbscan II topography; using the same measurement settings as the present study (floating BFS as a reference, measurement of posterior elevation in the central 5 mm), Rao et al21 and Fam et al23 both reported that, with Orbscan II topography, the posterior elevation optimal cutoff point to discriminate keratoconus and keratoconus suspect versus normal corneas was 40 m. Although this difference between the Pentacam rotating Scheimpflug camera and Orbscan II topography might be judged to be slight, it is clinically relevant. If we had used a cutoff point of 40 m, we would have missed 4% of keratoconus cases and 42% of subclinical keratoconus cases identified in this study. It is thus important for clinicians and refractive surgeons to be aware of this difference, because interpretation of posterior elevation data may vary with the instrument used. The tendency of the Pentacam rotating Scheimpflug camera to underestimate posterior corneal elevation in comparison with Orbscan II topography is supported by other findings. In this study, mean posterior elevation of normal corneas was 19.87 m with the rotating Scheimpflug camera, but Orbscan II topography has reported it to range from 28 to 30.47 m.23,24,41 Moreover, in 36 eyes with keratoconus, Quisling et al29 reported that mean posterior elevation was 34.86 m with the Pentacam rotating Scheimpflug camera and 48.50 m with Orbscan II topography. The reason underlying the disparity between these two methods in measuring posterior corneal elevation has yet to be elucidated. Although several factors, including tear film analysis, corneal scarring, and the software used to image the cornea, have been proposed as possible mechanisms, the different technology is probably a major factor.29 Orbscan II uses a hybrid system, based on slit scanning and Placido ring technology, to give a representation of the anterior corneal surface. For each acquisition, 40 independent images (20 scans to the right, 20 scans to the left) are captured, but not all images scan across the central cornea. The posterior corneal surface is then recreated digitally, through triangulation algorithms, from the previously generated anterior surface.26 Pentacam uses a rotating Scheimpflug camera, which captures 25 slit images on each acquisition, to give a representation of the corneal shape. All 25 Scheimpflug images involve the central cornea; however, posterior data are not measured directly, with this method either,
1538
because they are derived from anterior data based on BFS algorithm and corneal thickness. The performance of posterior corneal elevation in discriminating keratoconus and subclinical keratoconus might be overestimated in this study due to its design.42 Although consecutive patients were included, in case-control studies the test is performed in a group of patients already known to have the disease and in a group of normal patients, rather than in a relevant clinical population. Because a relevant clinical population is a group of patients covering the whole spectrum of the disease that is likely to be encountered in current or future use of the test, a more accurate estimate of the performance of posterior elevation for keratoconus screening would require a further study on a large number of candidates for refractive surgery. Moreover, the performance of posterior elevation may vary according to the area of measurement and to the setting of the instrument. In this study, posterior elevation was measured also in the central 3 mm and 7 mm of the posterior cornea. Measurements in these areas would not have increased the discrimination efficacy of this parameter (data not reported). However, subclinical keratoconus may rarely occur more peripherally, and calculation of the posterior elevation in a wider area that includes the corneal periphery may be necessary. With the Pentacam rotating Scheimpflug camera, other settings of reference surfaces (toric ellipsoid, ellipsoid) and other types of alignment are available, and future studies should investigate whether they may improve the efficacy of posterior elevation in discriminating keratoconus and subclinical keratoconus from normal corneas. Although it cannot be concluded from this study that posterior corneal elevation is sufficient alone as a single diagnostic index, it does seem to be very effective in discriminating keratoconus from normal corneas. Its efficacy is lower for subclinical keratoconus, and thus data concerning posterior elevation should be combined with curvature data in stratifying patients with this condition.
References 1. Krachmer JH, Feder RS, Belin MW. Keratoconus and related noninflammatory corneal thinning disorders. Surv Ophthalmol 1984;28:293–322. 2. Rabinowitz YS. Keratoconus. Surv Ophthalmol 1998;42:297– 319. 3. Wilson SE, Klyce SD. Screening for corneal topographic abnormalities before refractive surgery. Ophthalmology 1994; 101:145–52. 4. Nesburn AB, Bahri S, Salz J, et al. Keratoconus detected by videokeratography in candidates for photorefractive keratectomy. J Refract Surg 1995;11:194 –201. 5. Mamalis N, Montgomery S, Anderson C, Miller C. Radial keratotomy in a patient with keratoconus. Refract Corneal Surg 1991;7:374 – 6. 6. Ellis W. Radial keratotomy in a patient with keratoconus. J Cataract Refract Surg 1992;18:406 –9. 7. Seiler T, Quurke AW. Iatrogenic keratectasia after LASIK in a case of forme fruste keratoconus. J Cataract Refract Surg 1998;24:1007–9.
