Penetrating keratoplasty (PK) forms the cornerstone of corneal transplantation and in some circumstances remains the only option for ocular preservation or restoration of sight. Potential problems with mechanical trephination techniques involved in PKs include divergent recipient cut angles and convergent donor cut angles resulting in tissue deficit at the posterior corneal plane resulting in potential misalignment. These problems may be compounded in eyes with narrow palpebral apertures. Femtosecond wound construction allows improved centration without undercutting. This may also help reduce damage to the endothelium, with evidence from animal and human studies demonstrating preservation of cell counts [3, 40].

Even in conventional PK, suction-based trephination systems allow better stability during trephination, are faster, and facilitate rounder trephination with less slippage. There are problems however with intraocular pressure elevation, centration, and in eyes with reduced scleral support such as in aphakia. The application of suction, common to most femtosecond platforms, may also be associated with elevated IOP. A demonstrable reduction in intraocular pressure variation in femtosecond laser-assisted PK has been shown in the VisuMax system compared with manual trephination [3]. There is considerable variability in IOP elevation between platforms however, and this is discussed later in this chapter.

The data from clinical trials to date is relatively limited but increasing, summarized in Table 15.2. The variable results seen by Femto-assisted PK reflect the myriad outcome measures determined in individual trials or large case series including indication and graft size. Furthermore, many of the studies to date have evaluated the first commercially available platform, IntraLase, and data on many of the newer platforms is therefore restricted. Nonetheless, the principle considerations with regard to this technology relate to wound integrity and recovery, astigmatism and visual outcome, and endothelial cell preservation and rejection. These issues may well be addressed further by well-constructed randomized control trials.

The first major decision in choosing Femto-assisted penetrating keratoplasty over conventional grafting is better wound strength and alignment. The strength of the wounds constructed with femtosecond laser has been shown to be resistant to leakage even with less sutures [42]. The second outcome to contemplate (and related to the first) is astigmatism and visual recovery. A retrospective series directly comparing a straight-cut conventional PK and Femto-assisted PK (n = 20 in each group) suggested that there was less induced astigmatism (6.06 vs. 4.06 D; p 0.04) and faster visual recovery with Femto assistance, but the overall visual outcomes were similar (0.39 vs. 0.22; p = 0.8 LogMar) [33].

Femto-trephination can also facilitate novel and potentially more stable wounds with a theoretical reduction in astigmatism. Improved wound construction and alignment with a femtosecond laser assistance has been proposed with a number of methods including zigzag shapes, mushrooms, top hat (see above), dove and tail, decagonal and lock, and key designs among others [23, 26, 39, 46, 49, 58]. The advantages with shelved or stepped interfaces are a potential reduction in the number of sutures and faster postoperative recovery. Although a reduction in astigmatism by improved tissue apposition has been shown with zigzag configurations in the initial postoperative period (between 8.4 and 5.8 D at 4–6 months), this difference was not seen after 6 months [13, 25]. Other studies have failed to demonstrate an improvement compared to conventional surgery in astigmatism, albeit paradoxically with improvement in vision [27]. It is worth considering that the same early effect was seen with straight cuts as outlined above [33].

Third, there appears to be variable results in the reduction in endothelial cell loss compared to conventional PK. Kamiya and colleagues’ series with the VisuMax platform did not demonstrate a difference between Femto-PK vs. conventional PK [33]. Other clinical trials have shown postoperative endothelial cell counts in the range of 1200–2000 cells/mm² at 6–12 months [19, 30, 46]. Graft rejection rates in most series have been variable, but it is worth noting that few studies have follow-up data for greater than 12 months, compounded by the variable timing of suture removal. Larger series have suggested that complete suture removal can be achieved earlier than in conventional PK [7].

Finally, all three parameters must also be preceded with a fundamental question regarding the choice as to whether one should undertake PK over a lamellar procedure. This has not been fully addressed and will be considered in the following sections of this chapter.

