Ex Vivo Lenticule Customization for Stromal Lenticule Addition Keratoplasty

Abstract

Purpose: The purpose of this study was to explore the optimal shape of customized lenticules for stromal lenticule addition keratoplasty (SLAK) for off-centered ectasia.

Methods: Two different methods to create ex vivo models of eccentric-keratoconus were investigated. Twelve human corneas were used to create model 1 by a hyperopic photorefractive keratectomy (PRK), and model 2 by masked phototherapeutic keratectomy (PTK) on the anterior corneal surface, whereas both types received myopic ablation of the posterior surface. Keratoconus models underwent a modified femtosecond laser (FSL) flap-cut to create stromal pockets. Sixteen human corneas underwent FSL dissection to obtain four lenticule types: type I (planar) and type II (negative) lenticules were used without modifications, whereas type III (customized-planar), and type IV (customized-negative) lenticules underwent further masked-PRK to obtain an asymmetric bow-tie shape. Topographic, aberrometric analysis, and anterior segment optical coherence tomography (AS-OCT) were performed in all recipient corneas before and after lenticule implantation.

Results: Keratoconus model was successfully reproduced. Tomographic analysis showed a significant inferiorly decentered corneal steepening with coherent stromal thinning. Model 2 reproduced better the curvature of real keratoconus. Lenticules type I implantation induced a homogeneous corneal thickening, type III produced higher thickening in the inferior half of the cornea. Type II determined a maximal peripheral pachymetric increase, with a gradual reduction toward the center, and type IV presented an asymmetric peripheral thickening. Topographic assessment showed a cone apex flattening in all cases, but it was significantly higher in types II and IV. Customized lenticules improved significantly corneal surface regularity regarding types I and II.

Conclusions: The approach of customizing lenticules by increasing their asymmetry and tailoring the re-shaping effects, may improve SLAK outcomes in eccentric keratoconus.

Keratoconus is the most common corneal ectatic degeneration. It is characterized by progressive thinning, bulging, and distortion of the cornea and causes progressive changes in vision with increased myopia and myopic astigmatism, corneal irregularity, and visual loss.¹ Penetrating or lamellar corneal transplantation techniques, which present several drawbacks, such as graft rejection, failure, and slow visual recovery due to high levels of induced postoperative astigmatism, are the gold standard procedures for visual rehabilitation in advanced keratoconus.¹ 

Stromal lenticule addition keratoplasty (SLAK) has proven to be a promising minimally invasive treatment option for keratoconus, able to induce central corneal flattening through stromal tissue addition.^2,3 In this technique, central corneal curvature reduction is achieved utilizing peripheral tissue addition produced by intrastromal implantation of negative meniscus-shaped stromal lenticule produced by femtosecond laser (FSL) cut.^2–4 SLAK is still limited to central keratoconus because current FSL technology can produce only stromal lenticules with a symmetric thickness profile.³ 

As a result, only 25% of the eyes affected by keratoconus are suitable for this procedure.⁵ Moreover, the lack of lenticule shape customization algorithms based on cone specifications makes it difficult to optimize the curvature outcomes even in the central conus. 

Alió and colleagues, by investigating the feasibility of corneal stroma regenerative therapy with intrastromal implantation of large decellularized planar lenticules (with or without mesenchymal stem cells recellularization) in advanced keratoconus, observed some corneal flattening regardless of the cone apex location and the iso-planar shape of the implant.^6–8 

As already experienced in intra-corneal ring segment (ICRS) implantation procedures,^9,10 the flattening effect achieved with additive corneal surgery appears to be the result of a combination of corneal volume displacement and collagen fiber tensile strength redistribution.¹¹ The complex combination of both effects makes it difficult to understand the curvature changes required to develop the proper lenticule shape for the most common off-centered keratoconus. 

To clarify these concepts, in this study, we investigated the effect of different types of customized lenticule implantation on an ex vivo eccentric keratoconus model. 

To our knowledge, no validated methods are described in the literature, and only pilot studies describing the use of an excimer laser to model the corneal stroma in a keratoconus-like fashion have been reported.¹² 

Therefore, we proposed to simulate the decentered curvature of eccentric keratoconus, comparing two different technical approaches of donor cornea reshaping aiming at reproducing a realistic geometry in an ex vivo model, created by using excimer laser ablation of both anterior and posterior surfaces of the cornea. 

In addition, we investigated the use of excimer laser ablation to produce custom lenticules suitable for eccentric keratoconus correction. 

This study aims are to create a model that mimics only the corneal keratoconus shape and to explore the optimal lenticule shape that could make SLAK available for off-centered ectasia. 

Materials and Methods

Study Approval

The study protocol was drafted in agreement with the principles of the Declaration of Helsinki and was approved by the Institutional Review Board of the Department of Medicine and Ageing Sciences (University G. d'Annunzio Chieti-Pescara, Chieti, Italy). 

Creation of Keratoconus Eye Model, and Customization of ReLEx Lenticules

This was an ex vivo investigation of corneal curvature re-modeling by intrastromal lenticule implantation in human donor corneas. In this investigation, lenticules obtained from the hyperopic-refractive lenticule extraction (ReLEx) procedure were customized using excimer laser ablation to produce different asymmetric shaped patterns. They were implanted into FSL-intrastromal pockets (flocket) produced in re-shaped eye bank donor corneas. A model for eccentric keratoconus was created in recipient corneas through excimer laser ablation, to evaluate the flattening effect of lenticule intrastromal implantations. 

