Wang Sophia Y

Majid Iyad, Ruan Franklin Y, Wang Sophia Y

GPT-5 excelled at extracting clinical data from glaucoma notes, outperforming LLaMA and Mistral, showing its potential to improve ophthalmology data management and research.

Parekh Bela J, Stein Joshua D, Abuzaitoun Rebhi, Andrews Chris A, Branham Kari, Milentijevic Dejan et al.

This study found RPGR-RD patients experience progressive vision loss, with rapid visual field decline in rod-cone phenotype suggesting early intervention is crucial to preserve sight.

Zhao Yihan, Ravindranath Rohith, Hernandez-Boussard Tina, Stein Joshua D, Wang Sophia Y

Researchers studied bias mitigation in AI models predicting glaucoma surgery progression. Inprocessing methods consistently improved fairness across diverse patient groups, offering a balanced approach for clinical AI in ophthalmology.

Chan Fountane, Lin Wei-Chun, Tang Alan, Xu Benjamin Y, Wang Sophia Y, Boland Michael V et al.

Researchers developed methods to identify Normal Tension Glaucoma from EHRs. NLP-enhanced models outperformed ICD codes and structured data alone, significantly improving NTG case detection, which is crucial for research and clinical care.

Hallaj Shahin, Halfpenny William, Radgoudarzi Niloofar, Boland Michael V, Swaminathan Swarup S, Wang Sophia Y et al.

This study found most visual field test data elements are missing from standard medical terminologies, hindering data sharing. New concepts were proposed to LOINC to improve interoperability and data standards.

Koornwinder Abigail, Zhang Youchen, Ravindranath Rohith, Chang Robert T, Bernstein Isaac A, Wang Sophia Y

AI models combining EHR and RNFL OCT data best predicted glaucoma surgery, outperforming single data sources. This offers a powerful tool for personalized glaucoma progression prediction and clinical decision support.

Chen Maggie S, Ravindranath Rohith, Chang Robert, Zhou Yukun, Keane Pearse A, Wang Sophia Y

RETFound, an AI model, accurately predicted optic nerve cup-to-disc ratio and average RNFL thickness from fundus photos. This demonstrates its potential for glaucoma screening and monitoring, even without specific prior training.

Ravindranath Rohith, Naor Joel, Wang Sophia Y

AI models successfully used self-reported health data to predict glaucoma risk (AUROC up to 0.890). This could enable prescreening in low-resource settings, guiding referrals for specialized eye exams.

Ravindranath Rohith, Wang Sophia Y

AI models using EHR data effectively identified glaucoma patients without ophthalmic imaging, showing promise for automated, targeted screening referrals in primary care settings.

Radgoudarzi Niloofar, Hallaj Shahin, Boland Michael V, Stagg Brian, Wang Sophia Y, Xu Benjamin et al.

Kara Su, Yang Michael, Yeh Hsu-Hang, Sen Simmi, Hwang Hannah H, Wang Sophia Y

Deep learning models accurately identified MIGS procedures and precisely located the trabecular meshwork in surgical videos. This technology can enhance surgical training and patient care by providing automated feedback.

Ravindranath Rohith, Stein Joshua D, Hernandez-Boussard Tina, Fisher A Caroline, Wang Sophia Y

AI models predicting glaucoma surgery progression showed bias by sex, race, and ethnicity. Fairness varied by test population, highlighting the need for rigorous, population-specific evaluation before clinical use.

Barry Samuel, Wang Sophia Y

Machine learning models predicted glaucoma surgery outcomes from EHR data with 75.5% accuracy, especially for IOP. This could personalize treatment plans for patients.

Wang Sophia Y, Ravindranath Rohith, Stein Joshua D

Machine learning models, trained on multicenter EHR data, can reasonably predict glaucoma surgery need within a year, showing good generalizability, which could aid clinical decision-making.

Sun Michelle T, Pershing Suzann, Goldberg Jeffrey L, Wang Sophia Y

This study found Type 2 diabetes, especially with higher HbA1c and insulin use, increases the risk of glaucoma surgery in POAG patients, highlighting the importance of diabetes management.

Bernstein Isaac A, Fisher Ann Caroline, Singh Kuldev, Wang Sophia Y

This study found visual field loss is associated with higher odds of frailty in US adults, suggesting a potential link where frailty increases VFL risk or VFL contributes to frailty.

Tao Shiqi, Ravindranath Rohith, Wang Sophia Y

AI survival models, especially deep learning, accurately predict glaucoma progression to surgery using EHR data. This can help identify high-risk patients for earlier intervention.

Macri Carmelo, Wong Christopher X, Tu Samuel J, Casson Robert, Singh Kuldev, Wang Sophia Y et al.

A large study found higher systolic and pulse pressures increase primary open-angle glaucoma risk, suggesting blood pressure management may be clinically relevant for prevention.

Sun Michelle T, Singh Kuldev, Wang Sophia Y

This EHR study found glaucoma tube shunts failed more often than trabeculectomy. This suggests trabeculectomy may offer better long-term IOP control for suitable patients.

Sun Michelle T, Singh Kuldev, Wang Sophia Y

This study found that glaucoma management changes, especially medication adjustments, surgery, and laser, are more likely after visual field or OCT testing, highlighting their role in clinical decision-making.

Hu Wendeng, Wang Sophia Y

Transformer models successfully predicted glaucoma surgery progression from clinical notes, outperforming human review. This offers a promising AI tool to help clinicians tailor treatment for high-risk patients.

Wang Sophia Y, Tseng Benjamin, Hernandez-Boussard Tina

Deep learning models using both structured EHR data and free-text notes better predicted glaucoma surgery than structured data alone, highlighting the value of clinical notes for progression prediction.

Lee Eric B, Hu Wendeng, Singh Kuldev, Wang Sophia Y

This study found low blood pressure, not high BP or its medications, increases open-angle glaucoma risk, highlighting vascular dysregulation's role in glaucoma development.

Wang Sophia Y, Singh Kuldev

This review highlights that glaucoma progressing at low IOP requires confirming progression and further IOP lowering, often to single digits via trabeculectomy, for effective management.

Wang Sophia Y, Singh Kuldev, Stein Joshua D, Chang Robert T

This study found iStent implantation with cataract surgery significantly reduced glaucoma medication use compared to cataract surgery alone. This lessens medication burden, potentially improving patient adherence and outcomes.