Belghith Akram

Chuter Benton, Joshi Vedant, Hallaj Shahin, Walker Evan, Bowd Christopher, Belghith Akram et al.

A novel AI model for glaucoma detection showed multimodal imaging (CFP + OCT) was better than CFP alone, but not significantly better than OCT alone, suggesting OCT-based AI may suffice clinically.

Jalili Jalil, Huynh Justin, Walker Evan, Chuter Benton Gabriel, Bowd Christopher, Heinke Anna et al.

Fine-tuned vision-language models effectively detected glaucoma and predicted visual field loss from OCT images, showing promise for scalable AI decision support in glaucoma care, matching or exceeding specialized models.

Belghith Akram, Bowd Christopher, Weinreb Robert N, Jalili Jalil, Christopher Mark, Zangwill Linda M

3D deep learning models accurately detect glaucoma, outperforming GCIPL thickness. Cross-domain models, using synthetic data, perform similarly, suggesting broader applicability across different OCT devices.

Jiravarnsirikul Anuwat, Belghith Akram, Rezapour Jasmin, Micheletti Eleonora, Nishida Takashi, Moghimi Sasan et al.

This study found larger choriocapillaris dropout near the optic nerve is significantly linked to central visual field damage in glaucoma patients, regardless of myopia, which could help identify those at higher risk.

Jiravarnsirikul Anuwat, Belghith Akram, Rezapour Jasmin, Bowd Christopher, Moghimi Sasan, Jonas Jost B et al.

Diagnosing glaucoma in myopic eyes is challenging due to structural and visual field confounds, requiring careful interpretation of standard diagnostic tests for effective patient management.

Jalili Jalil, Jiravarnsirikul Anuwat, Bowd Christopher, Chuter Benton, Belghith Akram, Goldbaum Michael H et al.

GPT-4V analyzed fundus images for glaucoma detection and features. It showed promise, with variable agreement to experts, suggesting potential for screening but needing refinement for clinical use.

Jiravarnsirikul Anuwat, Belghith Akram, Rezapour Jasmin, Micheletti Eleonora, Nishida Takashi, Moghimi Sasan et al.

This study found that worsening choriocapillaris dropout correlates with progressive nerve damage in myopic glaucoma. OCT-A monitoring of MvD may help track glaucoma progression, especially in myopic patients.

Bowd Christopher, Belghith Akram, Rezapour Jasmin, Jonas Jost B, Hyman Leslie, Weinreb Robert N et al.

Wide-field OCT, combining optic nerve and macula scans, better detected retinal nerve fiber layer thinning rates in progressing glaucoma, improving monitoring for patients with or without high myopia.

Rezapour Jasmin, Walker Evan, Belghith Akram, Bowd Christopher, Fazio Massimo A, Jiravarnsirikul Anuwat et al.

This study found OCT parameters effectively detect glaucoma, but accuracy varies with myopia. Macula parameters (GCIPL) showed high diagnostic promise for glaucoma in highly myopic eyes.

Micheletti Eleonora, El-Nimri Nevin, Nishida Takashi, Moghimi Sasan, Rezapour Jasmin, Fazio Massimo A et al.

This study found choroidal microvasculature dropout (MvD) is more common and larger in myopic glaucoma eyes, correlating with worse central visual field damage. This highlights MvD's role in glaucoma progression, especially in myopia.

Chuter Benton, Huynh Justin, Bowd Christopher, Walker Evan, Rezapour Jasmin, Brye Nicole et al.

A deep learning model accurately assessed fundus photo quality, and filtering out low-quality images significantly improved automated glaucoma detection, enhancing efficiency and accuracy in clinical screening.

Bowd Christopher, Belghith Akram, Rezapour Jasmin, Christopher Mark, Jonas Jost B, Hyman Leslie et al.

A deep learning model combining OCT scans and texture data better diagnosed glaucoma, especially in highly myopic eyes, offering improved diagnostic accuracy.