Two different methods (defined below) were used to reproduce the eccentric keratoconus model: both were compared with a pool of topographies obtained from real keratoconus eyes to establish which was the most reliable. A total of 38 topographies of eyes affected by keratoconus corresponding to the intended parameters were collected from our database and assigned to the real keratoconus group. 

On the other hand, a total of 28 human eye bank donor corneas with scleral rims, non-suitable for human transplantation, were harvested from Fondazione Banca Degli Occhi Del Veneto Onlus (FBOV, Venezia, Italy) institute. All tissues had no evidence of stromal diseases, including opacities or irregularities and were stored for less than 48 hours, with a mean storage time of 30 hours. However, an additional 8 hours of storage in a dextran-enriched medium (Thin-C, Alchimia, Italy) was performed to minimize the storage-induced stromal edema. Tissues with a central corneal thickness (CCT), measured by US-Pachymetry, higher than 600 µm were excluded from the present study. 

Twelve corneas were assigned to the recipient group. Eight of them were sculpted to create two eccentric keratoconus models (4 for each type) with different specifications. After accurate topographic assessment (MS-39, CSO, Italy), the remaining four corneas were shaped with the most efficient technique capable of consistently reproducing the desired parameters of the keratoconus model (model 2) reported in the section of results (Supplementary Fig. S1). 

The remaining 16 donor corneas were assigned to lenticule production. 

Corneoscleral buttons for recipients and lenticules preparation were mounted onto artificial anterior chambers (Network Medical Products Ltd., Coronet House, UK). Internal pressure was standardized with a balanced salt solution (BSS) bottle at 180 cm of height. All surgeries were performed by two experienced surgeons (authors M.N. and J.A.B.). 

Corneal Recipient Preparation

Donor corneas assigned to the recipient keratoconus model group underwent excimer laser ablation re-shaping (MEL-80; Carl Zeiss Meditec, Germany) to simulate the curvature steepening and stromal thinning typical of keratoconus before flocket creation. 

A pool of corneal topographies from patients affected by eccentric keratoconus was used as a reference for ex vivo keratoconus model creation. Topographic stages K3-D3-T1 or T2 according to KDT (maximum keratometry [K_max], decentration, and thinnest pachymetry) staging scheme, as proposed by Tu et al.,¹³ were included. The intended curvature parameters were the following: 

• ○ K_max ranging from 60 and 70 diopters (D).
• ○ Decentration of the thinnest point of about 1.5 mm (ranging between 0.7 and 1.8 mm).
• ○ The thinnest pachymetry ranges between 380 and 490 µm.

The anterior surface reshaping was performed using an inferiorly decentered hyperopic photorefractive keratectomy (PRK) ablation in model 1 or phototherapeutic keratectomy (PTK) ablation with masking of the cone apex in model 2. Stromal thinning was reproduced using an inferiorly decentered myopic PRK ablation of the posterior corneal surface after corneal eversion in both models. 

Thinning Procedure of the Corneas: To begin with, the cornea was everted to expose the endothelium making the posterior surface facing upward, and then mounted onto the artificial anterior chamber. The endothelium was stained with trypan blue dye (Vision Blue; DORC International, Zuidland, The Netherlands), gently peeled off, and the surface dried with an ophthalmic sponge (Merocel Eye Spears; Medtronic Solan, Jacksonville, FL, USA). Approximately 240 µm of posterior stroma were removed using excimer laser ablation (MEL-80) with 1.5 mm inferiorly off-centered myopic PRK treatment (-18 D in total with an optical zone of 5 mm). After the ablation, the optical zone center was stained with gentian violet to easily identify the location of the thinnest point in the later phases. The cornea was then re-mounted onto the artificial anterior chamber in a reversed orientation with the anterior surface facing upward and the epithelium was removed through a blunt spatula. 

Model 1: A hyperopic PRK treatment (+5 D with an optical zone of 6 mm) was performed centered on the mark to reproduce an off-centered steepening of the anterior corneal curvature (Figs. 1A, 1B, 1C). The selected treatment area was 6.70 mm, the optical zone was 6 mm, and the transition zone was 0.7 mm. 

Figure 1

 

Keratoconus models. (A) Everted cornea is mounted onto the anterior chamber, Descemet membrane, and endothelium are removed before ablation. (B) Off-centered myopic ablation (-18 D) is performed on the posterior surface to simulate posterior ectasia and stromal thinning. (C) In model 1, the cornea is returned to its normal position and the anterior surface hyperopic ablation on the mark (+5 D, 5 mm optical zone) is performed. (D) In model 2, the cornea is returned to the normal position and the anterior surface PTK ablation is performed with a 4 mm mask (8 mm diameter and 100 µm depth).

Model 2: A PTK ablation of 100 µm and 8 mm diameter was performed centered on the mark with the exclusion of the central 4 mm area that was masked using a white paper disc placed onto the anterior corneal surface according to the mark position (see Figs. 1A, 1B, 1D). 

