Validation and application of a Novel 24-2 Rapid Protocol of Melbourne Rapid Fields AI-Enabled Online Perimeter in Glaucoma Screening
Kong G1,2, Sugihara K3, Thapa B4, Hosokawa M3, Okanouchi T3, Thapa S4, Vingrys A3,5
1Royal Victorian Eye And Ear Hospital, 2Centre of Eye Research Australia, Department of Ophthalmology, 3Department of Ophthalmology, Kurashiki Medical Center, 4Nepal Eye Hospital, 5Department of Optometry and Vision Sciences, The University of Melbourne
Biography:
A/Prof George Kong is a Melbourne-based glaucoma specialist and clinician-scientist focused on glaucoma treatment, digital innovation, and AI in healthcare. He is a consultant ophthalmologist at The Royal Victorian Eye and Ear Hospital and an Associate Professor at The University of Melbourne. He developed Melbourne Rapid Fields (MRF), an AI-enabled visual field testing system for glaucoma detection. He is committed to advancing safe, accessible, and evidence-based digital solutions in ophthalmology.
Abstract:
Purpose: The Melbourne Rapid Fields (MRF) AI-enabled web-browser perimetry software allows white-on-white increment threshold perimetry on any computer. This study evaluates the MRF 24-2 Rapid protocol (MRF24-2R) against the Humphrey Field Analyzer (HFA) for glaucoma screening.
Methods: MRF24-2R results were compared to HFA 24-2 SITA-Standard (HFA-SS) in 92 patients and HFA 24-2 SITA-Faster (HFA-SFR) in 131 patients for validation. Test-retest reliability and Coefficient of Reproducibility were assessed. MRF24-2R was later used in glaucoma screening in Nepal with 280 eyes screened.
Results: MRF24-2R showed high agreement for Mean Deviation (MD) with HFA-SS (ICC = 0.96) and HFA-SFR (ICC = 0.95). Bland-Altman analysis found a bias of -0.69 dB for MD and +0.89 dB for PSD. MRF24-2 showed excellent repeatability at retest for MD (ICC = 0.98). MRF24-2R was significantly faster than HFA-SS (170 ± 42 seconds vs. 346 ± 75 seconds, p<0.001) and marginally faster than HFA-SFR (192 ± 53 seconds, p=0.38). In Nepal, 70% of eyes were found to be normal/pre-perimetric, while 30% were noted to be abnormal (21% had early defect, 6% had moderate defect, and 3% had advanced defects). The screening in Nepal was conducted efficiently, providing timely results for glaucoma detection.
Conclusion: MRF24-2R provides a fast, portable, and reliable visual field test suitable for glaucoma screening and clinical management.
