Superior Rim Stability of the Lens Capsule Following Manual Over Femtosecond Laser Capsulotomy

Abstract

Purpose: Cataract surgery requires the removal of a circular segment of the anterior lens capsule (LC) by manual or femtosecond laser (FL) capsulotomy. Tears in the remaining anterior LC may compromise surgical outcome. We investigated whether biophysical differences in the rim properties of the LC remaining in the patient after manual or FL capsulotomy (FLC) lead to different risks with regard to anterior tear formation.

Methods: Lens capsule samples obtained by either continuous curvilinear capsulorhexis (CCC) or FLC were investigated by light microscopy, laser scanning confocal microscopy, and scanning electron microscopy; atomic force microscopy (AFM) was used to test the biomechanical properties of the LC. The mechanical stability of the LC following either of the two capsulotomy techniques was simulated by using finite-element modeling.

Results: Continuous curvilinear capsulorhexis produced wedge-shaped, uniform rims, while FLC resulted in nearly perpendicular, frayed rims with numerous notches. The LC is composed of two sublayers: a stiff epithelial layer that is abundant with laminin and a softer anterior chamber layer that is predominantly made from collagen IV. Computer models show that stress is uniformly distributed over the entire rim after CCC, while focal high stress concentrations are observed in the frayed profiles of LC after FLC, making the latter procedure more prone to anterior tear formation.

Conclusions: Finite-element modeling based on three-dimensional AFM maps indicated that CCC leads to a capsulotomy rim with higher stress resistance, leading to a lower propensity for anterior radial tears than FLC.

The goal of surgical capsulotomy is to create a round, well-centered, intact rim of the remaining anterior lens capsule (LC) before phacoemulsification. Anterior capsular tears represent a complication, which may interfere with the insertion and centration of the intraocular lens (IOL). A propagation of the rupture toward the lens equator is possible. Femtosecond laser (FL) technology has recently evolved to allow the performance of several important steps in cataract surgery, including capsulotomy. The laser creates a more consistent and more precisely and reproducibly sized and shaped capsulotomy opening than manual techniques. Counterintuitively, an increase in anterior capsular tears has been reported in FL interventions compared to manual capsulorhexis. Initially, this phenomenon was attributed to a learning curve effect of this relatively new technique.^1,2 However, recent research³ has established elevated anterior capsular tear rates across the learning curve and across various FL platforms. Because all anterior capsular tears have been observed in cases with complete capsulotomy, a systematic underlying root cause of the technology itself has been suspected.³ 

The current article analyzes this surprising vulnerability of the femtosecond laser capsulotomy (FLC) edge by comparing basic biomechanical and geometric properties of the capsulotomy edge in both FLC and manual continuous curvilinear capsulorhexis (CCC), based on scanning electron microscopy (SEM), confocal scanning laser microscopy (CLSM), and atomic force microscopy (AFM). While CLSM allows qualitative analysis of the protein composition of the edge,^1,2 AFM represents a unique tool to visualize, manipulate, and quantitatively assess structural and biomechanical characteristics of native biological tissues⁴ and yields quantitative information on capsular rim thickness and stiffness, as well as on edge geometry. Scanning electron microscopy confirmed the AFM findings of the edge geometry. Atomic force microscopy data were fed into a finite-element model (FEM), which was used to determine the effect of capsular rim biomechanics and geometry on intraoperative capsulotomy stability. 

Materials and Methods

Sample Harvesting

Exclusion criteria included patients younger than 18 or older than 90 years as well as chronic metabolic disease including diabetes mellitus. Ethics Committee approval was obtained. The research conducted followed the tenets of the Declaration of Helsinki. 

Manual Rhexis.

Lens capsule material obtained during CCC, performed by four experienced right-handed surgeons, was obtained from a total of 34 consecutive phacoemulsification interventions in 34 patients. Capsulorhexis was performed by manually puncturing the anterior LC and subsequently extending the perforation in a counterclockwise fashion by means of a tearing movement with a cystotome (three surgeons: PBH, DG, KG; refer to author list or acknowledgements) or forceps (one surgeon: SO; refer to acknowledgements). Excised LC material from the circular opening was placed in phosphate-buffered saline (PBS) immediately and stored for a maximum of 24 hours. Donor ages ranged from 58 to 87 years, with an average of 73 ± 8 years. 

