Risk of Normal Tension Glaucoma Progression From Automated Baseline Retinal-Vessel Caliber Analysis: A Prospective...

Area of Science:

• Ophthalmology
• Medical Imaging
• Glaucoma Research

Background:

• Normal tension glaucoma (NTG) is a progressive optic neuropathy.
• Assessing glaucoma progression risk is crucial for timely intervention.
• Retinal vasculature may serve as an indicator of glaucoma progression.

Purpose of the Study:

• To investigate the association between baseline retinal-vessel calibers, determined by deep learning, and the risk of NTG progression.
• To evaluate the predictive value of retinal vascular parameters for glaucoma development.

Main Methods:

• A prospective cohort study included 197 patients with NTG, followed for at least 24 months.
• Deep learning was used to compute central retinal arteriolar equivalent (CRAE) and central retinal venular equivalent (CRVE) from fundus photographs.
• Glaucoma progression was assessed by retinal nerve fiber layer (RNFL) thickness and visual field (VF) changes, analyzed using Cox proportional-hazards models.

Main Results:

• Over a mean follow-up of 34.36 months, 17.69% of eyes showed progressive RNFL thinning and 5.64% exhibited VF deterioration.
• Narrower baseline CRAE and CRVE were significantly associated with increased risk of RNFL thinning (HR per SD decrease: 1.36 and 1.35, respectively).
• Narrower baseline CRAE was also linked to an increased risk of VF deterioration (HR per SD decrease: 1.98).

Conclusions:

• A decrease in baseline CRAE or CRVE by one standard deviation was associated with over 30% increased risk of RNFL thinning and over 90% increased risk of VF deterioration.
• Baseline retinal vascular attenuation in NTG eyes predicts subsequent glaucoma progression.
• Deep learning-based retinal vasculature analysis shows clinical utility in NTG risk stratification.