An accurate topographic assessment (MS-39) of re-shaped recipient corneas was performed to determine which model better replicates the eccentric keratoconus shape according to the aforementioned parameters. 

The selected keratoconus model for recipient tissues underwent a modified FSL flap-cut procedure with a 500 kHz VisuMax FSL (Carl Zeiss Meditec, Jena, Germany), to produce the flocket according to the previously described method.^2,3,14 To obtain a stromal pocket at a depth of 120 µm from the anterior surface with a single small incision of 4 mm, the following parameters were adopted: lap diameter = 9.4 mm; lap thickness = 120 µm; hinge length = 24.61 mm; and side cut angle = 100 degrees. Laser settings included an energy cut index of 44 nJ and a spot distance of 4.2 µm for lamellar cuts and 1.6 µm for side cuts. Consequently, all recipient corneas were evaluated with anterior segment OCT (MS-39) to check the correct position and depth of the flocket before proceeding to the implantation. A blunt dissector was used to remove the residual tissue bridges in the intrastromal pocket after the FSL stromal cut. 

Lenticule Preparation

The effect of lenticule addition with four different geometric patterns was investigated. All lenticules were created using ReLEx-FLEx hyperopic procedure after corneal epithelium removal with a blunt spatula. The flap cut thickness was set to 120 µm and the diameter was set to 9 mm with a side cut angle of 45 degrees and a side cut length of 360 degrees. The negative meniscus lenticule was created through a +8 D hyperopic setting with a maximal peripheral thickness of 148 µm, a minimal central thickness of 30 µm, an optical zone diameter of 6 mm, and a transition zone of 0.7 mm. 

The first subset of lenticules with planar shape type I (planar lenticules) included the 120 µm stromal flaps obtained from ReLEx-FLEx procedure, as described by Alió and colleagues.⁶ 

The type II (negative lenticules) were extracted after FSL treatment reproducing the same specifications used in the previous clinical study realized by Mastropasqua and collaborators.² 

For these two lenticule types, a blunt dissector was used to gently separate the free flap (planar lenticule) from the underlying negative lenticule and the latter from residual stroma, without performing any excimer ablation on them. Lenticule types I and II are represented in Figure 2. 

Figure 2

 

Non-customized lenticules. (A, B) Type I lenticule, with a planar shape (B), is obtained from the 120 µm stromal flap (A; red arrows) of the ReLEx-FLEx procedure. (C, D) Type II lenticule, with a negative meniscus shape (D), is sculpted with a +8 D hyperopic FSL ReLEx-FLEx treatment (C; blue arrows).

Type III (or customized planar) and type IV (or customized negative) lenticules underwent further modification with excimer laser ablation before their removal to obtain an asymmetric shape (Fig. 3). 

Figure 3

 

Planar and negative lenticules customization. After placing a 10 mm-wide mask onto the inferior half of the lenticule (A), an astigmatic PRK treatment (B) is performed to produce a bowtie ablation of the superior half only.

In detail, after the aforementioned FLEx procedure, type III lenticules were created by placing a 10 mm-wide mask onto the inferior half of the flap (see Fig. 3A) and performing a +3.50 D, axis 180 degrees astigmatic PRK treatment with an optical zone diameter of 7.00 mm, to produce a bow-tie ablation of the superior half only (Fig. 3B). The thickest side of the lenticule located at 6 o’clock was marked with gentian violet to recognize its position after implantation. 

The lenticules type IV were produced with the same modality adopted for type III lenticules. After removal of the stromal flap, the anterior surface of the negative meniscus lenticule underwent +2.50 D, axis 180 degrees astigmatic masked PRK treatment with an optical zone diameter of 6.00 mm (see Fig. 3). The lenticules types III and IV are represented in Figure 4. 

Figure 4

 

Planar and negative customized lenticules. (A, B) Type III lenticule, with an asymmetric planar shape (B), is obtained through masked excimer laser ablation of the ReLEx-FLEx stromal flap (A). The thick red arrow indicates the masked zone, and the thin red arrow indicates the ablated zone. (C, D) Type IV lenticule, with an asymmetric negative meniscus shape (D), is obtained through masked excimer laser ablation of the stromal lenticule sculpted with the ReLEx-FLEx treatment after flap removal (C). The thick blue arrow indicates the masked zone and the thin blue arrow indicates the ablated zone.

With this procedure, two lenticules (1 planar/customized planar and 1 negative/customized negative) were created from each corneoscleral button and a total of eight lenticules for each type were obtained. All donor corneas underwent anterior segment optical coherence tomography (AS-OCT) assessment after lenticules sculpting (and before their removal) to confirm the achievement of the expected shape. 

Surgical Technique

The corneal flap was lifted after an accurate blunt spatula dissection from the underlying negative lenticule, which was, in turn, delicately separated from the residual stroma taking care of avoiding tissue damage. Specifically designed forceps were used to implant lenticules in recipient corneas flockets respecting their original orientation (anterior surface directed upward). Appropriate distention was achieved with forceps branch opening and fine adjustments were made later with surface manipulation through the blunt spatula. Lenticules were carefully centered, according to the corneal mark, on the conus apex. A thorough attention was needed for the radial orientation of the asymmetric lenticules (types III and IV) whose mark, corresponding to the thickest region, had to match with the conus apex. At the end of each procedure, the intraoperative microscope target cross was used to confirm proper lenticule distension and centration. The surgical procedure is illustrated in Figure 5. 

