ScholarDeck

AI / informaticsGraefe's archive for clinical and experimental ophthalmology = Albrecht von Graefes Archiv fur klinische und experimentelle Ophthalmologie · 2026

A deep-learning model for detecting choroidal metastases and predicting primary tumors from ultra-widefield fundus imaging.

Hyo Jin Seong, Choonghan Kim, Jinho Chang … Christopher Seungkyu Lee

Read paper

OphthalmologyAI / informatics

Deep-learning model detected choroidal metastases far more accurately than ophthalmologists.

A deep-learning model for detecting choroidal metastases and predicting primary tumors from ultra-widefield fundus imaging.

Hyo Jin Seong … Christopher Seungkyu Lee

Graefe's archive for clinical and experimental ophthalmology = Albrecht von Graefes Archiv fur klinische und experimentelle Ophthalmologie · 2026

Purpose

To develop and validate a deep-learning model for detecting choroidal metastasis and predicting primary cancer sites using ultra-widefield fundus photography (ultra-widefield fundus photography).

Methods

This retrospective cohort study utilized 719 ultra-widefield fundus photography images from 112 patients with choroidal metastasis and 288 normal photos from 288 patients treated at Severance Hospital between 2005 and 2023.

n = 288 patients

Results

correctly told a metastasis apart from a normal eye almost every time

Model

0.96AUROC

Ophthalmologists

0.69AUROC

More results

Ablation studies confirmed that fundus image features were the primary contributors to classification.

“

Conclusion · 1 of 2

The developed deep-learning model shows significant potential not only in detecting choroidal metastases but, more importantly, in predicting their primary cancer origins from ultra-widefield fundus photography images.

Conclusion · 2 of 2

This capability could serve as a valuable adjunct in clinical decision-making by guiding more targeted and efficient systemic evaluations, particularly in patients with undiagnosed primary cancers.

Post

A deep-learning model for detecting choroidal metastases and predicting primary tumors from ultra-widefield fundus imaging.

Deep-learning model detected choroidal metastases far more accurately than ophthalmologists.

Related papers