Fan Rui, Alipour Kamran, Bowd Christopher, Christopher Mark, Brye Nicole, Proudfoot James A et al.

A Vision Transformer (DeiT) detected glaucoma from fundus photos as well as ResNet-50 on internal data, but significantly better on external datasets, improving generalizability for clinical diagnosis.

Rezapour Jasmin, Bowd Christopher, Dohleman Jade, Belghith Akram, Proudfoot James A, Christopher Mark et al.

This study found macular GCIPL and GCC thickness correlate with glaucoma severity, not axial length, making them useful for monitoring glaucoma in highly myopic eyes.

Bowd Christopher, Belghith Akram, Rezapour Jasmin, Christopher Mark, Hyman Leslie, Jonas Jost B et al.

A novel OCT texture analysis (SALSA-Texture) improved glaucoma detection in high myopes compared to traditional thickness maps, offering a promising tool where standard segmentation is challenging.

Fan Rui, Bowd Christopher, Christopher Mark, Brye Nicole, Proudfoot James A, Rezapour Jasmin et al.

Deep learning accurately detected glaucoma from fundus photos, sometimes earlier than expert committees, suggesting potential for automated, standardized glaucoma detection in clinical trials and practice.

Rezapour Jasmin, Proudfoot James A, Bowd Christopher, Dohleman Jade, Christopher Mark, Belghith Akram et al.

This study found OCT BMO detection is significantly less accurate in highly myopic eyes, regardless of glaucoma status. Clinically, careful review and manual correction of BMOs are crucial for managing high myopes.

Bowd Christopher, Belghith Akram, Zangwill Linda M, Christopher Mark, Goldbaum Michael H, Fan Rui et al.

Deep learning analysis of OCTA images significantly improved glaucoma diagnosis compared to traditional methods using OCTA measurements or OCT nerve fiber layer thickness.

Proudfoot James A, Zangwill Linda M, Moghimi Sasan, Bowd Christopher, Saunders Luke J, Hou Huiyuan et al.

The STAGE model showed frequent testing (VF/OCT) can detect glaucoma progression within 18 months with <300 patients, enabling shorter, more efficient clinical trials for new therapies.

Bowd Christopher, Belghith Akram, Christopher Mark, Goldbaum Michael H, Fazio Massimo A, Girkin Christopher A et al.

Deep learning-identified, eye-specific OCT regions of interest better detect glaucoma progression than standard measurements, potentially improving early clinical monitoring.

Christopher Mark, Bowd Christopher, Proudfoot James A, Belghith Akram, Goldbaum Michael H, Rezapour Jasmin et al.

Deep learning accurately estimated visual field loss (10-2, 24-2) from macular OCT thickness maps, outperforming traditional methods. This can help clinicians personalize visual field testing.

Girkin Christopher A, Belghith Akram, Bowd Christopher, Medeiros Felipe A, Weinreb Robert N, Liebmann Jeffrey M et al.

Glaucoma-related lamina cribrosa deepening differs by race; European descent patients showed significantly greater posterior migration than African descent patients, impacting understanding of disease progression.

Christopher Mark, Nakahara Kenichi, Bowd Christopher, Proudfoot James A, Belghith Akram, Goldbaum Michael H et al.

Deep learning glaucoma detection is accurate, especially for moderate-to-severe disease. Combining diverse training data improves performance, suggesting AI's role in primary care glaucoma screening.

Bowd Christopher, Belghith Akram, Proudfoot James A, Zangwill Linda M, Christopher Mark, Goldbaum Michael H et al.

Gradient-boosting classifiers combining OCTA vessel density and OCT thickness improved early glaucoma detection compared to standard methods, offering better diagnostic accuracy.

Wong Brandon J, Moghimi Sasan, Zangwill Linda M, Christopher Mark, Belghith Akram, Ekici Eren et al.

In glaucoma patients, lower corneal hysteresis was linked to greater posterior lamina cribrosa displacement over time, suggesting it's a risk factor for progression.