Clinical Ocular Biomechanics: Where Are We after 20 Years of Progress?

Abstract

Purpose: Ocular biomechanics is an assessment of the response of the structures of the eye to forces that may lead to disease development and progression, or influence the response to surgical intervention. The goals of this review are (1) to introduce basic biomechanical principles and terminology, (2) to provide perspective on the progress made in the clinical study and assessment of ocular biomechanics, and (3) to highlight critical studies conducted in keratoconus, laser refractive surgery, and glaucoma in order to aid interpretation of biomechanical parameters in the laboratory and in the clinic.

Methods: A literature review was first conducted of basic biomechanical studies related to ocular tissue. The subsequent review of ocular biomechanical studies was limited to those focusing on keratoconus, laser refractive surgery, or glaucoma using the only two commercially available devices that allow rapid assessment of biomechanical response in the clinic.

Results: Foundational studies on ocular biomechanics used a combination of computer modeling and destructive forces on ex-vivo tissues. The knowledge gained from these studies could not be directly translated to clinical research and practice until the introduction of non-contact tonometers that quantified the deformation response of the cornea to an air puff, which represents a non-destructive, clinically appropriate load. The corneal response includes a contribution from the sclera which may limit corneal deformation. Two commercial devices are available, the Ocular Response Analyzer which produces viscoelastic parameters with a customized load for each eye, and the Corvis ST which produces elastic parameters with a consistent load for every eye. Neither device produces the classic biomechanical properties reported in basic studies, but rather biomechanical deformation response parameters which require careful interpretation.

Conclusions: Research using clinical tools has enriched our understanding of how ocular disease alters ocular biomechanics, as well as how ocular biomechanics may influence the pathophysiology of ocular disease and response to surgical intervention.

Keywords:

• Ocular biomechanics
• keratoconus
• laser refractive surgery
• corneal crosslinking
• glaucoma

The last 20 years have seen tremendous advancement in our technology and subsequent understanding of the fundamental biomechanics of healthy eyes, as well as those with pathology. Prior to 2005, the field of ocular biomechanics was confined to ex-vivo human studies, animal models, or computer simulations. Although much was learned from these investigations to inform future studies, translation to the clinic with a meaningful impact on direct patient care was at best, quite limited. The transformative event was the development of clinical devices to quickly assess biomechanical response in living subjects. The issue then became how to interpret the data acquired, which was difficult due to the complex nature of the measurements.

Biomechanical assessment involves quantifying the response to an applied load. Historically, this process was destructive in nature, which is clearly not appropriate for the clinic. Therefore, compromises had to be made, and both of the two commercial devices currently available utilize an air puff as the nondestructive load on the cornea (Figure 1). The cornea undergoes a macro-deformation from its natural convex shape through applanation to concavity during the loading portion of the cycle as the air pressure increases. This process is reversed during the unloading portion of the cycle, which begins as the air pressure peaks and is reduced. Once the cornea reaches its limit of concavity in the inward, forward direction, it subsequently moves outward through reducing concavity, a second applanation, and recovery of its convex shape. As the cornea first enters the concave phase in the forward direction, aqueous humor is displaced, and the sclera is engaged to resist this fluid motion.

Figure 1. Schematic diagram of air puff loading of the cornea with red arrows indicating direction of movement. The initial phase is Baseline Convex prior to deformation. The Loading Phase includes deformation through Inward Applanation (flattening) and into Concave phase in the Inward direction. The Unloading Phase includes initiation of recovery from Concave phase in the Outward direction through Outward Applanation to the Convex shape Recovered.

Involvement of the sclera in the deformation response was first demonstrated using paired donor corneas with air puff load.¹ One eye of the donor pair was mounted as a whole globe and in the fellow eye, a cornea-scleral rim was excised and mounted in a rigid artificial anterior chamber. Significantly greater deformation was measured in the whole globe mount from 10 mmHg to 50 mmHg than with the rigid artificial anterior chamber mount (Figure 2). The difference between deformation in the two eyes decreased as the internal pressure increased and the sclera of the whole globe stiffened with each pressure increment due to nonlinear biomechanical properties to be described in the next section (Figure 3). The lowest pressure had the greatest difference, indicating that the stiffer the sclera, the greater the resistance to fluid motion, which limited corneal deformation. This result might be misinterpreted as a stiffer corneal response. Follow up studies with finite element analysis, as well as measurement before and after a scleral stiffening procedure in donor eyes confirmed the key role of the sclera in resisting fluid displacement and limiting corneal deformation.^2,3 The scleral influence is important in both commercial devices, but the interpretation is different due to the distinct air puff strategy of each.

Figure 2. (A) Illustration of the difference between deformation of donor eye pairs under air-puff loading with one eye in a whole globe mount (WG) and the fellow eye in a rigid artificial anterior chamber mount (AC) under experimentally set internal pressures of 10 mmHg to 50 mmHg, showing greater deformation in WG mount for all pressures. (B) Plot of the difference in deformation between mounts showing nonlinear reduction in difference as the WG sclera stiffens with increasing pressure and exhibits behavior closer to the rigid AC mount. Reprinted from Metzler et al.¹ (C) Schematic of mechanism for scleral involvement in corneal deformation by air puff (gray arrow) which displaces aqueous that is resisted by the sclera, thus limiting fluid displacement and corneal motion. Adapted from Nguyen et al.²

Figure 3. Schematic example of a stress (load per unit area) versus strain (percent deformation or stretch) paradigm for the cornea. The cornea is an example of a nonlinear viscoelastic material with a stiffened response as strain increases, shown by the increasing slope in blue, and a different path during unloading that is shifted to the right. Corneal hysteresis (CH) is at the same strain or shape during loading and unloading and is represented by a vertical red line. The apparatus for uniaxial strip testing is shown in the inset with a strip of scleral tissue.

How does intraocular pressure affect biomechanical assessment?

