Post

New research · Ophthalmology
AJNR. American journal of neuroradiology · 16h
AI / informaticsAJNR. American journal of neuroradiology · 2026

Two-Step Semiautomated Classification of Choroidal Metastases on MRI: Orbit Localization via Bounding Boxes Followed by Binary Classification via Evolutionary Strategies.

Jeffrey S Shi, Bala McRae-Posani, Sofia Haque … Joseph Stember
Read paper
OphthalmologyAI / informatics

AI classifier distinguished choroidal metastases from normal orbits on brain MRI with high accuracy

Two-Step Semiautomated Classification of Choroidal Metastases on MRI: Orbit Localization via Bounding Boxes Followed by Binary Classification via Evolutionary Strategies.

Jeffrey S Shi … Joseph Stember
AJNR. American journal of neuroradiology · 2026
Background

The choroid of the eye is a rare site for metastatic tumor spread, and as small lesions on the periphery of brain MRI studies, these choroidal metastases are often missed.

Purpose

To improve their detection, we aimed to use artificial intelligence to distinguish between brain MRI scans containing normal orbits and choroidal metastases.

Methods

In step 1, we trained a localization network on 386 T2-weighted brain MRI axial slices from 97 patients.

n = 97 patients
Results

Area under curve of 0.93; 1.0 is perfect discrimination, 0.5 is chance

AUC 0.93 (95% CI 0.83 to 1.03)
null = 00.831.03
CI excludes the null - significant
More results

Our orbit localization model identified globes with 100% accuracy and a mean average precision (mAP) of intersection over union thresholds of 0.5-0.95 [mAP(0.5:0.95)] of 0.47 on held-out testing data.

More results

Similarly, the model generalized well to our step 2 data set, which included orbits demonstrating pathologies, achieving 100% accuracy and mAP(0.5:0.95) of 0.44. mAP(0.5:0.95) appeared low because the model could not distinguish left and right orbits.

“
Conclusion · 1 of 2

The semiautomated pipeline from brain MRI slices to choroidal metastasis classification demonstrates the utility of a sequential localization and classification approach, and clinical relevance for identifying small, "corner-of-the-image," easily overlooked lesions.

Conclusion · 2 of 2

Artificial Intelligence Level of Evidence: 5B.

Read paper
0 comments

No comments yet. Be the first.

Related papers

LatestFoundational
Cross-sectional
44.1%
had too much homocysteine in the blood, a far larger share than in dry AMD or healthy eyes
Cohort Study
92.3%
most laser-treated eyes had a favorable eye-structure outcome
Cohort Study
56%
lost three or more lines on the eye-chart vision test
Cohort Study
50%
older patients were more likely to lose three or more lines of vision
Study
81%
trials where mask air leaks were detected blowing toward the eyes
Cohort Study
13.5%
Central retinal artery peak blood-flow speed was 13.5% lower in diabetic patients.
Study
379 +/- 156 microm
lower thickness in the central retina, checked with an eye scan
Case Report
9 of 10 eyes
eyes had better measured vision after treatment