Femtosecond Laser-Assisted Keratoplasty: A Narrative Review of Technological Evolution, Clinical Evidence, Implementation and Future Directions

Abstract

Femtosecond laser-assisted keratoplasty (FLAK) enables programmable photodisruption, customized graft-host configurations, and image-guided corneal reconstruction. This narrative review synthesizes 107 scholarly publications spanning 2003–2026, identified through PubMed/MEDLINE and reference lists, alongside regulatory documentation. It examines mechanisms, technological development, clinical evidence, emerging approaches, and implementation. Evidence is interpreted primarily through final best-corrected visual acuity, graft survival, and reintervention, with incision precision treated as a surrogate. In penetrating keratoplasty (PK) and deep anterior lamellar keratoplasty (DALK), femtosecond laser (FSL) assistance may improve wound architecture and selected early outcomes, but durable patient-relevant superiority remains unproven. Randomized endothelial keratoplasty evidence identified poorer visual acuity and greater endothelial loss with early FSL-assisted donor preparation, despite reduced astigmatism. Selected Descemet membrane endothelial keratoplasty (DMEK) and reconstructive applications remain promising but insufficiently validated. Emerging approaches include stromal lenticule repair, additive reconstruction, limbal transplantation, and scaffold preparation. Their evaluation must separate procedural feasibility from comparative effectiveness. Postoperative healing, suture management, complications, and surgical experience can outweigh initial geometric precision. Historical economic evidence disfavors FSL-assisted endothelial graft preparation; contemporary cost-effectiveness remains uncertain. Selective implementation should consider clinical need, local expertise, shared infrastructure, and equitable access. Future evaluation should prioritize long-term patient outcomes, direct economic comparisons, and multicenter validation of lower-energy platforms, real-time imaging, artificial intelligence-assisted planning, and biomechanically informed incision design.

Introduction

Corneal transplantation remains essential for visual rehabilitation in eyes with irreversible corneal opacity or structurally significant corneal disease.¹ Over recent decades, the field has shifted from full-thickness replacement toward selective, lamellar, and indication-specific transplantation. Although penetrating keratoplasty (PK) was long the standard procedure for restoring corneal clarity,² deep anterior lamellar keratoplasty (DALK), Descemet stripping endothelial keratoplasty/Descemet stripping automated endothelial keratoplasty (DSEK/DSAEK), and Descemet membrane endothelial keratoplasty (DMEK) have expanded surgical options by replacing only the diseased layer while preserving uninvolved host tissue.^3–6 This transition reflects a central principle of modern corneal surgery: when the optical, biomechanical, and immunologic advantages of healthy host tissue can be retained, transplantation should be as anatomically selective as possible.

Despite this progress, conventional keratoplasty remains substantially dependent on surgeon-controlled tissue manipulation. In PK, vertical trephination produces a donor-host junction whose apposition and refractive effect are shaped by trephine alignment, graft-host sizing, tissue compression, and suture tension, potentially compromising wound biomechanics and inducing high or irregular postoperative astigmatism.⁷ In DALK, achieving a deep stromal plane adjacent to Descemet membrane remains technically demanding, with risks of microperforation and conversion to PK.⁸ In endothelial keratoplasty (EK), donor lenticule preparation, host-bed configuration, graft manipulation, attachment, and interface quality similarly influence visual recovery. These limitations have driven continued interest in technologies that enhance the precision, reproducibility, and geometric consistency of tissue preparation and donor-host apposition.

Femtosecond laser (FSL) technology is designed to address these unmet surgical needs by delivering ultrashort, tightly focused pulses that induce localized photodisruption, enabling planar, arcuate, and interlocking corneal incisions with limited collateral thermal injury.⁹ Initially transformative in refractive flap creation, FSL platforms have since been extended to anterior-segment procedures, including arcuate keratotomy, intrastromal corneal ring implantation, cataract surgery, and corneal transplantation.^10,11 In keratoplasty, customized wound architecture may improve donor-recipient geometric congruence, support reproducible dissection, and promote procedural standardization.^12–14 However, clinical translation remains variable and is influenced by laser energy and repetition rate, spot and line separation, patient-interface design, intraoperative imaging, corneal opacity and thickness, incision geometry, tissue-handling protocols, and the surgeon’s learning curve. Clinical reports of shaped FSL keratoplasty appeared in 2007, placing its development over nearly two decades.^15,16 This history makes durable visual and graft-survival outcomes particularly important when judging technical advances.

This review critically evaluates femtosecond laser-assisted keratoplasty (FLAK) across its distinct surgical applications. It summarizes FSL-tissue interactions and advances in laser platforms, patient interfaces, incision design, and image guidance, then assesses evidence across established and emerging corneal transplantation techniques with emphasis on clinically meaningful outcomes. It also addresses implementation barriers, cost, access, and future directions. The central question is not whether FLAK should replace conventional keratoplasty, but where its precision can improve clinical outcomes, reduce downstream risk, or enable procedures that are difficult to reproduce manually. The DALK–PK experience illustrates that an anatomical advantage may not produce better final vision or graft survival; this distinction also frames our assessment of FLAK.^17,18 ⁠

Methods and Scope of the Narrative Review

This narrative review draws on PubMed/MEDLINE and reference-list searching, with targeted source verification for this revision on 1 March 2026. Search concepts included femtosecond laser-assisted keratoplasty, corneal transplantation, penetrating keratoplasty, deep anterior lamellar keratoplasty, endothelial keratoplasty, intraoperative optical coherence tomography, patient interfaces, and corneal biomechanics. Boolean combinations are expressed with explicit grouping, for example: (“femtosecond laser”) AND (“endothelial keratoplasty” OR “DMEK” OR “DSAEK”). English-language literature and selected relevant non-English reports were considered. The cited scholarly literature spans 2003–2026; this publication span is distinct from a search execution date. The search supported thematic selection and does not constitute systematic evidence identification; completeness and freedom from selection bias cannot be assumed.

Selection was purposive and focused on clinical comparisons, experimental mechanisms, technical feasibility, and implementation relevant to corneal transplantation. Reviews provided context, while primary studies were prioritized for empirical claims. The final bibliography contains 107 scholarly publications and 7 regulatory or manufacturer documents; these are cited sources, not a systematic-review inclusion yield. No protocol registration, independent duplicate screening, or formal risk-of-bias assessment is claimed. Randomized comparisons inform treatment effects; observational studies inform associations and generalizability; laboratory studies and case series support mechanism and feasibility. Database and language restrictions, selective inclusion, and the absence of formal quality appraisal limit completeness and certainty. Figure 1 presents the review’s scope and interpretive framework.

Figure 1.

Scope map of this narrative review. The non-sequential arrangement relates literature sources, evidence appraisal, clinical applications, and outcome priorities. It does not represent systematic screening or a study-selection yield. FALK, femtosecond laser-assisted keratoplasty. This conceptual illustration was generated with assistance from ChatGPT (GPT-5.4, OpenAI) and reviewed and verified by the authors for accuracy.

Principles of FLAK

Fundamental Photophysical Mechanisms

The value of FSL technology extends beyond automating tissue cutting in keratoplasty; it transforms corneal separation into a preplanned laser-tissue interaction with defined location and geometry. Its ultrashort pulse duration, high peak power, and tight spatial focusing permit energy delivery within transparent ocular tissue without a mechanical blade.⁹ Unlike continuous-wave or longer-pulse lasers dominated by photothermal effects, FSL incisions are generated primarily through nonlinear, laser-induced optical breakdown and photodisruption.^9,19

Most ophthalmic FSL systems use near-infrared pulses, typically around 1050–1060 nm; 1053 nm is often cited because it traverses transparent ocular media with limited attenuation.²⁰ When focused within the corneal stroma, the pulse confines photons to a micrometer-scale focal volume, where photon density may exceed the thresholds for multiphoton and avalanche ionization, generating plasma composed of free electrons and ionized molecules.^9,19 Rapid plasma expansion produces shock waves and cavitation bubbles, whose expansion, interaction, and collapse mechanically cleave adjacent stromal lamellae along the intended focal plane.⁹

Photodisruption begins within the focal volume, but its mechanical effects extend beyond it through plasma expansion, shock waves, and cavitation. Ultrashort energy deposition limits thermal diffusion during the pulse; it does not eliminate collateral injury.^9,19 In keratoplasty, the residual stroma, Descemet membrane, and endothelium remain susceptible to mechanical and biological damage. High-speed imaging shows that bubble dynamics, pulse overlap, and tissue properties govern lamellar separation.⁹ FSL cutting is therefore described here as localized photodisruption, with limited thermal spread under appropriate conditions, instead of tissue evaporation or bulk thermal ablation.