de Sanctis et al 䡠 Keratoconus Detection by Posterior Corneal Elevation with Pentacam 8. Randleman JB, Russel B, Ward M, et al. Risk factors and prognosis for corneal ectasia after LASIK. Ophthalmology 2003;110:267–75. 9. Binder PS, Lindstrom RL, Stulting RD, et al. Keratoconus and corneal ectasia after LASIK. J Cataract Refract Surg 2005;35: 2035– 8. 10. Rabinowitz YS. Videokeratographic indices to aid in screening for keratoconus. J Refract Surg 1995;11:371–9. 11. Maeda N, Klyce SD, Smolek MK, Thompson HW. Automated keratoconus screening with corneal topography analysis. Invest Ophthalmol Vis Sci 1994;35:2749 –57. 12. Rabinowitz YS, Rasheed K. KISA% index: a quantitative videokeratography algorithm embodying minimal topographic criteria for diagnosing keratoconus. J Cataract Refract Surg 1999;25:1327–35. 13. Smolek MK, Klyce SD. Current detection methods compared with a neural network approach. Invest Ophthalmol Vis Sci 1997;38:2290 –9. 14. Maguire LJ, Bourne WM. Corneal topography of early keratoconus. Am J Ophthalmol 1989;108:107–12. 15. Rabinowitz YS, Garbus J, McDonnell PJ. Computer-assisted corneal topography in family members of patients with keratoconus. Arch Ophthalmol 1990;108:365–71. 16. Maguire LJ, Lowry JC. Identifying progression of subclinical keratoconus by serial topography analysis. Am J Ophthalmol 1991;112:41–5. 17. Rabinowitz YS, Nesburn AB, McDonnell PJ. Videokeratography of the fellow eye in unilateral keratoconus. Ophthalmology 1993;100:181– 6. 18. Li X, Rabinowitz YS, Rasheed K, Yang H. Longitudinal study of the normal eyes in unilateral keratoconus patients. Ophthalmology 2004;111:440 – 6. 19. Auffarth GU, Wang L, Volcker HE. Keratoconus evaluation using the Orbscan Topography System. J Cataract Refract Surg 2000;26:222– 8. 20. Tomidokoro A, Oshika T, Amano S, et al. Changes in anterior and posterior corneal curvatures in keratoconus. Ophthalmology 2000;107:1328 –32. 21. Rao SN, Raviv T, Majmudar PA, Epstein RJ. Role of Orbscan II in screening keratoconus suspects before refractive corneal surgery. Ophthalmology 2002;109:1642– 6. 22. Tanabe T, Oshika T, Tomidokoro A, et al. Standardized colorcoded scales for anterior and posterior elevation maps of scanning slit corneal topography. Ophthalmology 2002; 109:1298 –302. 23. Fam HB, Lim KL. Corneal elevation indices in normal and keratoconic eyes. J Cataract Refract Surg 2006;32:1281–7. 24. Sonmez B, Doan MP, Hamilton DR. Identification of scanning slit-beam topographic parameters important in distinguishing normal from keratoconic corneal morphologic features. Am J Ophthalmol 2007;143:401– 8. 25. Wilson SE. Cautions regarding measurement of the posterior corneal curvature. Ophthalmology 2000;107:1223.
26. Cairns G, McGhee CNJ. Orbscan computerized topography: attributes, applications, and limitations. J Cataract Refract Surg 2005;31:205–20. 27. de Sanctis U, Missolungi A, Mutani B, et al. Reproducibility and repeatability of central corneal thickness measurement in keratoconus using the rotating Scheimpflug camera and ultrasound pachymetry. Am J Ophthalmol 2007;144:712– 8. 28. Chen D, Lam AK. Intrasession and intersession repeatability of the Pentacam system on posterior corneal assessment in the normal human eye. J Cataract Refract Surg 2007;33:448 –54. 29. Quisling S, Sjoberg S, Zimmerman B, et al. Comparison of Pentacam and Orbscan IIz on posterior curvature topography measurement in keratoconus eyes. Ophthalmology 2006;113: 1629 –32. 30. Zadnik K, Barr JT, Edrington TB, et al, Collaborative Longitudinal Evaluation of Keratoconus (CLEK) Study Group. Baseline findings in the Collaborative Longitudinal Evaluation of Keratoconus (CLEK) study. Invest Ophthalmol Vis Sci 1998;39:2537– 46. 31. de Sanctis U, Missolungi A, Mutani B, Grignolo FM. Graft central thickness measurement by rotating Scheimpflug camera and ultrasound pachymetry after penetrating keratoplasty. Ophthalmology 2007;114:1461– 8. 32. Altman DG, Bland JM. Diagnostic tests 3: receiver operating characteristic plots. BMJ 1994;309:188. 33. Hosmer DW, Lemeshow S. Applied Logistic Regression. 2nd ed. New York: Wiley; 2000:160 –7. 34. Harrell FE Jr. Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis. New York: Springer; 2001:90 – 8. 35. Schwiegerling J, Greivenkamp JE. Keratoconus detection based on videokeratoscopic height data. Optom Vis Sci 1996; 73:721– 8. 36. Chastang PJ, Borderie VM, Carvajal-Gonzales S, et al. Automated keratoconus detection using the EyeSys videokeratoscope. J Cataract Refract Surg 2000;26:675– 83. 37. Twa MD, Parthasarathy S, Roberts C, et al. Automated decision tree classification of corneal shape. Optom Vis Sci 2005; 82:1038 – 46. 38. Tomlinson A, Schwartz C. The position of the corneal apex in the normal eye. Am J Optom Physiol Opt 1979;56:236 – 40. 39. Roberts C. Characterization of the inherent error in a spherically-biased corneal topography system in mapping a radially aspheric surface. J Refract Corneal Surg 1994;10: 103–11. 40. Belin MW, Khachikian SS. Keratoconus: it is hard to define, but . . . Am J Ophthalmol 2007;143:500 –3. 41. Wei RH, Lim L, Chan WK, Tan DT. Evaluation of Orbscan II corneal topography in individuals with myopia. Ophthalmology 2006;113;177– 83. 42. Lijmer JG, Mol BW, Heisterkamp S, et al. Empirical evidence of design-related bias in studies of diagnostic tests. JAMA 1999;282:1061– 6.
