Ultrastructural Basis Of Ectasia Stabilisation In Corneal Cross-Linking And Implications For The Optimisation Of Lenticule Insertion Methods.

The clinical management of keratoconus is increasingly broadening, from collagen cross-linking (CXL) to new lenticule-based additive procedures. While CXL is proven to halt disease progression, it remains unclear whether lenticule addition alone provides equivalent biomechanical stabilisation due to insufficient long-term data. However, it has proven to be effective in the visual rehabilitation of patients, even in cases of high refractive errors, unlike CXL. Here we characterise structural determinants of corneal biomechanics - specifically pressure-dependent recruitment of collagen ‘crimp’ (waviness) - to evaluate how these therapies can be optimised to halt disease progression while ensuring optimised refractive outcomes. 

Crosslinking, imaging, and analysis was conducted by Maryame Boutkhil, Dale Moulding, Abby Wilson and John Marshall at UCL Mechanical Engineering and the UCL Institute of Child Health (ICH) facilities. All lenticule addition surgeries were performed by Marwan Ghabra at The Harley Street Eye Centre in London. 

Corneas were imaged in the native state and after collagen CXL(n=10). CXL was performed using a high energy (14.4J/cm²) to account for the increased corneal porcine thickness. Additional corneas were imaged following delamination(n=2),pocket-based(n=2), and flap-based(n=2) procedures. All samples were imaged at increasing intraocular pressures from 5 to 45mmHg (5 mmHg increments).SHG imaging was performed using a Zeiss LSM980(820nm femtosecond laser,20×water-immersion objective, NA1.0) with forward signal acquisition.A custom designed chamber maintained corneal curvature, hydration and controlled IOP via PBS infusion and real-time pressure-sensing. Collagen recruitment was quantified by calculating the straightening rate across pressures. 

Collagen structure in native corneas varies with depth: ~78% of the stroma shows collagen crimp, with a period of crimp ranging ~17-24 µm, and the final 2% of the stroma is highly tortuous (period of the crimp~9 µm). With increasing IOP, mid-stromal straightening of collagen crimp occurred non-linearly, increasing by 2.38% per mmHg at 5-10 mmHg vs 0.11% per mmHg at 20-25 mmHg(p < 0.05), corresponding to a 21-fold higher straightening rate at low pressures. CXL corneas showed minimal change, with only a 7.5-fold fiber recruitment from low to high pressures. Pocket and delamination insertions preserved host fibre recruitment, whereas flap procedures disrupted it by releasing flap tension and over-straightening the residual stromal bed. 

Although CXL does not visibly change the cornea’s static structure, it changes tissue response to load. CXL works by creating covalent bonds between collagen fibres, which suppresses fibre straightening under load. These findings provide a structural basis for optimising different keratoconus treatments. Contrary to flap-based insertion techniques, pocket-based and delamination-based lenticule  insertion procedures better preserve the native stromal architecture and biomechanical integrity. However, alone it may not sufficiently reduce collagen fibre recruitment to stop disease progression. Crosslinking the corneas after lenticule surgeries suppresses recruitment and contributes to better mechanical stabilisation.