Mechanistic Basis of FSL Corneal Cutting

Corneal FSL surgery converts focal photodisruption into programmable three-dimensional incision patterns. The surgeon specifies the target depth, diameter, side-cut angle, lamellar plane, and wound profile, after which the laser delivers sequential pulses along the planned intrastromal trajectory.^14,21,22 With adequate visualization and docking, image-guided planning can adapt pulse placement to individual corneal contour and pachymetry, enabling micrometer-scale tissue separation.^14,22

The incision is generated not by a continuous beam but by the coalescence of densely spaced photodisruption events. Modern systems operate at kilohertz-to-megahertz repetition rates, while pulse energy, spot and line separation, stromal hydration, and opacity determine whether cavitation events create a smooth cleavage plane or leave residual tissue bridges.^9,23,24 With adequate pulse placement and overlap, microcavitation events merge into a continuous lamellar or side-cut plane, supporting reproducible tissue separation.⁹

This mechanism underscores that FLAK performance cannot be predicted from FSL use alone. Excessive pulse energy may enlarge cavitation bubbles and roughen the stromal interface, whereas insufficient energy or inadequate spot overlap can cause incomplete cleavage requiring additional manual dissection.²³ Patient-interface deformation, decentration, corneal scarring, edema, and planned cutting depth may further alter energy delivery and cutting accuracy. Thus, laser parameters, corneal optical quality, docking mechanics, and surgical design jointly determine the final graft-host interface.

Translational Advantages Arising from These Principles

The principal technical attraction of FLAK is that corneal incision architecture becomes a preoperative design parameter. Computer-controlled photodisruption can reproducibly generate the planned diameter, depth, side-cut angle, and lamellar contour according to programmed coordinates, although the achieved geometry requires verification. In principle, this may enhance donor-recipient congruence, reduce edge decentration or obliquity, and provide a more predictable structural basis for wound healing and visual rehabilitation.¹⁴

A second advantage is geometric design flexibility. Conventional trephination usually creates a vertical cylindrical wound, whereas FSL platforms can generate stepped, mushroom, top-hat, zig-zag, and other interlocking profiles. These configurations increase wound surface area, alter force distribution across the graft-host junction, and allow incision geometry to be selected according to the surgical goal, including endothelial preservation, anterior stromal replacement, biomechanical support, or complex reconstructive matching.^15,16,25

These mechanistic advantages may explain early clinical signals reported in comparative studies, including reduced early postoperative astigmatism, earlier suture removal, and faster initial visual recovery. Chamberlain et al reported less early corneal astigmatism and earlier suture removal after FLAK than after conventional PK.²⁶ Matched cuts can simplify selected surgical steps, but shorter open-sky time and reduced manipulation should be measured directly; they cannot be inferred from incision shape alone.

The clinical benefit of this control remains conditional. FLAK improves incision controllability, but final visual function depends on various factors mentioned earlier. The fundamental principles of FLAK therefore support its use most strongly when controlled three-dimensional geometry can solve a specific surgical problem. Figure 2 summarizes the basic workflow and mechanistic rationale for FLAK.

Figure 2.

Principle of FLAK. (A) Photodisruption produces focal plasma and cavitation bubbles that separate corneal tissue; the black arrow identifies a cavitation bubble at the laser focus. (B) Complementary donor and recipient incisions may improve graft–host apposition. The large curved blue arrows indicate the correspondence between donor and recipient incision profiles, and the small blue arrows highlight the complementary interlocking edges. (C) Graft placement, suturing, and subsequent healing determine the clinical course. The downward blue arrows indicate graft placement and progression to the sutured state, while the curved blue arrows in the lower image indicate the graft–host interfaces where healing occurs. The illustrated geometric and healing advantages are potential benefits; they do not establish freedom from collateral injury or comparative clinical superiority. This conceptual illustration was generated with assistance from ChatGPT (GPT-5.4, OpenAI) and reviewed and verified by the authors for accuracy.^9,14,22

Current Developments in Femtosecond Laser Platforms for Keratoplasty

Several FSL platforms have been adapted for corneal transplantation. Their performance depends on pulse delivery, patient-interface geometry, tissue properties, and image guidance.^14,21–24 Table 1 separates documented indications from clinical feasibility. FDA records explicitly include lamellar and penetrating keratoplasty for the FEMTO LDV Z8 and describe keratoplasty uses for IntraLase and VisuMax; VICTUS documentation uses broader corneal-incision language.^27–30 These records do not establish superiority or authorize every proposed incision pattern. A historical manufacturer statement reports CE marking for VICTUS penetrating keratoplasty, but does not verify every current software module.³¹ Platform names, research use, and general CE marks cannot substitute for jurisdiction-, model-, and module-specific labeling. The cited LenSx and CATALYS records describe cataract-related indications, not keratoplasty.^32,33

Table 1.

Documented Regulatory Indications and Limits of Interpretation

Platform/Model	Documented FDA Indication	CE Evidence for Keratoplasty	Interpretation for Incision Patterns
IntraLase FS/iFS	FDA device record describes lamellar and penetrating IntraLase-enabled keratoplasty and corneal harvesting.28⁠	No keratoplasty-specific CE certificate or current module labeling verified in the reviewed sources.	Broad keratoplasty wording does not verify every customized geometry. The historical FDA record is used for its indication text, not as evidence of current product availability.
VisuMax Laser Keratome	K100253 explicitly includes lamellar keratoplasty and penetrating keratoplasty/corneal harvesting.29⁠	No keratoplasty-specific CE certificate or current module labeling verified.	The cited clearance concerns the named model; it cannot be transferred automatically to VisuMax 800 or every incision profile.
LenSx Laser System	The reviewed FDA device record describes cataract-related capsulotomy, lens fragmentation, and corneal incisions.32⁠	No keratoplasty-specific CE documentation verified.	Keratoplasty is outside the indication documented in this record (off-label relative to that labeling). Current model-specific labeling requires separate verification.
CATALYS Precision Laser System	K141079 covers cataract surgery, including corneal arc cuts/incisions, capsulotomy, and lens fragmentation.33⁠	No keratoplasty-specific CE documentation verified.	Keratoplasty is outside the cataract indication in this clearance (off-label relative to that labeling). Arc-incision clearance is not keratoplasty clearance.
VICTUS Femtosecond Laser Platform	K200724 describes lamellar corneal resection and corneal incisions for cataract surgery or other ophthalmic treatment.30⁠	A 2014 manufacturer announcement reports CE marking for penetrating keratoplasty.31⁠	The broad FDA wording does not identify each keratoplasty pattern. The historical CE announcement is not a current certificate for every software module.
FEMTO LDV Z8	K213559 explicitly includes lamellar and penetrating keratoplasty; described modules include DALK and liquid-interface PK.27⁠	No keratoplasty-specific CE certificate or current module labeling verified.	Documented keratoplasty indication does not establish clinical superiority or clearance of every investigational pattern.
SCHWIND ATOS	No keratoplasty-specific FDA clearance was verified in the reviewed sources.	No keratoplasty-specific CE documentation verified.	Status remains undetermined here. Absence of a verified document is not proof of non-clearance and cannot establish an off-label classification.

Notes: This table distinguishes evidence in the cited records from current model-, software-, jurisdiction-, and pattern-specific authorization; it is not a comprehensive regulatory register. “Off-label” is limited to the particular cited labeling. Current instructions for use and CE certificates should be checked before clinical implementation.

Abbreviations: CE, Conformité Européenne; DALK, deep anterior lamellar keratoplasty; FDA, US Food and Drug Administration; PK, penetrating keratoplasty.

Differences in Laser Pulse Energy and Repetition-Rate Architecture

Current FSL platforms differ substantially in laser parameters that collectively influence cavitation bubble size, pulse overlap, residual tissue bridges, interface regularity, and the need for manual separation after laser delivery. Earlier microjoule-range IntraLase systems (Abbott Medical Optics/Johnson and Johnson Vision, CA, USA) operated with relatively higher pulse energies, enabling efficient photodisruption but potentially enlarging cavitation bubbles and compromising stromal bed smoothness. Serrao et al²⁴ showed that higher energies (0.65–0.75 μJ) with wider spot spacing and lower pulse overlap (5–6 μm) produced rougher stromal beds, whereas lower energies (≤0.55 μJ) with narrower spacing and greater overlap (4 μm) yielded smoother interfaces. These experiments concerned LASIK stromal beds and provide mechanistic evidence, not comparative keratoplasty outcomes.

A different trajectory is exemplified by the Femto LDV Z8 platform (Ziemer Ophthalmic Systems AG, Canton of Bern, Switzerland), which uses a low-energy, megahertz pulse-delivery strategy. High-density placement of low-energy pulses is intended to generate smaller cavitation bubbles and a more continuous cleavage plane. Ex vivo corneal transplantation experiments support evaluation of this architecture, but do not establish improved vision or graft survival.²¹ Thus, FSL should not be regarded as a uniform intervention, but as a family of devices whose tissue effects depend on pulse architecture, optical clarity, stromal hydration, and intended dissection depth. Figure 3 illustrates how laser parameters influence FSL cutting efficiency and quality.