Figure 5

 

Schematic representation of the surgical procedure. After intrastromal pocket creation (see text), the lenticule is positioned near the incision (A) of the recipient cornea, pushed inside the pocket (B), and when it reached the central position (C) proper distention is achieved by means of forceps opening (D).

Lenticule Implantation’s Main Outcome Measures

Topographic and aberrometric analysis and AS-OCT (MS-39) were performed in all recipient corneas before and after lenticule implantation. A specifically realized holder for the artificial anterior chamber was used to keep corneas in front of the instrument while scanning. 

Curvature changes were assessed through the average keratometry values in the central 3 mm zone (AvgK) and the K_max on the anterior surface of the cornea. In addition, the peripheral curvature in the 6 mm zones for the inferior and superior hemi-meridians and the symmetry index of the anterior curvature (SIf – Symmetry Index_FRONT) were recorded. The latter parameter is defined as the difference of curvature (expressed in diopters) of 2 circular zones with a radius of 1.5 mm centered on the vertical axis in the inferior and superior hemispheres. 

The pachymetry map was used to assess changes in the corneal apex (corresponding to the apex of the cone in keratoconus corneas) thickness, whereas vertical and horizontal AS-OCT scans were acquired to manually measure lenticule dimensions (central, minimal, and maximal lenticule thickness and lenticule diameter) through the caliper tool before extraction from donor corneas and after implantation in the recipient corneas. 

Corneal asphericity variations were assessed using the Q-value in the 6 mm peripheral-inferior zone (where the cone apex is located), the superior hemi-meridians, and the average Q-value of the 4 hemi-meridians. 

The root mean square per unit of area (RMS/A) of the anterior surface, defined as the deviation in regularity/aberrations of the corneal surface from the best fit asphero toric surface, was used as a shape index to quantitatively express corneal aberrations. 

A total of five study groups were determined: the recipient corneas before the implant were designated by (group 0), and the groups after lenticule implantation were defined by (groups 1, 2, 3, and 4 in which lenticules types I, II, III, and IV were implanted, respectively). Groups 1 and 2 were used as control groups, to verify the actual effect of lenticule customization. 

Statistical Analysis

Statistical analyses were conducted using MedCalc Software version 19.3.1 (MedCalc Software Ltd., Ostend, Belgium). Corneal apex thickness, AvgK, K_max, RMS/A, Q-values, SIf, and peripheral curvatures were treated as continuous variables and presented as mean ± standard deviation (SD). To detect departures from normal distribution, the D’Agostino-Pearson test was performed. Logarithmic transformation was applied when necessary to achieve normal distribution. 

Because of the small size of the study samples, the Mann-Whitney test was used to test for statistically significant differences between keratoconus models and real keratoconus parameters. 

Differences between data sets before and after lenticule implantation were assessed with paired samples t-tests. A value of P < 0.05 was considered statistically significant. 

Results

Keratoconus Model Results

An ex vivo keratoconus model was successfully reproduced in all corneas (X model 1 and Y model 2). An additional number of 4 corneas were assigned to Model 2, which reproduced better the corneal curvature of the real keratoconus group regarding model 1. Detailed data obtained from the real keratoconus group (Fig. 6, Supplementary Fig. S2) and both keratoconus models are reported in Table 1. 

Figure 6

 

Pre-implantation topographies of keratoconus models. (A, B) In model 1, in the anterior tangential map (A), a paracentral inferior oval steep area (corresponding to the cone apex) and a peripheral superior arcuate steep zone (corresponding to the superior edge of the anterior hyperopic excimer laser ablation) can be observed. The superior paracentral flattening between them is responsible for the high Symmetry Index of the anterior surface. The posterior elevation map (B) shows the inferior paracentral ectasia (corresponding to the cone apex). (C, D) In model 2, in the anterior tangential map (C), a paracentral inferior oval steep area (corresponding to the cone apex) can be observed. The posterior elevation map (D) shows the inferior paracentral ectasia (corresponding to the cone apex).

Table 1.

 

Mean Topographic Parameters of Real Keratoconus and Keratoconus Models

 

Topographic analysis showed a significant inferiorly decentered corneal steepening in both keratoconus models (see Fig. 6). Model 1 produced a smoother curvature regression from the cone apex toward the periphery compared to model 2 but was affected by a semilunar steepening in the superior half of the cornea corresponding to the shoulder of the ablation area (see Fig. 6A). Model 2 produced a sharper cone area with a more regular peripheral curvature compared to model 1 but the conus apex was affected by an isle of minimal flattening due to the total masking of that area during laser ablation (see Fig. 6C). 

Corneal thickness was significantly lower in model 1 compared to model 2 and real keratoconus (P < 0.01). Higher K_max and inferior curvature (i.e. lower radius of curvature) were found in model 2 compared to model 1 and the real keratoconus (P < 0.01 and P = 0.02, respectively). No significant differences in the central corneal curvature were observed between the two models and the real keratoconus group (P = 0.63). Model 1 showed superior peripheral curvature, asphericity, and SIf significantly higher compared to model 2, and the real keratoconus (P = 0.02, P = 0.01, and P < 0.01, respectively; see Table 1). 

