Review of preoperative tomographic and biomechanical corneal assessment for refractive surgery

Abstract

Screening prior to laser refractive surgery is important to detect abnormal preoperative topography, which is a significant risk factor for developing postoperative iatrogenic corneal ectasia. Pentacam Scheimpflug corneal tomography is effective in screening for ectasia, especially with the Belin/Ambrosio Enhanced Ectasia Display (BAD). However, Pentacam tomography only images corneal shape and it does not evaluate the dynamic biomechanical properties of the cornea. Corneal Visualization Scheimpflug Technology (Corvis ST) provides information on dynamic corneal biomechanical properties and supplements the data provided by Pentacam tomography. The Corvis Biomechanical Index (CBI) uses a combination of biomechanical parameters obtained from Corvis ST to screen for ectasia. Both Pentacam tomography and Corvis ST data are combined to give the Tomographical Biomechanical Index (TBI) which uses machine learning to provide a more accurate screening for early ectasia or keratoconus. The recommended cutoff value for CBI to differentiate normal versus keratoconic eyes is 0.5, with higher values indicating keratoconus. Out of the three indicators – BAD final D, CBI and TBI, TBI had the highest area under the receiver operating characteristic curve for diagnosing ectasia, with high sensitivity and specificity. Corvis ST is a useful tool which complements Pentacam tomographic results to enhance the diagnosis of clinical keratoconus, subclinical keratoconus, and forme fruste keratoconus. This is especially useful to determine the suitability for refractive surgery in patients.

Introduction

Laser refractive surgery is a commonly performed, safe, and effective procedure to correct refractive errors by altering the corneal shape. One of the major complications of laser refractive surgery is iatrogenic corneal ectasia, which has a reported incidence ranging from 0.04%–0.6% post-laser in situ keratomileusis (LASIK),[1,2,3] 0.020% post-photorefractive keratectomy (PRK), and 0.011% post-small incision lenticule extraction (SMILE).[4] The most significant risk factor for developing postoperative iatrogenic corneal ectasia is abnormal preoperative topography (such as forme fruste keratoconus [FFKC], keratoconus or pellucid marginal degeneration).[5] In order to reduce this risk, preoperative corneal tomography or topography is typically performed to detect these abnormalities before performing refractive surgery.

One such modality that is widely used and reliable in detecting abnormal corneal tomography is Pentacam Scheimpflug corneal tomography, which uses Scheimpflug imaging to measure anterior and posterior corneal tomography. In addition, the Belin/Ambrosio Enhanced Ectasia Display (BAD) uses pachymetric data and indices to provide highly sensitive and specific quantitative parameters for the screening of ectasia.[6,7] However, pentacam tomography is limited because it does not evaluate the dynamic biomechanical properties of the cornea.

Corneal Visualization Scheimpflug Technology (Corvis ST) provides information on dynamic corneal biomechanical properties by using a noncontact tonometer that delivers a symmetrically metered air pulse to the cornea to induce a controlled deformation, capturing high-speed images (up to 4000 frames per second) during the deformation process. Corvis ST analyses several parameters such as deformation amplitude ratio (DA Ratio), Integrated radius, Ambrosio Relational Thickness horizontal (ARTh), stiffness parameter at first applanation (SP-A1) and Corvis Biomechanical Index (CBI). Data from both Pentacam tomography and Corvis ST are integrated to give the Tomographical Biomechanical Index (TBI), which improves accuracy in early ectasia and keratoconus screening.[8,9]

Methods

This review analyses the current literature and evidence for the use of Corvis ST in conjunction with Pentacam tomography to enhance the screening of refractive surgery candidates for corneal ectasias. Case examples were included and informed written consent was obtained from the patients to use their anonymized data for publication in this journal.

Results and Discussion

Pentacam tomography is most commonly used in corneal screening before refractive surgery. The two main displays used in the screening for corneal ectasias are the four maps refractive display and the BAD.[10] In the four maps refractive display, the anterior and posterior elevation maps are usually displayed using a best fit sphere of 8 mm [Figure 1a and b]. The sagittal curvature map displays the topographic pattern of the cornea and may distinguish between regular cornea (spherical or astigmatic pattern) or irregular cornea. In the BAD display, to accentuate abnormalities, a 3.5-mm area around the thinnest location is excluded from the calculation of the best fit sphere to calculate the enhanced elevation map. The BAD display consists of two portions: the left are elevation maps and the right are pachymetric progression charts [Figure 1c and d]. At the bottom of the BAD display are 5 parameters with a final D index (BAD-D), which is predictive and separates normal (BAD-D < 1.6) from abnormal cornea. This final D value is based on regression analysis of 9 parameters, including these 5 parameters, anterior and posterior elevation at the thinnest point and Kmax. Patients with keratoconus also have thinner corneas and a faster, more abrupt increase of corneal thickness spatial profile and percentage thickness increase. This is measured by the average progression index, which for a normal cornea should be < 1.2. Another pachymetric parameter useful for diagnosing keratoconus is the Ambrosio’s Relational Thickness (ART) average and maximum values (ARTmax and ARTmin) with areas under the receiver operating characteristic (AUC) curves of 0.987 and 0.983 respectively.[11] ART is obtained by dividing the thinnest point by the pachymetric progression index.