Footnotes and Financial Disclosures Originally received: December 10, 2007. Final revision: February 6, 2008. Accepted: February 22, 2008. Available online: April 11, 2008.
Manuscript no. 2007-1574.
1
Department of Clinical Physiopathology, Ophthalmology Institute, University of Turin, Turin, Italy.
2
Cancer Epidemiology Unit, CeRMS and CPO Piemonte, University of Turin, Turin, Italy.
Financial Disclosure(s): The authors have no proprietary or commercial interest in any materials discussed in this article. Supported in part by the Department of Clinical Physiopathology, Turin University. Correspondence: Ugo de Sanctis, MD, PhD, Department of Clinical Physiopathology, Ophthalmology Institute, University of Turin, Via Juvarra 19, Turin 10121, Italy. E-mail: [email protected].
1539
Keratoconus is a noninflammatory disorder, characterized by corneal thinning and anterior protrusion.1,2 Detection of this disease is particularly important among patients considering refractive surgery, in whom the prevalence of keratoconus and subclinical keratoconus has been reported to be greater than in the normal population3,4; keratorefractive procedures may have unsatisfactory results and cause postoperative complications in these patients.5–7 Moreover, apart from excessive ablation of corneal tissue, unidentified subclinical keratoconus is considered to be the main cause of ectasia after LASIK.8,9 Clinical diagnosis of keratoconus in eyes with biomicroscopic, keratometric, and retinoscopic signs is not difficult.
1534
© 2008 by the American Academy of Ophthalmology Published by Elsevier Inc.
Moreover, advances in Placido disk-based computerized videokeratoscopy have lead to a variety of quantitative indices that have been found to be highly sensitive and specific in aiding with the diagnosis of keratoconus.10 –13 Diagnosis of forme fruste keratoconus or subclinical keratoconus is more challenging. The terms “forme fruste keratoconus” and “subclinical keratoconus” were introduced to indicate a very early preclinical stage of the disease, in eyes that do not show the classical keratometric, retinoscopic, or biomicroscopic signs, but show subtle topographic features similar to clinical keratoconus on videokeratoscopy.14 –18 However, an exact diagnosis of subclinical keratoconus is more difficult, because threshold criteria remain to be defined. ISSN 0161-6420/08/$–see front matter doi:10.1016/j.ophtha.2008.02.020
de Sanctis et al 䡠 Keratoconus Detection by Posterior Corneal Elevation with Pentacam Placido disk-based computerized videokeratoscopy only examines the anterior corneal surface. The advent of Orbscan slit-scanning topography (Orbtek, Inc., Salt Lake City, UT) meant that the posterior corneal curvature could also be studied in keratoconic eyes.19,20 In particular, several studies suggest that posterior corneal elevation may be used to detect this disease.21–24 However, the accuracy of the Orbscan system in the measurement of the posterior corneal surface has been criticized by some authors.25,26 The Pentacam (Oculus, Wetzlar, Germany) is a relatively new instrument that images the anterior and posterior corneal surfaces employing a rotating Scheimpflug camera. Measurements of corneal thickness and posterior elevation with the Pentacam rotating Scheimpflug camera have been reported to be highly reproducible and repeatable,27,28 but unlike those obtained with Orbscan topography,29 little is known about what constitutes normal or abnormal posterior corneal elevation measured with this method. The purpose of this study was to measure posterior corneal elevation in keratoconus, subclinical keratoconus, and normal corneas using the rotating Scheimpflug camera and, with this method, to estimate sensitivity and specificity of posterior elevation in discriminating keratoconus and subclinical keratoconus from normal corneas.
