Abstract
Conventional motion-resolved magnetic resonance fingerprinting (MRF) relies on retrospective sorting-based reconstruction, which fails to accurately capture irregular respiratory motion patterns. To better characterize liver tumor motion under free-breathing in radiotherapy, we propose a novel time-resolved cine MRF (cMRF-T) technique for dynamic multi-parametric quantitative imaging. As the first cine MRF framework for liver cancer, cMRF-T introduces a self-supervised motion navigation approach trained without the need for unattainable ground-truth tissue maps. cMRF-T features an innovative respiratory-informed low-rank-constrained registration (Relo-Reg) method for respiratory motion modeling. Relo-Reg exploits the intrinsic low-rank compressibility of organ motion and regularizes free-form deformation vector fields (DVFs) derived from highly undersampled sliding-window MRF images. cMRF-T was validated on digital phantom simulations (n=30) and an in vivo liver cancer patient cohort (n=33). It achieved superior tissue quantification accuracy, with significantly reduced normalized root-mean-square error (NRMSE) in T1, T2, and proton density (PD) maps across all anatomical structures (p < 0.001) compared to SOTA MRF methods. In addition, Relo-Reg demonstrated improved tumor motion modeling accuracy compared to SOTA registration methods such as VoxelMorph and TransMorph. In the in vivo study, cMRF-T produced sharp organ boundaries with minimal aliasing artifacts, achieving a liver–lung boundary motion trajectory error of 1.324 ± 0.872 mm. By enabling accurate, time-resolved quantitative imaging without supervised labels, cMRF-T establishes a scalable platform for free-breathing liver cancer motion management in oncology.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/CMMCA2026_011.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/forum?id=kEW6iEhd2G
BibTex
@InProceedings{LiuChe_Freebreathing_MICCAISAT2026,
author = { Liu, Chenyang AND Wang, Xiang AND Wang, Lu AND Wang, Peilin AND Ni, Yimin AND Zhang, Fan AND Lee, Victor Ho-Fun AND Cao, Peng AND Li, Tian AND Cai, Jing},
title = { { Free-breathing Time-resolved Cine MR Fingerprinting Using Self-supervised Motion Navigation for Liver Cancer Radiotherapy } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17278},
month = {pending},
page = {pending}
}