Corneal stromal diseases represent a significant cause of morbidity worldwide and one of the leading indications for keratoplasty. Among them, keratoconus is characterized by corneal ectasia resulting in irregular astigmatism and often causing loss of correctible visual acuity. Although the number of keratoplasty recipients for keratoconus has been decreasing over the years, secondary in part to corneal cross-linking and improved contact lens models, keratoconus remains one of the principal corneal diseases treated with keratoplasty, following only endothelial decompensation and repeat grafting.
9–13
DALK has been established as the procedure of choice for the surgical treatment of corneal stromal diseases. There is a consistent body of previous literature which shows that DALK produces visual outcomes and postoperative astigmatism that are similar to penetrating keratoplasty, but with a lower rejection rate, increased survival, and decreased endothelial cell decay.
14–20 In addition, DALK does not alter corneal biomechanical properties, whereas an overall reduction in corneal hysteresis and resistance factor is observed after penetrating keratoplasty.
21 Last, the cost-effectiveness ratio, as outlined in Singaporean and Dutch corneal transplant registry analyses, would also favor DALK over penetrating keratoplasty.
22,23
Despite proven superiority to penetrating keratoplasty for the treatment of corneal stromal diseases, the adoption of DALK has been somewhat suboptimal worldwide. In the 2019 EBAA Statistical Report, donor corneas used for anterior lamellar keratoplasty procedures accounted for 2.5% of domestic use.
13 This occurred despite corneal ectasias and thinning being the sixth most common surgical indication for keratoplasty in the United States overall. Consequently, 14% of penetrating keratoplasties in the United States are performed for corneal ectasia and thinning, thereby indicating gross underuse of DALK procedures. To paraphrase the 2019 EBAA statistical report, “The number of anterior lamellar keratoplasty … has been essentially flat over the last 8 years.”
13 The reasons behind DALK’s lack of popularity are multifactorial. Among them, the intrinsic difficulty of the surgical procedure and the long learning curve, often requiring prolonged unaccounted surgical time, seem to play a pivotal role. Popular DALK surgical techniques can be broadly divided into two categories: the ones in which cleavage of predescemetic or descemetic layer is obtained by injection of a foreign substance (e.g., air, saline solution, or ophthalmic viscoelastic material) and the ones where the deeper portion of the stroma is reached by manual layer-by-layer dissection.
2 The depth of placement of the injection cannula within the corneal stroma or depth of manual stromal dissection are largely based on personal surgical experience. Although surgeon's experience seems to affect outcomes minimally in academic settings, a definitive learning curve in single surgeon series with a decreased incidence of intraoperative complications over time has been reported.
24–26 Only referral centers performing high numbers of keratoplasties often have sufficient critical mass of surgical volume to overcome the DALK learning curve and ultimately achieve reproducible results.
In an effort to decrease intraoperator variability, standardize the surgical technique, and ultimately aid popularization of DALK we sought to conceive an inexpensive, simple, and versatile surgical device that may aid successful BB dissection. Our device does not require sterilization and uses disposable components (tubing and valve) that can be adapted to any DALK cannula of choice. Obtaining a consistent BB dissection could significantly decrease the overall surgical time and benefit efficiency in the operating room. Moreover, for surgeons needing to perform manual dissection DALK, the depth reached by the air cannula with this device could also be used a starting point to initiate manual dissection at the appropriate depth and rapidly reach the predescemetic or descemetic layer.
The depth of stromal insertion of the cannula for effective BB that could be extrapolated from our OCT experiment was roughly 150 µm from the endothelial side, which is in keeping with what was observed in previous studies.
5,6
The device captured a positive pressure differential when the BB was achieved versus a flat or negative pressure differential in unsuccessful cases. We have hypothesized that the reason behind these observations may lie in the amount of stromal tissue present above the DALK cannula. For a successful pneumatic dissection, the DALK cannula has to be positioned within the deeper layers of the stroma. In these cases, the remaining stroma may act as a hinged flap, producing a valve mechanism that is responsible for air entrapment in the system and consequent pressure increase. In contrast, a superficial placement of the cannula may not result in external compression by the overlying stroma therefore allowing air escape and creating a flat or negative pressure differential (
Fig. 2A). As a corollary to this observation, one could also postulate that not only the deep placement of the cannula is paramount in effective BB dissection, but also avoiding backtracking of air through a leaky path when the injection is performed, which is also a common observation in DALK surgery.
Another possible explanation could be that corneal stroma, because of its spongy macrostructure, would accommodate the small volume of air generated by the continuous air injection of the device, thereby not causing a pressure rise within the system.
27 Conversely, the predescemetic or descemetic layer would be less compliant and penetrable to air and produce a better sealing of the system, with a consequent increase in pressure and a positive pressure differential.
In an ex vivo setting, this device was able to signal to the surgeon correct intrastromal placement of the cannula corresponding with successful BB 94.1% of the time, which is superior to the success rate of BB dissection reported in the literature. The feedback from the device was given to the surgeon once intrastromal insertion of the cannula was completed. In an in vivo setting, we postulate that the surgeon would withdraw the cannula in case of negative signal from the device and attempt intrastromal introduction from a different point.
Presently, the only technological aid to assist cannula placement for BB dissection is offered by intraoperative anterior segment OCT (iAS-OCT). iAS-OCT produces real-time imaging of the anterior segment that could be of great help in guiding accurate placement of the cannula in BB DALK.
28–31 iAS-OCT can also guide manual dissection in DALK and contribute to early recognition and treatment of complications.
32,33 The repeatability, sensitivity, and specificity of iAS-OCT in BB-DALK have never been investigated. To date, this technology comes at a high cost, to the point that the cost effectiveness of iAS-OCT devices remains a challenge. In addition, real time imaging is often not coaxial with microscope focusing, requiring the surgeon to look away from the surgical field for an inconvenient additional focusing step. Last, DALK cannulae produce OCT shadowing, which decreases the quality of the imaging of the stromal bed under the cannula.
28 One of the advantages of our device would definitely be the lower cost of the machinery and consumables. The device signals the point of optimal penetration depth via an acoustic signal and does not require the surgeon to stop looking through the microscope binoculars at any point.
Newer, technologically advanced approaches based on evolution of OCT technology are under investigation. Shin et al.
34 showed promising results ex vivo and in a rabbit model using a custom-made 26G cannula with integrated M-mode swept source OCT in the tip. In addition, robotic fully automated or semiautomated needle insertion guided by volumetric OCT has been developed.
35 Robotic insertion and OCT-integrated cannulae have also been used in combination.
36 These cutting-edge technological approaches are presently still in early development.
In conclusion, this study proposes a simple prototypical pressure-based device that could indirectly measure depth of cannula placement in DALK by providing an acoustic signal to the surgeon once the deeper corneal stroma has been reached. The device could increase the success rate of DALK and flatten the learning curve for inexperienced surgeons. In addition, it may represent an inexpensive alternative to iAS-OCT. Future studies to further simulate surgical scenarios are in the pipeline. Additionally, we are planning a multicenter clinical trial in which the device would be used by surgeons with different levels of surgical experience with DALK.