Fat Fraction Mapping in the Presence of Magnetic Field Inhomogeneities With Phase‐Cycled bSSFP at 3 T

ABSTRACT Magnetic field inhomogeneities (Δ B 0 ) can cause fat–water swaps in proton density fat fraction (PDFF) mapping using phase‐cycled balanced steady‐state free‐precession (bSSFP). In this work, a graph‐cuts–based smoothness‐imposing optimization was developed to estimate Δ B 0 from phase‐cycled bSSFP data obtained in fat–water compartments. Data were acquired in phantoms spanning PDFF of 0%–100% to quantify robustness to Δ B 0 and to determine the effect of RF excitation angle ( α ) on PDFF estimation. The approach was further tested in healthy volunteer knee data ( n = 9) and chronic liver disease patients data ( n = 5). Phase‐cycled bSSFP‐based Δ B 0 and PDFF maps were compared with multi‐echo gradient‐echo (MEGRE) reference maps in phantoms and knees. In liver patients, PDFF maps were compared visually and by retrospective ROI analysis with reference quantitative Dixon (qDixon) PDFF maps. Estimated Δ B 0 maps matched the reference, with a root‐mean‐squared error (RMSE) of 9 ± 2 Hz for a Δ B 0 range of ±400 Hz in phantoms and 17 ± 4 Hz for a range of ±150 Hz in knees. PDFF mapping accuracy improved after Δ B 0 correction, with mean RMSE reduced from 35% ± 9% to 3.55% ± 0.92% and mean R 2 increased from 0.29 ± 0.3 to 0.99 ± 0 ( p < 0.0001) across the six phantom experiments (mean ± SD of experiments). PDFF quantification showed significant bias for α = 5°, 40°, and 45° (bias = 2.7%, −2.9%, and −5.8%; p < 0.05), whereas no significant bias was observed for 10° ≤ α ≤ 35° ( p > 0.05). Mean PDFF differences versus MEGRE reference were 2.4% ± 3%, −0.1% ± 0.8%, and 0.2% ± 0.3% for muscle, subcutaneous fat, and bone marrow, respectively. In liver patients, whole‐liver ROI analysis demonstrated a mean PDFF difference of 3% relative to qDixon. Fat–water swaps were successfully removed in knee and patient liver PDFF maps. These results demonstrate that the proposed approach accurately estimates Δ B 0 and mitigates fat–water swaps in phantoms and healthy volunteer knees, while the patient liver study demonstrates the technical feasibility of the proposed framework for phase‐cycled bSSFP‐based PDFF mapping in vivo.

Authors

Institutions

Publication Details

Journal
NMR in Biomedicine
Published
2026-11-16
DOI
https://doi.org/10.1002/nbm.70437
Primary Topic
Advanced MRI Techniques and Applications
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Fat Fraction Mapping in the Presence of Magnetic Field Inhomogeneities With Phase‐Cycled bSSFP at 3 T

Tom Hilbert, Naïk Vietti Violi, Jessica Bastiaansen, Eva S. Peper et al.
NMR in Biomedicine
Advanced MRI Techniques and Applications
article

Fat Fraction Mapping in the Presence of Magnetic Field Inhomogeneities With Phase‐Cycled bSSFP at 3 T

Tom Hilbert, Naïk Vietti Violi, Jessica Bastiaansen, Eva S. Peper, Adrian Thomas Huber, Berk Can Açikgöz, Adèle Mackowiak, Verena Carola Obmann, Giulia M. C. Rossi, Jean‐Baptiste Ledoux, Annika Haeussler
article en

Abstract

ABSTRACT Magnetic field inhomogeneities (Δ B 0 ) can cause fat–water swaps in proton density fat fraction (PDFF) mapping using phase‐cycled balanced steady‐state free‐precession (bSSFP). In this work, a graph‐cuts–based smoothness‐imposing optimization was developed to estimate Δ B 0 from phase‐cycled bSSFP data obtained in fat–water compartments. Data were acquired in phantoms spanning PDFF of 0%–100% to quantify robustness to Δ B 0 and to determine the effect of RF excitation angle ( α ) on PDFF estimation. The approach was further tested in healthy volunteer knee data ( n = 9) and chronic liver disease patients data ( n = 5). Phase‐cycled bSSFP‐based Δ B 0 and PDFF maps were compared with multi‐echo gradient‐echo (MEGRE) reference maps in phantoms and knees. In liver patients, PDFF maps were compared visually and by retrospective ROI analysis with reference quantitative Dixon (qDixon) PDFF maps. Estimated Δ B 0 maps matched the reference, with a root‐mean‐squared error (RMSE) of 9 ± 2 Hz for a Δ B 0 range of ±400 Hz in phantoms and 17 ± 4 Hz for a range of ±150 Hz in knees. PDFF mapping accuracy improved after Δ B 0 correction, with mean RMSE reduced from 35% ± 9% to 3.55% ± 0.92% and mean R 2 increased from 0.29 ± 0.3 to 0.99 ± 0 ( p < 0.0001) across the six phantom experiments (mean ± SD of experiments). PDFF quantification showed significant bias for α = 5°, 40°, and 45° (bias = 2.7%, −2.9%, and −5.8%; p < 0.05), whereas no significant bias was observed for 10° ≤ α ≤ 35° ( p > 0.05). Mean PDFF differences versus MEGRE reference were 2.4% ± 3%, −0.1% ± 0.8%, and 0.2% ± 0.3% for muscle, subcutaneous fat, and bone marrow, respectively. In liver patients, whole‐liver ROI analysis demonstrated a mean PDFF difference of 3% relative to qDixon. Fat–water swaps were successfully removed in knee and patient liver PDFF maps. These results demonstrate that the proposed approach accurately estimates Δ B 0 and mitigates fat–water swaps in phantoms and healthy volunteer knees, while the patient liver study demonstrates the technical feasibility of the proposed framework for phase‐cycled bSSFP‐based PDFF mapping in vivo.

NMR in BiomedicineVol. 39(12)
University of Bern (CH), University Hospital of Bern (CH), Centre Hospitalier Universitaire Vaudois (CH), Zug Cantonal Hospital (CH), Luzerner Kantonsspital (CH), École Polytechnique Fédérale de Lausanne (CH), University of Lausanne (CH)
Advanced MRI Techniques and Applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon