This series of articles summarises an afternoon of presentations, plus a debate, on the topic of corneal ectasia that was held as part of the 29th Cologne Adventsymposium, a meeting organised by Laserforum e.V. and recently held in Cologne, Germany. Chaired by Drs Omid Kermani and Georg Gerten, the presentations each considered the condition from a different perspective and are available online.
Part 1, published in the June 2022 issue of Ophthalmology Times Europe®, considered the contribution of eye rubbing to the development of keratoconus. In this second instalment, we look at developments in diagnostics and classification methods for ectatic corneal diseases. This is based on the presentation that ophthalmologist Prof. Renato Ambrósio Jr. gave during the one-day event.
Reviewed by Dr Omid Kermani and Prof. Renato Ambrósio Jr.
There have been many advancements in investigative technologies in corneal refractive screening and diagnostics over the past twenty years. Different multimodal imaging approaches can provide important information on the state of the cornea. In his presentation at the 29th Cologne Adventsymposium, Prof. Ambrósio explained that applications for corneal ectasia include screening and early diagnosis, prognosis, classifying and staging the disease.
He stated that when contemplating treatment, auxiliary tests are relevant for optimising manifest refraction, contact lens fitting, customised surgical planning and for detailed follow-up postsurgery. The data are also crucial for patient orientation, education and treatment guidance.
In diagnosing ectasia, Prof. Ambrósio believes it is important to think ahead. For example, monitoring corneal topography enables subclinical ectasia to be detected prior to the development of slit-lamp signs and loss of distance-corrected vision.1 However, to enhance the sensitivity further, there is a need for more advanced techniques for three-dimensional tomography and biomechanical assessments of the cornea.
Prof. Ambrósio pointed out that, when it comes to the available technology platforms, some of the diagnostic approaches for ectasia are integrated in the same machine. For example, the Oculus Pentacam is a rotating Scheimpflugcorneal and anterior segment tomography device; newer models include optical biometry for measuring the axial length andocular wavefront.
Thereby, anterior chamber depth and volume; anterior and posterior corneal elevation; corneal thickness mapping; total corneal and corneal surface refractive power; and corneal aberrometry can be combined with the measurement of axial length in the Pentacam AXL for enhancing IOL planning.
Furthermore, Prof. Ambrósio described that such data can be integrated with the ocular wavefront in the Pentacam AXL-Wave to provide a complete understanding of the optical properties of the eye. He added that these data are essential for detecting other issues besides corneal ectasias, such as angle-closure glaucoma, and are necessary for making clinical decisions such aswhether to proceed with lens surgery.
For diagnosing keratoconus and screening for ectasia specifically, Prof. Ambrósio pointed out, the Pentacam Belin/Ambrósio display (BAD) characterises the deviation from normalityin the standard andenhanced elevation of the front and back surfaces, along with the thickness spatial profile. The BAD also includes a final “D value,” basedon a linear regression analysis, to enhance the accuracy ofectasia detection even in subclinical cases.
Next, Prof. Ambrósio moved on to discussing forme fruste. When it comes to classifying keratoconus, there is no consensus on the definition of forme fruste.2 Prof. Ambosio referred to a systematic review of the literature, which revealed that there is also tremendous variability in the criteria for defining forme fruste keratoconus and keratoconus suspect based on topography.3
He said that in a very asymmetric ectasia case, the clinically normal eye is often considered forme fruste keratoconus. Many studies have addressed such cases when developing and testing novel diagnostic approaches for detecting mild or subclinical ectasia.4 Nevertheless, Prof. Ambrósio believes it is vital to understand that topographic criteria are insufficient to separate forme fruste from a normal eye (Figure 1).
He added that it is essential to have an objective means of distinguishing normal and keratoconus eyes. Subclinical cases fall in between the two extremes, and corneal topography is not yet sufficient to distinguish.