Targeted replacement of the anterior stromal layers of the cornea by anterior lamellar keratoplasty may involve the superficial layers (by manual dissection or microkeratome) or deeper layers (through manual dissection or the use of a big bubble). Deep anterior lamellar keratoplasty (DALK) has the advantage of facilitating an extraocular procedure and theoretically reduces the risk of both rejection and endothelial cell damage where the endothelium is unaffected. To date, outcomes from large national datasets in the United Kingdom and Australia have indicated worse visual outcomes and survival than for penetrating keratoplasty [21, 31]. Advocates of DALK argue that the published data relates to (relatively) historical series when more surgeons were undergoing the learning curve, and this may be borne out by higher rates of primary graft failure in the early weeks posttransplantation for this group. This issue is one of contention however and an ophthalmic technology assessment undertaken by the American Academy of Ophthalmology published in 2011 (albeit predating the work by Coster and colleagues) found that there was no difference between DALK and PK with regard to visual acuity, but endothelial cell counts were better preserved with DALK [50].

An apparent advantage with DALK is cost-effectiveness. Two cost-utility analyses to date have demonstrated that despite the higher costs of undertaking DALK (in part due to increased operative time), there are financial longer-term benefits [35, 63]. The rates of lamellar (including anterior lamellar surgery) continue to rise internationally and we will therefore consider how the femtosecond laser may enhance the application of this procedure.

In addition to the problems seen in Femto-PK relating to wound configuration and astigmatism, the major consideration relating to femtosecond-DALK, like conventional DALK, is the technical challenge presented with creating a clear graft-host interface and leaving the minimal amount of residual tissue bed. This difficulty in theory should be circumvented by the automated lamellar dissection offered with femtosecond laser, which may offer a smoother interface. A caveat however is that with deeper dissection, there is enhanced light scatter and potentially a less smooth surface. Differences in collagen disruption have also been determined with the construction of differential laser firing during refractive lenticule construction [51]. Whether this will have an influence on lamellar construction in the context of keratoplasty is yet to be determined. Visual outcomes in DALK may in part be explained by the thickness of the host tissue bed [5]. Notwithstanding the effects of scatter, Femto-dissection could create potentially thinner dissections without the inherent risk posed by manual dissection alone.

Although microkeratomes facilitate anterior cuts, pneumatic dissections for deep lamellar procedures carry a risk of perforation. The femtosecond laser offers theoretical improvement in control in creating an interface during tissue separation. Laboratory data shows that the interface created by femtosecond lasers is smoother, and as previously discussed, low-energy high-frequency platforms may offer an enhanced role for Femto-assisted DALK [51–53]. Femto-assisted lamellar trephination also lends itself to treating superficial pathology and no difference was seen at 12 months follow-up when comparing the visual acuity of Femto-anterior lamellar keratoplasty cut at <250 μm and >250 μm [1].

There is a paucity of clinical trials evaluating Femto-DALK, but a number of series have been published regarding the technique. Both Farid and Price separately presented a zigzag adaptation of a “debulking” big-bubble technique where a zigzag side cut is combined with a 300 or 250 m lamellar dissection, respectively. This facilitates a big-bubble separation of the residual stroma [24, 47]. Buzzonetti and colleagues also describe an adaptation of Anwar’s original big-bubble procedure [4] whereby the IntraLase femtosecond laser is employed to facilitate a side cut of 50 μm and a lamellar cut of 100 μm anterior to the thinnest point [11, 12]. This in turn is supplemented by the injection of air to separate Descemet’s in a technique termed Intra-Bubble. Outcomes from a 1-year case series (n = 11) by the same group determined that the best-corrected distance visual acuity was 0.52 ± 1 with a refractive outcome of −1.50 ± 1.7 diopters (D) sphere and 2.00 ± 2.6 D cylinder (two attempts were converted to PK at outset) [12]. Longer follow-up also shows mean BCVA of 0.3 ± 0.1 at 24 months (n = 12) and a mean cylinder of 1.7 ± 1.4 D [56].

The bespoke interfaces used in femtosecond-penetrating keratoplasty such as a top-hat configuration may improve the speed of recovery in femtosecond-DALK [14]. The integrity of zigzags, top-hat, and mushroom configurations has been evaluated in an experimental model to determine burst pressure with these cuts [36]. Although the pressure required to induce wound burst was variable, direct comparisons were not undertaken, and it is therefore difficult to draw definitive conclusions regarding the optimal technique.

