OncoPrep Documentation¶
OncoPrep is a neuro-oncology MRI preprocessing pipeline built on Nipype workflows. It follows NiPreps patterns (fMRIPrep, sMRIPrep) for BIDS-Apps compatibility, supporting DICOM→BIDS conversion, tumor segmentation with Docker-based multi-model ensembles, and BIDS derivative outputs.
Note
This is NOT a clinical tool. OncoPrep is intended for research use only.
Key features¶
BIDS-compliant preprocessing — anatomical registration, skull-stripping, template normalization
nnInteractive default segmentation — zero-shot 3D promptable foundation model (no Docker required; Isensee et al., 2025)
Multi-model ensemble segmentation — 14 BraTS-challenge Docker models
Segmentation fusion — majority voting, SIMPLE, and BraTS-specific fusion
Radiomics — PyRadiomics-based quantitative feature extraction
VASARI — automated VASARI feature extraction and radiology report generation (vasari-auto)
Surface processing (planned) — FreeSurfer + GIFTI/CIFTI-2 derivatives
HTML reports — fMRIPrep-style quality-assurance reports
User Guide
- Installation
- Quick Start
- Tutorial: End-to-End Neuro-Oncology Preprocessing
- Prerequisites
- Step 1 — DICOM to BIDS conversion
- Step 2 — Preprocessing
- Step 3 — Tumor segmentation
- Step 4 — Radiomics (optional)
- Step 5 — Group-Level ComBat Harmonization (multi-site studies)
- Step 6 — Quality control with MRIQC (temporarily disabled)
- Step 7 — Reports
- Using the Python API
- End-to-End Reference
- Troubleshooting
- Command-Line Interface
- Segmentation
- Group-Level ComBat Harmonization
- VASARI Feature Extraction & Radiology Reports
- Docker Usage
- HPC / Singularity Deployment
Reference
About