Intraocular pressure (IOP) is constantly loading (i.e. exerting a force per unit area upon) the globe. As IOP increases, a stiffer response is produced due to the nonlinear material properties of the cornea and sclera, shown in Figure 3. The stress is defined as the force per unit of cross-sectional area experienced by the tissue when loaded in a test where a strip of ocular tissue is stretched. The test apparatus is illustrated in the Figure 3 inset with an excised strip of scleral tissue. Strain is defined by the non-dimensional deformation or percent stretch as the tissue is pulled. The tangent elastic moduli (proportional to stiffness) are calculated as the slope of the stress-strain curve at a given value of strain. The slope increases as the load increases, which is characteristic for this nonlinear behavior. Therefore, IOP is critical in describing and assessing biomechanical behavior in the cornea and sclera. A weaker cornea under greater IOP may produce a stiffer response than a stronger cornea under lower IOP. It has been reported that the strongest predictor of deformation amplitude under an air puff load is the IOP,⁴ followed by stiffness and thickness. The stress distribution is a function of both thickness and curvature, which can be quantified in the cornea using Hoop Stress: (1) $$\sigma =P•R/2t,$$(1) where σ is the stress, P is the IOP, R is the radius of curvature, and t is the corneal thickness. This equation predicts that higher stress is associated with a thinner and flatter cornea with higher IOP. It is important to note that this equation was derived for a sphere with uniform thickness and radius of curvature, which is clearly an approximation when applied to the cornea. However, this approximation can be quite useful to estimate simple relationships, which can then be validated with more complicated mathematical approaches such as finite element modeling.

Ocular response analyzer and the cornea visualization Scheimpflug technology

Ocular Response Analyzer

Quantification of the biomechanical deformation response in the Ocular Response Analyzer (ORA; Reichert, Depew, NY) is with an electro-optical system that detects both applanation events.⁵ The air puff is customized to each individual, such that higher IOP generates a greater maximum air pressure. This customization is accomplished by detecting the first applanation (A1) and shutting down the internal signal to the piston. Inertia in the piston is what causes the air pressure to continue to increase as the sclera actively resists fluid displacement.⁶ The cornea never reaches its limit of deformation before the air pressure begins to decrease. Hysteresis is simply a difference in the biomechanical path during loading and unloading (Figure 3). In other words, the magnitude of the load and corresponding deformation is different when the load is increasing than when it is decreasing, with a shift in the unloading pathway. In the ORA, the difference in air pressure between A1 and the second applanation (A2) is defined as Corneal Hysteresis (CH), which represents the viscoelastic nature of the cornea and sclera. The most important feature for interpretation is that it represents energy dissipation (or damping), and there is no direct correspondence with corneal stiffness.⁶

CH is a dynamic parameter that will change as IOP increases or decreases. An increasing IOP directly causes CH to decrease and vice versa. Therefore, when monitoring an individual patient, it is recommended to always evaluate CH in the context of IOP. The influence of the sclera on CH has been demonstrated in contralateral eyes of subjects where one eye underwent a scleral buckle procedure, stiffening the sclera and reducing CH compared to the fellow eye without intervention.^7,8 This result is extremely important in the interpretation of CH as an ocular parameter, rather than purely a corneal parameter. The load is applied to the cornea and the entire eye responds to dissipate the energy.

Cornea visualization with Scheimpflug Technology

Cornea Visualization with Scheimpflug Technology (Corvis ST; Oculus, Wetzlar, Germany) quantifies the biomechanical deformation response through analysis of 140 images acquired with a high-speed system using Scheimpflug geometry that captures the deformation process over 30 ms (Figure 4).⁹ Analysis of these images allows for repeatable and reproducible measurements of the corneal deformation response along with the potential to evaluate corneal elastic parameters.^6,10,11 In contrast to the ORA, the spatial and temporal profiles of air puff generated by the Corvis ST are consistent from patient to patient. Therefore, every cornea reaches its limit of deformation, which is the beginning of the Oscillation Phase shown in Figure 5. As the air pressure continues to increase past this limit, additional tissues become engaged, inducing Whole Eye Movement (WEM) that is resisted by the orbital structures. This response has enabled the Corvis ST to evaluate the compliance of orbital soft tissues in thyroid eye disease.¹³ Once the pressure reaches its consistent maximum, the unloading phase begins. The dynamic corneal response (DCR) parameters produced by the Corvis ST that are associated with the loading phase are predominantly elastic in nature, up to and including the point of highest concavity, which is the time of greatest resistance by the sclera to fluid displacement. This is the core concept of stiffness parameter at highest concavity (SP-HC), which is the load at first applanation (air pressure minus IOP) divided by displacement from first applanation (A1) to highest concavity (HC; Figure 4(A)) and has been shown to capture the scleral response.^2,12 Other DCRs that are associated with corneal stiffness, are relatively independent of IOP, and are associated with the shape of the cornea during deformation include Deformation Ratio 2 mm (DA Ratio), which is the central deformation divided by the average of the two lateral deformations (Figure 5(A)) and is maximum near A1, as well as Integrated Inverse Radius (IIR), which is concave curvature integrated between A1 and second applanations.¹⁴ Additional elastic parameters associated with corneal stiffness are Stiffness Parameter at First Applanation (SP-A1), which is load at applanation divided by displacement from undeformed to A1 (Figure 4(A)),¹² and the Stress-Strain Index (SSI), which determines stiffness from a finite element model, validated with clinical data.¹⁵

Figure 4. (A) A series of anterior surfaces of the deforming cornea plotted throughout the air puff showing initial displacement (short thick red arrow) from the undeformed surface to first applanation (A1), marked by a horizontal red line, and further displacement (longer thick red arrow) from first applanation to the highest concavity (HC) limit of corneal deformation. (B) A series of frame captures from a video of the cornea deforming under an air puff with red arrows indicating direction of motion.