Methods Patient Inclusion This prospective, case-control study included patients with keratoconus, subclinical keratoconus, and candidates for refractive surgery with normal corneas. They were enrolled among consecutive patients examined at the Cornea Service of Turin University Ophthalmology Institute between April 2006 and June 2007 by a single cornea specialist. An eye was diagnosed as having keratoconus if it met the criteria of the Collaborative Longitudinal Evaluation of Keratoconus study.30 The ocular findings that defined keratoconus were (1) an irregular cornea determined by distorted keratometry mires, distortion of the retinoscopic or ophthalmoscopic red reflex (or a combination of the two) and (2) at least 1 of the following biomicroscopic signs: Vogt’s striae; Fleischer’s ring of ⬎2 mm arc; corneal scarring consistent with keratoconus. Keratoconus cases with a history of corneal surgery or with extensive corneal scarring were excluded from the study. An eye was diagnosed as having subclinical keratoconus if it was the fellow eye of a patient with keratoconus and showed the following features: (1) normal-appearing cornea at slit-lamp biomicroscopy, keratometry, retinoscopy, and ophthalmoscopy; (2) inferior–superior asymmetry and/or bow-tie pattern with skewed radial axes, detected on tangential Placido disk-based videokeratographs (CSO EyeTop, Florence, Italy), which were displayed using an absolute scale (1.5 dioptric steps); and (3) no history of contact lens wear, ocular surgery, or trauma. Candidates for refractive surgery included patients who underwent preoperative screening for myopia (⬍7 diopters) and/or astigmatism (⬍4 diopters) with normal corneal examination findings. Patients with history of ocular surgery and trauma or with any suspicious for subclinical keratoconus, keratoconus, or pellucid marginal degeneration at Placido disk-based videokeratography were excluded. For contact lens wearing patients, they were asked to stop wearing contact lenses for ⱖ3 weeks for rigid contact lenses and 1 week for soft contact lenses before assessment. After stopping contact lens
use for the recommended period, patients who still showed apparent corneal warping were also excluded. Informed consent was obtained from all patients after the purpose and characteristics of the study had been explained. When the study was begun, review by the local ethics committee was not required. The tenets of the Declaration of Helsinki were followed for all study procedures.
Pentacam Topography and Data Calculation For the purpose of the study, only one eye from each patient was examined. In patients with bilateral keratoconus and in candidates for refractive surgery, the eye to be examined was chosen using a random numbers table. In patients with subclinical keratoconus, the fellow eye with keratoconus was excluded from the analysis. All topographic examinations were performed using the Pentacam rotating Scheimpflug camera, software version 1.15 (Oculus). A single expert examiner, who was blind to the clinical condition of the patient, acquired Pentacam images in the manner described elsewhere.31 Briefly, the patient’s chin was placed on the chin rest and the forehead against the forehead strap. The patient was asked to open both eyes and stare at the fixation target. The examiner adjusted the joystick until alignment was perfect; then, using the automatic release mode, 25 single images were captured by the rotating Scheimpflug camera within 2 seconds for each eye. Images quality was checked, and for each eye only 1 examination with high-quality factor was recorded. For posterior corneal elevation measurement, a best-fit sphere (BFS) was used as reference surface. The sphere that best fit the posterior corneal surface was automatically generated by the software, with the float option over a 9-mm fit. The float option means that the BFS fits any direction of the corneal surface, rather than being set to fit only one direction, such as apex or axial alignment. A 9-mm fit was chosen because a larger fit may lead to a nonuniform BFS in eyes with few data points in the corneal periphery. The posterior elevation maps, which represent the radial distance between the sphere and the posterior corneal surface, were then displayed with 2.5-m colorcoded scales. On these maps, posterior elevation was measured as the maximum value above the BFS in the central 5 mm of the posterior cornea.
Statistical Analyses Eyes with keratoconus and those with subclinical keratoconus were compared with normal corneas in separate series of analyses. One-way analysis of variance with the least significant difference procedure was used to compare mean posterior corneal elevation values. Receiver operating characteristic (ROC) curves were used to determine the overall predictive accuracy of the test as described by the area under the curve. These curves are obtained by plotting sensitivity against 1⫺specificity, calculated for each value observed. An area of 100% implies that the test perfectly discriminates between groups. We also used this approach to calculate specificity, sensitivity, and positive [sensitivity/(1⫺specificity)] and negative [(1⫺sensitivity)/specificity] likelihood ratios (LR) for cutoff points of posterior corneal elevation selected a priori, and to identify posterior corneal elevation cutoff points that maximized sensitivity and specificity in discriminating keratoconus and subclinical keratoconus from normal corneas.32 Logistic regression was used to support the cutoff points identified with the ROC curves analysis and to check for model validity.33 For this purpose, the goodness-of-fit of the model, with the posterior elevation included as a continuous variable, was estimated using r2, and its internal validation was evaluated by bootstrapping analysis.34 Analyses were carried out using the software STATA9 and R (StataCorp, College Station, TX).