Prof. Ambrósio’s opinion is that forme fruste keratoconus should not be defined as a topometric criterion butrather as ahigh susceptibility for ectasia development and progression. This was presented for the first time at the American Academy of Ophthalmology’s annual meeting in 2008.5
Prof. Ambrósio also discusseddistinguishing normal from keratoconus eyes through segmental or micro-layer tomography. He explained that this is possible with very-high-frequency ultrasound (VHFUS) and spectral domainoptical coherence tomography (OCT) and allows ophthalmologists to visualise structures such as the epithelium, the flap, the endothelium and the Bowman’s layer and identify thinning patterns in the cornea.6
Hespoke of the various AI algorithms thatare available to examine epithelial thickness. The Artemis software for analysing arc scan data with VHFUS includes analgorithm developed by Dr Dan Reinstein and coworkers;7 analysis of OCT epithelial pattern deviation can distinguish keratoconic corneas from contact lens warpage.8
Another useful tool, according to Prof. Ambrósio, is the ultra-high-speed Scheimpflug imaging from the Oculus Corvis ST, which can generate data to separate a normal cornea from a weaker one, measuring stiffness by deformation response. He said that current work combines tomography and biomechanics, using AI.
The tomography and biomechanical index (TBI) is a parameter that characterises biomechanical susceptibility directly, Prof. Ambrósio explained. For example, in a case with typical topography and relatively normal BAD-D, an abnormal TBI means that the patient is highly susceptible to developing ectasia.
Hewent on to discuss the original TBI study, which included 94 cases with typical topography based on strict criteria from very asymmetric ectasia cases along with the unoperated ectatic eyes from 72 of these cases.9 In some cases, the fellow eye had received cross-linking, rings or transplants.
Eyes randomly selected from 480 normal subjects and from 204 keratoconus patients (one eye per subject) were also analysed. The TBI had near-perfect accuracy (AUC = 1) for detecting ectasia in the frank ectatic cases (72 unoperated ectaticeyes + 204 keratoconic) and a sensitivity higher than 90% for detecting abnormalities in the normal eyes of very asymmetric ectasia patients, while optimising the cutoff criteria with a specificity higher than 95%.
Bilateral forme fruste was possible, Prof. Ambrósio explained, as demonstrated in a report of twin sisters,10 only one of whom had clinical keratoconus in one eye while the TBI was abnormal in all four eyes. However, ectasia can also occur in one eye only, as in one patient (Figure 2) who was followed after having a corneal ring procedure.11 The left eye was normal, even when analysed using VHFUS for epithelial thickness and the Saad–Gatinelscore from the Bausch + Lomb Orbscan. Interestingly, the TBI was normal.
Prof. Ambrósio stated that this case supports the global consensus that ectasia can occur unilaterally due to mechanical forces whereas keratoconus is a bilateral disease. In other words, true unilateral keratoconus does not exist.
Rounding up his talk, Prof. Ambrósio emphasised that multimodal imaging for corneal ectasia requires the combination of AI and ancient intelligence to provide a customised evaluation of biomechanical susceptibility and for diagnostic applications such as screening, staging, prognosing, classifying and monitoring progression, and planning individualised treatments.
Prof. Ambrósio’s presentation was debated with the other presenters—Prof. Damien Gatinel (France), Prof. Dan Reinstein (UK), Dr Tim Archer (UK) and Prof. Farhad Hafezi (Switzerland)—and with members of the online audience.
Dr Kermani: Biomechanics is a controversial topic. How specific and sensitive is the biomechanical analysis of the cornea and how do you see the future of biomechanics in the diagnostics of keratoconus?
Prof. Ambrósio: Imaging will definitely evolve more and more, and biomechanics provides very relevant data. OCT can provide information about corneal properties and reveal the segmental layer shape.
I look forward to the development of novel imaging tools for biomechanical assessment, such as OCT elastography, which may be less impacted by IOP. It is key to recognise that IOP influences biomechanical measurements of the cornea and impacts the deformation response.