Laser Rhexis.

Of a total case series of 80 FL-assisted cataract interventions, 18 consecutive cases were included. Donor ages ranged from 65 to 82 years, with an average of 73 ± 7 years. All operations were performed by one surgeon (CHM; refer to author list) using the Victus femtosecond laser platform (Bausch & Lomb, Aliso Viejo, CA, USA). Lens capsule samples were stored immediately in PBS for a maximum of 72 hours. 

Lens Capsule Mounting

Samples were washed for 5 minutes with 2% Triton X-100 to remove all cellular elements while leaving the extracellular matrix intact. The LCs were subsequently washed in PBS and pipetted in a folded manner onto a poly-L-lysine–coated culture dish (TPP Techno Plastic Products AG, Trasadingen, Switzerland) or coverslips. Samples were centrifuged at 3000g for 5 minutes to firmly attach the LC. The immobilized LCs were kept in PBS for either immediate use or storage. Samples were not fixed in order to avoid shrinkage of the LC due to cross-linking. 

Light Microscopy

The circular LC (Fig. 1A) segments removed during capsulotomy were flat-mounted onto culture dishes for examination by light microscopy. As reported earlier,⁵ isolated LC segments curl, whereby the lens epithelial side faces outward and the anterior chamber side inward (Fig. 1B), allowing for side-specific flat-mounting (Figs. 1C, 1D). 

Figure 1

 

Schematic drawing of the LC orientation. (A) Sagittal view of the lens during capsulotomy with the LC, lens epithelium (LE), lens fibers (LFs), and zonulae (Zo). A circular segment of the LC (star) is removed by either manual or femtosecond laser capsulotomy. Mechanical stress on the capsulotomy rim (arrow) can result from manipulation with the intraocular instruments or is due to the pressure onto the LC. The anterior LC samples obtained during cataract surgery curl when floating in PBS (B) with the Ep facing outward and the ACh facing inward. For examination, the LC samples are flat-mounted onto slides or dishes with either the Ep up (C) or the Ach side up (D). Ach, anterior chamber; Ep, epithelial side.

Scanning Electron Microscopy

Specimens mounted on coverslips were subsequently fixed in 4% glutaraldehyde overnight. Samples were washed in ethanol and dehydrated in a graded ethanol series. After critical point drying, samples were sputter-coated with platinum to a nominal thickness of 3 to 5 nm and examined with a Hitachi S4800 FEG scanning electron microscope (Tokyo, Japan) at 5-kV accelerating voltage. 

Atomic Force Microscopy

Lens capsules were mounted on culture dishes and measured while submerged in PBS to maintain physiological conditions. The measurements were performed by using an ARTIDIS-T system based on a FlexAFM V2 with a 10 μm piezo mounted on a Zeiss Axiovert 135 TV microscope (Zeiss AG, Oberkochen, Germany). For the LC edge measurements, the ARTIDIS-O system with the same scan head and an additional 100 μm piezo in the sample stage (Nanosurf AG, Liestal, Switzerland) was used. Before performing force spectroscopy measurements, the cantilevers' spring constant was determined by using thermal tune and the Sader method,⁶ and the deflection sensitivity was calibrated on a culture dish. Force spectroscopy was performed with a load of 3.1 nN at an indentation speed of 16 μm/s. 

Lens Capsule Side Determination by Stiffness.

Force spectroscopy maps (25 × 25 μm, 20 × 20 pixels) of both sides of the LC were recorded by using ARTIDIS-T with HYDRA6V-200WG cantilevers (k = 0.081 N/m; AppNano, Mountain View, CA, USA). 

Topography and Stiffness of the LC Rim.

Force spectroscopy maps (100 × 100 μm, 100 × 100 pixels) were recorded for LC edge profiling and thickness measurements by using ARTIDIS-O with Prototype SD-AXL-CONT (k = 0.2 N/m, tip-height = 30 μm, Nanosensors; Nanoworld AG, Neuchatel, Switzerland) cantilevers. Atomic force microscopy images of 50 × 50 μm of the LC edge topside were taken with a DNP S-10 (k = 0.06 N/m, tip-height = 2.5–8 μm; Bruker AFM Probes, Camarillo, CA, USA). 

