📚 Proceedings: Image-Based Prediction of Retinal Disease Progression

Image-Based Prediction of Retinal Disease Progression

These proceedings document the outcomes of two significant challenges—DIAMOND and MARIO—organized in conjunction with the 27th International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2024), held on October 10, 2024, in Marrakesh, Morocco.

Chaired by Gwenolé Quellec, Mostafa El Habib Daho, and Rachid Zeghlache, this volume brings together cutting-edge work in the area of retinal disease progression prediction using deep learning on multimodal images.

DIAMOND Challenge

  • Focused on predicting center-involved diabetic macular edema (ci-DME) onset one year in advance.
  • Utilized ultra-wide-field color fundus photography (UWF-CFP) from the EviRed cohort.
  • Featured 6 selected papers from 8 submissions.

MARIO Challenge

  • Focused on predicting the progression of neovascular age-related macular degeneration (nAMD) within three months.
  • Relied on optical coherence tomography (OCT) imaging data.
  • Featured 15 selected papers from 17 submissions.

Both challenges were hosted on the Codabench platform, promoting transparent AI development through containerized submissions and open access to training datasets (MARIO only).

Topics addressed in this volume include:

  • Deep learning ensemble approaches
  • OCT-based progression prediction
  • Multimodal data fusion
  • Model calibration and generalization across domains

Each submission underwent peer review by experts in the field, assessing novelty, clarity, and relevance.

The event was a testament to the growing role of AI in precision ophthalmology and the practical challenges of working with regulated medical data.

Link to proceedings: Springer Link image