Figure 5. Superimposed frames at different time points extracted from a single examination, showing (A) cornea in the predeformation phase (pseudocolored blue), at maximal corneal deformation or Highest Concavity (pseudocolored red), and at maximal whole eye movement (pseudocolored white); (B) cornea at maximum deflection with illustration of displacement from predeformation anterior surface arc (blue line); and (C) correction for whole eye motion by aligning all corneal positions to that at predeformation. Note the example shown is the exam with the greatest magnitude of whole eye motion (WEM) in a large dataset for illustration of multiple phases, and therefore represents an extreme value of WEM with the temporal progression from blue to red to white shown at the edges of the imaging window. Also note that maximal whole eye motion occurred after corneal deflection had recovered. (D) Phases of Deformation with the consistent air puff of the Corvis ST. Corneal Apex Displacement (blue) and Whole Eye Motion (green) are directly measured and subtracted to calculate Corneal Deflection (red). Adapted from Roberts et al.¹²

Interpretation of the elastic parameters must consider the nature of the calculation. As described, SP-HC, SP-A1, and SSI all represent different algorithms for stiffness. Therefore, an increase in SP-HC, SP-A1, and SSI indicates an increase in these different measures of stiffness. In contrast, a decrease in DA Ratio, Peak Distance and IIR indicates an increase in stiffness since these are shape parameters of the cornea during deformation and a lower value indicates greater resistance to a change in shape. Peak Distance describes the width of the deformation defined by the bending “peaks” at the edges, which increases with increasing depth (Figure 5(B)). IIR represents the area under the curve of inverse radius vs time between applanation events, so it is a cumulative response rather than a single point. Inverse radius is simply curvature with lower concave curvature indicating a stiffer response. These shape parameters are illustrated in Figure 5. WEM is also illustrated along with the difference in Deformation Amplitude (DA) quantifying cornea apex displacement and includes cornea motion plus WEM compared to Deflection Amplitude (DeflA) which quantifies cornea motion only, with WEM removed. It is important to note, as previously stated, that central DA and DeflA are strongly influenced by IOP,^4,14 and therefore should always be interpreted in the context of IOP.

Biomechanics of keratoconus

Keratoconus (KCN) is a disease of the cornea that 20 years ago was thought be an overall weak cornea based on uniaxial strip testing studies that were not designed to determine biomechanical asymmetry, or a regional difference in properties.¹⁶ If a corneal strip with weakness in only one focal region were stretched in this tensile test, it would be measured as overall weak compared to normal corneas. It has since been learned that keratoconus is characterized by a focal area of weakness^17,18 that is associated with a focal area of thinning and increased curvature. It has been proposed that this focal weakness is the initiating event in a susceptible cornea, resulting in a cycle of biomechanical decompensation with increased strain and thinning in the weak area, leading to a focal increase in stress. The curvature will increase in response to reduce the stress according to Equation (1)(1) $$\sigma =P•R/2t,$$(1) , which further modifies the stress distribution, and the cycle continues (Figure 6).¹⁹ The ultimate result is that the keratoconic cornea demonstrates asymmetry of biomechanical properties, thickness, and curvature. Therefore, the primary event in development of keratoconus would be asymmetry in biomechanical properties, with thickness and curvature secondary to the underlying structural changes.

Figure 6. The biomechanical cycle of decompensation in ectasia is shown with the initiating event the focal reduction in elastic modulus that leads to a redistribution of stress, which generates thinning as the cornea strains, leading to another redistribution of stress. The cornea responds with an associated focal increase in curvature that again redistributes the stress. The cycle continues as the ectasia progresses. Adapted from Roberts and Dupps.¹⁹

If focal biomechanical weakening is the primary event in the development of keratoconus, then it should be detectable prior to alteration in both thickness and curvature patterns. The salient feature would be asymmetry of biomechanical properties, rather than overall weakening. This focal weakness in keratoconus has indeed been reported both ex-vivo and in-vivo using Brillouin Microscopy, a noncontact optical device which measures very small frequency shifts in the scattered light due to spontaneous acoustic phonons that are related to the biomechanical properties of the tissue.^17,18 Brillouin Microscopy is in the process of commercialization for clinical use but remains a laboratory tool as of the timing of the current review. As previously stated, current clinical devices provide a global assessment of biomechanical response and cannot evaluate asymmetry.

Assessment of keratoconus with the ORA and with the Corvis ST

Significant differences have been reported in biomechanical parameters between diagnosed keratoconus and healthy controls. In the original paper introducing the ORA and CH,⁵ it was shown that CH is significantly lower in KCN, which can be interpreted that the viscoelastic response is lower. In other words, the cornea/globe is less able to dissipate energy. Although there is a significant difference in the means, there is also a large overlap in the distribution of CH between KCN and healthy subjects, indicating that the global parameter of CH would not be appropriate for screening or detection of KCN. In addition, the ORA produces a plot of the air pressure and infrared applanation waveforms from which waveform parameters are derived that can be used to improve keratoconus detection (Figure 7).²⁰ However, waveform parameters are no longer included in the newest models of ORA.

Figure 7. Sample Ocular Response Analyzer waveforms from (A) a healthy individual and from (B) an individual with keratoconus, showing blunted peaks in keratoconus. Peaks occur when the cornea is applanated. The y-axis is number of photons aligned with the detector of the device, and the x-axis is time (ms). The green trace is air pressure created by the piston of the device. The red and blue traces are the unfiltered and filtered, respectively, number of photons reaching the detector. (C) Two sets of waveforms generated by the Ocular Response Analyzer from a patient with keratoconus before cornea crosslinking (red) and one year after cornea crosslinking (blue). Note the difference in the applanation pressures does not change, despite the greater peaks in the post-op blue waveforms, indicating stiffening of the cornea. Adapted from Roberts.⁸

The Corvis ST has shown great success in the detection of KCN using elastic parameters, leading to the conclusion that elastic deterioration is a stronger feature of keratoconus than alteration in viscoelasticity. This is consistent with the focal disruption of lamellar orientation reported in keratoconus,²¹ which would link collagen ultrastructure with weaker elastic properties. An index termed Corvis Biomechanical Index (CBI) was developed to differentiate keratoconus from normal corneas using a combination of DCRs, which resulted in area under the receiver operating characteristic curve greater than 0.98.²² Other studies have confirmed this performance. When biomechanical parameters are combined with corneal tomographic parameters that can capture asymmetry, the Tomographic Biomechanical Index (TBI) has been shown to further improve KCN detection.²³ The problem remains to detect, define, and interpret sub-clinical keratoconus when the CBI and TBI parameters produce differing predictions. One study of subjects with a clinical diagnosis of ectasia in one eye and normal topography/tomography in the contralateral eye showed abnormal CBI in both eyes of 12 subjects.²⁴ However, it remains problematic to interpret an abnormal CBI in the presence of normal topography/tomography, but in the absence of ectasia in the contralateral eye.