Atomic Force Microscopy Data Analysis and Statistical Analysis

Force-distance curves were analyzed with the NuoAnalyzer (Nuomedis AG, Liestal, Switzerland), using the Oliver-Pharr model⁷ to calculate the elastic (Young's) modulus of the sides and rim of the LC and the quasi-height. Significance was tested by using the Student's t-test. Atomic force microscopy topography data were analyzed with Gwyddion (Necas D and Klapetek P, Masaryk University Kamenice, Brno, Czech Republic). 

Immunostaining

Lens capsules were mounted on coverslips and incubated with a polyclonal antibody to laminin at 1:200 in PBS (L9393; Sigma-Aldrich Corp., St. Louis, MO, USA) and with a monoclonal antibody to the 7S domain of collagen IV (J3-2; Sigma) overnight.⁸ The sample was then rinsed with PBS and incubated with secondary antibodies for an hour (Alexa488, Alexa568; Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA). The sample was rinsed again and mounted onto glass slides with Mowiol. Confocal images were recorded by using an LSM700 inverted microscope (Zeiss AG, Oberkochen, Germany). 

Finite-Element Modeling

Finite-element modeling was used to compare the mechanical stability of CCC and FLC edges in the patients from data derived from the analysis of the surgical LC samples removed during capsulotomy. The geometric information and distribution of material properties were extracted from AFM and SEM measurements (Figs. 2A, 2B). The modeled segments were 104 μm long and divided into an upper (collagen IV, yellow) and a lower (laminin, red) segment (Figs. 2D, 2E), at a thickness ratio of laminin to collagen of 1:3, as found in AFM and CLSM measurements. 

Figure 2

 

Diagrams showing the lateral view of the lens capsule after CCC (A) and after FLC. (B). The nanobiomechanical properties and the morphologic characteristics of the edges created by each of the rhexis methods were obtained by investigating the surgical samples obtained during cataract surgery (red circle). The modeling of the edges remaining in patients (black box) was based on the light, SEM, and AFM microcopy geometry data from the surgical samples. (C) Frontal view of surgical situation: the main deformation that a segment of the rim undergoes owing to the displacement of the surgical instrument against the rim is along the circumferential direction. (D, E) Meshes used for the modeling of the CCC segment and the FLC segment, respectively. A homogeneous displacement of the same magnitude was applied on each of the parallel lateral walls of the segments in the y-direction. The red color in the diagrams depicts the stiffer, laminin-positive side of the LC, whereas the yellow color depicts the 7S collagen IV–positive anterior chamber side. Δy, displacement in y-direction.

A circumferential stretch of the capsular edge was defined as the most relevant stress for capsular tear formation: any radial extension, by contact with the rhexis margin, exposes the adjacent regions to a stretch in the circumferential direction, which is, however, seen as a longitudinal stretch when a small edge segment is considered (Fig. 2C). 

Two 3D meshes of hexahedral elements of second order (Figs. 2D, 2E) were created with Salome 7.5.1 (Open Cascade SAS, Guyancourt, France). A conformal interface between the collagen IV and laminin was implemented,^5,8 by sharing the nodes at the interface to improve the accuracy of the calculation. Finite-element modeling was done with Elmer-FEM 7.0 (CSC, Kajaani, Finland) by using an isotropic linear elastic material model. The Young's modulus (E) and Poisson's ratio (υ), used in the constitutive equation of both domains, were derived for a bilayered composite under uniaxial strain (Equation 1). The overall modulus E_Total is a linear combination of the corresponding material property (E or υ) of the two layers, weighted by the volume fraction (Vf) of each layer, where E_Total = 1.5 MPa⁹ and ν_Total = 0.47.¹⁰ 

The monolayer E_Total = 1.5 MPa corresponds to the modulus in the strain region 0% to 10% of the stress-strain curve.⁹ E_Collagen was assumed to be 0.85 MPa.¹¹ The derived bulk material properties were 4 times higher on the laminin than on the collagen IV side. A displacement equivalent to 10% strain was applied homogeneously on the lateral walls in the y-direction The rest of the surfaces were displacement free, except for two nodes on the back surface, on which fixed Dirichlet boundary conditions were imposed in the x- and z-directions (Δx = 0, Δy = 0). 