1535
Ophthalmology Volume 115, Number 9, September 2008
Figure 1. Distribution of posterior elevation (PE) in normal corneas and eyes with subclinical keratoconus and keratoconus.
Results One eye of 75 patients (45 men, 30 women) with keratoconus, 25 patients (18 men, 7 women) with subclinical keratoconus, and 64 candidates for refractive surgery (33 men/31 women) with normal corneas were analyzed. The mean ages were 40⫾15, 35⫾14, and 43⫾14 years, respectively, in patients with keratoconus, subclinical keratoconus, and normal corneas. Mean posterior corneal elevation was 100.7⫾49.2 in keratoconus, 39.9⫾15.0 in subclinical keratoconus, and 19.8⫾6.37 in normal eyes. The distribution of posterior corneal elevation in the three groups is summarized in Figure 1. Differences versus normal eyes were statistically significant for both keratoconus (P⬍0.001) and subclinical keratoconus (P ⫽ 0.01). Figure 2 reports the distribution density of posterior corneal elevation in the three groups, to illustrate the overlap between normal corneas, and eyes with subclinical keratoconus and keratoconus. Figure 3 compares results of the ROC curve analyses for keratoconus and subclinical keratoconus with those relating to normal corneas. The ROC graph showed that posterior corneal elevation discriminated keratoconus from normal corneas almost
perfectly. The area under the curve was close to 1.0 (0.99) for keratoconus and 0.93 (95% confidence interval [CI]: 0.87– 0.98) for subclinical keratoconus. Table 1 reports the specificity, sensitivity, and positive and negative LRs, identified by cutoff points of posterior corneal elevation selected a priori, in discriminating between eyes with keratoconus, subclinical keratoconus, and normal corneas. Based on the ROC curves, the optimal cutoff point to identify eyes with keratoconus was estimated to be 35 m. This cutoff point was associated with a sensitivity of 97.3% (95% CI, 90.7–99.7) and a specificity of 96.9% (95% CI, 89.3–99.6). To identify eyes with subclinical keratoconus, the optimal cutoff point was estimated to be 29 m. This cutoff point was associated with a sensitivity of 68% (95% CI, 46.5– 85), a specificity of 90.8% (95% CI, 81–96.5), and a LR⫹ of 7.39 and a LR⫺ of 0.35. With logistic regression analysis, cutoff points were similar to those identified by ROC analysis (38 m for keratoconus and 32 m for subclinical keratoconus). In comparing keratoconus and normal eyes, the model had a good fit to the data (r2 ⫽ 0.93), and r2 changed only marginally after validation through bootstrapping analysis. Similarly, we found an r2 of 0.63 for the comparison between subclinical keratoconus and normal eyes; r2 remained unchanged after validation.
Discussion The study showed that posterior corneal elevation measured with the Pentacam rotating Scheimpflug camera is higher in eyes with keratoconus or subclinical keratoconus than in normal corneas, and that posterior elevation is a useful index for discriminating this disease. Past approaches to estimating the performance of diagnostic indices for keratoconus have frequently used the classical method of data splitting, in which the data are arbitrarily or randomly divided and then used to develop a classification method (training set), or to estimate the generalization error of a trained classifier (testing set). So as to use the available data more efficiently, in this study the
Figure 2. Distribution density of posterior corneal elevation (PE) in normal corneas, and in eyes with subclinical keratoconus and keratoconus.
1536
de Sanctis et al 䡠 Keratoconus Detection by Posterior Corneal Elevation with Pentacam
Figure 3. Receiver operator characteristic curves for keratoconus and subclinical keratoconus versus normal corneas.
