Post

New research · Ophthalmology
Translational vision science & technology · 2d
Animal / preclinicalTranslational vision science & technology · 2026

Automated Deep Learning Quantification of Avascular Area and Intravitreal Neovascularization in Retinal Flatmounts of Rodent Oxygen-Induced Retinopathy Models.

Neal S Shah, Aniket Ramshekar, Bright Asare-Bediako … M Elizabeth Hartnett
Read paper
OphthalmologyAnimal / preclinical

Reviewers preferred the AI model's neovascularization outlines in most comparisons

Automated Deep Learning Quantification of Avascular Area and Intravitreal Neovascularization in Retinal Flatmounts of Rodent Oxygen-Induced Retinopathy Models.

Neal S Shah … M Elizabeth Hartnett
Translational vision science & technology · 2026
Purpose

To develop a single deep learning model that quantifies the retinal avascular area (avascular area) and intravitreal neovascularization (intravitreal neovascularization) in rodent oxygen-induced retinopathy (oxygen-induced retinopathy) models.

Methods

We assessed intergrader reliability and agreement at metric and pixel levels.

n = 325 images
83.3%
Results
83.3%
Human reviewers picked the deep-learning model's abnormal blood vessel outline as best in 83.3% of comparisons.
n = 325 images
More results

Intergrader reliability was high for percent avascular area (mouse intraclass correlation coefficient [ICC] = 0.840; rat ICC = 0.971), moderate for rat percent intravitreal neovascularization (ICC = 0.509), and low for mouse percent intravitreal neovascularization (ICC = -0.082).

More results

Metric-level correlation was strong in rat oxygen-induced retinopathy (percent avascular area r = 0.979; percent intravitreal neovascularization r = 0.943) and for mouse percent avascular area (r = 0.957), but weak for mouse percent intravitreal neovascularization (r = 0.265).

“
Conclusion

Our deep learning model supports automated rat oxygen-induced retinopathy analysis while maintaining mouse performance and may improve reproducibility of oxygen-induced retinopathy measurements.

Read paper
0 comments

No comments yet. Be the first.

Related papers

LatestFoundational
AI / Informatics
0·962 for OCT
MerMED-FM was very accurate at diagnosing diseases using eye scans
AI / Informatics
0.98
Artificial intelligence's eyelid-height measurements matched doctors' manual measurements almost perfectly
AI / Informatics
92.1%
combining eye scans and photos correctly told benign from cancerous lesions apart nearly every time
Cohort Study
28.6%
recurred locally in more than 1 in 4 patients over years of follow-up
Cohort Study
-0.395
excision group's post-op eyelid fullness score was lower - greater improvement
Cohort Study
86.7%
of infants probed after 12 months still had unresolved tear duct blockage
Observational
16%
eyelid tissue in rosacea patients showed less of this key repair-signaling protein inside cell nuclei
Cohort Study
25%
about 1 in 4 treated cases had symptoms return after improving