Results

Clinical Outcome

Clinically, among the 34 manually performed capsulorhexes no anterior or posterior tears were observed. The 18 femtolaser capsulotomies were also completed without anterior or posterior ruptures. Over the course of the complete series of 80 FL-assisted cataract interventions, no posterior capsular ruptures were observed. In two cases an anterior tear became apparent during the removal of viscoelastic material from the capsular bag after the FLC. In both cases no posterior propagation of the tear was appreciated, and the IOL was successfully inserted into the capsular bag. Both these cases were not part of the 18 consecutive LC samples examined for material properties. 

Lens Capsule Rim Geometry

The diameter of the circular LC surgical samples measured with a stage micrometer under the light microscope were 4.2 ± 0.3 mm (n = 34) for CCC and 5.2 ± 0.5 mm diameter (n = 18; with 16 × 5.4 mm, 1 × 3.7 mm, and 1 × 4.8 mm) for FLC (Figs. 3A, 3C), making the FLC specimens significantly wider in diameter (P < 0.0001). Mounted with the epithelial side up (Fig. 1C), light microscopy evidenced a regular, wedge-shaped edge (Fig. 3B), which was observed in all of the 34 CCC samples. The pointed side of the wedge was always at the anterior chamber side of the surgical sample, whereas the inclining surface was oriented toward the epithelial side (n = 34; Fig. 2A). The width of the wedge was 66 ± 17 μm (n = 46 measurements from 34 CCC samples). Rim geometry was independent of the surgeon involved and independent of whether the CCC had been created with a cystotome or a forceps. Atomic force microscopy height measurements revealed that the anterior LC has a thickness of 31.5 ± 1.1 μm and confirmed the width and angle of the wedge of the CCC samples with approximately 60 μm and an angle of 30°. Owing to the incidence angle of the laser, FLC created LC samples with an edge angle of 80°. A frayed rim was observed in all of the 18 FLC samples (Fig. 3D). 

Figure 3

 

Dark field and phase-contrast micrographs of the circular segments of the anterior lens capsule obtained during cataract surgery. The samples in (A, B) were obtained by CCC, the samples in (C, D) by FLC. The samples were folded and mounted onto glass or culture dishes. The edges of the samples (boxed in [A, C]) are shown at higher power in (B, D). The CCC created a very uniform, wedge-shaped edge of approximately 65 μm width (B). The FLC created frayed, nearly perpendicular edges (D). Scale bars: 1 mm (A, C); 100 μm (B, D).

Scanning electron microscopy confirmed the light microscopy images: CCC-derived LC samples had a very smooth, regular wedge-shaped edge (Figs. 4A, 4B). Femtosecond laser capsulotomy samples were irregularly notched (Fig. 4C, black and white arrows; 4D) along the entire rim and rows of regular circular perforations toward the interior (Fig. 4C) that penetrate the LC completely. The perforations comprised a pattern of at least three parallel rows. The diameter of these perforations ranged from 1.53 to 5.88 μm and averaged 3.66 ± 1.29 μm (n = 35). The spacing between the perimeters of two consecutive perforations along the circumferential direction of the rim was 1.3 ± 0.19 μm, while the spacing between the perimeters of two adjacent holes of neighboring rows averaged 2.53 ± 0.43 μm. 

Figure 4

 

Scanning electron microscopy micrographs showing the edges of lens capsule samples obtained from cataract surgeries. The samples in (A, B) were obtained by continuous curvilinear capsulorhexis, the samples in (C, D) by FLC. The samples had been mounted on glass slides and an overview of an entire sample is shown in (A). High-power view of the edge of the sample (boxed in [A]) is shown in (B). The edge has a very uniform and smooth wedge shape. A low-power view of the edge of an FLC is shown in (C). The edge is irregular and shows numerous perforations created by the laser beams. A high-power view (D) shows the straight border of the LC samples and the curvature created by the laser beam. Scale bars: 500 μm (A); 10 μm (B, C); 2.5 μm (D).

