Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: May 29, 2026

Combining Reflectance Confocal Microscopy with Optical Coherence Tomography for Noninvasive Diagnosis of Skin Cancers via Image Acquisition
09:37

Combining Reflectance Confocal Microscopy with Optical Coherence Tomography for Noninvasive Diagnosis of Skin Cancers via Image Acquisition

Published on: August 18, 2022

Multimodal Deep Learning for Benign vs Malignant Eyelid Lesion Classification Using Optical Coherence Tomography and

Weronika Jakubowska1,2, Clément Playout1,2, Renaud Duval1,2

  • 1Department of Ophthalmology, Université de Montréal, Montreal, Quebec, Canada.

Ophthalmic Plastic and Reconstructive Surgery
|May 27, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same authorSame journal

Ethmoiditis Mimicking Acute Dacryocystitis-An Underreported Pathology.

Ophthalmic plastic and reconstructive surgery·2026
Same author

Retinal lesion annotation on fundus imaging: an interobserver variability study.

Scientific reports·2026
Same author

Long-Term Sequelae of Retinopathy of Prematurity-A Scoping Review.

Children (Basel, Switzerland)·2026
Same author

Impact of AI-enhanced three-dimensional OCT scans on disease activity assessment in patients with nAMD: the RAZORBILL study.

The British journal of ophthalmology·2026
Same author

DNAi: an open-source AI tool for unbiased DNA fiber analysis.

Nucleic acids research·2026
Same author

Bilateral symmetric lower eyelid multinucleate cell angiohistiocytomas.

Eye (London, England)·2026

Deep learning models using optical coherence tomography (OCT) and clinical photos accurately classify eyelid lesions. Multimodal fusion of OCT and photos offers the best diagnostic performance for distinguishing benign from malignant growths.

Area of Science:

  • Ophthalmology
  • Artificial Intelligence
  • Medical Imaging

Background:

  • Periocular lesions require accurate diagnosis to differentiate benign from malignant conditions.
  • Current diagnostic methods may involve invasive procedures, highlighting the need for non-invasive alternatives.

Purpose of the Study:

  • To assess deep learning models for classifying eyelid lesions as benign or malignant.
  • To evaluate the performance of models using optical coherence tomography (OCT) volumes, clinical photographs, and their combined (multimodal fusion) data.
  • To compare these models against histopathology as the gold standard.

Main Methods:

  • A prospective cohort study included 65 patients with 71 periocular lesions.
  • Spectral-domain OCT and slit-lamp photography were used for imaging before biopsy.

Related Experiment Videos

Last Updated: May 29, 2026

Combining Reflectance Confocal Microscopy with Optical Coherence Tomography for Noninvasive Diagnosis of Skin Cancers via Image Acquisition
09:37

Combining Reflectance Confocal Microscopy with Optical Coherence Tomography for Noninvasive Diagnosis of Skin Cancers via Image Acquisition

Published on: August 18, 2022

  • Three Vision Transformer models were developed: OCT-only, photograph-only, and multimodal fusion.

    • Main Results:

    • The multimodal fusion model achieved the highest accuracy (83.2%) and sensitivity (91.8%).
    • The photograph model showed strong performance with 82.3% accuracy and 91.7% AUC.
    • The OCT model demonstrated promising results with 73.9% accuracy and 82.5% AUC.

    Conclusions:

    • Deep learning analysis of OCT volumes is feasible for eyelid lesion classification.
    • Integrating OCT and clinical photography via multimodal AI enhances diagnostic accuracy.
    • Multimodal AI shows potential as a scalable, non-invasive tool for lesion triage, potentially reducing reliance on biopsies.