Figure 1.

(a) Right eye Pentacam 4 maps refractive display showing a normal astigmatic cornea: K1:40.2D, K2:41.8D, sagittal curvature map: symmetric bow-tie pattern, normal anterior/posterior elevation maps, normal pachymetric map with thinnest pachymetry 514 µm, (b) Left eye Pentacam 4 maps refractive display showing a normal astigmatic cornea: K1:40.3D, K2:41.5D, sagittal curvature map: symmetric bow-tie pattern, normal anterior/posterior elevation maps, normal pachymetric map with thinnest pachymetry 511 µm, (c) Right eye Pentacam Belin/Ambrosio Enhanced Ectasia Display (BAD) showing no significant difference in the difference map (between the anterior/posterior elevation maps and enhanced anterior/posterior elevation maps) and a normal D value: 0.57, (d) Left eye Pentacam BAD showing no significant difference in the difference map (between the anterior/posterior elevation maps and enhanced anterior/posterior elevation maps) and a normal D value: 0.32

Most studies show that Pentacam tomography is effective in detecting keratoconus and subclinical keratoconus, with the AUC for BAD-D >0.8.[12,13,14,15,16,17] Ruiseñor Vázquez et al. showed that Pentacam BAD-D score had an AUC of 0.95 in keratoconus and 0.93 in subclinical keratoconus, with a sensitivity of 85.5% and 89.2% in keratoconus and subclinical keratoconus, respectively.[16] However, Toprak et al. showed that while the AUC for BAD-D score in mild keratoconus was 0.999, it was only 0.788 in subclinical keratoconus.[18] For FFKC though, Bae et al. reported than BAD-D and ART were not significantly different between normal and fellow eyes of unilateral FFKC patients.[19] Hence, this indicates that Pentacam tomography may not be able to detect FFKC from normal corneas.

Figure 2a and b are examples of a patient with obvious signs of advanced keratoconus on Pentacam tomography with Kmax of 69D, raised anterior and posterior elevation maps corresponding to the inferior steepening pattern seen on the keratometric map and a thinnest pachymetry of 411 µm. However, Figure 2c and d show the patient with FFKC in the fellow eye with normal tomography, normal anterior and posterior elevation maps, normal keratometric and pachymetric maps and a normal D value of 1.37. Hence, Pentacam tomography is unable to differentiate between normal and FFKC eyes.

Figure 2.

(a and b) Right eye: advanced keratoconus on Pentacam with Kmax of 69D, raised anterior and posterior elevation maps corresponding to the inferior steepening pattern seen on the keratometric map, thinnest pachymetry of 411 µm and D value of 16.93 (normal <1.6), (c and d) Left eye: 4 maps refractive display showing a normal astigmatic cornea: K1:42.3D, K2: 44.5D, sagittal curvature map: symmetric bow-tie pattern, normal anterior/posterior elevation maps, normal pachymetric map with thinnest pachymetry 516 µm

Corvis ST provides biomechanical information about the cornea, supplementing the data provided by Pentacam tomography to differentiate between FFKC/subclinical keratoconus from normal corneas. In this noncontact tonometer, a symmetrically metered air pulse is delivered to the cornea and the deformation in the corneal shape is measured with ultra high-speed Scheimpflug technology. The corneal dynamic response has different phases: the first applanation phase, the highest concavity phase and the second applanation phase. The dynamic corneal response parameters generated include DA Ratio which measures the ratio of deflection and eye movement at apex versus 2 mm from apex, integrated radius which refers to the area under inverse concave radius curve (1/radius of curvature), ARTh which refers to pachymetry (thinnest)/pachymetry progression and SP-A1. For DA Ratio and Integrated radius, the lower the index value, the stiffer the cornea while for ARTh and SP-A1, the higher the index value, the stiffer the cornea. The CBI screens for ectasia using a combination of 6 biomechanical parameters obtained from the Corvis ST. In normal eyes, CBI is usually very close to zero, while in ectasia, CBI is usually close to 1. A CBI score between 0.2 and 0.8 indicates a moderate risk for ectasia. A study by Vinciguerra et al. recommended a CBI cut-off of 0.5 to differentiate between normal and keratoconic eyes.[20] Pentacam tomography and Corvis ST data are combined to give the TBI which uses machine learning to provide a more accurate screening for early ectasia or keratoconus.