Ocular biomechanics and cornea crosslinking

The asymmetric nature of KCN is important to consider when developing and planning biomechanical-based treatments such as cornea crosslinking (CXL) using riboflavin with UVA light as the photosensitizer, as well as intrastromal ring placement. These two approaches are based on different mechanisms to treat the pathology. Intrastromal rings generate an immediate modification of the stress distribution due the change in curvature pattern induced by rings. This new stress distribution can interrupt the cycle of decompensation. It has been reported that there is no change in properties with intrastromal rings with both ORA²⁵ and Corvis ST,²⁶ so the long-term biomechanical response is strictly due to the change in stress distribution. In contrast, CXL is designed to increase the stiffness of the cornea by increasing the number of crosslinks in order to shift the stress-strain curve to the left and to interrupt the cycle of decompensation by biomechanically stabilizing the cornea to prevent further progression.²⁷ This goal has been achieved in multiple reports as CXL gained popularity,²⁸ and regression over several years is one possible outcome.²⁷

Is it possible to measure changes in stiffness and viscoelasticity after CXL that drive the outcomes of cessation of progression to potential regression? Interestingly, it has been reported that there is no change in CH after one year following CXL,²⁹ despite a positive clinical outcome. This is likely due to changes in viscosity that mask changes in elasticity or stiffness, which is illustrated in Figure 8 with a simple viscoelastic model that shows hysteresis can be constant with either high or low elasticity and/or viscosity.³⁰ Therefore, it is difficult to interpret changes in viscoelasticity after CXL. However, CXL is designed to increase corneal stiffness, which is an elastic parameter. Alterations in the waveforms that accompany CH measurement have shown significant changes indicating a qualitative increase in stiffness (Figure 7(C)).²⁹

Figure 8. Resulting plot of hysteresis as a function of viscosity from inset viscoelastic model if springs are equal with three values of elastic constant. At lower viscosities, hysteresis increases as elastic constant decreases. At higher viscosities, hysteresis decreases as elastic constant decreases. A horizontal line of constant hysteresis shows that hysteresis is not unique, but a function of elasticity and viscosity. Adapted from Glass, et al.³⁰

On the other hand, quantitative studies of stiffness response using the Corvis ST have shown significant changes after CXL in multiple DCRs, all associated with increased stiffness, up to four years after treatment indicating a long term result.^31–33 Specific DCRs with changes include significant increases in SSI and SP-A1, as well as significant decreases in IIR and DARatio. Although these elastic parameters can be used to evaluate biomechanical response to CXL treatment, the question that remains is whether outcomes can be predicted after CXL given measurable differences in baseline biomechanical parameters. More research is needed.

Biomechanics of laser refractive surgery

In the early days of Laser Vision Correction (LVC) more than 20 years ago, the biomechanical response of the cornea to a change in its structure due to tissue ablation by an Excimer laser³⁴ was not recognized.³⁵ This incomplete understanding of the response impacted both design of ablation profiles and variability of outcomes. Ablation profiles were initially spherical, based on Munnerlyn’s formula,³⁶ and then moved to aspheric with a modified Munnerlyn for the purpose of smoothing transition zones. The magnitude of programmed corrections of the multiple laser manufacturers were adjusted empirically based on differences in targeted and actual outcomes of large datasets of cases. The aspheric profile was shown to reduce variability in outcomes, likely a combination of reduced biologic and biomechanical responses.

This empirical algorithm adjustment approach was followed in the early 2000s by an intense effort to use newly developed optical wavefront sensors to evaluate not only 2nd order sphere and cylinder, but also higher order aberrations that impact blur and affected visual quality, including 3rd order coma and 4th order spherical aberration. Wavefront-guided procedures were initially designed to generate aberration-free outcomes and subsequently to reduce induced aberrations in the optical zone surrounded by a transition zone. Wavefront-guided procedures were ultimately abandoned when it was shown that reduction in induced aberrations could be produced without using a wavefront sensor.³⁷

Interestingly, the primary aberration that could not be eliminated was spherical aberration, which is the result of increased curvature in the mid-periphery driven to a large part by the biomechanical response of the entire cornea to a change in its central structure.³⁸ Lamellae run limbus-to-limbus, and the loss of tension in the peripheral lamellae that are severed centrally by the removal of tissue leads to an increase in the peripheral swelling pressure,³⁹ peripheral stromal thickening,⁴⁰ and an outward “pulling” force that causes biomechanical central flattening and peripheral steepening (Figure 9(A,B)). The central flattening occurs whether the procedure is a myopic procedure intended to flatten the central cornea (Figure 9(C)) which is therefore biomechanically enhanced, or a hyperopic procedure intended to steepen the central cornea (Figure 9(D)) which is in biomechanical opposition to the ablation profile. It was shown in a cornea culture model that the vertical cuts which severed lamellae in flap creation were predominantly responsible for the biomechanical response, and that the interlamellar horizontal cut had little effect.⁴¹ Although these results were in the context of a flap, they also apply to the lamellae that are severed during tissue removal to create the refractive effect. The biomechanical response increases as the cornea stiffens with age, so greater central flattening occurs with older individuals. Greater biomechanical flattening means less laser-induced tissue removal is required in myopic treatments with increasing age, but greater laser-induced tissue removal is required in hyperopic treatments with increasing age where the ablation profile is intended to increase central curvature in opposition to the biomechanical response. Epithelial remodeling follows the biomechanical response and leads to thickening in areas of lower stromal curvature and thinning in areas of greater curvature,⁴² which blunts the intended correction in both myopic and even more so, hyperopic treatments due to the nature of the corneal surface shape produced.