Figure 3.

Interacting determinants of corneal photodisruption. The central panel illustrates focal plasma, cavitation bubbles, and intrastromal separation beneath an intact surface. (A) Pulse energy must exceed the local breakdown threshold; excess energy can enlarge bubbles and increase collateral injury. (B) Repetition rate and scan speed jointly determine pulse placement. (C) Spot spacing influences overlap, tissue bridges, and bubble interactions. (D) Focal spot size and tissue clarity affect optical breakdown. The schematic depicts tissue separation rather than surface evaporation or bulk material removal; no universal optimum is implied.^9,19,23,24 This conceptual illustration was generated with assistance from ChatGPT (GPT-5.4, OpenAI) and reviewed and verified by the authors for accuracy.

Differences in Patient-Interface Design

The patient interface is another key determinant of FLAK performance, influencing eye stabilization, corneal deformation, intraocular pressure, and the optical path of laser delivery. Applanation-based systems, such as the IntraLase iFS 150, use a flat contact lens to stabilize and flatten the cornea during cutting. Menabuoni et al³⁴ demonstrated that this approach can generate complex donor and recipient wound geometries, including an anvil-profile penetrating keratoplasty.

Mechanical flattening can alter corneal geometry and the relationship between planned and achieved incisions. Primary OCT observations and experimental comparisons of flat and liquid interfaces support this concern.^14,21 In porcine-eye experiments, Liu et al²¹ found lower intraoperative pressure elevation with a liquid interface and less endothelial damage at higher energy settings. Donner et al³⁵ compared donor trephination using liquid and applanated interfaces, addressing donor preparation without measuring postoperative vision. Clinical benefit remains uncertain: Klimek et al‘s retrospective study of 121 eyes found no adjusted difference in postoperative astigmatism between liquid-interface femtosecond laser-assisted PK (FL-PK) and vacuum-trephine PK (6.7 ± 4.2 D vs 6.7 ± 3.8 D; adjusted P = 0.42).³⁶ Reduced docking deformation is consequently a plausible technical advantage, not evidence of superior visual rehabilitation.

Differences in Incision-Design Patterns

The ability to generate three-dimensional wound profiles is a distinctive advantage of FSL technology in keratoplasty. These profiles are intended to increase graft-host contact area, improve apposition, and redistribute tensile forces. The top-hat configuration is particularly notable because its posterior flange expands the internal wound surface area. Steinert et al¹⁵ reported greater biomechanical stability with this design than with conventional trephination. By resisting intraocular pressure-induced wound separation while permitting a larger posterior endothelial-bearing surface, the posterior flange provides a mechanistic rationale for top-hat PK. Experimental wound-strength studies demonstrate configuration-dependent mechanical effects; these surrogate results cannot establish long-term graft survival.²⁵

Additional incision geometries have been developed to enlarge wound surface area, enhance interface stability, and support disease-specific reconstruction. Menabuoni et al³⁴ created an anvil-shaped incision with the IntraLase iFS 150, while Farid et al¹⁶ reported early PK experience with a zig-zag incision designed to improve wound stability and healing predictability. More recently, complex lamellar surgery has incorporated integrated multi-step designs. Abusayf et al²² used the Ziemer LDV Z8 to combine a mushroom-shaped wound profile with an intrastromal tunnel for big-bubble formation in a one-step technique. Palevski et al³⁷ described a half-top-hat approach for keratoglobus, adding the top-hat component only to the donor cornea to reinforce an extremely thin recipient periphery. These examples demonstrate the feasibility of customized geometry in anatomically challenging eyes; comparative clinical benefit remains uncertain.

Clinical benefits of complex incision architecture remain uncertain. Baqalaqil et al pooled nine retrospective studies and reported favorable best-corrected visual acuity (BCVA), refractive, and endothelial associations.³⁸ Selection bias, overlapping sources of confounding, variable suture protocols, and differences in follow-up remain after pooling. A larger pooled sample does not confer randomized certainty. These limitations apply to favorable and unfavorable observational results alike. Figure 4 illustrates a restricted set of geometries supported by the cited primary literature; the drawings indicate structural concepts, not a ranking of clinical effectiveness.^15,16,25,34

Figure 4.

Representative incision configurations used in FLAK (taking PK as an example): (A) conventional incision; (B) mortise-and-tenon incision; (C) zigzag incision; (D) mushroom incision; (E) top-hat incision; (F) anvil incision; (G) dovetail incision; and (H) Christmas tree incision. These laser-created profiles are designed to increase graft–host contact area, improve wound apposition, and enhance biomechanical stability compared with conventional straight incisions. Precise and customizable interlocking wound patterns may facilitate better graft alignment, more stable healing, and reduced postoperative astigmatism. This conceptual illustration was generated with assistance from ChatGPT (GPT-5.4, OpenAI) and reviewed and verified by the authors for accuracy.

Image Guidance and the Transition Toward Feedback-Controlled Corneal Reconstruction

The incorporation of image guidance marks a recent and conceptually important stage in FSL platform evolution. Earlier workflows with systems such as the IntraLase iFS 150 and WaveLight FS 200 (WaveLight GmbH/Alcon Laboratories, Inc., TX, USA) primarily relied on preoperative measurements, surface landmarks, and programmed cutting patterns.³⁹ Although these systems can reproduce planned incisions with high consistency, they do not provide real-time optical coherence tomography (OCT)-based verification of tissue response, residual stromal thickness, scar-Descemet membrane relationships, or whether photodisruption occurs at the intended plane. Technical descriptions of the WaveLight FS 200, for example, emphasize rapid flap creation and customizable geometry instead of integrated real-time intraoperative OCT (iOCT) guidance.³⁹

This distinction is important because many corneas requiring transplantation have optical or biomechanical abnormalities that limit the reliability of static planning nomograms.^14,21,22 Keratoplasty-related FSL platforms are therefore evolving toward lower-distortion coupling, denser and more predictable pulse placement, and intraoperative imaging feedback.²² In this paradigm, image guidance does more than document the procedure; it may refine centration, verify depth, identify residual or incomplete cleavage planes, and facilitate safer transitions between laser and manual steps.^14,22

Overall, FSL platform development in corneal transplantation is shifting from automated cutting toward precision-guided corneal reconstruction. The key trend is not the superiority of any single commercial platform, but the integration of optimized laser delivery, reduced collateral tissue effects, lower-distortion interfaces, individualized three-dimensional wound architecture, and intraoperative verification.¹⁴ Whether this convergence makes FLAK a consistently superior and scalable strategy will depend on linking engineering advances to well-defined indications, standardized reporting, and meaningful patient-centered outcomes.

Advances in the Application of Femtosecond Lasers Across Common Keratoplasty Techniques

The clinical relevance of FLAK depends strongly on the specific transplantation procedure.⁴⁰ In PK, FSL is used mainly to redesign the full-thickness graft-host junction. In DALK, its potential value lies in improving the predictability of deep stromal dissection and wound configuration while preserving the host endothelium.^41,42 In EK, the key issue is whether laser-based preparation can enhance donor or recipient tissue precision without inducing interface irregularity or endothelial trauma.⁴³ Accordingly, current evidence should not be interpreted as a simple comparison between “laser” and “manual” keratoplasty, but in relation to the surgical layer replaced, the tissue at risk, and the procedural step targeted by laser assistance.

Femtosecond Laser-Assisted Penetrating Keratoplasty

PK remains indispensable when disease affects the full corneal thickness, when stromal scarring coexists with endothelial dysfunction, or when lamellar surgery is unsuitable. Conventional PK relies on mechanical trephination, typically creating a vertical cylindrical graft-host junction.¹⁴ Although robust and widely available, its final wound profile is influenced by trephine alignment, donor-recipient sizing, tissue compression, manual handling, wound healing, and suture tension, which may cause graft-host mismatch, delayed wound stability, and regular or irregular postoperative astigmatism. Femtosecond laser-assisted PK (FL-PK) was developed to overcome these limitations by generating reproducible three-dimensional wound profiles that increase appositional surface area and may distribute mechanical stress more favorably across the graft-host interface.²⁵

Early attempts to improve the PK wound did not focus solely on femtosecond cutting. Pini et al⁴⁴ proposed laser welding as an adjunct to, and potentially a partial substitute for, conventional suturing in PK. Their approach used indocyanine green staining and 810-nm diode laser irradiation to generate selective thermal bonding of corneal tissue. Buzzonetti et al⁴⁵ subsequently reported early pediatric experience with FSL combined with laser welding in PK and cataract surgery, suggesting that laser-mediated wound stabilization might be useful in eyes in which conventional suture-related problems are particularly consequential. These studies are best viewed as proof-of-concept work demonstrating the feasibility of laser-mediated wound stabilization.

