Automated Identification of Lesion Activity in Neovascular Age-Related Macular Degeneration

Chakravarthy, Usha; Goldenberg, Dafna; Young, Graham; Havilio, Moshe; Rafaeli, Omer; Benyamini, Gidi; Loewenstein, Anat

doi:10.1016/j.ophtha.2016.04.005

Cited by 96 publications

(50 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…[53][54][55][56] Prahs et al looked at AI to support therapy decisions for intravitreal injection. This has been done by training MLCs using OCT imaging and analysing different features of the scan, particularly retinal fluid.…”

Section: Assessment Of Age-related Macular Degenerationmentioning

confidence: 99%

“…This has been done by training MLCs using OCT imaging and analysing different features of the scan, particularly retinal fluid. [53][54][55][56] Prahs et al looked at AI to support therapy decisions for intravitreal injection. They found that their deep CNN was able to correctly predict the need for anti-VEGF therapy in 95% of the casessimilar to an average human grader in Chakravarthy et al's study.…”

Section: Assessment Of Age-related Macular Degenerationmentioning

confidence: 99%

“…They found that their deep CNN was able to correctly predict the need for anti-VEGF therapy in 95% of the casessimilar to an average human grader in Chakravarthy et al's study. 54,55 Schlegl et al also designed a MLC (using 1200 OCT images), with a mean accuracy of 0.94 and an AUROC of 0.92. 56 Though in its preliminary stages, predictive treatment technology would be a useful addition to clinical practice.…”

Section: Assessment Of Age-related Macular Degenerationmentioning

confidence: 99%

See 2 more Smart Citations

Current state and future prospects of artificial intelligence in ophthalmology: a review

Hogarty

Mackey

Hewitt

2018

Clinical Exper Ophthalmology

144

View full text Add to dashboard Cite

Artificial intelligence (AI) has emerged as a major frontier in computer science research. Although AI has broad application across many medical fields, it will have particular utility in ophthalmology and will dramatically change the diagnostic and treatment pathways for many eye conditions such as corneal ectasias, glaucoma, age-related macular degeneration and diabetic retinopathy. However, given that AI has primarily been driven as a computer science, its concepts and terminology are unfamiliar to many medical professionals. Important key terms such as machine learning and deep learning are often misunderstood and incorrectly used interchangeably. This article presents an overview of AI and new developments relevant to ophthalmology.

show abstract

Section: Assessment Of Age-related Macular Degenerationmentioning

confidence: 99%

Section: Assessment Of Age-related Macular Degenerationmentioning

confidence: 99%

Section: Assessment Of Age-related Macular Degenerationmentioning

confidence: 99%

See 1 more Smart Citation

Current state and future prospects of artificial intelligence in ophthalmology: a review

Hogarty

Mackey

Hewitt

2018

Clinical Exper Ophthalmology

144

View full text Add to dashboard Cite

show abstract

“…Thus, researchers have started to question the feasibility of a manual review of OCT images in clinical practice, and suggest automated computational analysis of SD-OCT data to aid clinical management and decision making 17 .…”

Section: Introductionmentioning

confidence: 99%

Evaluating the impact of vitreomacular adhesion on anti-VEGF therapy for retinal vein occlusion using machine learning

Waldstein

Montuoro

Podkowinski

et al. 2017

Sci Rep

View full text Add to dashboard Cite

Vitreomacular adhesion (VMA) represents a prognostic biomarker in the management of exudative macular disease using anti-vascular endothelial growth factor (VEGF) agents. However, manual evaluation of VMA in 3D optical coherence tomography (OCT) is laborious and data on its impact on therapy of retinal vein occlusion (RVO) are limited. The aim of this study was to (1) develop a fully automated segmentation algorithm for the posterior vitreous boundary and (2) to study the effect of VMA on anti-VEGF therapy for RVO. A combined machine learning/graph cut segmentation algorithm for the posterior vitreous boundary was designed and evaluated. 391 patients with central/branch RVO under standardized ranibizumab treatment for 6/12 months were included in a systematic post-hoc analysis. VMA (70%) was automatically differentiated from non-VMA (30%) using the developed method combined with unsupervised clustering. In this proof-of-principle study, eyes with VMA showed larger BCVA gains than non-VMA eyes (BRVO: 15 ± 12 vs. 11 ± 11 letters, p = 0.02; CRVO: 18 ± 14 vs. 9 ± 13 letters, p < 0.01) and received a similar number of retreatments. However, this association diminished after adjustment for baseline BCVA, also when using more fine-grained VMA classes. Our study illustrates that machine learning represents a promising path to assess imaging biomarkers in OCT.