Figure 9. (A) Schematic representation of a pre-operative cornea showing lamellae under tension with interlamellar cohesive forces and swelling pressure. (B) The same cornea post refractive surgery after tension-bearing lamellae are severed, which increases the swelling pressure in the remaining lamellar segments and causes a thickening in the periphery that is linked to central flattening. (C) Red area shows shape of tissue removed for a myopic profile to produce flattening with radius of curvature, R2, greater than R1, and (D) Red area shows shape of tissue removed for a hyperopic profile to produce steepening with R2 less than R1. A and B reprinted from Ruberti et al.³⁸

Biomechanical response to laser vision correction

The biomechanical response to LVC falls into two major categories – that which affects every stable outcome and that which leads to biomechanical decompensation and iatrogenic ectasia. The current discussion will focus on interpretation of the corneal response to structural alteration, which affects the post-op surface shape and thus, every outcome. Three major surgical approaches are currently being performed. The first encompasses various versions of Photorefractive Keratectormy (PRK), or corneal surface ablation, including transepithelial PRK (tPRK) and other techniques for epithelial removal grouped as Advanced Surface Ablation (ASA). The second approach is Laser in situ Keratomileusis (LASIK) with various mechanisms for creating a lamellae-severing flap under which the ablation occurs, including a microkeratome or a femto-second laser (FS-LASIK). The third is small incision lenticule extraction (SMILE) which is intended to preserve the anterior region of the stroma, reported as the strongest layer of the cornea due to lamellar interweaving.⁴³ All three procedures are illustrated in Figure 10. The anterior layer is removed in PRK and it is the region where loss of lamellar tension occurs in LASIK, due to the flap. SMILE uses a femtosecond laser to create a lenticule of tissue, which is physically removed through a small side cut, leaving the anterior region predominantly intact. Corneal tissue is removed in all three approaches, whether at the surface, under a flap, or under a cap, meaning that the cornea is thinner after all three approaches. Many publications have compared outcomes between the procedures. However, procedures are difficult to compare biomechanically without bias since the surgical parameters use different optical zones, transition zones, flap/cap thicknesses, and resulting residual stromal beds (RSB), as well as different tissue removal techniques between ablation and lenticule creation, all of which impact the biomechanics of the post-operative cornea.

Figure 10. (A) Central profile of tissue removal (red) for a myopic surface ablation with anterior peripheral lamellae no longer in tension. (B) Top view of surface ablation without a flap or a cap. (C) Central profile of tissue ablated (red) in myopic Laser-Assisted Keratomileusis (LASIK) with a flap overlying the ablated region and peripheral lamellae without tension. Severed lamellae in anterior flap region can no longer bear tension. (D) Top view of LASIK with near circumferential severing of lamellae in flap region with the presence of a hinge, often nasal. (E) Central profile of lenticule creation for tissue removal (red) in myopic Small Incision Lenticule Extraction (SMILE) with a cap overlying the lenticule and lamelle in region of cap under reduced tension than pre-operatively due to longer arclength on posterior cap than anterior residual stromal bed. Peripheral lamellae in region of lenticule are without tension. (F) Top view of SMILE with small side cut for lenticule removal, often superotemporal.

Despite the difficulties in comparing procedures, the ORA, the Corvis ST, and computer simulations have been used to evaluate the biomechanical impact of the different LVC procedures in multiple studies, partially summarized in a meta-analysis that included ORA and Corvis ST.⁴⁴ All biomechanical parameters were affected after LVC, indicating the predominant biomechanical impact on the cornea was the tissue removed, rather than the specific procedure performed. It was also reported that CH was preserved to a greater extent in SMILE than LASIK or FS-LASIK, but was comparable to Femtosecond Lenticule Extraction (FLEx) or PRK. FLEx is a procedure where a femtosecond laser is used to both create a lenticule as well as a flap to allow the lenticule to be lifted from the stromal bed. The similarity of a flapless procedure to that with a flap may seem confusing, as well as the lack of comparability found between LASIK and FLEx, both of which have a flap that severs lamellae. However, FLEx and SMILE both rely on a similar tissue removing technique in the creation of a lenticule, while LASIK and PRK rely on ablation with an Excimer laser. In addition, CH does not evaluate stiffness or strength (as mistakenly reported in the meta-analysis), since CH is a viscoelastic parameter. It may be that lenticule creation preserves the viscoelasticity to a greater extent than Excimer ablation. It was also reported that the difference in CH between SMILE and FS-LASIK was greater after post-operative month 12. The meta-analysis reported no difference between SMILE and FS-LASIK using a subset of DCRs from the Corvis ST. Unfortunately, the DCRs chosen in the published studies did not include the newest ones that are less dependent on IOP and are more reliable indicators of stiffness, so this result cannot be considered conclusive. Using the new DCRs of Integrated Inverse Radius and DA Ratio 2 mm in a comparison of tPRK to FS-LASIK with the Corvis ST, it was found that there was greater change associated with a reduction in stiffness in FS-LASIK than in tPRK. In other words, less weakening was produced with tPRK.⁴⁵

In a paired eye comparison, 5 subjects received SMILE in one eye and FLEx in the fellow eye, so that the only difference between eyes was the presence of a flap since both procedures relied on lenticule creation.⁴⁶ Patient-specific finite element models were generated for the 10 pre-operative eyes. The tissue representing the lenticule was removed in each simulation, and an inverse approach was used to iterate the biomechanical properties in the model until the post-operative tomography was achieved within a specified error limit. The flap procedure (FLEx) produced 49% (range 2–87%) greater mean reduction in stromal collagen fiber stiffness within the flap region compared to the contralateral cap region with SMILE. After simulating an increase of IOP in the model, lower stresses and deformations were shown in the residual stromal bed in the eyes that had received SMILE.

One cannot consider the cap region in SMILE as comparable to the pre-operative anterior region of the cornea, since there are two distinct biomechanical regions in terms of stress and strain in the post-operative cornea. There is a mismatch in the arclength between the posterior cap, which is greater than the arclength of anterior residual stromal bed (Figure 10). In other words, the back of the cap is longer than the front of the bed. Therefore, the lamellar tension in the region of the cap is reduced compared to the stromal bed and reduced compared to pre-operatively. In LASIK, the anterior lamellae are severed, so they can no longer bear tension. This change leads to a more dramatic stress disparity between the flap region and the bed. PRK results in only one central biomechanical region relative to stress, which likely drives the smaller changes in biomechanical parameters. Ultimately, the amount of tissue removed is the primary biomechanical impact on the cornea, as previously stated. The location of that tissue, whether at the surface, under a flap, or under a cap is secondary.