Comparative studies have examined whether customized wound architecture translates into clinical benefit. Baqalaqil et al‘s meta-analysis of nine retrospective studies reported favorable pooled associations for BCVA, astigmatism, and endothelial preservation.³⁸ These estimates remain vulnerable to indication, baseline prognosis, surgeon experience, and postoperative management differences; pooling cannot remove unmeasured confounding. In contrast, Klimek et al‘s single-center retrospective comparison of 121 eyes found no adjusted postoperative astigmatism advantage for liquid-interface FL-PK.³⁶ The two reports therefore provide differing observational findings, not a randomized result opposed to observational evidence. Moreover, a significant within-group visual change in one arm but not the other is not evidence of a significant between-group treatment effect. Earlier synthesis by Peng et al also found that favorable early signals did not establish superior BCVA at 12 months.⁴⁶ Variation in wound configuration, graft sizing, suture removal, and follow-up can explain heterogeneity but cannot be used to dismiss null findings. The evidence supports possible early benefits in selected settings while leaving durable visual superiority uncertain. Future comparisons should prespecify final corrected vision and graft survival, measure postoperative interventions, and standardize assessment relative to suture removal.

Endothelial preservation and graft survival remain more difficult to interpret. Available comparative evidence does not support the assertion that FL-PK consistently causes less endothelial cell loss than conventional PK. In the meta-analysis by Baqalaqil et al,³⁸ FL-PK was associated with better postoperative endothelial cell density preservation than conventional PK, although heterogeneity and the predominance of non-randomized studies limit the strength of this conclusion. Individual observational studies also vary: Kamiya et al⁴⁷ reported no significant difference in endothelial cell density change between FL-PK and conventional PK, whereas top-hat or mushroom femtosecond-assisted configurations showed lower or numerically lower endothelial cell loss than conventional PK in selected cohorts.^48,49 These observations do not establish non-inferiority, which requires an appropriate design and a prespecified margin. A protective effect is mechanistically plausible because laser-assisted trephination may improve graft-host fit, reduce manual rotational trauma, and allow larger posterior donor profiles in some designs. Nevertheless, endothelial cell density alone is insufficient to define the clinical value of FL-PK. Long-term graft clarity also depends on indication, donor quality, wound configuration, suture management, immune rejection, intraocular pressure, surgical complexity, and the learning curve.

The technical rationale for FL-PK may be strongest when wound geometry itself is expected to change the surgical problem: pediatric keratoplasty, traumatic or highly irregular host beds, ectatic disease requiring biomechanically favorable apposition, and eyes in which earlier wound stability or suture removal would be clinically meaningful. For routine optical PK in otherwise anatomically regular eyes, current evidence does not yet support a universal conclusion that FL-PK is superior to well-performed conventional trephination. Future trials should stratify by incision geometry, interface type, suture protocol, indication, and endothelial risk to determine where femtosecond precision produces a durable clinical advantage beyond an early technical benefit.

Femtosecond Laser-Assisted Deep Anterior Lamellar Keratoplasty

DALK is conceptually attractive because it replaces diseased corneal stroma while preserving the recipient Descemet membrane and endothelium, thereby largely avoiding endothelial graft rejection. Its principal limitation is technical: the surgeon must dissect sufficiently deep to remove diseased stroma while avoiding Descemet membrane perforation.⁵⁰ In keratoconus, anterior stromal scarring, and other disorders with preserved endothelium, this step may determine whether lamellar surgery is completed or converted to PK.⁵⁰ Thus, in DALK, FSL assistance is not merely a means of circular trephination, but a strategy to standardize deep stromal geometry, guide the dissection plane, and facilitate big-bubble formation.⁵¹

Du et al pooled nine non-randomized comparative studies of FSL-assisted DALK (FL-DALK) and manual DALK in keratoconus.¹³ They reported better corrected distance visual acuity at 1–6 months and fewer intraoperative Descemet membrane perforations (OR = 0.53). These associations are clinically relevant but remain susceptible to case selection, differences in technique and surgeon experience, and heterogeneous follow-up. They do not establish a durable visual or graft-survival benefit and should be interpreted below the randomized evidence.

In the FEDS randomized trial, Sorkin et al studied 100 eyes of 97 participants with keratoconus or corneal scarring and assessed outcomes through 15 months.⁵² FL-DALK did not show a significant advantage in the primary astigmatism outcomes or convincing superiority across secondary functional and anatomical outcomes. Follow-up losses and imprecision limit exclusion of modest effects, but the trial does not support routine adoption to improve final vision. This randomized comparison carries greater weight for treatment effects than pooled non-randomized early benefits; absence of statistical superiority is not proof of equivalence.

Technique-specific studies support feasibility. Alió et al compared postoperative healing patterns after laser-assisted and manual DALK, providing direct evidence about wound morphology without establishing superior long-term patient outcomes.⁵³ Abusayf et al described iOCT-guided creation of mushroom-shaped donor-recipient junctions and intrastromal tunnels in one laser session; big-bubble separation was achieved in all 10 clinical eyes.²² Villarrubia et al also reported mushroom-profile FL-DALK experience.⁵⁴ These small series justify further evaluation of reproducibility, conversion risk, and learning curves. Successful dissection and favorable wound images alone cannot demonstrate comparative safety or visual superiority.

FL-DALK may help standardize difficult lamellar steps, although improved final BCVA has not been demonstrated in randomized comparison. Potential applications include steep ectatic corneas, scars with preserved endothelium, and cases requiring controlled residual stromal thickness. These are proposed selection criteria, not validated subgroup benefits. The remaining questions are whether these technical advantages reduce conversion to PK, shorten the learning curve, improve patient-reported visual function, and maintain endothelial safety over long follow-up. These questions require adequately powered, procedure-specific trials with standardized indications and incision profiles.

Femtosecond Laser-Assisted Endothelial Keratoplasty

EK has transformed the management of endothelial failure by selectively replacing dysfunctional endothelium and adjacent posterior tissue while preserving most host cornea. DSEK/DSAEK and DMEK are now preferred alternatives to PK for Fuchs endothelial corneal dystrophy, pseudophakic bullous keratopathy, and other forms of isolated endothelial decompensation.^55,56 In this context, the rationale for FSL use is narrower and more demanding than in PK or DALK. Laser assistance must improve donor preparation, recipient-bed configuration, descemetorhexis, or targeted posterior stromal removal without increasing interface roughness, posterior surface irregularity, graft detachment, or endothelial cell loss.⁵⁷

Randomized clinical evidence already identified important limitations of early FSL-assisted EK. Cheng et al‘s Dutch multicenter trial randomized 80 eyes to FSL-assisted DSEK (FL-DSEK; termed FLEK in the original report) or PK.⁵⁸ At 12 months, postoperative corrected acuity was poorer after FL-DSEK (20/70 vs 20/44), endothelial cell loss was greater (65% vs 23%), and graft dislocation occurred in 27.8% of FL-DSEK eyes, despite less refractive astigmatism. The trial compared complete operations, not isolated cutting methods; insertion trauma and graft preparation both matter. Its optical follow-up and subsequent interface studies support posterior interface quality as an explanation, without proving a single causal mechanism.^59,60 These findings help explain limited adoption of early laser-prepared posterior grafts, but do not establish why all later development slowed. Subsequent in vitro work by Titiyal et al found rougher interfaces and greater endothelial damage with FSL than microkeratome preparation (29.49 ± 6.91% vs 13.28 ± 3.62%).⁶¹

A more targeted direction is the use of FSL as a reconstructive tool instead of as a routine substitute for microkeratome donor preparation. Lim et al⁶² introduced intraoperative OCT-guided femtosecond laser-assisted DMEK (FL-DMEK) for eyes with deep residual stromal scarring or endothelial decompensation after previous DALK. In this approach, real-time iOCT guidance was used to plan an endothelial-side vertical cut with depths ranging from 96 to 329 μm, while a non-applanating liquid interface minimized tissue distortion. The scarred posterior tissue could then be selectively removed before DMEK grafting. Thus, FSL may be more valuable in complex EK when it enables selective removal of abnormal posterior stromal tissue that would otherwise be difficult to delineate manually.

Clinical comparisons in DSAEK further reinforce the need for caution. Vetter et al⁶⁰ compared curved-interface FSL-assisted DSAEK with microkeratome-assisted DSAEK and found worse postoperative BCVA in the FSL group (0.48 ± 0.20 vs 0.33 ± 0.11 logMAR, P = 0.038). The FSL group also showed greater posterior corneal surface irregularity relative to an ideal paraboloid, with a significantly higher root mean square error (9.9 ± 2.2 μm vs 5.7 ± 2.2 μm, P < 0.001). These data support the view that posterior interface quality is a central determinant of visual recovery after EK while their observational design limits causal attribution to the laser alone.

