Authors
Michael Ferraro, Sirisha Tadimalla, Fidel Navarro Salazar, Eireann Cosgriff, Annette Haworth, Jonathan Sykes
Published in
Medical physics. Volume 53. Issue 11.
Abstract
Respiratory-resolved magnetic resonance imaging (4D-MRI) provides superior soft-tissue contrast for characterizing respiratory motion in radiotherapy, but its geometric fidelity depends on the reconstruction and binning strategy. As with 4D-CT, peak-to-peak amplitude derived from binned 4D-MRI is prone to underestimation. However, as acquisitions span many respiratory cycles, the standard deviation (SD) of motion, required for probability-based planning approaches such as mid-ventilation, may be estimated accurately, provided binning-induced bias is corrected.
To develop and validate a motion SD estimator for equal-probability amplitude-binned 4D-MRI data that corrects for systematic underestimation caused by intra-bin motion variability.
4D-MRI data were acquired on a dynamic motion phantom driven by five patient-derived respiratory traces. Images were reconstructed using both a standard non-uniform fast Fourier transform (NUFFT) and the compressed sensing-based XD-GRASP algorithm with varying numbers (2-20) of equal-probability respiratory amplitude bins. An SD estimator based on the law of total variance was implemented to correct for intra-bin variance. The accuracy of peak-to-peak amplitude and SD was evaluated against ground-truth physical phantom motion, with performance under deformable motion tested on a digital phantom (CoMBAT).
Both reconstruction methods underestimated true peak-to-peak motion. While the inter-bin SD calculation also underestimated motion variability, the proposed correction substantially mitigated this bias. It yielded robust SD estimates with median residual underestimation of 0.13 mm (5 bins) and 0.11 mm (10 bins) across patient traces. In CoMBAT, the absolute bias of the total SD estimate was below 3% from five bins onward.
The proposed estimator provides an accurate measurement of respiratory motion SD from amplitude-binned 4D-MRI by accounting for intra-bin variability. This supports the reliability of 4D-MRI for planning strategies dependent on motion probability distributions. Additionally, the confirmed underestimation of peak-to-peak amplitude highlights a fundamental limitation in defining internal target volumes directly from binned 4D-MRI.
PMID:
43033365
Bibliographic data and abstract were imported from PubMed on 15 Nov 2026.
Read full publication at:
Please sign in
to see all details.
Advertisement
Stats
- Recommendations n/a n/a positive of 0 vote(s)
- Views 7
- Comments 0