Current Issue

---

In this week’s issue

• Experts created standardized grading scales to improve detection and management of ocular toxicities from emerging cancer therapies.
• ChatGPT o1 Pro performed comparably to glaucoma specialists in diagnosing glaucoma from visual field and OCT-RNFL data.
• Early-onset high myopia is strongly driven by identifiable genetic mutations, with sequencing and polygenic risk scores helping clarify its genetic basis.
• Impaired fasting glucose and diabetes are linked to significantly increased risk of third, fourth, and sixth cranial-nerve palsies.
• Use of carbonic anhydrase inhibitors in patients with IIH is associated with a significantly increased risk of kidney-stone formation.

https://open.spotify.com/episode/2IV8G6qk3rQw9JbWRjClNk?si=H3k-O3pMQuShQ9-Pq2cKXw 

JAMA Ophthalmology

Consensus on oncology drug-related ocular adverse event reporting

Let’s all get on the same page. Antibody-drug conjugates (ADC) are a class of targeted cancer therapies that consist of monoclonal antibodies. A common ocular adverse effect of ADCs is corneal pseudomicrocysts, which can lead to eye pain and blurred vision in over 90% of patients on ADCs. The current grading scales used for adverse ocular effects are ambiguous and lack standardization. In this study, ophthalmologists and oncologists from 11 academic centers and the FDA formed a task force to review the literature and develop a consensus on new ocular adverse effect grading scales, with emphasis on clear terminology, differentiating signs and symptoms, using photographs, and providing clear drug-dosing recommendations. Six adverse effect grading scales were developed: eye symptoms, visual acuity, cornea, conjunctiva/sclera, anterior chamber, and retina/posterior segment. The development of these grading scales should provide ophthalmologists and oncologists with a more objective and consistent measurement of ocular adverse effects. These new ocular adverse effect grading scales have facilitated safe and effective evaluation of eye toxicity from ADCs.

American Journal of Ophthalmology (AJO)

ChatGPT’s accuracy in glaucoma detection using VF and OCT

Can AI reliably interpret glaucoma testing the way specialists do? Recent advancements in artificial intelligence and large language models are opening new possibilities in ophthalmic diagnostics. Glaucoma assessment relies on careful interpretation of visual field (VF) testing and optical coherence tomography (OCT)-derived retinal nerve fiber layer (RNFL) measurements. This cross-sectional study enrolled 204 biobank participants across multiple ancestries and polygenic risk strata to evaluate whether a multimodal large language model, ChatGPT o1 Pro, can interpret VF and OCT-RNFL data and compare results with a masked glaucoma specialist consensus. ChatGPT demonstrated high diagnostic performance with 96% sensitivity, 83.7% specificity, and a 99% negative predictive value when identifying glaucoma from numeric and VF data. Accuracy and area under the receiver operator curve were similar across European and non-European groups and across high vs. low/intermediate polygenic risk, indicating no significant subgroup bias. These findings suggest that ChatGPT may diagnose glaucoma similarly to specialists using only VF and OCT data.

British Journal of Ophthalmology 

What genetic clues help identify early-onset high myopia?

Seeing myopia’s roots in our DNA. Early-onset high myopia (eo-HM), defined as onset before age 7 with binocular spherical equivalent refraction of less than -6.00 diopters, has been linked to genetic factors and carries a greater risk of lifelong visual morbidity. In this cross-sectional study of 100 patients with high myopia, researchers used whole-exome sequencing (WES) and polygenic risk scores (PRS) to map the mutation landscape of eo-HM. Five de novo variants (COL2A1, ARR3, GLRA2, ZEB1, HDAC8) and four novel candidate myopia genes (TCF7L2, AIPL1, INPP5E, SALL4) were uncovered. WES identified likely pathogenic variants in 35% of participants, but mutation detection was significantly higher in eo-HM compared with late-onset HM (56.5% vs 16.7%). Mutations in retina-related genes were significantly more frequent among eo-HM cases, and carriers demonstrated worse visual acuity and characteristic fundus pathology. PRS showed moderate ability to distinguish eo-HM from later-onset disease. Together, these findings expand the mutational spectrum of early-onset high myopia and support integrating WES with PRS to identify high-risk children earlier and guide targeted surveillance.

