Construction of a whole-brain segmentation pipeline for 0.35T neonatal MRI

Abstract

BackgroundWhole-brain segmentation of neonatal low-field MRI is needed for regional volumetric analysis, but established segmentation systems are designed around different field strengths, contrasts, and age ranges.

MethodsWe first applied complete end-to-end pipelines, including FreeSurfer, FSL-based processing, Infant FreeSurfer, FastSurfer, BrainSuite, SPM12/CAT12, iBEAT v2, the dHCP structural pipeline, and M-CRIB-S, directly to 0.35T neonatal T2-weighted images. Because those systems did not produce a complete 47-region output, we next tested released end-to-end AI segmentation models, including SynthSeg, FastSurfer, FastSurferVINN, and QuickNAT. These were highly sensitive to input format, contrast, voxel geometry, and the high-field training domain. We then decomposed the task into skull stripping, registration, super-resolution, atlas propagation, and tissue refinement and tested the available components individually.

ResultsNo complete end-to-end pipeline or released AI model delivered a usable 47-region segmentation in direct application. Stage-wise testing showed that preprocessing, reconstruction, registration, atlas propagation, and expectation-maximization each solved only part of the problem. Their final assembly produced complete 47-region outputs across the neonatal cohort.

ConclusionsDirect application of established pipelines was insufficient for 0.35T neonatal T2-weighted MRI. A stage-wise combination of compatible components provided the workable solution.

Detector-Stage Failures Account for Most of the Increase in the External Deficit of a Box-Prompted Glioma Segmentation Pipeline on BraTS-Africa: A Three-Arm Decomposition Across Two External Cohorts
Aykut Kaşkaya, Baha Şen
Diagnostics (Basel, Switzerland)1 Oct 2026Peer-reviewedCC BYFull textNeurologyEurope PMCDOAJ
Automatic segmentation of histological zones in rat distraction osteogenesis using a self-configuring deep learning pipeline
Leonie Dittrich, Jan-Moritz Ramge, Claudia Neunaber +1
Scientific Reports21 Oct 2026Peer-reviewedCC BY 4.0Full textGeneral medicineCrossrefDOAJ
An integrated fuzzy multi-criteria decision making and mixed-integer linear programming optimization framework for project risk assessment and response strategy selection: A case study of a liquefied petroleum gas pipeline construction project
Mohammad Senisel Bachari, Asgar Khademvatani, Mahdi Iranfar
PLOS One24 Sep 2026Peer-reviewedCC BY 4.0Full textNeuroscience Materials & polymersPLOS
Understanding vulnerability to reproductive, maternal, neonatal and child health outcomes in Senegal using a population segmentation
Palash Singh, Finou Thérèse Mendy, Camille Kramer Courbariaux +12
BMC Public Health28 Oct 2026Peer-reviewedCC BY 4.0Obstetrics, gynaecology & paediatricsCrossref
TexTNet: a texture-guided hierarchical architecture for prostate cancer Gleason segmentation in whole slide images
Muhammad Arshad, Wang Chengliang, Muhammad Wajeeh Us Sima +3
Frontiers in medicine19 Oct 2026Peer-reviewedCC BYFull textGeneral medicineEurope PMCPMCEurope PMC

Cite

BibTeX RIS

BibTeX

@misc{sun2026construction,
  title = {{Construction of a whole-brain segmentation pipeline for 0.35T neonatal MRI}},
  author = {Z. Sun and J. Huang and X.-N. Zuo and H. He and G. Yu},
  howpublished = {medRxiv},
  year = {2026},
  month = sep,
  doi = {10.64898/2026.09.23.26363741},
  url = {https://www.medrxiv.org/content/10.64898/2026.09.23.26363741v1},
  note = {CC BY-NC-ND},
}

RIS

TY  - UNPB
TI  - Construction of a whole-brain segmentation pipeline for 0.35T neonatal MRI
AU  - Z. Sun
AU  - J. Huang
AU  - X.-N. Zuo
AU  - H. He
AU  - G. Yu
PY  - 2026
DA  - 2026/09/23
PB  - medRxiv
DO  - 10.64898/2026.09.23.26363741
UR  - https://www.medrxiv.org/content/10.64898/2026.09.23.26363741v1
AB  - BackgroundWhole-brain segmentation of neonatal low-field MRI is needed for regional volumetric analysis, but established segmentation systems are designed around different field strengths, contrasts, and age ranges.  MethodsWe first applied complete end-to-end pipelines, including FreeSurfer, FSL-based processing, Infant FreeSurfer, FastSurfer, BrainSuite, SPM12/CAT12, iBEAT v2, the dHCP structural pipeline, and M-CRIB-S, directly to 0.35T neonatal T2-weighted images. Because those systems did not produce a complete 47-region output, we next tested released end-to-end AI segmentation models, including SynthSeg, FastSurfer, FastSurferVINN, and QuickNAT. These were highly sensitive to input format, contrast, voxel geometry, and the high-field training domain. We then decomposed the task into skull stripping, registration, super-resolution, atlas propagation, and tissue refinement and tested the available components individually.  ResultsNo complete end-to-end pipeline or released AI model delivered a usable 47-region segmentation in direct application. Stage-wise testing showed that preprocessing, reconstruction, registration, atlas propagation, and expectation-maximization each solved only part of the problem. Their final assembly produced complete 47-region outputs across the neonatal cohort.  ConclusionsDirect application of established pipelines was insufficient for 0.35T neonatal T2-weighted MRI. A stage-wise combination of compatible components provided the workable solution.
ER  -

Abstract

Related

Cite