The topographic assessment was repeated after FSL stromal pocket creation and no significant changes in the aforementioned parameters were observed (P > 0.2 for all thickness and curvature parameters). Residual tissue bridges were removed successfully through a blunt dissector. No damage to the stomal bed and stromal cap occurred and, on AS-OCT scans, interfaces were slightly hyper-reflective with a mean pocket depth of 153 ± 16 µm (Fig. 7A). 

Figure 7

 

Anterior segment OCT scan of recipient cornea (keratoconus model 2) after stromal pocket creation (A) and lenticule implantation (B–E). (A) The pocket is slightly hyper-reflective and its depth is constant along the entire surface. (B) Type 1 lenticule with a planar profile. (C) Type II lenticule with a negative meniscus profile. (D) Type III lenticule with an asymmetric planar profile. (E) Type IV lenticule with an asymmetric negative meniscus profile. The regular lenticules interfaces and the absence of folds can be appreciated in all cases. Thick and thin red arrows (D, E) indicate the masked and the ablated areas of customized lenticules, respectively.

Implanted Lenticules Results

Each recipient cornea was implanted with all four lenticule types in random order. Topography was repeated after every lenticule removal to verify the recovery of pre-implantation parameters, which was confirmed in all cases. Lenticules parameters derived from AS-OCT assessment before extraction are reported in Table 2. 

Table 2.

 

Mean Dimensions of Each Lenticule Type Before Extraction

 

Lenticule extraction and implantation with proper centering and distension of the tissues inside the intrastromal pocket were obtained in all cases. The regular lenticules interfaces and the absence of folds and voids were documented with AS-OCT. Lenticule profiles were clearly appreciable in contrast with the surrounding recipient corneal stroma and they were properly distended inside the pocket throughout their extension (Figs. 7B–E). 

Lenticule Dimension Outcomes

Lenticule dimensions were re-assessed after implantation. Negligible variations of central, minimal, and maximal thicknesses were detected (P = 0.23, P = 0.28, and P = 0.36, respectively), whereas lenticule diameter was found to be slightly inferior to the pre-extraction values in all lenticule types (P = 0.06). Morphologic evaluation of recipient corneas on AS-OCT scans revealed regularly distended lenticules inside the stromal pocket without folds or fluid retention (see Figs. 7B−E). 

Pachymetric Parameters Outcomes

The pachymetric assessment showed an overall increase of corneal thickness in the lenticule addition area in all cases with different patterns dependent upon each lenticule shape. Planar lenticules induced a homogeneous corneal thickening, whereas customized planar lenticules produced higher thickening in the inferior half of the cornea consistent with the lenticule thickness profile and its rotational orientation. Negative lenticules determined a maximal increase of pachymetry at their periphery, with a gradual reduction toward the center. The peripheral thickening was asymmetric (more pronounced in the inferior half of the cornea) after customized negative lenticule implantations with the thicker area of the lenticule positioned inferiorly. In both customized and non-customized negative meniscus lenticules, the area of maximal thickening had an annular shape and was located at 1.5 mm of radius (3.0 mm diameter) from the thinnest point in the pachymetric map, corresponding to the cone apex. Mean corneal apex thickness observed in group 1 and group 3 was significantly higher compared to groups 2 and 4, respectively (P < 0.01). 

Topographic Parameter Outcomes

In all cases of lenticule implantation, corneal topography showed a flattening of the cone apex regardless of the lenticule type (Fig. 8). Flattening was confirmed by a significant reduction of K_max in the central cornea which was more pronounced in groups 2 and 4 compared to groups 1 and 3 (P < 0.01). 

Figure 8

 

Topographies before and after lenticule implantations. (A) The keratoconus model tangential anterior map shows a paracentral inferior steep area corresponding to the cone apex. (B–E) All the differential maps after lenticule implantations show the reduction of the cone apex curvature and the peripheral steepening, that is annular and pericentral after type II (C) and type IV (E) lenticules implantation, whereas type I (B) and type III (D) lenticules induce a wider steepening. However, both type III (D) and type IV (E) lenticules implantation result in a less pronounced steepening in the superior hemisphere compared to type I (B) and type II (C), respectively.

An increase of anterior curvature was detected in correspondence of negative lenticules periphery where the maximal intrastromal pocket spacing was produced by the thickest part of the lenticule. Although type II lenticules caused a uniform annular steepening along the whole periphery (see Fig. 8C), customized negative lenticules (type IV) induced a lower curvature increase in the upper peripheral half of the cornea where the thinner part of the lenticule was located (see Fig. 8E). Asymmetric curvature change was documented by lower SIf values in group 2 compared to group 4 (P = 0.02). 

In addition, planar lenticules (types I and III) induced a peripheral steepening that was less pronounced but wider compared to negative lenticules (see Figs. 8B, 8D). 