The holes on the third row sometimes formed larger cavities either by the fusion with those on the second row or the ones to their side. These cavities formed the valleys (Fig. 4C, white arrows) of a wavy profile at the cut edge with a diameter of 5.48 ± 2.27 μm. The corresponding peaks or protuberances (black arrows) displayed a height of 6.14 ± 4.05 μm (n = 6) and a width of 11.24 ± 1.24 μm. On top of the peaks of the wavy profile, at the cut edge, there were smaller peaks with amplitudes of approximately 2 μm (Fig. 4C, rectangle; 4D). 

Atomic force microscopy measurements showed that the peaks of the wavy profile (Fig. 5A, long white arrows) at the cut edge had a height and width of 9.23 ± 5.07 μm and 12.34 ± 2.93 μm (n = 7), respectively (Figs. 5B, 5C). The perforations had a diameter of 2.04 ± 1.24 μm (n = 34), while the valleys (short white arrows) of the wavy profile at the cut edge had a diameter of 3.77 ± 2.76 μm (n = 6). The presence of small peaks (1.01 ± 0.29 μm) (Fig. 5A, black arrows) on top of this profile was characteristic. 

Figure 5

 

(A) Peaks (long white arrows), valleys (short white arrows), holes and small peaks (black arrows) found on the FLC samples by AFM. (B) Superolateral perspective of “a segment” similar to the one modeled, where all the features mentioned in (A) can be appreciated in 3D. (C) Sequence of magnification displays, which show the detailed microstructural features that are left by the laser beams. The micro structural features left by the laser beam are smaller than the laser beam.

Mechanics

Atomic force microscopy stiffness measurements confirmed previous reports⁸ showing that the epithelial side of the LC has a 4 times higher stiffness than the anterior chamber side (P < 0.001; n = 30). No stiffness difference was found between CCC and FLC samples. Atomic force microscopy was also used to create a stiffness profile across the entire wedge after CCC (Fig. 6A). The stiffness profile after CCC expectedly showed a higher stiffness at the epithelial fourth of the wedge, in comparison to the other three-quarters corresponding to the anterior chamber side, reflecting the laminin and collagen IV domains (Fig. 6B). 

Figure 6

 

Stiffness profile of the LC after CCC. The SEM micrograph shows the wedge-shaped edge of the CCC sample (A). The stiffness overlay in (B) was across the edge as shown by the stippled line in (A). Staining of the LC edge with antibodies to laminin showed that side-specific labeling was restricted to the epithelial layer of the LC (C). Staining of the same sample with an antibody to the 7S domain of collagen IV a345 showed that the epithelial side was not labeled; rather, the staining was very prominent at the anterior chamber side (D). Double-labeling shows selective staining by either antibody (E). Scale bars: 100 μm (A); Image size: 30 μm × 80 μm; (C, D, E).

Protein Composition

Staining across the edge after CCC showed a distinct and strong laminin signal at the epithelial side of the LC (Fig. 6C), where the collagen IV marker was virtually absent (Fig. 6D). We refer to this side of the LC as the laminin sublayer. Collagen IV 7S was present across the wedge in a gradient that showed a stronger staining toward the anterior chamber side (Figs. 6D, 6E), which corresponds to the stromal side of a basement membrane (BM). We refer to this side of the LC as the collagen IV sublayer. 

Finite-Element Modeling

The rim segments from CCC and FLC showed different stress distributions attributed to their particular edge profile, with maximum stress ratios between the laminin and collagen IV domain (Fig. 7) of 1.3 (CCC) and 2 (FLC). Continuous curvilinear capsulorhexis had a homogeneous stress distribution in contrast to the FLC edge profile, with marked stress concentrations in the valleys of the notched profile (Fig. 8). The laminin domain yielded higher stress values by a factor of 7 in comparison to the collagen domain owing to its higher Young's modulus (Figs. 8A–C, 8E). A decreasing gradient profile in both the collagen and the laminin sides was observed (Figs. 8D, 8F), with the highest value at the point of the lowest radius of curvature. However, the gradient was steeper on the laminin side (Fig. 8F). 