performance of posterior elevation was estimated using the whole dataset, and model validity was assessed by logistic regression and bootstrapping analysis. The overall predictive accuracy of posterior elevation, as described by the area under the ROC curve, was high for both keratoconus and subclinical keratoconus; values ⬎0.90 indicate outstanding discrimination of a test.32 Posterior elevation was very effective for the discrimination of keratoconus. The cutoff point of 35 m showed high sensitivity and specificity (97.3% and 96.9%, respectively), values that are comparable or above those obtained with other topographic indices derived from Placido disk-based videokeratography.10 –13,35–37 However, unlike posterior elevation, which derives from a single data, most topographic indices derived from Placido disk-based videokeratography include multiple parameters, require integration of the data into a
decision-making process, such as neural network or automated tree classification or are based on a more sophisticated polynomial analysis.11–13,35–37 The posterior elevation was less effective in discriminating subclinical keratoconus than it was in discriminating keratoconus. The cutoff point of 29 m had 68% sensitivity and 90.8% specificity. Although it has been suggested that an increase in posterior elevation may be the earliest sign of subclinical keratoconus,21 indices derived from Placido diskbased videokeratography may be more sensitive in discriminating this condition, which is in large part defined on the basis of topographic patterns produced with this method. However, in eyes with subclinical keratoconus, the Pentacam rotating Scheimpflug camera may add useful information to Placido disk-based videokeratography. Axial curvature analysis, provided by the latter method, suffers from some limitations when
Table 1. Specificity, Sensitivity, Positive and Negative Likelihood Ratios Identified by Cutoff Points of Posterior Corneal Elevation Selected a priori (eyes with keratoconus or subclinical keratoconus vs normal corneas) Keratoconus vs Normal Corneas
Subclinical Keratoconus vs Normal Corneas
Cutoff Point (m)
Sensitivity (%)
Specificity (%)
LR⫹
LR⫺
Sensitivity (%)
Specificity (%)
LR⫹
LR⫺
10 15 20 25 30 35 40 45
100 100 100 100 97.3 97.3 96.0 92.0
0 20.0 55.4 75.4 90.8 96.9 98.5 100
1 1.25 2.24 4.06 10.5 31.6 62.4 —
— 0 0 0 0.03 0.03 0.04 0.08
100 100 100 92.0 68.0 64.0 44.0 32.0
0 20.0 55.3 75.4 90.8 92.3 98.5 100
1 1.25 2.24 3.74 7.37 8.32 28.6 —
— 0 0 0.11 0.35 0.39 0.57 0.68
LR⫺ ⫽ negative likelihood ratio; LR⫹ ⫽ positive likelihood ratio.
1537
Ophthalmology Volume 115, Number 9, September 2008 studying an abnormal shape. Axial curvature analysis is based on the assumption that the reference axis used to generate the maps is the same as the visual axis and the corneal apex. In many normal eyes, the corneal apex and the corneal sighting point do not correspond,38 and Placido disk-based videokeratography may generate incomplete or sometimes misleading pictures.39,40 Because the cornea is analyzed around a point other than its center, a normal aspherical surface may generate an asymmetrical bow-tie pattern or inferior steepening, which are the commonest patterns seen in subclinical keratoconus. In such eyes, the Pentacam rotating Scheimpflug camera may help to distinguish normal from abnormal corneal shapes, because this system directly acquires elevation points from the corneal surfaces, which are independent of both the visual axis and the corneal apex. In this study, the cutoff points of posterior elevation values measured with the Pentacam rotating Scheimpflug camera were lower than those reported using Orbscan II topography; using the same measurement settings as the present study (floating BFS as a reference, measurement of posterior elevation in the central 5 mm), Rao et al21 and Fam et al23 both reported that, with Orbscan II topography, the posterior elevation optimal cutoff point to discriminate keratoconus and keratoconus suspect versus normal corneas was 40 m. Although this difference between the Pentacam rotating Scheimpflug camera and Orbscan II topography might be judged to be slight, it is clinically relevant. If we had used a cutoff point of 40 m, we would have missed 4% of keratoconus cases and 42% of subclinical keratoconus cases identified in this study. It is thus important for clinicians and refractive surgeons to be aware of this difference, because interpretation of posterior elevation data may vary with the instrument used. The tendency of the Pentacam rotating Scheimpflug camera to underestimate posterior corneal elevation in comparison with Orbscan II topography is supported by other findings. In this study, mean posterior elevation of normal corneas was 19.87 m with the rotating Scheimpflug camera, but Orbscan II topography has reported it to range from 28 to 30.47 m.23,24,41 Moreover, in 36 eyes with keratoconus, Quisling et al29 reported that mean posterior elevation was 34.86 m with the Pentacam rotating Scheimpflug camera and 48.50 m with Orbscan II topography. The reason underlying the disparity between these two methods in measuring posterior corneal elevation has yet to be elucidated. Although several factors, including tear film analysis, corneal scarring, and the software used to image the cornea, have been proposed as possible mechanisms, the different technology is probably a major factor.29 Orbscan II uses a hybrid system, based on slit scanning and Placido ring technology, to give a representation of the anterior corneal surface. For each acquisition, 40 independent images (20 scans to the right, 20 scans to the left) are captured, but not all images scan across the central cornea. The posterior corneal surface is then recreated digitally, through triangulation algorithms, from the previously generated anterior surface.26 Pentacam uses a rotating Scheimpflug camera, which captures 25 slit images on each acquisition, to give a representation of the corneal shape. All 25 Scheimpflug images involve the central cornea; however, posterior data are not measured directly, with this method either,
1538
because they are derived from anterior data based on BFS algorithm and corneal thickness. The performance of posterior corneal elevation in discriminating keratoconus and subclinical keratoconus might be overestimated in this study due to its design.42 Although consecutive patients were included, in case-control studies the test is performed in a group of patients already known to have the disease and in a group of normal patients, rather than in a relevant clinical population. Because a relevant clinical population is a group of patients covering the whole spectrum of the disease that is likely to be encountered in current or future use of the test, a more accurate estimate of the performance of posterior elevation for keratoconus screening would require a further study on a large number of candidates for refractive surgery. Moreover, the performance of posterior elevation may vary according to the area of measurement and to the setting of the instrument. In this study, posterior elevation was measured also in the central 3 mm and 7 mm of the posterior cornea. Measurements in these areas would not have increased the discrimination efficacy of this parameter (data not reported). However, subclinical keratoconus may rarely occur more peripherally, and calculation of the posterior elevation in a wider area that includes the corneal periphery may be necessary. With the Pentacam rotating Scheimpflug camera, other settings of reference surfaces (toric ellipsoid, ellipsoid) and other types of alignment are available, and future studies should investigate whether they may improve the efficacy of posterior elevation in discriminating keratoconus and subclinical keratoconus from normal corneas. Although it cannot be concluded from this study that posterior corneal elevation is sufficient alone as a single diagnostic index, it does seem to be very effective in discriminating keratoconus from normal corneas. Its efficacy is lower for subclinical keratoconus, and thus data concerning posterior elevation should be combined with curvature data in stratifying patients with this condition.