Average keratometry did not change significantly in group 2 compared to group 0 (P = 0.12). A reduction of AvgK, instead, was observed with planar and customized lenticule, with a mean difference between pre- and post-implantation of 4.05 D (P < 0.01) with type I, of 4.4 D (P < 0.01) with type III, and of 3.2 D (P = 0.02) with type IV. 

Corneal asphericity variations resulted differently depending on the type of the lenticule implanted. Customized negative lenticule presented the highest effect on asphericity, especially on the inferior hemi-meridian, with a mean difference of 1.15 between pre- and post-implantation Q-values (P < 0.01). Type II lenticules induced a comparable effect on inferior asphericity (P = 0.58), whereas the superior Q-value became more negative after implantation, resulting in a less pronounced average Q-value increase compared to group 4 (P = 0.04). Planar and customized planar lenticules had a lower effect on corneal asphericity compared to negative lenticules, affecting less the inferior hemi-meridian. Both of them induced a similar increase in inferior Q-value, whereas the average and superior Q-values increased more in group 1 than in group 3 (P = 0.01 and P = 0.02, respectively). 

The regularity of the anterior corneal surface, expressed by RMS/A, was significantly influenced by lenticule customization. Type I lenticules did not induce any variation in RMS/A value, although a slight reduction was detected in group 2 (P = 0.12). Customized negative and, especially, customized planar lenticules, induced a significant reduction of the RMS/A (P = 0.03 and P = 0.01, respectively). The topographic parameters of each group are reported in (Table 3). 

Table 3.

 

Mean Topographic Parameters of Each Study Group

 

Discussion

The possibility of corneal remodeling by tissue addition was already proposed at the beginning of the refractive surgery era back in the 1950s. According to the “thickness law” introduced by Barraquer, central corneal flattening can be achieved by removing tissue from the central cornea or by peripheral addition.¹⁵ The rapid evolution of subtractive techniques with the introduction of the excimer laser,^16–18 led to the early abandoning of the tissue addition idea for refractive purposes. 

The introduction of small incision lenticule extraction (SMILE) not only innovated the field of refractive surgery but also opened new paths for innovative surgical approaches to corneal pathologies. Stromal lenticules created by FSL have accurate geometry, smooth surfaces, preserved cell vitality, and undamaged collagen structure.^19–21 Moreover, FSL can be used to fashion intrastromal dissection planes minimally affecting the biomechanics of the anterior stromal layers.²² 

Intra-stromal implantation of refractive lenticules derived from myopic SMILE has been proposed as a tissue addition-based approach to steepen the cornea for treating presbyopia, hyperopia, and aphakia.^23–25 Nearly contemporary to Ganesh et al.²⁶ and Alió et al.^6,27 we hypothesized the use of lenticule implantation to treat keratoconus by improving the anterior corneal shape. 

Ganesh et al. proposed the implantation of donut-shaped stromal lenticules obtained by a 3 mm central punching of cryopreserved myopic lenticules combined with collagen cross-linking progressive keratoconus.²⁶ The result of the addition of stromal tissue in the mid-periphery and around the cone caused a relative flattening in the center and modification in the corneal shape to a less hyper-prolate shape.²⁶ The theoretical mechanism of action of this technique is thought to be partially similar to the ICRS because both techniques involve the addition of volume and local elevation in the mid-periphery.²⁸ 

In our first ex vivo study, we showed that the implantation of stromal lenticules obtained from hyperopic-FLEx treatment,² whose thickness profile gradually increases peripherally toward the edge of the optical zone, induces central corneal flattening.² The implantation of such shaped additional stromal tissue, besides decreasing the curvature, may provide biomechanical support to the recipient cornea thanks to the central stromal tissue addition, whereas, at the same time, the creation of the intrastromal pocket is believed to minimally alter the biomechanical strength.²³ 

On these bases, we investigated the morphological and refractive effects of intrastromal implantation of negative meniscus-shaped stromal lenticules, obtained from eye bank donor corneas, in the treatment of advanced non-progressive central keratoconus.³ Topographical analysis showed that all eyes had a detectable reduction of central anterior corneal curvature, indicating a significant relative flattening of the cone, with negligible effects on the posterior corneal curvature.³ Correspondingly, a significant improvement of corneal asphericity (Q) values was noted, indicating a less hyperprolate shape, along with a significant increase in visual acuity and consistent reduction of myopic refractive error and manifest astigmatism.³ The main limitation of this study is that only patients with central keratoconus were included because only symmetric lenticules can be produced with the available FSL cutting algorithm at present. 

Alió and colleagues⁶ attempted for the first time to combine stromal keratophakia with stromal regenerative therapy (by seeding mesenchymal stem cells on decellularized human stromal lenticules) to treat patients with advanced keratoconus. They used lenticules with a planar profile and homogeneous thickness of 120 µm with a 9 mm diameter. Independently from recellularization, they observed a significant improvement in anterior keratometric and pachymetric parameters after planar lenticule implantation, witnessed by the reduction in avgK and K_max and an increase of minimal CCT.^7,29 

The principle of the SLAK technique is to reshape the anterior corneal profile by stromal tissue addition, increasing thickness, and improving the anterior corneal curvature, which is the main determinant of visual impairment in keratoconus. Although remarkable results were obtained in central keratoconus with symmetric lenticules,^2,3,30 to our knowledge, a customized stromal tissue addition to improve the outcomes in eccentric ectasia has not been described yet in a model of eccentric keratoconus. The currently available FSL platforms do not allow the creation of stromal dissections with desired asymmetric profiles. Therefore, we resorted to a masked excimer laser ablation to reshape FSL-derived stromal lenticules. 