show abstract

“…Recent advances in machine learning for image analysis could be a solution to reduce subjectivity and possibly reduce errors in fields requiring image assessment, if they are found to be unacceptably high. 31,32 However, there is a distinction to be made between errors in data entry, which was the focus of this audit, and misdiagnoses by clinicians, which is a much larger issue and beyond the scope of this study. Selection bias may also be present because participation in this data quality review was entirely voluntary.…”

Section: Discussionmentioning

confidence: 99%

Assessing the Accuracy of a Large Observational Registry of Neovascular Age-Related Macular Degeneration

Nguyen

Leung

Nguyen

et al. 2020

Retina

View full text Add to dashboard Cite

PURPOSE To evaluate the accuracy of an observational database that tracks real-world treatment outcomes for neovascular age-related macular degeneration. METHODS We audited 245 randomly sampled eyes from 189 patients with 3,356 visits from 11 doctors in the Fight Retinal Blindness! DATABASE Sex, birth year, previous treatments received, treatment, and visual acuity were validated against the clinical notes. Error rates, the proportion of missed visits (the number of visits present in the patient record but not entered into Fight Retinal Blindness!), the level of agreement using Cohen's kappa () and intraclass correlation coefficients, and positive and negative predictive values were calculated. A visual acuity error was defined as an absolute difference of 5 letters. RESULTS The overall error rate was 3.5% (95% confidence interval: 3.1-3.9). The error rate for visual acuity was 5.1% (95% confidence interval: 4.2-5.9) and <5% for the remaining fields. The level of agreement for each field ranged from good to excellent ( or intraclass correlation 0.75). The positive predictive value and negative predictive value for visits were 99% and 89%, respectively. The proportion of missed visits was 10.2%. CONCLU-SION Accuracy of the Fight Retinal Blindness! database was good (>95%). The rate of missed visits was high, possibly due to the high burden of retrospective data entry or patients switching practitioners during treatment. Mark (2020). Assessing the accuracy of a large observational registry of neovascular age-related macular degeneration. Retina, 40(5):866-872.Purpose: To evaluate the accuracy of an observational database that tracks real-world treatment outcomes for neovascular age-related macular degeneration.Methods: We audited 245 randomly sampled eyes from 189 patients with 3,356 visits from 11 doctors in the Fight Retinal Blindness! database. Sex, birth year, previous treatments received, treatment, and visual acuity were validated against the clinical notes. Error rates, the proportion of missed visits (the number of visits present in the patient record but not entered into Fight Retinal Blindness!), the level of agreement using Cohen's kappa (k) and intraclass correlation coefficients, and positive and negative predictive values were calculated. A visual acuity error was defined as an absolute difference of $5 letters.Results: The overall error rate was 3.5% (95% confidence interval: 3.1-3.9). The error rate for visual acuity was 5.1% (95% confidence interval: 4.2-5.9) and ,5% for the remaining fields. The level of agreement for each field ranged from good to excellent (k or intraclass correlation $ 0.75). The positive predictive value and negative predictive value for visits were 99% and 89%, respectively. The proportion of missed visits was 10.2%.Conclusion: Accuracy of the Fight Retinal Blindness! database was good (.95%). The rate of missed visits was high, possibly due to the high burden of retrospective data entry or patients switching practitioners during treatment.

show abstract

Automated Identification of Lesion Activity in Neovascular Age-Related Macular Degeneration

Cited by 96 publications

References 14 publications

Current state and future prospects of artificial intelligence in ophthalmology: a review

Current state and future prospects of artificial intelligence in ophthalmology: a review

Evaluating the impact of vitreomacular adhesion on anti-VEGF therapy for retinal vein occlusion using machine learning

Assessing the Accuracy of a Large Observational Registry of Neovascular Age-Related Macular Degeneration

Contact Info

Product

Resources

About