Figure 7

 

Distribution of von Mises stress on the frontal side of the edge remaining in the patient after femtosecond laser capsulotomy. Stress concentrations are present in the valleys of the wavy profile, with higher levels of stress on the laminin side than on the collagen side, reaching a maximum value of 1.56 MPa on the second valley from the left. The rest of the complete valleys (from left to right) on the laminin side also reached high stress values with a maximum of 1.46 MPa (first), 1.50 MPa (second), 0.93 MPa (third), 0.95 MPa (fourth), and 1.18 MPa (fifth), respectively; the collagen side showed values of 0.74, 0.79, 0.5, 0.52, and 0.63 MPa, respectively.

Figure 8

 

Comparison of the von Mises stress distribution between the CCC edge and the FLC edge. The lateral side of both edges (A, B) shows the higher stress value on the laminin side than on the collagen side. The CCC edge (A, C, E) shows a homogeneous stress distribution in each domain as compared to the FLC edge (B, D, F). The lateral side (D) and the inferior side (F) of the FLC edge show the gradient of stress characteristic of the stress field surrounding a semielliptical surface crack (F). The maximum von Mises stress values on the laminin and collagen domains on the CCC edge segment were 0.37 and 0.28 MPa, respectively, while on the FLC edge segment they were 1.56 and 0.78 MPa, respectively.

The von Mises criterion is a measure that is used to predict failure of composite materials. The maximum von Mises stress on the bulk of the CCC and FLC segments are not significantly different, and were 0.37 and 0.38 MPa, respectively. The ultimate strength of the anterior LC under uniaxial loading is 4.11 MPa,⁸ and thus 11 times higher. According to the von Mises criterion, neither the CCC segment nor the FLC segment is close to bulk failure. The maximum von Mises stresses found throughout the whole collagen side of the FLC edge were 3 times higher than in the CCC edge. An even greater difference was found on the laminin side between the two segments, with a ratio of 4.2 (Fig. 8). 

Discussion

Cataract surgery is one of the oldest surgical procedures known,¹² and the most commonly performed contemporary ophthalmic operation with an estimated 6 to 19 million interventions every year worldwide.¹² Critical steps of the technique involve the incision to access the anterior chamber, circular opening of the anterior capsule (capsulotomy), phacoemulsification and aspiration of the crystalline lens, and finally, the insertion of an artificial intraocular lens into the capsular bag.¹³ 

Introduced by Neuhann^14,15 and Gimbel,¹⁵ CCC is a capsulotomy technique by which a circular opening is torn into the anterior LC with a cystotome or intraocular forceps. It is regarded as the gold standard for anterior capsule opening. Animal models have shown a surprising stability of the CCC-created rim,^16,17 attributed to a smooth margin and the absence of irregularities.^18,19 The present data, based on LM, SEM, AFM, and confocal microscopy, confirmed a smooth and regular rim microstructure for all 34 cases of CCC samples examined. Owing to the flat mounting of the LC samples on cover slips, or dishes, the smooth, wedge-shaped rim with a characteristic angle of roughly 30° became particularly obvious. The pointed side of the wedge from the excised LC samples was always directed toward the stromal side (Fig. 2). Accordingly, the pointed side of the wedge of the remaining LC in the patient is directed toward the epithelial side (Fig. 2). Immunostaining and AFM confirmed previous observations that the LC, like all other basement membranes, has two distinct sublayers⁸: the epithelial one-fourth of the LC marked by high stiffness and abundance in laminin, and is therefore referred to as the laminin sublayer; and the anterior chamber (stromal) three-fourths sublayer, which is 4 times weaker and is characterized by an abundance of collagen IV 7S, and is therefore referred to as the collagen IV sublayer. 

