ScholarDeck

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

Imaging-based machine learning for the diagnosis and prognosis of uveal melanoma: a systematic review and meta analysis.

Andres Bravo-Gonzalez, Pablo Dominguez-Ruiz, Maria J Buitrago-Gonzalez … Pedro F Salazar

Read paper

OphthalmologyAI / informatics

Ultrasound-based AI models diagnose uveal melanoma more specifically than fundus-based models.

Imaging-based machine learning for the diagnosis and prognosis of uveal melanoma: a systematic review and meta analysis.

Andres Bravo-Gonzalez … Pedro F Salazar

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

Background

Uveal melanoma (uveal melanoma) is the most common primary intraocular malignancy in adults and carries a high risk of metastasis and poor prognosis when diagnosed late.

Purpose

To systematically review and meta-analyze the performance of published machine learning/deep learning algorithms using ocular imaging on detecting and predicting prognosis of uveal melanoma.

Methods

A systematic search was conducted in PubMed, Scopus, Web of Science, Embase, and IEEE Xplore up to June 2025 for English and Spanish publications since 2012.

n = 4,600 patients

Results

ultrasound AI correctly ruled out non-melanoma cases more often than fundus-photo AI

78.21%

Sensitivity

94.28%

Specificity

More results

Thirteen diagnostic studies met inclusion criteria.

Pooled sensitivity was 78.21% and specificity 94.28%.

0.984 (ultrasound; Δ0.073, bootstrap 95% CI − 0.101 to 0.143).

More results

Six prognostic studies (up to 4,600 patients) reported AUCs between 0.71 and 0.92 for predicting metastasis, survival, or enucleation, though all lacked external validation.

“

Conclusion · 1 of 3

Machine learning and deep learning models show strong diagnostic performance and emerging prognostic value in uveal melanoma. However, reproducibility and real-world validation remain limited.

Conclusion · 2 of 3

New foundation models such as RETFound and VisionFM, trained on large multimodal eye datasets, could improve standardization, explainability, and cross-center generalization.

Conclusion · 3 of 3

As these models evolve, they have the potential to become essential tools in clinical practice, accelerating the translation of AI into reliable, routine implementation in ocular oncology.

Post

Imaging-based machine learning for the diagnosis and prognosis of uveal melanoma: a systematic review and meta analysis.

Ultrasound-based AI models diagnose uveal melanoma more specifically than fundus-based models.

Related papers