Integrative Machine Learning Approaches for Differentiating Pediatric Multiple Sclerosis from Non-Inflammatory Disorders via Optical Coherence Tomography - Takeaways - MDSpire

Integrative Machine Learning Approaches for Differentiating Pediatric Multiple Sclerosis from Non-Inflammatory Disorders via Optical Coherence Tomography

  • By

  • Chaojun Chen

  • Sahar Soltanieh

  • Sajith Rajapaksa

  • Farzad Khalvati

  • E. Ann Yeh

  • April 21, 2026

  • 0 min

Share

  • 1

    Early diagnosis of pediatric multiple sclerosis (POMS) is crucial for improving long-term outcomes through timely therapeutic intervention.

  • 2

    Optical coherence tomography (OCT) serves as a non-invasive imaging tool that captures high-resolution details of the retina, aiding in MS diagnosis.

  • 3

    The study evaluated deep learning and machine learning models using OCT data to differentiate POMS from non-inflammatory neurological conditions.

  • 4

    The early fusion multimodal model outperformed unimodal models, achieving an AUC of 0.90 and accuracy of 87% in diagnosing POMS.

  • 5

    Integrating OCT-derived features with raw 3D OCT images enhances diagnostic performance by capturing complementary patterns associated with MS.

Original Source(s)

Integrative Machine Learning Approaches for Differentiating Pediatric Multiple Sclerosis from Non-Inflammatory Disorders via Optical Coherence Tomography

  • by Chaojun Chen, Sahar Soltanieh, Sajith Rajapaksa, Farzad Khalvati, E. Ann Yeh

  • April 21, 2026

Related Content

Electroencephalographic abnormalities and clinical phenotypes in children with autism spectrum disorder: a single center cohort study

Perampanel use in pediatric autoimmune encephalitis-related seizures and epilepsy: a retrospective case series

Feasibility testing of a home-based exercise intervention in children with cerebral palsy who are ambulant—a study protocol of the HOME-EX study