In keratoconus, it is thought that biomechanical deterioration occurs before tomographic changes or clinical signs.[21] In a study by Ambrósio et al., the AUC for diagnosing ectasia was statistically higher for TBI (0.996), than BAD-D (0.956) and CBI (0.936).[22] Specifically, in patients with clinical ectasia in one eye but apparently normal topography in the fellow eye, AUC for detecting ectasia for TBI was 0.985, higher than BAD-D (0.839) and CBI (0.822).[22] Optimal TBI cutoff values differ when detecting clinical ectasia (cut-off value of 0.79 provides 100% sensitivity and 100% specificity) versus subclinical ectasia/FFKC (cutoff value of 0.29 provides 90.4% sensitivity and 96% specificity).[22] In subclinical keratoconus eyes, DA Ratio, TBI and CBI were all found to be significantly higher as compared to controls.[23] ARTh, SPA1 showed significantly lower values in subclinical keratoconus eyes.[23] In normal eyes, ARTh was significantly higher and integrated radius and CBI significantly lower as compared to FFKC eyes.[24] While comparative analysis between Pentacam tomography and Corvis ST parameters showed that Corvis ST parameters had a lower AUC when differentiating keratoconus from normal eyes, similar AUC was observed between CBI (AUC = 0.785, sensitivity 63.2%, and specificity 80.3%) and BAD-D (AUC = 0.757, sensitivity 52.6%, and specificity 80.3%) (P ≥ 0.0477).[25] In a study involving four groups– normal, bilateral keratoconus, unilateral keratoconus and FFKC group, BAD-D, CBI and TBI were evaluated and the sensitivity and specificity of these indices were evaluated using receiver operating characteristic (ROC) curves.[26] TBI showed the highest diagnostic efficiency in normal versus keratoconus or FFKC eyes (AUC 0.965) and CBI showed the highest diagnostic efficiency when comparing normal versus FFKC.[26] By providing biomechanical data, Corvis ST enhances the information provided by Pentacam tomography, which may be especially useful in FFKC/subclinical keratoconus. The type of refractive surgery performed appears to affect the corneal biomechanical strength to different degrees, with LASIK causing the highest reduction in corneal biomechanical strength, followed by SMILE, PRK or laser-assisted sub-epithelial keratectomy (LASEK).[27]

The CBI was developed using the data from the South American and Caucasian populations. However, several studies have reported ethnic differences in ocular measurements, including corneal thickness, corneal curvature, BAD parameters, anterior chamber depth, corneal diameter, and axial length.[28,29,30,31,32,33] Boyd et al. compared Pentacam indices between 100 North American (New York, USA) and 49 Chinese (Chongqing, China) patients and found that among all variables, corneal diameter influenced pachymetric progression and BAD-D the most, especially in patients of Chinese ethnicity.[30] As for Corvis ST parameters, there are significant differences in deformation response parameters between white and Chinese populations, particularly in ARTh and SP-A1 variables.[34] Specifically, lower values of SP-A1 (indicating a softer cornea) and lower ARTh (indicating a thinner cornea and/or a faster thickness increase toward the periphery) were seen in Chinese patients.[34] In addition, it was found that Chinese patients tend to have smaller corneal diameters.[33] Our previous study showed that eyes with smaller diameters have higher BAD-D and a lower ARTmax and ARTh, and this likely leads to overestimating ectasia risk prerefractive surgery.[35] Recently, a new CBI and TBI optimized for East Asian populations (cCBI and cTBI) was developed to overcome this limitation with increased sensitivity and specificity in the detection of keratoconus.[36] Figure 1a-d are examples of Pentacam tomography of a female 30 year old Chinese patient refractive surgery candidate (manifest refraction R: −5.25/−1.00 × 15°; L: -5.25/−0.75 × 10°) showing normal with-the-rule regular astigmatism in both eyes and low BAD-D scores of 0.57 (right eye) and 0.32 (left eye). However, Corvis ST CBI scores were moderately high (0.42 in the right eye and 0.41 in the left eye). TBI scores were also moderately high in the left eye– 0.34, which is higher than the cut-off value[22] of 0.29 to detect ectasia [Figure 3a and b]. Using these values, this patient would be at higher risk of ectasia following refractive surgery. However, using the new cCBI and cTBI which is more appropriate in this Chinese patient, the cCBI fell within the normal range (0.05 in both the right and left eyes) and cTBI also fell within the normal range (0.10 in the right eye and 0.22 in the left eye) [Figure 3c and d].

Figure 3.

(a and b) Right eye: Corneal Visualization Scheimpflug Technology CBI scores: moderately high (0.42 in the right eye and 0.41 in the left eye). TBI scores: moderately high in the left eye– 0.34 [same patient as in Figure 1a–d], (c and d) The new cCBI and cTBI: cCBI: 0.05 in both the right and left eyes, cTBI: 0.10 in the right eye and 0.22 in the left eye [same patient as in Figure 1a–d]

Conclusion

Corvis ST is a useful tool which complements Pentacam tomography to enhance the diagnosis of clinical keratoconus, subclinical keratoconus and FFKC. This is especially useful to determine suitability for refractive surgery in patients.

Declaration of patient consent

This study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki and its amendments. The authors certify that they have obtained all appropriate patient consent forms. In the form, the patients have given their consent for their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal identity, but anonymity cannot be guaranteed.

Data availability statement

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.

Conflicts of interest

The authors declare that there are no conflicts of interests of this paper.

Funding Statement

Nil.