It has to be noted that one of the main limitations of ex vivo studies investigating intrastromal lenticule addition for treating corneal ectasia is the fact that ex vivo models of keratoconic corneas have not been established yet. Pedrotti and colleagues investigated, ex vivo, the feasibility of intrastromal lenticule insertion to restore corneal shape in a model of ectatic human cornea.¹² Assuming that ectasia in keratoconus starts from the posterior stroma, they realized a posterior corneal surface ectasia model with excimer ablation aimed at achieving the flattening of the posterior surface by myopic lenticule implantation. In their model, excimer laser ablation on the endothelial side only produced paracentral stromal thinning, and a minimal nonpredictable increase in anterior corneal curvature was obtained by high-pressure inflation of the artificial anterior chamber (AAC) with BSS up to 100 mm Hg. 

Despite posterior flattening, the implantation of myopic lenticules induced an increase in the anterior corneal curvature with a trend to a more prolate profile.¹² 

In the present study, we aimed to provide a standardized and controlled keratoconus corneal model from donor sclero-corneal discs, that takes into account both the pachymetric reduction and the localized increase in anterior surface curvature, typical of the disease, obtained by posterior and anterior excimer laser ablations, without the need to modify AAC pressure beyond physiological values. 

Additionally, we compared the effects on corneal curvature of intrastromal implantation of different lenticule profiles, in a such created ex vivo model of eccentric keratoconus. 

The first issue to be addressed was the creation of a reliable eccentric keratoconus model using excimer laser corneal reshaping. We compared the tomographic results of two different ablation procedures with a pool of real eccentric keratoconus topographies from our patients’ database. In both models, a myopic PRK on the endothelial side was performed to achieve the cone apex thinning. To obtain the apical steepening, the anterior corneal surface underwent a hyperopic PRK in model 1, and a PTK with a paper disc mask on the conus apex in model 2. In both cases, the typical inferior paracentral oval pattern on the anterior tangential map was obtained. Model 2 showed higher K_max values and a steeper curvature progression from the cone to the periphery and it more accurately reproduced the profile of a real keratoconus because it did not alter significantly the superior hemi-cornea. On the other hand, the anterior hyperopic excimer ablation performed in model 1 resulted in a superior paracentral flattening (higher SIf) and peripheral steepening (higher 6 mm superior peripheral curvature), which may affect keratometric indices (see Fig. 6). In addition, the higher cone apex thickness observed in model 2 was more consistent with curvature parameters referring to real keratoconus. For these reasons, model 2 was chosen to produce recipient corneas for lenticule implantations. 

In the actual ex vivo model, four different lenticule profiles were therefore evaluated. The decellularization protocol was not considered, because the influence of the presence or absence of the corneal keratocytes will not have any effect on this level of the study. In future investigations, this procedure will be taken into consideration. The planar lenticule (type I) and the negative meniscus shaped lenticule (type II), produced by FSL only, were already tested in the previous clinical studies that proved their efficacy in the treatment of keratoconus.^2,3,6–8 The customized type III and type IV lenticules were obtained from the formers by further excimer laser ablation. As expected, all lenticules improved corneal pachymetry proportionally to the amount of the tissue addition. 

Although this ex vivo model simulates eccentric keratoconus conditions, it offers an environment significantly different from the in vivo application. The in vivo phenomena, such as the post-implantation wound healing processes, inflammatory reaction, keratocyte activity and collagen remodeling, and nerve degeneration as well as regeneration processes, occurring in the early and late period after surgery cannot be evaluated in such an ex vivo study. 

The topographic assessment showed detectable reductions in central anterior corneal curvature, pointing out a noteworthy relative flattening of the cone apex achieved in all cases after implantation, although it was higher with negative and customized negative lenticules, compared to planar and customized planar ones. Two principles may be responsible for these observations. The addition of volume around the cone base, which is greater with type II and type IV lenticules, displaces the anterior surface upward and reduces the curvature of the cone apex according to the “law of thickness.”¹⁵ Second, the insertion of a spacing element between collagen lamellae induces an increased tension of stromal fibers that are physically rerouted around the added tissue. As a consequence, an arc-shortening effect is produced and the cone apex becomes relatively flatter.²⁸ 

Types I, III, and IV lenticules also led to a reduction in AvgK, whereas negligible differences were observed after type II lenticule implantation. This can be explained by curvature changes induced by the thickest part that superiorly falls in the paracentral corneal region. It follows a keratometric redistribution with inferior flattening and superior steepening that results in an AvgK almost unchanged. Differently, customized negative lenticules produced a less curvature increase in the paracentral cornea thanks to their lower local tissue volumes, due to customized ablation, whereas planar and customized planar lenticules, having a wider diameter and a more regular profile, affect paracentral corneal curvature much less. 