It is not possible to recover pre-operative strength using an accelerated CXL protocol after tissue removal. This variation of LVC is termed “Extra” and has been attempted after PRK, LASIK, and SMILE. However, it does not restore the pre-operative biomechanical parameters. Rather, CXL with specific irradiation and exposure parameters has been shown to reduce the weakening from tissue removal, which is critically necessary for a refractive effect.⁴⁷

Biomechanics of glaucoma

Glaucoma is a group of ocular diseases that is associated with progressive optic neuropathy and that leads to the loss of visual function. Although the pathophysiology of glaucoma is complex, and its risk factors are multifactorial, elevated IOP is widely regarded as the most important variable in the incidence⁴⁸ of and in the progression⁴⁹ of the disease. Moreover, the only known mechanism to slow the progression of glaucomatous visual field degradation is to lower IOP.⁵⁰ There is little consensus, however, on what constitutes “elevated” IOP, and the threshold is likely patient-dependent. For example, one-third to one-half of people with glaucoma present with an IOP below 20 mmHg,^51,52 yet the incidence of glaucoma among people with ocular hypertension is only one percent per year.^53,54 Clearly, there are structural elements of the eye that allow some individuals to withstand elevated IOP without neuropathy, while others lose vision with relatively low IOP.

Ocular biomechanical parameters may mediate the disjointed relationship between IOP and glaucomatous optic neuropathy. In glaucoma, insult to the axons of retina ganglion cells as they exit the globe through the pores of the lamina cribrosa initiates a complex array of molecular pathways that leads to neuronal death.⁵⁵ Since the lamina cribrosa is thought to be the primary site of glaucoma pathology, its biomechanical parameters and those of the adjacent posterior sclera are likely a key variable in risk stratification for the disease.^56,57 Clinical assessment of the lamina cribrosa and of the posterior sclera is difficult, however, even with emerging technology such as spectral domain optical coherence tomography enhanced depth imaging.⁵⁸ Moreover, there is no commercially available software that can segment the lamina cribrosa for quantification and for comparison against a non-existent normative database. These limitations currently limit the utility of assessing the biomechanical parameters of the lamina cribrosa and of the posterior sclera in the clinic. However, both the cornea and the sclera are more clinically accessible than both the lamina cribrosa and the isolated posterior sclera, since the entire ocular shell is involved in the biomechanical response to a macro-indentation of the cornea over a short time scale with fluid displacement. This makes air puff induced deformation of the cornea appealing in current clinical devices to measure biomechanical parameters in the context of assessing glaucoma risk.

Central corneal thickness

The association between the structural and the biomechanical parameters of the cornea and the incidence of and the progression of glaucoma are well established. Interest in the role that central corneal thickness (CCT) plays in the incidence of glaucoma started in 2002 when the Ocular Hypertension Treatment Study first reported that a thin central cornea is an independent risk factor for the conversion of ocular hypertension to open-angle glaucoma.⁵⁹ Since then, other groups have replicated the importance of the CCT measurement when assessing an individual’s risk for developing open-angle glaucoma^60,61 or risk for progression of preexisting disease.^62,63 In contrast, population-based studies on patients from West Africa⁶⁴ and from East Asia⁶⁵ have found no difference in CCT between individuals with glaucoma and those without, calling into question the predictive role of CCT in the management of, or in the long-term prognosis of open-angle glaucoma.⁶⁶ Therefore, it has been proposed that CCT, a structural feature of the eye, may be a proxy for biomechanical parameters that may be more directly related to the pathophysiology of glaucoma.

Corneal hysteresis

One such parameter may be CH. If it is assumed the cornea’s ability to dampen energy is a representation of the entire corneoscleral shell’s ability to dampen energy,⁸ then CH may be an important variable to consider when assessing glaucoma risk. IOP fluctuates both on a longer time scale (e.g. diurnal IOP variation)^67,68 and on a shorter time scale (e.g. IOP pulsates with every heartbeat and spikes with eye movements and blinks).^69,70 Two large-scale randomized clinical trials, the Advanced Glaucoma Intervention Study⁷¹ and the Collaborative Initial Glaucoma Treatment Study,⁷² have found that high levels of IOP variation are associated with progressive visual field loss in patients with open-angle glaucoma, suggesting that the energy associated with even transient episodes of high IOP can expand the scleral canal and can subsequently increase strain on the lamina cribrosa.⁷³ Thus, the eye’s ability to dissipate the mechanical stresses associated with short- and long-term changes in IOP may be crucial to preserving the integrity of RGC axons as they pass through the lamina cribrosa.

In support of this theory, there is strong evidence of a link between low CH and glaucomatous optic neuropathy. In cross-sectional studies, many groups have reported lower CH in patients with open-angle glaucoma than in control patients (Table S1). In a special case, Anand and colleagues studied biomechanical parameters in patients with asymmetrical open-angle glaucoma.⁷⁴ They found not only that that eyes with more visual loss had lower CH than the fellow eyes, but also that low CH was more strongly associated with visual field loss than were corneal resistance factor, corneal-compensated IOP, Goldmann-correlated IOP, CCT, and refractive error. These results have since been replicated in patients with asymmetrical normal tension glaucoma, where low CH, but not low CCT, was associated with visual field loss in the worse eye.⁷⁵