In selected DMEK settings, FSL-enabled steps may have a more favorable risk-benefit profile. Einan-Lifshitz et al⁶³ and Sorkin et al⁶⁴ reported that femtosecond-enabled descemetorhexis or FL-DMEK was associated with lower graft detachment rates, fewer rebubbling procedures, and lower endothelial cell loss than manual DMEK in their comparative series. Although final BCVA improvement was generally similar between groups, improved graft attachment and reduced postoperative intervention are clinically meaningful outcomes in DMEK, where detachment and rebubbling are common drivers of postoperative burden. However, these non-randomized comparisons are susceptible to temporal and learning-curve effects and do not establish long-term graft-survival superiority.

In summary, FSL use in EK should be regarded as indication- and step-specific. For routine DSAEK donor lenticule preparation, current evidence shows no consistent benefit and raises concerns about posterior surface roughness and endothelial injury. In DMEK, especially in complex post-DALK or reconstructive eyes, FSL may facilitate descemetorhexis, posterior stromal excision, and graft–host apposition. Table 2 summarizes selected FLAK studies published in 2021–2025; the earlier randomized EK trial is discussed above.^{13,21,22,37,38,52,65–76}

Table 2.

Selected FLAK Studies Published in 2021–2025

Study/Year	Design and Sample	Population/Material	Procedure and Purpose	Findings and Evidential Limits
Gogri et al 202165	Ex vivo comparative experiment; 20 human donor corneas	Donor corneoscleral tissues unsuitable for transplantation	DALK (preparation/cutting only) — Assess whether an FSL-created stromal channel facilitates BB formation during DALK.	A laser-created stromal channel facilitated type 1 big-bubble formation in donor tissue. This does not establish a clinical perforation or visual benefit.
Pedrotti et al⁠ 202166	Ex vivo comparative experiment; 20 human donor corneas	Human donor corneas unsuitable for transplantation	DALK (preparation/cutting only) — Determine whether the sequence of FSL-created tunnel and lamellar incisions influences type 1 BB formation.	Creating the tunnel before the lamellar incision favored type 1 big-bubble formation. The experiment does not establish the optimal clinical sequence.
Liu et al⁠ 202221	Animal-tissue experiment; 30 porcine eyes	Fresh porcine eyes	PK (preparation/cutting only) — Compare liquid-interface and flat-interface Ziemer Z8 FL-PK under different energy settings.	The liquid interface reduced pressure elevation and, at higher energy settings, endothelial denudation. These are experimental surrogate outcomes.
Farid et al⁠ 202267	Ex vivo technique study; 13 human donor corneas	Human donor corneas mounted on an artificial anterior chamber	DALK (preparation/cutting only) — Develop an FSL technique for creating a smooth deep posterior lamellar dissection for DALK.	Smooth deep posterior dissection was feasible in donor tissue. Clinical safety, conversion risk, and final visual outcomes were not established.
Gerten et al⁠ 202268	Uncontrolled clinical series; 10 eyes of 10 patients	Eyes with stromal disease requiring keratoplasty	DALK — Report the feasibility of a novel CCF-DALK technique intended to reduce the technical difficulty of DALK.	Complete Descemet membrane baring and limited endothelial loss were reported at 6 months. A small uncontrolled series cannot establish comparative benefit.
Feldhaus et al⁠ 202269	Ex vivo feasibility study; 4 human donor corneas	Human ex vivo corneal tissue	DMEK (preparation/cutting only) — Evaluate the precision and safety of FSL-assisted descemetorhexis and corneal wound healing in human ex vivo specimens.	Precise descemetorhexis edges with limited endothelial damage were reported ex vivo. Edema-related stromal injury requires further evaluation.
Malyugin et al 202370	Comparative clinical study; 61 eyes of 61 patients	Keratoconus	DALK — Assess novel FSL software enabling precise intrastromal tunnel creation for BB air injection with real-time microscope-integrated OCT guidance.	OCT-guided tunnel creation was associated with improved big-bubble formation and selected operative/endothelial outcomes. Comparative design and possible confounding require attention.
Wylęgała et al 202371	Observational clinical cohort; 32 patients	Anterior corneal layer pathologies	ALK — Evaluate the clinical outcome, efficacy, and safety of FL-ALK.	Favorable clinical outcomes were reported after FL-ALK. An observational cohort cannot establish comparative safety or graft-survival superiority.
Sorkin et al 202352	Randomized controlled trial; 100 eyes of 97 participants	Keratoconus or corneal scarring	DALK — Compare FL-DALK with standard manual DALK in eyes with keratoconus or corneal scarring, with astigmatism outcomes assessed at 15 months.	No significant superiority in primary astigmatism outcomes was demonstrated at 15 months. Follow-up loss and imprecision limit inference; the result does not prove equivalence.
Sorkin et al 202374	Long-term comparative clinical cohort; 58 eyes of 52 patients	Fuchs endothelial corneal dystrophy	DMEK — Evaluate 5-year safety and efficacy outcomes of FL-DMEK in patients with Fuchs’ endothelial dystrophy.	Visual improvement persisted through 5 years. Endothelial loss differences favored FL-DMEK at earlier time points, but not consistently later; non-randomized comparison limits causality.
Diaz-Valle et al 202475	Case report; 2 eyes of 1 patient	Schnyder corneal dystrophy	DALK — Evaluate the surgical effect of Pre-descemetic FL-DALK with simulated pachymetry.	Visual acuity improved without reported complications in one bilateral case. Generalizability and uncommon risks remain uncertain.
Abusayf et al 202422	Clinical feasibility series; 10 eyes of 9 patients	Keratoconus; Corneal scar; Corneal Dystrophy; etc	DALK — Evaluate the feasibility of iOCT-guided Ziemer LDV Z8 FSL cuts for mushroom graft-host geometry, lamellar cuts, and intrastromal tunnels in DALK.	Integrated mushroom, lamellar, and tunnel cuts were feasible; big-bubble formation occurred in all 10 clinical eyes. No comparative functional superiority was established.
Lin et al 202472	Clinical case series; 3 eyes of 3 patients	Ocular surface failure caused by bilateral ocular chemical burns	Lamellar corneal-limbal keratoplasty — Assess the feasibility and efficacy of FSL-assisted large-diameter lamellar corneal-limbal keratoplasty in bilateral chemical injury-related ocular surface failure.	Ocular surface and visual improvement were reported. The uncontrolled series included deterioration after immunosuppression withdrawal in one patient; superiority was not tested.
Palevski et al 202537	Clinical case series; 2 eyes of 2 patients	Keratoglobus	PK — Describe a modified technique utilizing FL-PK for the surgical management of keratoglobus with an extremely thin peripheral cornea.	Donor-only top-hat grafting improved structural and visual outcomes in two eyes followed for 12–24 months. This supports feasibility in selected eyes.
Du et al⁠ 202313	Meta-analysis; 9 non-randomized comparative studies, 1713 eyes	Keratoconus	DALK — Compare the efficacy and safety of FL-DALK and manual-trephination DALK for keratoconus.	Early corrected acuity and perforation outcomes favored FL-DALK. Non-randomized evidence does not establish durable visual or graft-survival superiority.
Baqalaqil et al 202538	Meta-analysis; 9 retrospective studies, 1940 eyes	Keratoconus; Corneal scar; Corneal Dystrophy; Herpetic Disease, etc	PK — Compare visual, refractive, endothelial, and complication outcomes between FLAK and conventional PK.	Pooled BCVA, astigmatism, and endothelial outcomes favored FL-PK. All included studies were retrospective; selection bias and confounding limit causal interpretation.
Li et al 202373	Comparative observational cohort; 63 eyes of 56 patients	Keratoconus	MILK — To compare the outcomes of FL-MILK for mild-to-moderate keratoconus and advanced keratoconus.	Corneal stabilization and flattening were reported over 24 months, with greater flattening in steeper eyes. Observational severity-group comparisons do not isolate a laser effect.
Zhang et al 202476⁠	Observational cohort; 50 eyes of 47 patients, including 23 unoperated eyes	Keratoconus	MILK — Evaluate the efficacy and safety of RGP-CLs wear after FL-MILK in progressive and advanced keratoconus.	Rigid gas-permeable contact lenses provided visual rehabilitation after FL-MILK. Fitting outcomes and surgical effectiveness should be interpreted separately.

Notes: Years follow the first publication date used by the source citation; online and issue dates may differ. Experimental samples and clinical eyes are not interchangeable units of evidence.

Abbreviations: ALK, anterior lamellar keratoplasty; BB, big bubble; BCVA, best-corrected visual acuity; CCF, clear cornea femtosecond; DALK, deep anterior lamellar keratoplasty; DM, Descemet membrane; DMEK, Descemet membrane endothelial keratoplasty; FLAK, femtosecond laser-assisted keratoplasty; FL-, femtosecond laser-assisted; FSL, femtosecond laser; iOCT, intraoperative optical coherence tomography; IOP, intraocular pressure; MILK, minimally invasive lamellar keratoplasty; PK, penetrating keratoplasty; RGP-CL, rigid gas-permeable contact lens.

