Treating Ectasia Without Keratoplasty

Can ophthalmic innovation make penetrating keratoplasty for keratoconus obsolete? A Clinical Research Symposium at the 2026 ESCRS Annual Congress examined several promising techniques that could make it possible.

Farhad Hafezi MD, PhD discussed a technique he has developed known as epithelial map-guided phototherapeutic keratectomy combined with customised epi-on cross-linking, or ELZA-PACE.

ELZA-PACE for treating keratoconus or post-refractive ectasia begins with anterior segment OCT epithelial mapping and topography/tomography. The surgeon then performs a phototherapeutic keratectomy to remove a limited epithelial area over the cone, followed by collagen cross-linking (CXL) with riboflavin and customised UV-A irradiation.

Understanding the early clinical results

“To date, we have performed ELZA-PACE in almost 300 eyes at the ELZA Institute in Zurich, Switzerland, of which more than 100 have now completed one-year follow-up,” Dr Hafezi reported. “A consistent pattern has emerged: cone-centred flattening is apparent immediately after the procedure (the photochemical effect is essentially immediate) and continues as the epithelium remodels, with topographic stability typically reached by month six.”

A flattening of 4–5 D within the first month is common and, because central flattening is accompanied by compensatory mid-peripheral steepening (the ‘coupling’ effect), the overall reduction in corneal asymmetry is often even greater, typically 8–12 D and up to 20 D in selected cases, he noted.

Functionally, this regularisation translates into a reduction of vertical coma and gains in corrected distance visual acuity (CDVA) over 6–12 months. In one representative case, CDVA improved from 0.4 to 1.0 logMAR with glasses within a week of treatment. Crucially, this was achieved without removing any stromal tissue, preserving the corneal thickness profile. ELZA-PACE therefore regularises the cornea and improves vision, rather than merely stabilising the disease, he explained.

Noting the procedure’s progression

ELZA-PACE is the result of more than two decades of CXL research. The original Dresden protocol (epi-off, 3.0 mW/cm² for 30 minutes, 5.4 J/cm²) normally halts ectasia progression and typically does little to correct the irregularities on the corneal surface. Epi-on approaches initially underperformed because the epithelium absorbs riboflavin and UV light while consuming oxygen—the rate-limiting component of the photochemical reaction. Higher fluences, pulsed UV delivery, and penetration enhancers have since closed the efficacy gap between epi-on and epi-off CXL.

The first generation of customised CXL used a topography-guided, eye-tracking UV source to deliver more energy over the cone apex; this improved flattening to roughly 3–4 D but remained moderately predictable. ELZA-PACE is the second generation: instead of shaping only the UV beam, it also models oxygen availability and riboflavin distribution. An epithelial map-guided PTK removes the epithelial cells over the cone, so the three overlapping gradients—riboflavin concentration, oxygen availability, and UV fluence—all reinforce each other at the cone apex.

This also distinguishes ELZA-PACE from excimer-plus CXL protocols (Athens, Cretan, STARE-X, TREK) that achieve regularisation through stromal ablation. In ELZA-PACE, the stroma remains untouched, with the entire effect coming from cross-linking. That tissue preservation matters particularly in younger patients.

Establishing the CXL part of the treatment

In contrast to cross-linking, the PTK step is deliberately minimal. A custom-offset, custom-sized ablation zone, typically oval and a few millimetres across, is placed over the cone tip region and removes only the local epithelial thickness, without any stromal tissue removal.

“We then saturate the cornea with a penetration-enhancer-assisted, hypo-osmolar 0.1% riboflavin solution,” Dr Hafezi explained. “Over the denuded cone, the stroma undergoes rapid saturation. Under the intact peripheral epithelium, penetration is slower, and this creates a riboflavin gradient. Because the epithelium also consumes oxygen and absorbs UV light, oxygen and UV gradients form in parallel, all centred on the epi-off window over the cone.”

UV-A delivery uses the C-eye device (EMAGine AG, Switzerland) through an accelerated, pulsed protocol at 18 mW/cm² across the whole treatment zone. The protocol is followed by an additional fluence boost confined to the central cone region, resulting in a considerably higher cumulative fluence at the cone tip than in the periphery. Cross-linking is therefore strongest where the cornea is weakest: the cone flattens, the mid-periphery steepens in compensation, and the cornea becomes more regular—all photochemically, without ablating stroma, he said.

Farhad Hafezi MD, PhD, FARVO is Medical Director at the ELZA Institute of Zurich, Switzerland, and Professor at the Universities of Geneva, Switzerland; New York, US; Los Angeles, US; and Wenzhou, China. farhad@hafezi.ch.