Longitudinal evidence also supports the importance of CH as a variable in assessing glaucoma risk. Over the course of nearly four years, Susanna and colleagues found that the CH values of patients who converted from ocular hypertension to primary open-angle glaucoma (mean ± standard deviation: 9.5 ± 1.5 mmHg) were significantly lower than patients who did not convert (10.2 ± 2.0 mmHg), even when controlling for age, IOP, CCT, visual field status, and medical glaucoma therapy.⁷⁶ Each 1 mmHg reduction in CH was associated with a 21% increased risk of conversion. Interestingly, there were no differences between the groups in IOP or in CCT. CH does not lose clinical relevance after a glaucoma diagnosis. Zhang and colleagues reported that among 186 eyes with treated glaucoma, eyes with low CH (CH for all eyes at baseline was 9.2 ± 1.8 mmHg) were more likely to lose thickness of the retinal nerve fiber layer at a faster rate than eyes with high CH, both before and after controlling for the effects of age, race, IOP, and CCT.⁷⁷ If low CH is associated with rapid glaucomatous tissue thinning, then it is unsurprising that it is also associated with the progressive loss of visual field sensitivity. Both retrospective^78,79 and prospective^80,81 observational cohort studies have demonstrated that patients with low CH lose visual field sensitivity at a faster rate than those with high CH, even in patients with relatively well-controlled IOP.⁸²

Changes to corneal hysteresis

CH is not a static measurement because it is influenced by external factors that may change over time in patients with glaucoma. There is a consistently reported positive relationship between CCT and CH, with thick corneas tending to dissipate energy better than thin corneas.^78,80,83,84 The cornea thins slowly with age,⁸⁵ possibly contributing to a reported decrease in CH values over time;⁸⁶ however, this thinning likely has little influence on the progression of glaucoma.^87,88 Unlike CCT, IOP can change quickly over time. As previously introduced, there is a strong negative correlation between CH and IOP.^83,84,86 Eyes with high IOP do not dissipate energy well, an undesirable situation in patients with uncontrolled glaucoma.

The primary goal of glaucoma therapy is to lower IOP, but a secondary benefit is that glaucoma treatments also increase CH. Low CH values in patients with chronic angle-closure glaucoma partially recover after medical and surgical interventions.⁸⁹ Similar effects have been reported in patients who receive only medical therapy. In a 2012 retrospective study, Agarwal and colleagues found that CH increases in patients with open-angle glaucoma after 1.5 months of prostaglandin analogue therapy.⁹⁰ After adjusting for baseline IOP, patients with low baseline CH experienced a higher prostaglandin-mediated reduction in IOP than patients with relatively high baseline CH. This result was later replicated by another group in patients who had undergone selective laser trabeculoplasty after failing medical therapy.⁹¹ In a 2015 prospective study, Bolivar and colleagues also reported an increase in CH 6 months after the initiation of prostaglandin analogues to treat newly diagnosed open-angle glaucoma.⁹² Interestingly, the magnitude of the increase in CH was unrelated to the magnitude of IOP reduction or to the baseline CH value. This finding suggests that prostaglandin analogues, at least partially, increase CH through a mechanism that is independent from IOP reduction. Specifically, prostaglandin analogues bind to prostaglandin F receptors located in ciliary body, trabecular meshwork, episclera, sclera, and cornea.^93,94 F-receptor activation leads to enhanced matrix metalloproteinase activity,^95,96 which in turn causes extracellular matrix expansion and reduces collagen in the same structures.^97,98 These two morphological changes increase the transmissibility of the uveoscleral outflow pathway^99,100 and possibly enhance the eye’s ability to dissipate energy. Clinicians should not think of CH as a static contributor to glaucoma pathophysiology, but rather as dynamic risk factor that changes with the disease and with its treatment. Unlike CCT, CH should therefore be measured periodically in patients with glaucoma or in patients with suspected glaucoma and should be interpreted in the context of the IOP.

Stiffness of ocular tissues

The mechanism governing the association between CH in the anterior segment and loss of optic nerve structure and function in the posterior segment is unknown, but emerging evidence suggests that stiffness of ocular tissues makes a large contribution. For example, patients with open-angle glaucoma have stiffer corneas than healthy patients,^101–103 and Qassim and colleagues recently reported eyes of glaucoma suspects with both thin and stiff corneas are at high risk of damage to the optic nerve head and of visual field loss (Figure 11).¹⁰⁴ This finding is not universal, however, as several groups have reported either that patients with primary open-angle glaucoma have more deformable corneas than healthy control patients,¹⁰⁵ or that there is no difference in corneal stiffness between the two groups.^106–109 The possible globe-softening effects of topical prostaglandin analogues and/or the Corvis-ST-generated corneal stiffness parameters chosen to represent corneal stiffness may be the sources of these disparate reports of corneal stiffness in glaucoma.

Figure 11. Three-dimensional plots of central corneal thickness (CCT) and of corneal stiffness parameter (SP-A1) versus (A) structural progression and (B) functional progression in glaucoma suspects. Structural progression is quantified with Optical Coherence Tomography (OCT) retinal nerve fiber layer (RNFL) thinning and functional progression is quantified with visual fields. Reproduced Qassim et al.¹⁰⁴ Copyright (2021) with permission from Elsevier.

The cornea is not the structure that supports retinal ganglion cell axons as they exit the globe, so measuring its stiffness is likely a marker for the stiffness parameters of other ocular structures that contribute more directly to the pathophysiology of open-angle glaucoma. Wells and colleagues demonstrated a significant association between CH and optic nerve head compliance in patients with glaucoma but not in control patients.¹¹⁰ The authors interpreted these findings as suggesting that glaucoma alters the biomechanics of ocular tissues to become stiffer and less viscoelastic; however their cross-sectional study design did not allow them to determine whether a relatively stiff eye contributed to glaucoma pathophysiology or was a consequence of it. Similarly, Lanzagorta-Aresti and colleagues reported that treatment of patients with glaucoma and with ocular hypertension with a prostaglandin analogue allows the lamina cribrosa to move anteriorly to its original position and that this movement is significantly associated with CH when controlling for IOP.¹¹¹ They too suggested that their results support the idea that low CH is related to a stiff eye, reducing the capacity of the eye to dampen strain on the lamina cribrosa.