Eye
Glycemic status and risk of cranial nerve palsy

Your cranial nerves may be counting carbs! Individuals with diabetes commonly present with cranial nerve palsies (CNP). However, the gradient of risk for CNP with glycemic state, especially prediabetes, is poorly understood. This cohort study evaluated data from 4,067,842 adults across South Korea from 2009-2018 to determine the effect of fasting glucose status and duration of diabetes on incidence of third, fourth, and sixth CNP. Individuals were categorized as non-diabetic, impaired fasting glucose (IFG), newly diagnosed diabetes, diabetes <5 years, or diabetes ≥5 years. Worsening glycemic status showed a stepwise rise in CNP risk, from a modest increase in IFG to more than a 2.5-fold increase in long-standing diabetes (≥5 years). Adjusted hazard ratios were 1.098 for IFG, 1.779 for newly detected diabetes, 1.921 for diabetes <5 years, and 2.571 for diabetes ≥5 years. These results highlight that even mild hyperglycemia carries increased CNP risk and underscore the importance of early glycemic optimization in reducing microvascular cranial neuropathies.

Journal of Neuro-Ophthalmology

Kidney stone formation and idiopathic intracranial hypertension

Headaches or kidney stones? Carbonic anhydrase inhibitors are the mainstay of treatment for idiopathic intracranial hypertension (IIH), but growing evidence suggests they may come with an unwanted side effect: kidney stones. This study takes a closer look at that risk by examining a large retrospective cohort of patients with IIH who were either treated with CAIs or left untreated. A total of 23,182 treated patients were compared with 50,080 controls, and kidney stone rates were tracked over three intervals: within 1 year, 1-2 years, and 2-3 years. After matching both groups for baseline characteristics, CAI users consistently showed a higher risk of developing kidney stones. The risk was elevated at 1 year (RR 1.49), peaked at 1-2 years (RR 2.24), and remained higher at 2-3 years (RR 1.51). Patients taking at least 500 mg of acetazolamide showed a similar pattern, with a sustained increase in stone risk. Lastly, men were more susceptible, while Black or African American patients had comparatively lower risk. Overall, these findings suggest that while CAIs are effective for managing IIH, they may also raise the likelihood of kidney stone formation, and patients should be counseled on this side effect prior to starting this medication.  

Collaborative Normal Tension Glaucoma Study (CNTGS)

No pressure… but we may need to lower the pressure. Prior to the CNTGS, there was no consensus regarding the treatment of normal tension glaucoma - glaucomatous optic nerve damage in the setting of normal range intraocular pressures (IOP). In this 1998 study, patients with normal tension glaucoma were randomized into a treatment cohort (pressure lowered by >30%) and an untreated, control cohort to determine if lowering IOP played a protective role in patients with normal tension glaucoma.

Key Points

• At follow-up, mean IOP was significantly lower in the treatment group compared to the control (10.6 mm Hg vs 16.0 mm Hg, P < 0.0001)
• When using the defined endpoints of optic disc changes or visual field loss, survival analysis showed significantly greater progression of glaucomatous disease in the untreated control (P < 0.0001)
• Cataracts developed more often in the treatment group, likely due to surgical intervention (38% vs 14%, P = 0.0011)
• A follow-up, intention-to-treat analysis initially found no difference in visual field survival analysis outcomes between cohorts, but after excluding patients with cataracts (i.e., a treatable problem), they found favorable outcomes for the treatment group (P = 0.0018)

Overall, the CNTGS is a landmark study because it highlights the role of pressure-lowering therapy in preventing disease progression of normal tension glaucoma. Importantly, the visual field benefits were only identified once excluding patients for visual changes secondary to cataract formation. Still, the clinical relevance of the CNTGS showed that treating IOP, even in patients with normal IOP, reduces glaucomatous visual field damage in patients with NTG.