Glaucoma Field Defect Analyser (GFDC)

Decision making in glaucoma often relies on subjective interpretation of perimetry data. For instance, registering blind and sight impaired patients is a clinically decision that is officially guided by visual acuity and perimetry, but ophthalmologists often disagree on the severity of visual field defects and patients miss out on care and support as a result.

GFDC (https://gfdc.app) is a computer vision algorithm that applied validated Hodapp-Parrish-Anderson criteria automatically to grade glaucomatous field defects as mild, moderate, or severe (or 'no defect'). In our validation study, GFDC exhibited 100% accuracy (npj Digital Medicine). We are now exploring clinical applications of GFDC to facilitate and improve clinical and research work involving glaucoma patients.

Certification of visual impairment

In the UK, certification of visual impairment (CVI) is a necessary precursor for sight impaired ('partially sighted') and severely sight impaired ('blind') patients to access social support and benefits from local authorities. Certification is based on visual acuity and perimetry parameters but is ultimately a clinical decision requiring referral by a consultant ophthalmologist. Many patients miss out on registration due to arbitrary factors and the subjectivity of assessment.

We used GFDC to facilitate semi-automated screening of glaucoma patients for CVI eligibility by non-specialists. Missed registration was worryingly frequent and occurred for reasons including administrative failure, frailty, and mental health diagnoses. Our full report is available in the British Journal of Ophthalmology and was awarded the 2025 Best Paper Prize by the Royal Society of Medicine Ophthalmology Section.

GFDC can be used as a free tool to prompt CVI referral, with or without a consultant ophthalmologist in the loop. This could improve the accuracy and fairness in provision of support for severely sight impaired patients.

Page updated

Google Sites

Report abuse