Scleral biomechanics

The sclera is the physical link between the cornea and the lamina cribrosa and is the main load-bearing structure of the eye,¹¹² prompting several groups to begin investigating the biomechanical parameters of the sclera in the context of glaucoma. Finite-element modelling of the sclera has suggested that it contributes to the deformation of the lamina cribrosa in response to elevated IOP.¹¹³ Scleral stiffness, in particular, is modelled to play a central role in maintaining the mechanical integrity of the lamina cribrosa and of the optic nerve head,¹¹² with a stiff sclera potentially conveying larger stresses onto the lamina cribrosa than a compliant sclera.^57,114 Animal studies differ on whether a stiff sclera contributes to^102,115 or is protective against¹¹⁶ open-angle glaucoma. Finite-element and animal modelling are likely limited in their ability to predict the exact role scleral stiffness plays in glaucoma pathophysiology in living human patients, however, prompting the need for human studies.

As previously introduced, clinicians and researchers were unable to measure scleral stiffness in-vivo until the introduction of the Corvis ST and the calculation of SP-HC. Since SP-HC is a relatively new parameter, the link, if there is one, between SP-HC and glaucoma risk remains unclear. Vinciguerra and colleagues recently published results suggesting that a low SP-HC value, which represents a compliant sclera, is associated with advanced visual field defects in patients with open-angle glaucoma.¹⁰⁸ This report must be interpreted with caution, however, because many patients with open-angle glaucoma in the study were being actively treated with prostaglandin analogues, which have been shown to decrease scleral stiffness.¹¹⁷ In the same paper that identified thin and stiff corneas as a risk factor for progression from being a glaucoma suspect to being a glaucoma patient, Qassim and colleagues did not find that that SP-HC was a risk factor for conversion.¹⁰⁴ Although this longitudinal study controlled for medical glaucoma therapy, it did not have healthy control cohort, so it remains uncertain how SP-HC differs between healthy patients, patients who are glaucoma suspects, and patients who have glaucoma. It is also unclear how SP-HC interacts with other variables, such as race, age, and CH, known to influence overall glaucoma risk.

In addition to the Corvis ST, the ORA may also provide valuable information on the biomechanical response of the sclera. Parameters beyond CH can be derived from the infrared waveform generated during a measurement.¹¹⁸ This waveform has two peaks positioned between a central trough (Figure 12). The first peak occurs when the cornea applanates as it deforms inward in response to an air puff, and the second peak occurs when the cornea applanates a second time as it recovers from a concave posture as the air puff magnitude decreases. The biomechanical properties of the sclera appear to govern the shape of the second peak of the waveform. The association between the second peak and scleral parameters was first demonstrated by a study conducted on 18 patients who had received a unilateral scleral buckle for the treatment of retinal detachment found that the cornea recovered from concavity to its natural convex shape faster in the buckled eye with the stiff sclera than in the fellow eye.⁷ This difference was detected in 11 of the 38 waveform-derived parameters developed by the manufacturer,¹¹⁸ 3 of which were related to the shape of the first and second peaks, and 8 of which were related to only the second peak. CH was also significantly lower in the operated eye (9.0 ± 1.8 mmHg) than in the fellow eye (10.1 ± 1.8 mmHg), despite no difference between the two in IOP measured with Goldmann Applanation Tonometry. If the Ocular Response Analyzer is able to assess the biomechanical response of the sclera, as this paper suggests, then it provides an opportunity to better understand how the sclera influences glaucoma pathophysiology in living human eyes.

Figure 12. Definitions of Waveform Parameters from the Ocular Response Analyzer, including (A) path 1 and path 2, indicated by the arrows around Peak 1 and Peak 2, respectively; p1area and p2area, indicated in gray for Peak 1 and Peak 2 respectively; and h1 and h2, the heights of the respective Peaks, with w1 and w2 indicating the width of the respective peaks. (B) Additional parameters include: uslope1 and dslope1, indicating the slopes of the rising and falling sides of Peak 1 with a 25% baseline; uslope11 and dslope11, indicating the same with a 50% baseline. The rising and falling slopes of Peak 2 are dslope2 and uslope2, respectively for a 25% baseline; and dslope21 and uslope21 indicate the same with a 50% baseline. The y-axes are number of photons aligned with the detector of the device, and the x-axes are time (ms). The green trace is air pressure created by the piston of the device. The red traces are the number of photons reaching the detector. Figure adapted from Luce and Taylor.¹¹⁸

There is emerging evidence that second-peak parameters may be associated with glaucoma risk. Aoki and colleagues recently investigated the relationship between waveform parameters and progressive visual field loss in 101 eyes with treated glaucoma.¹¹⁹ They found that CH (9.1 ± 1.1 mmHg in the cohort) and second-peak waveform parameters (e.g. path1, path2, and w2) were associated with visual field progression. The association was stronger for the waveform parameters than for CH, and there were no associations between CCT or IOP and visual field progression. These results should be interpreted carefully: all the patients in the study were on medical therapy, the sample size was relatively small, and there was no control group. Nevertheless, they represent in-vivo evidence that scleral biomechanical parameters affect glaucoma progression, and they may spur future studies that seek to harness the complete utility of the ORA as a tool to assess glaucoma risk.

Conclusion

Understanding of corneal and scleral biomechanics is far greater now than 20 years ago, with the advent of new clinical tools and subsequent strong interest by the scientific and clinical communities, leading to many published studies. However, there is still a great deal to study and learn in the applications addressed in this review, including keratoconus, laser refractive surgery, and glaucoma. It is also clear that our interpretation of biomechanical response should not be isolated to the specific region affected by an intervention but should recognize that the entire cornea responds even if there is only a regional intervention, and that the entire eye responds to corneal deformation with an air puff. Interpretation also depends on IOP and the air puff strategy. The two commercial devices currently available provide complimentary biomechanical parameters that evaluate distinct responses. New devices to assess spatially resolved biomechanical responses are under development and have the potential to improve the detection and management of ocular diseases, to create a new generation of clinical studies, and to inform more sophisticated computer simulations.

Disclosure statement

Phillip T. Yuhas has no competing interest to declare. Cynthia J. Roberts is a consultant for Oculus Optikgeräte GmbH and Ziemer Ophthalmic Systems AG. She received an honorarium from Heidelberg Engineering, Inc.