Sutureless techniques have also been described with variable mean uncorrected and best-corrected visual improvement [8, 67]. Evidence from retrospective series has indicated a faster visual recovery when comparing mushroom configurations with conventional straight cuts undertaken with the IntraLase system but with no difference in overall visual recovery or astigmatic outcome, similar to the findings in Femto-assisted PK [55]. Although conventional DALK is undertaken to prevent endothelial cell loss, it is worth considering the potential effects of the femtosecond laser due to the application of energy in the host bed.

The uptake for Femto-DALK has been limited in part due to the technical difficulty of achieving a safe dissection in the context of severe ectatic disease and reflected by the absence of controlled trials to date. This is further highlighted when the surgeon is faced with existing posterior stromal scarring, a situation that complicates previous hydrops. Limitations in visualizing the cornea in real time by OCT and Scheimpflug imaging raise legitimate concerns about proceeding with femtosecond laser-assisted surgery following docking, as small movement may have catastrophic consequences on the already-friable host bed. High-resolution intraoperative OCT has been shown to enhance the ability to undertake DALK safely by conventional methods [22]. It is hoped that recent improvements in imaging platforms attached to femtosecond platforms may offer an improvement in this regard and a safer option to undertake femtosecond-DALK.

In contrast to anterior lamellar keratoplasty, the uptake of Femto-assisted dissection of graft material for endothelial keratoplasty has been more widely adopted. Endothelial keratoplasty, like its anterior counterpart, potentially facilitates smoother and more accurate cutting of the desired tissue bed. This is particularly important when minimizing the residual stromal bed transplanted.

Manual or microkeratome dissection has traditionally been employed for Descemet’s stripping automated endothelial keratoplasty (DSAEK). The accuracy of depth of microkeratome cuts may be less consistent than with Femto-assisted dissection. Femtosecond laser-assisted ablations have been shown to have a mean deviation in attempted depth of 15 μm [44]. Further adaptations such as double-pass techniques may consistently achieve sub-150 μm grafts with improved visual acuity however [9].

Femtosecond beds have been shown to be smooth under histological evaluation [17, 43]. Another study evaluating the effects of the IntraLase 30 kHz femtosecond laser has demonstrated the mechanical microkeratome may improve the depth and smoothness of the cut [32]. The “rougher” interface created by the femtosecond laser was postulated as having a potentially improved interface for maintaining adherence – however, the rate of graft dislocation has previously been shown to be as high as 20 % when undertaking Femto DSAEK [16]. Whether this relates to surgical technique or the smooth tissue bed created remains to be elucidated. Furthermore, the type of laser may influence the interface created and another comparisons using the IntraLase platform found rougher surfaces were created with the femtosecond laser compared to microkeratomes [45].

Endothelial cell loss is an important consideration in judging the safety of Femto-assisted endothelial keratoplasty. Both the aforementioned studies comparing smoothness of the interface created by microkeratome found no difference in the reduction of endothelial cell count [32, 45]. Inverse cutting techniques have also been proposed as a means of safely maintaining endothelial cell counts when creating lamellar cuts [29]. A study comparing 50 vs. 150 μm dissection in rabbits using the Wavelight FS200 however showed significantly higher rates of endothelial cell damage and apoptosis with thinner cuts [41]. A large randomized control trial also found that rates of endothelial cell loss were higher with Femto-assisted endothelial keratoplasty compared to conventional penetrating keratoplasty (1200 vs. 2150 cells/mm² at 3 months) [18]. The difficulty in comparing two separate techniques and by laser and conventional means makes this more difficult to interpret. Furthermore, the authors concede that the method by which the graft was inserted (by forceps with a folded graft) will have contributed to the endothelial cell attrition. The limits by which safe dissection can be achieved need to be evaluated further. This of course represents a challenge for undertaking the very thin cuts needed to facilitate ultrathin Descemet’s stripping endothelial keratoplasty (UT-DSEK).

Outcomes of randomized clinical trials involving femtosecond laser-assisted endothelial keratoplasty are also sparse but summarized in Table 15.3. It is interesting to note that the visual acuity was reduced in the large trial comparing conventional PK and may also reflect the quality of the interface [20].