All FLs operate at near-infrared wavelengths and produce ultrashort pulses of laser light.²⁰ The localized delivery of the laser energy results in the transformation of tissue into plasma,²¹ a process that can be used to mechanically dissect tissues through the confluence of individual perforations.^20,22 The FL technique has been proposed to be particularly suitable for creating a capsulotomy of exact and optimum size, macrostructural rotundity, and centration with improved consistent IOL/capsule overlap, as compared to manual capsulorhexis.^23,24 Indeed, in our sample, all capsulotomies reached exactly the intended size: 16 samples had a diameter of 5.4 mm. In the two remaining samples, the laser settings had been reduced to an intended capsulotomy size of 3.7 and 4.8 mm, owing to small pupil size. Capsulotomy diameter was thus close to 5.5 mm, regarded by many authors as the ideal capsulotomy size in conjunction with the predominant contemporary IOLs, while the mean capsulotomy diameter was roughly 1 mm below this value for CCC.^25–27 

The rims of the FLC-excised LCs appear macroscopically smooth, sharp, and continuous. However, magnification of the microstructure by means of SEM and AFM analyses demonstrated microgrooves and multiple rows of perforations, corresponding to the numerous FL impulses, by which the FLC was created. Our data are consistent with earlier reports with very similar morphologies.^3,28,29 The FLC margin has an angle of approximately 80° that corresponds to the incidence angle of the laser beam onto the LC. This angle is in agreement with the geometric relationship between the axial incision of the laser and the curved LC.²⁴ The almost perpendicular edge after FLC is very different to the smooth, wedge-shaped 30° LC rim following CCC (Fig. 2). 

To compare the stability of LC rims after the two capsulotomy methods more objectively, AFM data from our study were used to transform the characteristic frayed and smooth profiles of CCC and FLC into three-dimensional models, which could then be used as the basis for FEM. Through FEM, we could link the microirregularities of the FLC to stress concentration in characteristic stress confluence zones (Fig. 7), suggesting a higher risk of capsular tear as compared to CCC with the smooth margin and even stress distributions along the entire length of the rim (Fig. 8). 

For the calculation of the stress distribution, a linear model was chosen, since we were interested in studying whether there were reliable differences between the two segments already within the lowest tissue strain range (0%–10%). A linear model is considered accurate within this strain range. For these small strains the failure point was not reached; but FLC and CCC segments each showed a distinct mechanical performance. At 10% of strain, both segments appeared to be at a similar risk of failure according to the ratio between the LC bulk ultimate strength and the corresponding developed bulk stress (Table). However, microstructural irregularities (notches) caused by the laser in the FLC segment experience high magnitudes of local stress, leading to potential sites for defects where the collagen and laminin networks are prone to initial ruptures. When the network rupture grows to the size of a crack or tear in the LC, propagation to the posterior capsule side will ensue. 

Table

 

Comparison of CCC and FLC Segments' Bulk Stress Versus Bulk UTS and Maximum (Local) Edge von Mises Stresses (Edge Stress) of FLC Versus CCC

 

After a tear is formed, the propagation and ultimate rupture is governed by the mechanics of the fiber networks (Fig. 9). Neither the ultimate tensile strength (UTS) of the laminin and collagen fiber networks in the LC, nor the stress value at which an initial tear is formed, have been reported. As shown in Figure 9, initial tear stress values can be either lower (case 1) or higher (case 2) than bulk UTS when there is no tear present, and the ultimate strength in the bulk can be reached earlier due to tear propagation. Hence, FLC notches, the preferential sites for tear formation and propagation, reduce the bulk UTS. Just like notched paper rips more easily than a clean paper edge, an FLC edge is less stable than a CCC edge. Considering the high rate with which the local stress in the notches of the FLC segment increases linearly to 1.56 MPa when the segment is deformed (Table), and extrapolating this result into the nonlinear region of the stress-strain curve, the FLC rim may be more likely to fail than the CCC rim (Fig. 9), which showed similar low stress levels both at the edge and at the bulk (Table). 

Figure 9

 

Stress-strain curves extracted from the bulk and the edge area from CCC and FLC simulations indicate that the bulk of both specimens, as well as the CCC edge, show very similar behavior. In contrast, stress increases much faster in the notches cut by the laser in FLC specimens. Additionally, the bulk failure level is decreased for the FLC edge (green arrows) as compared to CCC edges (violet arrows) owing to the presence of many microscopic defects, the frayed edge of the FLC, which is absent in the CCC edge.