References 1. Krachmer JH, Feder RS, Belin MW. Keratoconus and related noninflammatory corneal thinning disorders. Surv Ophthalmol 1984;28:293–322. 2. Rabinowitz YS. Keratoconus. Surv Ophthalmol 1998;42:297– 319. 3. Wilson SE, Klyce SD. Screening for corneal topographic abnormalities before refractive surgery. Ophthalmology 1994; 101:145–52. 4. Nesburn AB, Bahri S, Salz J, et al. Keratoconus detected by videokeratography in candidates for photorefractive keratectomy. J Refract Surg 1995;11:194 –201. 5. Mamalis N, Montgomery S, Anderson C, Miller C. Radial keratotomy in a patient with keratoconus. Refract Corneal Surg 1991;7:374 – 6. 6. Ellis W. Radial keratotomy in a patient with keratoconus. J Cataract Refract Surg 1992;18:406 –9. 7. Seiler T, Quurke AW. Iatrogenic keratectasia after LASIK in a case of forme fruste keratoconus. J Cataract Refract Surg 1998;24:1007–9.
de Sanctis et al 䡠 Keratoconus Detection by Posterior Corneal Elevation with Pentacam 8. Randleman JB, Russel B, Ward M, et al. Risk factors and prognosis for corneal ectasia after LASIK. Ophthalmology 2003;110:267–75. 9. Binder PS, Lindstrom RL, Stulting RD, et al. Keratoconus and corneal ectasia after LASIK. J Cataract Refract Surg 2005;35: 2035– 8. 10. Rabinowitz YS. Videokeratographic indices to aid in screening for keratoconus. J Refract Surg 1995;11:371–9. 11. Maeda N, Klyce SD, Smolek MK, Thompson HW. Automated keratoconus screening with corneal topography analysis. Invest Ophthalmol Vis Sci 1994;35:2749 –57. 12. Rabinowitz YS, Rasheed K. KISA% index: a quantitative videokeratography algorithm embodying minimal topographic criteria for diagnosing keratoconus. J Cataract Refract Surg 1999;25:1327–35. 13. Smolek MK, Klyce SD. Current detection methods compared with a neural network approach. Invest Ophthalmol Vis Sci 1997;38:2290 –9. 14. Maguire LJ, Bourne WM. Corneal topography of early keratoconus. Am J Ophthalmol 1989;108:107–12. 15. Rabinowitz YS, Garbus J, McDonnell PJ. Computer-assisted corneal topography in family members of patients with keratoconus. Arch Ophthalmol 1990;108:365–71. 16. Maguire LJ, Lowry JC. Identifying progression of subclinical keratoconus by serial topography analysis. Am J Ophthalmol 1991;112:41–5. 17. Rabinowitz YS, Nesburn AB, McDonnell PJ. Videokeratography of the fellow eye in unilateral keratoconus. Ophthalmology 1993;100:181– 6. 18. Li X, Rabinowitz YS, Rasheed K, Yang H. Longitudinal study of the normal eyes in unilateral keratoconus patients. Ophthalmology 2004;111:440 – 6. 19. Auffarth GU, Wang L, Volcker HE. Keratoconus evaluation using the Orbscan Topography System. J Cataract Refract Surg 2000;26:222– 8. 20. Tomidokoro A, Oshika T, Amano S, et al. Changes in anterior and posterior corneal curvatures in keratoconus. Ophthalmology 2000;107:1328 –32. 21. Rao SN, Raviv T, Majmudar PA, Epstein RJ. Role of Orbscan II in screening keratoconus suspects before refractive corneal surgery. Ophthalmology 2002;109:1642– 6. 22. Tanabe T, Oshika T, Tomidokoro A, et al. Standardized colorcoded scales for anterior and posterior elevation maps of scanning slit corneal topography. Ophthalmology 2002; 109:1298 –302. 23. Fam HB, Lim KL. Corneal elevation indices in normal and keratoconic eyes. J Cataract Refract Surg 2006;32:1281–7. 24. Sonmez B, Doan MP, Hamilton DR. Identification of scanning slit-beam topographic parameters important in distinguishing normal from keratoconic corneal morphologic features. Am J Ophthalmol 2007;143:401– 8. 25. Wilson SE. Cautions regarding measurement of the posterior corneal curvature. Ophthalmology 2000;107:1223.
