May 18, 2022

The most important ophthalmology research updates, delivered directly to you.

In this week's issue

  • Artificial intelligence can diagnose nuclear sclerotic and cortical cataracts similar to an ophthalmologist
  • Targeting KAT5, a lysine acetyltransferase, may help alleviate inflammation in allergic conjunctivitis
  • Primary rhegmatogenous retinal detachments were more likely to present as mac-off during the pandemic as compared to the year prior to March 2021
  • Cystatin C among potential biomarkers was found to exhibit sufficient discriminatory power for detecting sight-threatening diabetic retinopathy

Do you despise spring skies when they cause itchy eyes?


Can we identify a new target that could bring relief to those pesky allergies? Allergic conjunctivitis (AC) is a response to allergens contacting the eye, causing vascular permeability and infiltration of inflammatory cells into the conjunctiva. Clinical symptoms of AC include ocular pruritus, redness, and watery discharge. In this study, Luo et al. explore the role of KAT5, a lysine acetyltransferase known to have a role in inflammation in AC. An experimental mouse model of AC (EAC) was developed by sensitizing female BalB/c mice to ragweed pollen, and severity of EAC was evaluated with a clinical score and measures of vascular permeability and inflammatory markers. In EAC mice, KAT5 overexpression promoted more inflammation through increased CD11c+ dendritic cells and CD4+ T cell infiltration into the conjunctiva. Treatment with a KAT5 inhibitor (NU9065) was shown to attenuate this inflammation. The data from this study support that inhibition of KAT5 has a protective effect on ocular inflammation in EAC mice, and indicate that KAT5 may be a promising therapeutic target for AC prevention and treatment in future studies.

Automated cataract diagnosis? The prospect of deep learning AI


Building better diagnosticians... literally. A new study suggests artificial intelligence can diagnose and classify age-related cataracts as accurately as trained ophthalmologists, if not better. Detecting cataract progression and learning to attribute symptoms to cataract as opposed to coexisting conditions is a difficult skill to master. Researchers designed and tested deep learning models to diagnose and classify age-related cataract in a quantitative way that resembles the methodology used by human experts. DeepLensNet was trained to detect the three most common types of cataracts using 18,999 images (6333 triplets) from the Age-Related Eye Disease Study dataset and classify its severity. The machine’s performance was then compared to 14 ophthalmologists. DeepLensNet diagnosed and classified nuclear sclerotic and cortical cataracts with significantly superior accuracy than ophthalmologists. For posterior subcapsular cataracts, their accuracy was similar. Performance in the artificial intelligence group was not affected by pupil size, whereas the performance of ophthalmologists seemed to be worse in smaller pupils. This research suggests artificial intelligence could play a substantial role in screening and improving accessibility of cataract diagnosis without pupil dilation. However, this study was limited in that it could not assess the severity by which the cataracts were affecting vision.

Primary rhegmatogenous retinal detachments: a pandemic hurdle

American Journal of Ophthalmology

Social detachment and retinal detachments both worse during the same year? Primary rhegmatogenous retinal detachment (RRD) is an ophthalmic emergency characterized by a full-thickness retinal break from vitreous traction with vitreous fluid accumulating in the subretinal space. It is classified as macula-sparing (mac-on) versus macula-involving (mac-off), with mac-off yielding a poorer visual prognosis. This single-center, retrospective cohort study investigated changes in clinical trends for primary RRD presentation over the first full year of the COVID-19 pandemic. The pandemic cohort for this study consisted of 952 patients (March 9, 2020 to March 7, 2021), with 872 patients as controls (March 11, 2019 to March 8, 2020). A total of 60.92% of pandemic cohort patients presented with mac-off RRDs compared to 48.17% of control patients. Pandemic cohort patients exhibited significantly worse final best-corrected visual acuity (0.30 logarithm of the minimum angle of resolution) compared with the control cohort (0.18 logarithm of the minimum angle of resolution). The COVID-19 pandemic negatively impacted the clinical course of patients with primary RRD. With fluctuating numbers of COVID-19 cases, this devastating change in RRD trends will hopefully be lessened.

Diagnostic biomarkers in sight-threatening diabetic retinopathy

JAMA Ophthalmology

Could biomarkers tell us which diabetics will develop visual impairment? There are 537 million people in the world with diabetes, many of whom live in resource-restricted low-income countries. In these challenging situations, a biomarker to help detect which patient is at highest risk of developing sight-threatening diabetic retinopathy (STDR) would help efficiently allocate resources like retinal screenings, provide healthcare to those who need it most, and improve visual outcomes. A cross-sectional multicenter study was performed using clinical data from 538 participants in India and the UK between 2018-2021 analyzing whether 13 previously verified biomarkers could distinguish STDR from no diabetic retinopathy in patients outside of age, disease, duration, ethnicity, and HbA1c. Cystatin C, age, disease duration, ethnicity (in the UK), and HbA1c exhibited near-acceptable discrimination power for detecting STDR in the UK and India (AUC 0.0779, 215 patient in the UK; AUC 0.696, 208 patients in India). Cystatin C, with other common patient parameters, was found to exhibit sufficient discriminatory power for detecting STDR and could be a practical, cost-efficient tool in triaging diabetic patient for retinal screening, if validated with further studies in the future. 


Risk factors for deterioration of PACD following lens extraction 

Journal of Clinical Medicine

What risk factors in primary angle closure lead to postoperative deterioration following lens extraction? Lens extraction is commonly used in primary angle closure (PAC), primary angle closure suspect (PACS), and primary angle closure glaucoma (PACG) as an effective treatment for primary angle closure disease (PACD). Researchers analyzed the post-operative outcomes through retinal nerve fiber layer (RNFL) thickness and visual field measurements in patients who underwent lens extraction surgery. After a mean postoperative follow up period of 3.5 years, 17% of patients with PACS/PAC experienced structural progression (decrease in RNFL thickness) and no patients experienced visual field progression. 66.7% of patients in the PACG group experienced structural progression postoperatively and 33% displayed visual field progression. In all groups (PACS/PAC/PACG), no risk factors were identified for structural progression. The risk factors for visual field progression in PACG included high postoperative IOP fluctuations and thinner preoperative RNFL thickness. Limitations of this study include small sample size (77) and length of follow up to detect visual field progression (mean 3.5 years). 

Question of the Week

A 39-year-old woman presents with a "droopy left eyelid and double vision that gets progressively worse late in the day". On further examination, her diplopia goes away as her ptosis gets worse.
Which of the following is the most appropriate next step?

A. Ptosis surgery
B. Pupillary tests
C. Ice test
D. CT scan to rule out CVA

Bonus question: Why did her diplopia resolve as her ptosis got worse?
Keep scrolling for answer or click here

Helpful Links

Quiz Answer: C
Quiz Answer Explained

Please note that there was an error in last week's question of the week. Anti-VEGF therapy was included as part of the treatment for a patient with VKH. The imaging findings did not show clear CNV and therefore anti-VEGF therapy would likely not have been necessary in that patient.
Twitter Twitter
Website Website
Share with a friend Share with a friend
Copyright © 2022 The Lens Newsletter LLC, All rights reserved.