A similar trend was observed also for asphericity changes following tissue addition. In the keratoconus model, the inferior hemi-meridian was the most aspheric due to the presence of the cone apex. Ideally, it would be desirable to flatten the cone without affecting the superior corneal curvature to obtain a close to normal corneal asphericity. Negative meniscus lenticules had the greatest effect in flattening the cone apex, but they increased the paracentral superior corneal curvature hampering the improvement in average Q-value. Although customized planar lenticules had a lower effect on the superior curvature, they affected the corneal asphericity the least because they also induced a lower K_max reduction. Planar lenticules showed a midway behavior, with an average effect between the previous ones, whereas customized negative lenticules proved to be the most effective in restoring a close-to-normal corneal asphericity. The significant cone apex flattening, combined with a lower superior steepening compared to negative lenticules, can be accounted as an explanation of this phenomenon. 

Interestingly, lenticule customization had a strong impact on the shape index RMS/A, improving anterior corneal surface regularity compared to the best fit asphero toric surface. The masked excimer laser ablation gave all customized lenticules an asymmetric bowtie profile allowing to add less tissue in the superior half of the stromal pocket (oppositely to the cone apex), where no pachymetric and curvature changes are needed. Customized planar lenticules were the most effective in restoring a more physiological corneal profile, probably due to the absence of a focal thickening in the paracentral superior cornea and to their wider area that positions the edge of the lenticule out from the 8 mm area explored with the RMS/A index. 

The concept of adding a higher amount of tissue around the cone apex, exploiting the law of thickness, and the arc-shortening effect,^15,28 proved again to be the most effective in reducing its curvature. The main issue was the impact that negative meniscus lenticules had on the central corneal curvature, which is crucial for visual performance. Planar lenticules partially prevent this problem at the expense of lower cone flattening. Lenticule customization allowed us to overcome these drawbacks and, at the same time, to exert remarkable effects on corneal asphericity and regularity. Customized negative lenticules induced the highest improvement in corneal asphericity, resulting in a less hyperprolate shape, whereas customized planar lenticules proved to be the most effective in restoring a corneal profile, closer to the best fit asphero toric surface. 

Improving regularity and reducing the excessive curvature in the central cornea, along with increasing the stromal thickness, is probably the most challenging goal of the intrastromal addition procedures aiming to re-shape the keratoconic corneas, having a profound impact on non-correctable high-order aberrations, and therefore on the corrected visual acuity of patients. 

Despite an ideal lenticule shape suitable for every keratoconic eye cannot be realized, considering the wide range of asymmetry of keratoconus morphologies, the approach to customize lenticules by increasing their asymmetry and tailoring the re-shaping, refractive, and structural effects in eccentric cones may be feasible. Corneal remodeling should be certainly further investigated to develop customization algorithms, hopefully, supported by technology advancements to optimize their reproducibility. In addition, an ex vivo model can only partially reproduce the complex impairment of corneal structure and geometry induced by keratoconus. Creating this ex vivo keratoconus model by removing of Bowman membrane, as well as the endothelium, and Descemet from the corneal center does not change the fact that the stromal collagen structure is impaired in real keratoconus. The aim of this study was not to reproduce the corneal tissue characteristics or biomechanic behavior, but only to explore the shape and thickness changes preliminarily to optimize lenticule reshaping for further in vivo investigation in animals, and humans, in which we will study possible changes in thickness, curvature, aberrometry, and the corneal biomechanics in real pathological corneas after adding such tissue. 

Because the process of lenticule production might be complex and requires a specific setup, other approaches can be considered as eye bank lenticule production, cryopreservation, and delivery of customized lenticule. 

Future studies to assess the functional impact of custom-made lenticule implantations in patients with eccentric keratoconus and the wound healing process after excimer laser ablation are, therefore, needed. If providing satisfying best corrected visual acuity and contact lens tolerance, such minimally invasive additive procedures may reduce the need for the more invasive approaches, such as traditional keratoplasty. 

Acknowledgments

The authors, their families, their employers, and their business associates have no financial or proprietary interest in any product or company associated with any device, instrument, or drug mentioned in this study. The authors have not received any payment as reviewers, consultants, or evaluators of any of the instruments, devices, or drugs mentioned in this study. 

Author Contributions: Mario Nubile was the principal investigator, principal surgeon, study concept and design, analysis and interpretation of data, writing and critical revision of the manuscript. Luca Cerino was responsible for writing the manuscript, analysis and interpretation of data, writing and critical revision of the manuscript, tables, and figures. Niccolò Salgari was responsible for the study concept and design, analysis and interpretation of data, critical revision of the manuscript, and the figures. Mona El Zarif was responsible for the writing in the introduction, critical revision of the manuscript and the figures, and final version preparation. Jorge L. Alió del Barrio was the co-principal investigator, and co-principal surgeon for the critical revision of the manuscript. Manuela Lanzini was responsible for the interpretation of data and critical revision of the manuscript. Michele Totta was responsible for the interpretation of data and critical revision of the manuscript. Leonardo Mastropasqua was responsible for interpretation of data and critical revision of the manuscript. 

Disclosure: M. Nubile, None; J.L. Alio del Barrio, None; L. Cerino, None; N. Salgari, None; M. El Zarif, None; M. Totta, None; M. Lanzini, None; L. Mastropasqua, None