26. Cairns G, McGhee CNJ. Orbscan computerized topography: attributes, applications, and limitations. J Cataract Refract Surg 2005;31:205–20. 27. de Sanctis U, Missolungi A, Mutani B, et al. Reproducibility and repeatability of central corneal thickness measurement in keratoconus using the rotating Scheimpflug camera and ultrasound pachymetry. Am J Ophthalmol 2007;144:712– 8. 28. Chen D, Lam AK. Intrasession and intersession repeatability of the Pentacam system on posterior corneal assessment in the normal human eye. J Cataract Refract Surg 2007;33:448 –54. 29. Quisling S, Sjoberg S, Zimmerman B, et al. Comparison of Pentacam and Orbscan IIz on posterior curvature topography measurement in keratoconus eyes. Ophthalmology 2006;113: 1629 –32. 30. Zadnik K, Barr JT, Edrington TB, et al, Collaborative Longitudinal Evaluation of Keratoconus (CLEK) Study Group. Baseline findings in the Collaborative Longitudinal Evaluation of Keratoconus (CLEK) study. Invest Ophthalmol Vis Sci 1998;39:2537– 46. 31. de Sanctis U, Missolungi A, Mutani B, Grignolo FM. Graft central thickness measurement by rotating Scheimpflug camera and ultrasound pachymetry after penetrating keratoplasty. Ophthalmology 2007;114:1461– 8. 32. Altman DG, Bland JM. Diagnostic tests 3: receiver operating characteristic plots. BMJ 1994;309:188. 33. Hosmer DW, Lemeshow S. Applied Logistic Regression. 2nd ed. New York: Wiley; 2000:160 –7. 34. Harrell FE Jr. Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis. New York: Springer; 2001:90 – 8. 35. Schwiegerling J, Greivenkamp JE. Keratoconus detection based on videokeratoscopic height data. Optom Vis Sci 1996; 73:721– 8. 36. Chastang PJ, Borderie VM, Carvajal-Gonzales S, et al. Automated keratoconus detection using the EyeSys videokeratoscope. J Cataract Refract Surg 2000;26:675– 83. 37. Twa MD, Parthasarathy S, Roberts C, et al. Automated decision tree classification of corneal shape. Optom Vis Sci 2005; 82:1038 – 46. 38. Tomlinson A, Schwartz C. The position of the corneal apex in the normal eye. Am J Optom Physiol Opt 1979;56:236 – 40. 39. Roberts C. Characterization of the inherent error in a spherically-biased corneal topography system in mapping a radially aspheric surface. J Refract Corneal Surg 1994;10: 103–11. 40. Belin MW, Khachikian SS. Keratoconus: it is hard to define, but . . . Am J Ophthalmol 2007;143:500 –3. 41. Wei RH, Lim L, Chan WK, Tan DT. Evaluation of Orbscan II corneal topography in individuals with myopia. Ophthalmology 2006;113;177– 83. 42. Lijmer JG, Mol BW, Heisterkamp S, et al. Empirical evidence of design-related bias in studies of diagnostic tests. JAMA 1999;282:1061– 6.
Footnotes and Financial Disclosures Originally received: December 10, 2007. Final revision: February 6, 2008. Accepted: February 22, 2008. Available online: April 11, 2008.
Manuscript no. 2007-1574.
1
Department of Clinical Physiopathology, Ophthalmology Institute, University of Turin, Turin, Italy.
2
Cancer Epidemiology Unit, CeRMS and CPO Piemonte, University of Turin, Turin, Italy.
Financial Disclosure(s): The authors have no proprietary or commercial interest in any materials discussed in this article. Supported in part by the Department of Clinical Physiopathology, Turin University. Correspondence: Ugo de Sanctis, MD, PhD, Department of Clinical Physiopathology, Ophthalmology Institute, University of Turin, Via Juvarra 19, Turin 10121, Italy. E-mail: [email protected].
1539