Skip to main content
Skip to main content
Back to Grand Rounds
Grand RoundsWeekly Evidence Brief

Ophthalmology

Edition

30-Second Takeaway

  • A glaucoma PRS stratifies lifetime risk and predicts need for intensified therapy.
  • Microperimetry detects more functional responders than LLVA in RPGR gene therapy trials.

Week ending June 13, 2026

Five recent studies with clinical implications for glaucoma risk stratification, vision-CVD links, AI ultrasound screening, GA progression metrics, and RPGR trial endpoints

Glaucoma polygenic risk score stratifies lifetime risk and predicts treatment escalation

OPHTHALMOLOGYJun 12, 2026

In 402,739 FinnGen participants, a top-performing glaucoma PRS conferred HR 3.32 for glaucoma (≥90th vs 20–80th percentile). Lifetime glaucoma risk rose from 2.5% (<1st percentile) to 31.9% (95–99th percentile) and 45.3% (≥99th percentile) by age 85. PRS and family history independently contributed to risk; PRS effect size changed minimally after adjusting for family history. Those in the highest PRS decile had greater 20-year cumulative rates of medication escalation (60.6% vs 38.4%), laser, and surgery.

Visual impairment correlates with diabetes and cumulative cardiovascular risk in US Hispanic/Latino adults

JAMA NETWORK OPENJun 12, 2026

In 3,288 Hispanic/Latino adults aged ≥40, age-standardized habitual VI prevalence was about 7.1%. Diabetes was associated with best-corrected VI (OR 4.65), and higher cumulative CVD risk factor counts showed graded associations with VI. VI prevalence varied by Hispanic/Latino background and measurement method, with higher rates in some subgroups. Cross-sectional design limits causal inference; temporal and confounding mechanisms remain unresolved.

SonoEye ultrasound foundation model achieves high screening sensitivity and Eye-RADS stratification

NPJ DIGITAL MEDICINEJun 12, 2026

SonoEye, trained on 215,356 image-text pairs, achieved 98.3% screening sensitivity and mean accuracy 96.3% across 18 eye diseases. The Eye-RADS four-tier framework produced strong internal agreement (Cohen's kappa 0.808) and moderate external agreement. Incorporating age improved performance in elderly populations but not younger cohorts. Model outputs include attention-based visualizations and structured reports to aid interpretability.

References

Numbered in order of appearance. Click any reference to view details.

Additional Reads

Optional additional studies from this edition.

Edition context

Clinical signal

  • Consider PRS-informed risk discussion and earlier surveillance for patients in high PRS percentiles.
  • Treat cross-sectional VI–cardiovascular associations as hypothesis-generating, not causal.
  • Validate SonoEye locally before deployment; monitor elderly-specific performance differences.