Skip to content
DraconiticResearch
Search Sources API
203,203 studies31 sourcesUpdated 56 min ago
PreprintarXiv licencearXiv

Quantitative T2* Fidelity in Accelerated Radial Ultrashort-Echo-Time MRI of the Lumbar Cartilage Endplate: A Synthetic Computational Validation Study

  • Yousef Riyazifar
Open at sourcePDFDOICite

Abstract

The lumbar cartilage endplate (CEP) is a short-T2* structure whose composition has been linked to disc health. Ultrashort-echo-time (UTE) MRI enables direct CEP visualization, but acceleration and reconstruction may alter quantitative T2* even when image quality improves. We evaluated how radial undersampling and reconstruction strategy affect CEP T2* fidelity in a fully synthetic setting. A 96 x 96 sagittal lumbar phantom with five CEP levels (ground-truth T2*: 24, 22, 20, 18, and 16 ms) was simulated with six UTE echoes, eight receive coils, and radial sampling. Acceleration factors R=1,2,4,6,8 used 302, 151, 76, 50, and 38 spokes, respectively. Reconstructions were CG-SENSE, joint multi-echo FISTA, and direct model-based estimation from k-space. Twenty independent complex-noise realizations were evaluated per acceleration. Direct model-map RMSE increased from 0.172 ms at R=1 to 0.800 ms at R=4, 1.472 ms at R=6, and 2.256 ms at R=8. At R=4, joint-FISTA had lower image NRMSE than the direct model (0.132 vs. 0.203), while the direct map had lower level-wise T2* RMSE than CG-SENSE and joint-FISTA ROI fits (0.800 vs. 1.448 and 1.288 ms; estimators differed). At R=8, CG-SENSE ROI fitting had the lowest T2* RMSE (1.501 ms), whereas the direct model showed systematic underestimation (bias, -2.136 ms; CCC, 0.723). Image fidelity and quantitative biomarker fidelity therefore diverged under radial acceleration. In this synthetic framework, R=4 was a promising operating point for direct CEP T2* mapping, whereas R=8 produced substantial bias and spatial instability. Scanner, phantom, and in vivo validation are required before clinical translation.

Related

  1. Photoemission from Semi-Infinite Crystals: First-Principles Scattering States and Quantitative Ag(111) Validation
    Tyler Wu, Truman Idso, Siddharth Karkare +1
    arXivPreprintCC BY 4.0Full textApplied physics & instrumentationCondensed matterarXiv
  2. Reliability of AI/ML Computational Searches for Magnetic Materials: Databases, Validation, and Synthesizability
    Vladimir Antropov, Subhadip Pradhan, Kirill D. Belashchenko
    arXivPreprintCC BY 4.0Full textCondensed matterarXiv
  3. Quantum computational resources and validation protocols for a three-mode non-Gaussian trilinear Hamiltonian
    Niccolò Laurora, Matteo Bina, Giulia Ferrini +1
    arXivPreprintarXiv licenceFull textQuantum physics & quantum informationarXiv
  4. Validation of a Computational Respiratory System Model for Mechanical Ventilation
    Carlotta Hennigs, Charlott Danielson, Franziska Bilda +7
    arXivPreprintCC BY 4.0Full textGeneral biologyDynamical systemsarXiv
  5. Quantitative Ultrasound Perfusion Biomarker: Cross-Modality Validation with Laser Speckle Imaging
    Haowei Tai, Oscar Duran-Garnica, Karen Villarreal +2
    arXivPreprintarXiv licenceFull textBiological & medical physicsarXivPMC

Cite

BibTeXRIS
BibTeX
@misc{riyazifar2026quantitative,
  title = {{Quantitative $T_2^*$ Fidelity in Accelerated Radial Ultrashort-Echo-Time MRI of the Lumbar Cartilage Endplate: A Synthetic Computational Validation Study}},
  author = {Yousef Riyazifar},
  howpublished = {arXiv},
  year = {2026},
  month = nov,
  doi = {10.48550/arxiv.2611.00005},
  eprint = {2611.00005},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2611.00005},
}
RIS
TY  - UNPB
TI  - Quantitative T_2^* Fidelity in Accelerated Radial Ultrashort-Echo-Time MRI of the Lumbar Cartilage Endplate: A Synthetic Computational Validation Study
AU  - Yousef Riyazifar
PY  - 2026
DA  - 2026/11/01
PB  - arXiv
DO  - 10.48550/arxiv.2611.00005
UR  - https://arxiv.org/abs/2611.00005
AB  - The lumbar cartilage endplate (CEP) is a short-T_2^* structure whose composition has been linked to disc health. Ultrashort-echo-time (UTE) MRI enables direct CEP visualization, but acceleration and reconstruction may alter quantitative T_2^* even when image quality improves. We evaluated how radial undersampling and reconstruction strategy affect CEP T_2^* fidelity in a fully synthetic setting. A 96 x 96 sagittal lumbar phantom with five CEP levels (ground-truth T_2^*: 24, 22, 20, 18, and 16 ms) was simulated with six UTE echoes, eight receive coils, and radial sampling. Acceleration factors R=1,2,4,6,8 used 302, 151, 76, 50, and 38 spokes, respectively. Reconstructions were CG-SENSE, joint multi-echo FISTA, and direct model-based estimation from k-space. Twenty independent complex-noise realizations were evaluated per acceleration. Direct model-map RMSE increased from 0.172 ms at R=1 to 0.800 ms at R=4, 1.472 ms at R=6, and 2.256 ms at R=8. At R=4, joint-FISTA had lower image NRMSE than the direct model (0.132 vs. 0.203), while the direct map had lower level-wise T_2^* RMSE than CG-SENSE and joint-FISTA ROI fits (0.800 vs. 1.448 and 1.288 ms; estimators differed). At R=8, CG-SENSE ROI fitting had the lowest T_2^* RMSE (1.501 ms), whereas the direct model showed systematic underestimation (bias, -2.136 ms; CCC, 0.723). Image fidelity and quantitative biomarker fidelity therefore diverged under radial acceleration. In this synthetic framework, R=4 was a promising operating point for direct CEP T_2^* mapping, whereas R=8 produced substantial bias and spatial instability. Scanner, phantom, and in vivo validation are required before clinical translation.
ER  -