search

raidionics / Raidionics

Software for automatic segmentation and generation of standardized clinical reports of brain tumors from MRI volumes
https://raidionics.github.io/
BSD 2-Clause "Simplified" License
31 stars 4 forks source link

Multilabel (BRATS-like) output for single-time point post-operative glioma segmentation model #76

Open gusSCIMOV opened 3 months ago

gusSCIMOV commented 3 months ago

Hi Raidionics Team I've been working a time ago with many of the Colab notebooks added in "processing backend for performing segmentation and computation of standardized report (RADS)" and it has been quite useful

I just wondering if I might get a BRATs-like multilabel output (Edema, enhancing, necrosis) using just the post-operative scans using the MRI_GBM_Postop_FV_1p or FV_3p, instead of the whole tumor mask since that is an important part of therapy response assessment

Thanks, I really appreciate your help

andreped commented 3 months ago

@dbouget Just FYI, this is related to this Issue thread: https://github.com/raidionics/Raidionics/issues/67#issuecomment-2117494776

dbouget commented 3 months ago

Hi @gusSCIMOV,

At the moment, we do not have such annotations at our disposal to train in-house models. The next iteration of the BraTS challenge will include early post-operative segmentation tasks, and will provide the desired class-annotations (i.e., necrosis, enhancing, and edema).

We'll start including such models in Raidionics after the MICCAI conference, scheduled early October 2024, so stay tuned!