We also know that the opposite is true: corneal properties affect IOP measurements. I also believe that corneal properties may reflect scleral properties and thereby the optic nerve’s susceptibility to glaucoma progression.
Eventually, corneal properties can also translate to the risk of peripheral retinal disease because, if you think about the very thin sclera, the extraocular muscles may lead to more pressure inside the eye and then you can have more issues with the peripheral retina or retinal detachment.
These are things that we are looking at in prospective studies, and I think integrated multimodal imaging is the future; it is not that one technique will replace everything.
Epithelial thickness is significant for any refractive surgery, for pre-and postoperative analysis. The epithelium reacts to the corneal shape so, with reference to the evolution of keratoconus, there is a biomechanical susceptibility that is triggered by genetics and the environment. Of course, all this is linked to changes in the stroma, which reflect on the epithelium.
The beauty of modern technology is that we can help patients in an unprecedented way. In the past we were only able to do a transplant, but now we have cross-linking rings and many other procedures that can rehabilitate vision and stop progression.
Today I consider keratoconus to be a specialty for refractive surgeons and part of the refractive surgery world. The major paradigm shift is that we think not only about the rehabilitation of vision but also about the progression of the disease.
Dr Kermani: To ask our other speakers: Farhad, Damien, Timothy, do you use biomechanics diagnostics in your keratoconus patients or is this is not yet a practical diagnostic tool in your office?
Dr Archer: We have had the Corvis ST in the clinic for 4 or 5 months now, and I think that what Renato has said is the key, in that all the information these devices can add is helping us detect more cases. Indeed, adding biomechanical data will help; it is noisier than some of the other variables that we use, but it contributes.
I wanted to comment on something else that Renato said about the epithelium being reactive to the stroma. I totally agree. The great thing about the epithelium is that it is so sensitive because it is a thin layer, 50 microns or so. Such a thin layer easily reveals those very small (1–3 micron) changes that are harder to see on the stromal surface, which makes epithelial mapping a sensitive method for detecting things early.
Prof. Gatinel: I use biomechanics for refractive surgery evaluation more than for keratoconus follow-up. It provides a tie-breaker when you have topographies resembling forme fruste.
I use the optical response analyser (ORA) system, which I like, and I look more at the signal shape than at the actual numbers, which are influenced by IOP. When it shows low and irregular peaks, this may be indicative of a soft cornea. This reinforces my need to persuade the patient to stop rubbing their eyes and avoid sleeping on their face.
I think those two things significantly reduce the risk of ectasia post LASIK or PRK, so I use biomechanics, but not for keratoconus because by the time keratoconus has developed it is too late.
Dr Kermani: Farhad, I know that you are more interested in keratoconus than in refractive surgery but is biomechanical analysis a helpful tool in keratoconus assessment in the very early stages or in progression?
Prof. Hafezi: Yes, I strongly believe it is. The question is, how reliable is the technology that we have now? And how is it evolving?
I used the ORA for 3 or 4 years, and since 2014 I have been using the Corvis. I do not base my final clinical assessment on Corvis data for either refractive surgery planning or keratoconus assessment, but, frankly, sometimes I am puzzled when I put together topography, tomography and biomechanics data as delivered by the Corvis.
Where I use biomechanic measurements, a lot is for screening. I ask every adult keratoconus patient to bring their children if they have any. I find it highly interesting to see an 8-year-old with quasi-normal tomography but with a highly relevant pathological CBI.
I think the next step might be Brillouin or OCT-based elastography and very soon we will do the first OCT elastography measurement in humans, which should be an order of magnitude more precise than a Corvis ST measurement.
Dr Kermani: With the increasing sensitivity of biomechanics, new devices could provide helpful screening in children and young adults, particularly those with keratoconus in the family or with allergies that are accompanied by chronic eye rubbing.