Interpretation of Comparative Evidence

Patient Outcomes and the Limits of Geometric Precision

The central distinction is between a better starting wound and a better clinical course. The laser determines initial incision geometry, but healing, suture loosening or infection, early or asymmetric suture removal, inflammation, intraocular pressure, and postoperative adherence continue to alter refraction and graft health.^7,14 Persistent eye rubbing is a plausible additional modifier that should be recorded, not assumed to explain a treatment difference. Smoother interfaces and better apposition are relevant mechanistic measures; their clinical value requires demonstrated improvements in final corrected vision, graft survival, or reintervention. Conversely, the suggestion that cavitation-related micro-irregularities might favor integration remains a hypothesis and cannot be invoked as evidence of benefit.

DALK vs PK provides a historical caution. Although preserving host endothelium offers an anatomical and immunologic rationale, Keane et al‘s Cochrane review of two randomized trials found insufficient evidence of better final BCVA or graft survival with DALK; limited sample sizes and imprecision preclude a conclusion of equivalence.^⁠17⁠ Coster et al‘s Australian registry analysis reported poorer DALK survival and visual outcomes than PK for keratoconus during the study period.^⁠18⁠ Registry comparisons reflect case mix, practice patterns, and surgical experience and cannot isolate postoperative events as the cause. This precedent does not prove that current DALK or FLAK is inferior. It shows why a mechanistically attractive procedure must be evaluated through clinical outcomes and implementation, with attention to the learning curve.

Beyond these limitations, several additional findings warrant emphasis. First, apparent reductions in immunologic rejection should not be attributed to the FSL incision alone, but may reflect patient selection, intraoperative and postoperative management, and medication protocols.⁷⁷ Second, endothelial cell loss is procedure-dependent. In PK, endothelial reduction is influenced by baseline tissue status and surgical manipulation. In EK, laser-related interface roughness, cavitation effects, graft handling, and rebubbling may affect endothelial outcomes; however, the underlying mechanisms differ between DSAEK/DSEK and DMEK and should not be generalized across procedures.^64,78

Outcome Hierarchy and Sources of Heterogeneity

The heterogeneity across FLAK studies is best interpreted as the result of multilayered interactions among laser-platform configuration, intraoperative decision-making, and outcome assessment, with no single factor explaining all variation. Laser-tissue interaction, docking precision, centration, suction stability, image guidance, and cut completion are closely linked to the surgeon’s familiarity with device-specific nomograms, intraoperative judgment during imperfect dissections, willingness to convert from lamellar to full-thickness transplantation, and experience with complex wound designs. Interpretation is further limited by the methodological fragility of current evidence, as most studies are small, retrospective, single-center series from highly specialized institutions, reducing generalizability. In comparative DALK studies, femtosecond assistance may improve wound architecture or anatomical healing without consistently yielding superior long-term visual or refractive outcomes compared with conventional surgery.^53,79

Future comparisons should prioritize final BCVA, graft survival, repeat transplantation, and other reinterventions, followed by patient-reported visual function and recovery burden. Endothelial cell density, rejection, graft attachment, and refractive stability remain important secondary outcomes; incision accuracy, big-bubble formation, and interface roughness are mechanistic or procedural measures. Trials should prespecify clinically meaningful differences, report confidence intervals, and define follow-up relative to suture removal. Observational analyses require adjustment for indication, baseline acuity, donor quality, platform, surgeon experience, suture management, and postoperative complications. Registries can assess uncommon harms and generalizability but cannot replace randomization for treatment effects. Table 3 presents this hierarchy and the major variables requiring control.

Table 3.

Patient Outcomes, Surrogates, and Implementation Priorities

Priority/Domain	Patient-Relevant Question	Current Interpretation	Variables Requiring Control
Final corrected vision	Does the procedure improve final BCVA or patient-reported visual function?	Durable superiority of the studied laser applications is unproven. Early refractive improvement cannot substitute for final vision.	Baseline acuity and indication; follow-up after suture removal; opacity; comorbidity; visual rehabilitation.
Graft survival	Does the graft remain clear and functional; is repeat transplantation avoided?	Long-term comparative survival data remain insufficient. Endothelial density or wound strength alone does not establish survival benefit.	Donor quality; rejection; inflammation; infection; intraocular pressure; learning curve; loss to follow-up.
Reintervention and burden	Are rebubbling, resuturing, repeat grafting, and recovery burden reduced?	Selected DMEK cohorts suggest fewer interventions, but randomized confirmation and standardized definitions are needed.	Suture complications and early removal; graft attachment; surgeon thresholds; postoperative adherence and eye rubbing.
Secondary clinical measures	Are refraction, rejection, and endothelial preservation improved?	Effects depend on the procedure and workflow. Early FL-DSEK illustrates that less astigmatism can coexist with poorer acuity and greater endothelial loss.	Laser step; graft insertion; medication; donor preparation; platform; timing of assessment.
Procedural/mechanistic surrogates	Are cuts accurate, interfaces regular, and dissection reproducible?	These establish technical performance or feasibility. Their relationship to vision, survival, and reintervention must be tested.	Docking; pulse placement; opacity; tissue hydration; incision profile; big-bubble technique; manual completion.
Implementation across priorities	Is any incremental clinical benefit worth the resources and opportunity cost?	Historical FL-DSEK economics were unfavorable; contemporary procedure-specific cost-effectiveness is uncertain. Tiered access is a proposed framework.	Acquisition; interfaces and disposables; training; volume; eye-bank workflow; repeat procedures; access inequity.

Notes: Priorities distinguish patient outcomes from mechanistic measures; a null superiority test does not demonstrate equivalence.

Abbreviations: BCVA, best-corrected visual acuity; DMEK, Descemet membrane endothelial keratoplasty; FL-DSEK, femtosecond laser-assisted Descemet stripping endothelial keratoplasty.

Emerging Surgical Approaches Enabled by Femtosecond Laser-Assisted Keratoplasty

Beyond PK, DALK, and EK, FSL technology is being extended to therapeutic, tectonic, additive, and tissue-engineering-related corneal procedures.⁸⁰ These applications are not intended to replace standard keratoplasty, but to use reproducible, high-precision cutting to address limitations not fully resolved by conventional transplantation.⁸¹ Current evidence remains preliminary, consisting mainly of small clinical series, technical reports, and early follow-up studies. Accordingly, these approaches should be regarded as indication-specific extensions of FLAK with unestablished comparative effectiveness. Figure 5 illustrates their technical principles.

Figure 5.

FSL-assisted corneal reconstruction strategies. In corneal stromal lenticule transplantation, FSL is used to prepare a customized lenticule that fills and seals stromal defects. In Bowman layer transplantation, FSL creates an intrastromal pocket for graft placement to stabilize keratoconic corneas. In FSL-assisted G-SLET for limbal stem cell deficiency, laser-made microtunnels guide limbal micrografts and support re-epithelialization. For tissue-engineered corneas, FSL micropatterning enables scaffold modification and cell guidance, promoting organized corneal regeneration. This conceptual illustration was generated with assistance from ChatGPT (GPT-5.4, OpenAI) and reviewed and verified by the authors for accuracy.

Corneal Stromal Lenticule Repair in Therapeutic Keratoplasty

Therapeutic keratoplasty is frequently employed for the management of corneal abnormalities, including infectious corneal ulcers and even corneal perforations. In these settings, the immediate objective is not refractive rehabilitation or maximal optical clarity, but restoration of tectonic integrity, maintenance of anterior chamber stability, and control of infection or inflammation. Conventional options encompass sealing the perforation using tissue adhesive, conducting amniotic membrane transplantation or conjunctival flap coverage, and ultimately, performing large-diameter PK. However, every method presents its own limitations. For example, adhesives might only offer temporary support, whereas amniotic membranes may fail to provide adequate structural integrity for larger defects. Conjunctival flaps could potentially impede subsequent optical rehabilitation, and PK may induce surgical trauma in eyes affected by inflammation, vascularization, or infection.^82,83 Against this background, stromal lenticules obtained during small-incision lenticule extraction (SMILE) have been reframed from refractive-surgery by-products into potentially useful biological implants.

Corneal stromal lenticule transplantation (CSLT) uses autologous or allogeneic SMILE-derived stromal lenticules as tectonic patches to fill or cover areas of ulceration, thinning, or perforation.^83,84 The anatomical rationale is straightforward. These lenticules consist mainly of organized stromal collagen and extracellular matrix, without an epithelial or endothelial cell layer, and therefore may have relatively low immunogenicity while preserving native corneal collagen architecture. In practice, the recipient defect can be debrided or regularized, and the lenticule can then be sutured, tucked, layered, or placed as an onlay depending on defect depth and shape. The implant functions first as a mechanical spacer that restores thickness and then as a stromal scaffold for host cell migration, matrix remodeling, and epithelial resurfacing. Compared with glue or amniotic membrane alone, it provides a more substantial structural component; compared with PK, it limits tissue replacement and may preserve the eye for later elective optical reconstruction.

