lgg_regrowth: Biologically conditioned deep learning for glioma growth prediction

20260 citationsOthergreen Open Access

Authors

Max Wartha · Johannes Kepler University of Linz

Petra Böhm · Johannes Kepler University of Linz

Baran Atli · Johannes Kepler University of Linz

Matthias Gmeiner · Johannes Kepler University of Linz

Martin Aichholzer · Johannes Kepler University of Linz

Carlo Serra · University of Zurich

Michael Sonnberger

Abstract

Training, evaluation, and inference code for a biologically conditioned 3D U-Net that predicts diffuse lower-grade glioma growth patterns by integrating IDH mutation status, histological subtype, and cortical architecture type via Feature-wise Linear Modulation (FiLM).

Topics & Keywords

Publication Details

Published in: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.19268238

Command Palette

lgg_regrowth: Biologically conditioned deep learning for glioma growth prediction

Authors

Abstract

Topics & Keywords

Publication Details