PRecision Intraocular Surgery through real-time phototoxicity dosimetry, Mitochondrial stress profiling, and AI-based risk assessment

Intraocular surgery relies on intense focal illumination to visualise delicate micrometre-scale ocular structures, including corneal endothelium, internal limiting membrane, and neurosensory retina; yet surgical light itself is a recognised source of iatrogenic damage. Current safety standards rely on static exposure thresholds that ignore patient-specific vulnerability, surgical geometry, dye-assisted procedures, and cumulative exposure dynamics. Although phototoxic damage is a known potential complication of intraoperative light exposure, beyond these overt phototoxic events the exposure thresholds capable of eliciting subclinical cellular and functional damage remain largely undefined. Consequently, surgeons remain without real-time information on the effective phototoxic burden reaching ocular tissues. PRISM addresses this unmet clinical need through a translational framework integrating optical engineering, mitochondrial biology, artificial intelligence, and clinical ophthalmology. The project investigates whether clinically relevant sub-toxic surgical illumination induces early mitochondrial and oxidative stress responses before tissue damage becomes detectable. PRISM will develop an integrated optoelectronic sensing module prototype for ophthalmic illuminators enabling real time reconstruction of intraoperative photonic exposure parameters. Experimentally reconstructed surgical illumination paradigms, including filters, dark intervals, and dye-assisted conditions, will be reproduced in automated biological platforms using ocular cell models and ex vivo tissues to identify early functional and inflammatory signatures associated with distinct exposure conditions. A prospective observational clinical study on corneal endothelial transplantation and macular surgery will systematically record all relevant surgical parameters, including light-related exposure variables, and prospectively follow patients to detect structural and functional signs compatible with phototoxic damage. The photonic and operative variables are systematically recorded in the observational cohort and, in parallel, experimentally reproduced in WP3 to identify which combinations are associated with early mitochondrial and oxidative stress. Convergence between the two levels is therefore established at the level of exposure conditions rather than biological endpoints, and the resulting integrated evidence will inform biologically grounded AI models that combine operative exposure conditions, experimentally validated stress-inducing paradigms, and clinical outcomes to generate patient-specific biologically informed exposure-risk modelling. By transforming surgical light from an empirical parameter into a measurable and biologically interpretable variable, PRISM aims to establish the foundations for safer and more personalized intraocular surgery.