OREHAS: A fully automated deep-learning pipeline for volumetric endolymphatic hydrops quantification in MRI
Published:
Recommended citation: Caterina Fuster-Barceló, Claudia Castrillón, Laura Rodrigo-Muñoz, Victor Manuel Suárez-Vega, Nicolás Pérez-Fernández, Gorka Bastarrika, Arrate Muñoz-Barrutia. "OREHAS: A fully automated deep-learning pipeline for volumetric endolymphatic hydrops quantification in MRI." arXiv preprint arXiv:2601.18368 (2026). https://arxiv.org/abs/2601.18368
OREHAS: A fully automated deep-learning pipeline for volumetric endolymphatic hydrops quantification in MRI
Published in: arXiv preprint
Authors: Caterina Fuster-Barceló, Claudia Castrillón, Laura Rodrigo-Muñoz, Victor Manuel Suárez-Vega, Nicolás Pérez-Fernández, Gorka Bastarrika, Arrate Muñoz-Barrutia
Abstract
This work presents OREHAS (Optimized Recognition & Evaluation of volumetric Hydrops in the Auditory System), a fully automatic pipeline for volumetric quantification of endolymphatic hydrops from routine 3D-SPACE-MRC and 3D-REAL-IR MRI.
The system integrates slice classification, inner ear localization, and sequence-specific segmentation into a single workflow that computes per-ear endolymphatic-to-vestibular volume ratios directly from whole MRI volumes, eliminating the need for manual intervention.
Trained with only 3 to 6 annotated slices per patient, OREHAS generalized effectively to full 3D volumes, achieving strong segmentation results and closely matching expert annotations in external validation while clearly outperforming the clinical syngo.via software in volumetric agreement.
These results show that reliable and reproducible endolymphatic hydrops quantification can be achieved from standard MRI using limited supervision, providing a robust basis for large-scale studies and clinically aligned volumetric analysis.