Continuous curvilinear capsulorhexis and FLC segments have the same material properties. Differences in stress concentrations between CCC and FLC are thus a result of their peculiar edge geometries. The von Mises criterion can be used to predict failure of composite materials. The CCC edge has a maximum von Mises stress ratio of 1.3 between the laminin and collagen domains, meaning that there is only a marginal difference in the stress developed by both layers. Owing to the different geometry, the FLC profile showed a ratio of 2, denoting a 2-fold difference in the stress developed by both layers. Furthermore, compared to CCC, FLC showed 3 times higher stress values on its stromal side and 4.2 times higher values on its epithelial side. In other words, at the relatively low stress level of an extension of 10% used in our model, the stress resistance of the CCC was roughly 4 times higher than that of the FLC. In a surgical situation, higher extensions can be expected. The present study had an experimental approach and the data cannot be transferred directly into the clinical setting. However, the relatively high von Mises ratio for the comparison of CCC and FLC suggests an increased propensity of the FLC for anterior tears. 

Two anterior capsular tears were indeed observed in our clinical series. This finding is in line with increased rates of anterior radial tears reported by different groups, including experienced surgeons and across all available femtosecond devices.³ Our measurements and modeling results suggest the notched margin of the FLC as a possible systematic cause for LC failure,³ independent of the surgeons' learning curve for this new technique.² Implantation of the IOL into the capsular bag was possible without posterior tear extension in both cases of anterior tears in our series, however, suggesting FLC allows a less traumatic management of anterior tears than would be expected in traditional phacoemulsification, mainly due to the relative ease with which the pretreated nucleus may be removed in FLC. 

A major concern in our study was that FLC specimens from only one laser system (Victus, Bausch & Lomb) were analyzed and may hence not be representative of the technology itself. However, thorough analysis of the literature shows that for almost all commercially available laser systems (LenSx, Alcon Laboratories, Inc.; Catalys, OptiMedica Corp.; LensAR, LensAR, Inc.) similar FLC rim geometries are reported when using SEM.^3,24,28–30 No SEM images are available for the Ziemer LDV 8 (Ziemer Ophthalmic Systems AG, city, state, country), but we expect similar rim morphology for this system, since the principle of shooting ultrafast individual laser beams into the LC, to create perpendicular and frayed edges, is the same as for the other systems. Since our stability simulations are based on rim morphologies, we can expect that the data that we generated in our study essentially apply to all available FLC systems and are not restricted to the laser system (Victus) that we used in our study. 

With regard to the future of FLC, improved laser settings that would lead to more stable capsular margins are desirable. Our data also suggest that an improved design of the FLC perforation pattern could be reached by a tighter spacing of individual laser pulses together with an increased beam waist. The resulting smoother margin can be expected to be less prone to tear formation. Alternatively, innovative laser qualities, such as thermal tissue treatment instead of microbursts, might create a smoother microstructural appearance of the capsulotomy margin. 

In conclusion, manual capsulorhexis and FLC created different edge architectures of the capsular opening. When local stress values are considered, the notches of the frayed profile of the FLC created sources of potential tears (Fig. 9). Continuous curvilinear capsulorhexis, in turn, created a surprisingly even edge geometry at the microscopic level. We propose that the FLC is closer to failure than the CCC at the examined level of strain (Fig. 9). 

Acknowledgments

The authors thank the following surgeons: David Goldblum, MD, Konstantin Gugleta, MD, Selim Orgül, MD (Department of Ophthalmology, University of Basel, Basel, Switzerland); Katharina Schaedler for tissue preparation (Pallas Kliniken, Aarau, Switzerland); Josef Flamer, MD, for kind support and consulting throughout the entire project (Department of Ophthalmology, University of Basel, Basel, Switzerland); Imaging Core Facility, (Biozentrum, Basel, Switzerland); Zentrum für Mikroskopie, (Biozentrum, Basel, Switzerland); and Edoardo Mazza, PhD, and Marco Pensalfini (ETH, Zurich, Switzerland) for essential support and counseling. 

Supported by Swiss Nano-Tera Initiative; Comission for Technology and Innovation (CTI), Switzerland; and University of Basel, Nachwuchsfoerderung DMS2136. The authors alone are responsible for the content and writing of the paper. 

Disclosure: M. Reyes Lua, None; P. Oertle, None; L. Camenzind, None; A. Goz, None; C.H. Meyer, None; K. Konieczka, None; M. Loparic, None; W. Halfter, None; P.B. Henrich, None