Early clinical evidence supports the feasibility of this tissue-sparing strategy, but not its routine adoption. Min et al⁸³ reported that CSLT sealed selected corneal ulcers, with no clinically evident immune rejection or infection during 6 months of follow-up and visual improvement in some eyes; however, several cases still required additional PK or scleral patching. In eyes with corneal perforation, Abd Elaziz et al similarly reported successful closure using SMILE-derived lenticules, with maintenance of corneal integrity and no reperforation or infection during 12 months of follow-up.⁸⁴ These results are encouraging, particularly when donor corneal tissue is scarce or inflammation makes PK high risk. Nevertheless, published series remain heterogeneous in etiology, defect size, infectious control, lenticule preparation, adjunctive antimicrobial therapy, and fixation method. CSLT is therefore best positioned as a bridging, tectonic, or tissue-preserving intervention. In a related concept, Jiang et al⁸⁵ evaluated therapeutic keratoplasty with intrastromal lenticule implantation and reported restoration of globe integrity with maintenance or improvement of visual function in treated eyes, further illustrating how FSL-generated stromal implants may support constructive corneal repair.

Structural Reconstruction and Corneal Stromal Augmentation

For keratoconus, post-refractive surgery ectasia, and focal non-inflammatory stromal thinning, the surgical aim in selected eyes is increasingly shifting from tissue replacement toward restoration of corneal shape and biomechanical stability. In these eyes, the host endothelium may be healthy and the central cornea may still retain useful optical clarity. The value of FSL technology lies in its ability to create stromal pockets, lamellar channels, and customized implantation planes with controlled depth and geometry.⁸⁶ This enables the feasibility of corneal stromal augmentation, a procedure that involves reinforcing or reshaping the cornea by adding biological tissue to the weakened area, instead of removing the native cornea and replacing it with a full-thickness or deep lamellar graft.

Bowman layer transplantation (BLT) is an additive approach for selected progressive keratoconus. A thin donor layer is inserted into the stroma to flatten or reinforce the cornea while retaining the host endothelium. FSL assistance can create the receiving pocket or prepare a graft with stromal inclusion.^87,88 Estrada-Mata et al reported three-year outcomes of the latter approach; longer follow-up of conventional Bowman layer inlay transplantation informs the underlying procedure but cannot establish an incremental laser benefit.^87,89 These distinctions are important because graft preparation, pocket creation, and disease severity vary between series. BLT may be considered when conventional stabilization or contact-lens rehabilitation is inadequate, with the aim of delaying more extensive transplantation. It is not established as a substitute for DALK or PK in eyes requiring optical tissue replacement.

A broader additive category includes stromal keratophakia, allogeneic or SMILE-derived stromal lenticule implantation, and localized lamellar tissue augmentation.^90–92 Here, preoperative tomography, pachymetry, epithelial thickness mapping, and refractive analysis can define both implant geometry and the intrastromal receiving space. FSL can then create a pocket or lamellar bed at a planned depth, allowing the implant to be positioned in relation to the cone, the thinnest point, or the area of biomechanical weakness.^93,94 Compared with synthetic intrastromal corneal ring segments, biologic additive grafts offer different design variables. Their mechanical effect may also be more distributed than that of a rigid segment. This flexibility is attractive in post-surgical ectasia, traumatic localized thinning, and highly irregular corneas in which a standardized implant may not match the underlying geometry.

The systematic review by Riau et al⁹¹ suggests that additive stromal procedures may improve uncorrected or corrected vision, reduce corneal steepness, and increase corneal thickness, but several translational questions remain unresolved. First, the long-term spatial stability, transparency, and remodeling behavior of implanted stromal tissue require longer follow-up. Second, tissue source, decellularization protocols, storage conditions, thickness profiles, and sterility standards have not been harmonized. Third, refractive predictability remains less mature than in established corneal refractive procedures. Future evaluation of FSL-assisted structural keratoplasty should therefore extend beyond short-term topographic improvement and should include corneal biomechanics, interface quality, implant clarity, host remodeling, and the need for subsequent DALK, PK, or repeat additive surgery.^91,94

Reconstruction of the Ocular Surface and Corneal Limbus

Limbal stem cell deficiency (LSCD) represents a critical ocular surface disorder. It impairs the renewal of corneal epithelium and subsequently induces the proliferation of corneal neovascularization, leading to the transformation of the transparent cornea into an opaque conjunctiva-like structure, ultimately culminating in profound visual impairment. Simple limbal epithelial transplantation (SLET) and glueless SLET (G-SLET) use small limbal tissue explants to repopulate the ocular surface and have become important strategies for unilateral LSCD.^95–97 FSL-assisted G-SLET adapts this concept into a more standardized pocket- or tunnel-based procedure. The laser can create peripheral stromal tunnels or micro-pockets to secure limbal micrografts without sutures or fibrin glue, and it may also assist in harvesting consistently sized limbal tissue from the donor eye, potentially reducing donor-site trauma and improving explant uniformity.^98–100

The clinical experience of this technology remains early but promising. Malyugin et al^95,100 reported that FSL-assisted G-SLET supported reconstruction of a stable corneal epithelial surface during 6 to 12 months of follow-up and was accompanied by some improvement in visual function in early cases. The potential advantage of this approach is that it provides physical fixation of limbal tissue while limiting manipulation of the central optical zone. However, major uncertainties remain, including donor-eye safety, optimal tunnel depth, micrograft number, graft distribution, postoperative epithelial behavior, and durability beyond the first postoperative year. FSL-assisted G-SLET should therefore be considered a precision-oriented extension of ocular surface reconstruction requiring comparison with conventional SLET.

Applications Associated with Tissue Engineering

In tissue-engineered corneal surgery, FSL is better understood as a three-dimensional microfabrication tool than as a conventional cutting device. It can be used to prepare stromal sheets of controlled thickness and shape from allogeneic or xenogeneic corneas, which may then be combined with decellularization protocols to create scaffolds designed to retain aspects of native stromal architecture.¹⁰¹ Micropatterning of scaffolds is a proposed extension of this manufacturing role. Effects on cell migration, alignment, and integration require direct corneal validation; general regenerative-medicine principles alone do not demonstrate a clinical laser benefit.^102,103 These functions are conceptually important because they link laser precision with cell behavior, matrix organization, and regenerative repair.

The clinical translation of FSL-assisted tissue engineering remains at an early stage. Most evidence derives from laboratory studies, proof-of-concept models, and limited reconstructive applications instead of large clinical trials.^103–105 Translation requires sterility, biocompatibility, and structural stability; acellular scaffold evaluation centers on structural integrity and host-cell compatibility. The material’s optical transparency, long-term durability, and immune compatibility are essential factors in ensuring the restoration of patients’ visual quality post-transplantation. Potential FSL applications include the fabrication of standardized biological patches, refinement of the scaffold’s microstructure, and facilitation of the repair of complex matrix defects. Whether this approach can mature into a clinically scalable alternative to donor corneal transplantation will depend on interdisciplinary work linking laser physics, matrix biology, biomaterials engineering, and long-term clinical validation.

Cost-Effectiveness and Value-Based Implementation of FLAK

FLAK economic evidence is sparse, but not absent. In a Dutch one-year evaluation of 118 patients, van den Biggelaar et al compared PK, early FL-DSEK, and microkeratome DSAEK.¹⁰⁶ Mean total costs were €6,674, €12,443, and €7,072, respectively; the proportions meeting the combined visual-improvement and astigmatism endpoint were 44%, 43%, and 52%. FL-DSEK was more costly and less effective on the point estimates. The DSAEK cohort was non-randomized, and these historical costs and workflows cannot be transferred directly to current platforms. Nevertheless, this is direct evidence against assuming that technical precision offsets added expenditure. Contemporary economic evaluations specific to FL-PK, FL-DALK, and image-guided reconstructive applications remain needed.

Economic Evidence and Local Resource Use

Femtosecond cataract surgery provides a related caution: Bénard et al‘s trial-based cost-utility analysis did not establish favorable value for routine laser assistance.¹⁰⁷ That finding is an analogy, not a FLAK cost estimate. A contemporary FLAK evaluation should include acquisition and maintenance, disposable interfaces, staff training, operating time, eye-bank preparation, complications, and subsequent procedures. Shared platforms may reduce allocated capital costs, but do not remove consumable costs or establish clinical benefit. Local volume, alternative uses of equipment, and reimbursement determine affordability; final vision, graft survival, and patient-reported function determine effectiveness. Both are needed before incremental cost-effectiveness can be assessed.

Access to Resources and Tiered Implementation Pathways

We propose a tiered implementation framework, not a validated economic model. Resource-rich referral centers may evaluate selected complex cases with prospective outcome and cost collection. Intermediate-resource systems may use shared platforms and regional eye-bank or referral networks where volume and expertise justify access. Resource-limited settings should preserve reliable conventional PK and manual lamellar services and refer exceptional reconstructive cases when appropriate.¹⁰⁸ Selection should identify the specific technical problem, an achievable patient-relevant benefit, and a feasible alternative. Proposed uses in difficult DALK, irregular scars, or complex graft matching remain hypotheses requiring comparison. Access should be based on clinical need and expected value, not device availability alone.

Evidence Required Before Broader Adoption

Routine acquisition or routine substitution is difficult to justify without a demonstrated incremental benefit. Clinical innovation in selected cases can proceed through structured evaluation, but technical success should not be presented as cost-effectiveness. Studies should report reintervention, visual rehabilitation, resource use, and uncertainty over a clinically relevant horizon. The tiered framework is intended to make selective access testable while protecting established transplantation capacity.

Other Challenges Facing the Clinical Implementation of FLAK

Clinical implementation also depends on optical access, docking, surgical experience, and reliable postoperative care. These constraints can affect both technical completion and patient outcomes.

Technical Challenges Inherent to the Technology Itself

FLAK remains subject to important technical constraints. Because photodisruption depends on precise optical focusing, marked corneal opacity, dense stromal scarring, or edema may reduce laser penetration and compromise tissue separation or interface regularity.^9,21 Parameter selection and docking stability also influence clinical outcomes. Inadequate surgical planning or protocol selection may lead to incomplete corneal separation, residual tissue bridges, interface adhesion, stromal-bed roughening, or endothelial injury.^21,61

Primary imaging and experimental studies show that docking and tissue deformation can alter incision geometry.^14,21 Clinical adoption also requires expertise in keratoplasty, laser planning, and management of incomplete dissection. The FEDS trial and comparative DALK series underscore the need to measure conversion, complications, and learning effects instead of presuming that automation removes operator dependence.^52,79,109 Image guidance can help verify the operative plane, but does not control subsequent wound healing or suture management. These limitations should be included when counseling patients and evaluating new workflows.

Limitations of the Current Evidence Base

Evidence remains fragmented across indications and platforms. Randomized results, including the FEDS and Dutch EK trials, do not establish durable patient-relevant superiority for the studied laser applications.^52,58 Favorable pooled observational estimates remain subject to selection bias, heterogeneous surgical protocols, and confounding.^13,38 Newer low-energy or reconstructive approaches cannot inherit benefits from older studies merely because they share the FLAK label. Conversely, early-platform complications should not be assumed to quantify current risk. This narrative synthesis is additionally limited by purposive selection, a single principal database, and no formal risk-of-bias assessment. Conclusions therefore distinguish observed associations, mechanistic plausibility, and outcomes still requiring direct comparison.

Potential Safety Concerns

Safety must be evaluated by procedure and laser step. Experimental studies document interface irregularity and endothelial damage, while the early randomized EK trial identified substantial endothelial loss and graft dislocation.^21,58,61 These outcomes reflect the complete workflow, including tissue preparation and insertion. Gao et al reported no infection or immunologic rejection during two years of follow-up after minimally invasive lamellar keratoplasty, but that series cannot exclude rare harms or establish comparative safety.¹¹⁰ Complications reported after other femtosecond procedures cannot be attributed to keratoplasty without procedure-specific evidence. Reliable safety estimates require denominators, standardized event definitions, device and procedure details, and adequate follow-up.

Future Directions in FLAK Development

Future FLAK development should move beyond the assumption that increasingly complex laser-generated incisions are inherently beneficial, toward an intelligent, biologically grounded, and implementation-aware workflow that links preoperative imaging, intraoperative guidance, and postoperative biomechanics to indication-specific surgical design.^14,22 Immediate priorities include integrating low-distortion liquid interfaces with real-time iOCT to reduce uncertainty in abnormal corneas, improve centration, and verify cutting depth and three-dimensional geometry before irreversible manual steps.^22,111 Equally important are multicenter registries and pragmatic trials that report not only visual outcomes, but also learning curves, reintervention rates, graft-related complications, resource use, and value-based endpoints.^43,112 A plausible application is in complex reconstruction, failed-graft rescue, and patient-specific constructive surgery, where platform capabilities can be matched to a defined anatomical or geometric problem. Such technical matching still requires patient-outcome validation.^14,22

Artificial Intelligence-Assisted Planning and Intraoperative Correction

Artificial intelligence (AI) is likely to influence FLAK first through decision support under surgeon oversight.¹¹³ Preoperative integration of corneal tomography, pachymetry, anterior segment OCT, ocular history, and prior outcome data may help estimate dissection depth, patient-interface selection, graft profile, and perforation risk in keratoconus, stromal scarring, or previously grafted eyes.¹¹³ Intraoperatively, AI-enabled image analysis could assist positioning, delineate tissue layers, detect incomplete photodisruption, and alert the surgeon before manual dissection. The near-term opportunity is therefore a closed-loop workflow linking preoperative imaging, laser planning, and iOCT-based verification, subject to independent clinical validation and surgeon oversight.¹¹¹

New Laser Physics and Wavelength Engineering

As noted above, laser-tissue interaction is a key determinant of FLAK outcomes. Continued reductions in pulse energy, optimization of repetition rate and spot overlap, and potential use of alternative wavelengths or shorter-pulse architectures may reduce cavitation-related collateral effects.⁹ This is particularly relevant for endothelial applications, where future systems must balance complete tissue separation with interface smoothness and endothelial safety.²¹ Translational studies should therefore move beyond dissection feasibility to examine how photodisruption alters extracellular-matrix ultrastructure, inflammatory signaling, postoperative wound remodeling, and endothelial cell viability.^9,21,101

Biomechanical Optimization of Incision Geometry

Classical top-hat, zig-zag, and mushroom profiles were designed to increase wound surface area and resist wound dehiscence, but future geometries should be guided more explicitly by corneal biomechanics.^15,16,25 The aim should be to distribute stress more physiologically across the donor-host junction, reduce shape mismatch, and lessen postoperative astigmatism after suture adjustment and long-term remodeling.^14,114 This may require patient-specific incision design tailored to the individual eye, especially in ectatic, pediatric, traumatic, or highly asymmetric corneas.^14,114

What Next-Generation Studies Should Concentrate on

Future studies should prespecify final BCVA, graft survival, and repeat intervention as principal endpoints, together with patient-reported visual function. Recovery trajectories, standardized rejection events, endothelial cell loss, graft detachment, and resource use provide complementary information.¹¹² Technical measures should be analyzed as possible mediators of clinical outcomes. Economic evaluations should estimate incremental costs and quality-adjusted outcomes with sensitivity analyses and a stated perspective. Multicenter trials and registries should report surgeon experience, conversions, suture events, postoperative care, and losses to follow-up so that efficacy, feasibility, and real-world value can be distinguished.

Conclusions

FLAK provides programmable incision geometry, lamellar dissection, and selected reconstructive capabilities after nearly two decades of clinical development. However, the available randomized evidence has not established durable superiority in final corrected visual acuity or graft survival for the applications studied; observational improvements and mechanical precision cannot establish those benefits. Early laser-assisted EK also demonstrated that reduced astigmatism can coexist with poorer acuity and greater endothelial loss. The DALK–PK precedent reinforces the need to evaluate the entire clinical course, including learning effects, suture events, wound healing, and postoperative care. Selective use may be reasonable when a defined technical bottleneck in complex PK, DALK, stromal augmentation, or posterior reconstruction cannot readily be addressed manually, but these proposed indications require comparative validation. Cost and access must remain part of that decision: historical FL-DSEK economic results were unfavorable, and contemporary value is uncertain. A tiered approach can concentrate evaluation in experienced centers, support shared regional access, and preserve conventional services where resources are constrained. Lower-impact laser delivery, real-time imaging, AI-assisted planning, and individualized geometry should be judged by long-term vision, graft survival, reintervention, and patient-reported function before broader adoption.

Funding Statement

This work was supported by the Beijing Chao-Yang Hospital Golden Seeds Foundation (CYJZ202550, to Yifan Du).

AI Declaration

Figures 1–5 were newly created as conceptual schematics using Python/Matplotlib with AI-assisted coding and image generation/editing support from ChatGPT (GPT-5.4, OpenAI). No clinical photographs or experimental measurements were generated or altered, and the replacement figures do not incorporate image elements from the original figure files.

The authors have reviewed all figures and confirmed the originality and accuracy of their content. The authors have checked the terms of use applicable to the AI tool and confirmed the suitability of the generated images for publication. The authors confirm that they have the right to publish these images and have obtained any necessary permissions. The authors take full responsibility for the integrity of the entire content, including the accuracy of the illustrations and associated references. All image versions and records of the prompts used have been retained and will be made available upon request.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.