The Role of Free-breathing GRASP MRI in Accurate Phase Matching with 4D-CT for Motion Representation in Liver Cancer Radiotherapy

Fei Liu; Guohua Wu; Jianrong Dai; Jiayun Chen; Junchao Li; Shengqi Chen

arxiv: 2606.08065 · v1 · pith:VHBKOQTWnew · submitted 2026-06-06 · ⚛️ physics.med-ph

The Role of Free-breathing GRASP MRI in Accurate Phase Matching with 4D-CT for Motion Representation in Liver Cancer Radiotherapy

Junchao Li , Shengqi Chen , Guohua Wu , Jianrong Dai , Jiayun Chen , Fei Liu This is my paper

Pith reviewed 2026-06-27 18:59 UTC · model grok-4.3

classification ⚛️ physics.med-ph

keywords GRASP MRI4D-CTliver SBRTrespiratory motionimage fusionmotion assessmentradiotherapy planning

0 comments

The pith

Free-breathing GRASP MRI represents liver tumor motion accurately only in the mid-respiratory phases of 4D-CT.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The study tests whether free-breathing GRASP MRI can capture the full range of breathing motion in liver tumors for radiation planning. Researchers fused GRASP MRI images with each phase of 4D-CT scans from 54 patients and measured how well they matched using cross-correlation scores. Good matches occurred only in the middle breathing phases from 30% to 60%, best at 50%, while extreme inhale and exhale phases matched poorly. This indicates that GRASP MRI alone misses important motion at the ends of the breathing cycle, so combining it with 4D-CT is needed for accurate tumor targeting in liver cancer radiotherapy.

Core claim

Free-breathing GRASP MRI cannot independently represent organ motion across all respiratory phases; it accurately characterizes motion only within the mid-respiratory phases (30%-60%), with optimal performance at the 50% phase. When used as a delineation standard in liver SBRT, GRASP MRI should be combined with 4D-CT or dynamic imaging modalities to ensure comprehensive motion assessment and accurate target volume definition.

What carries the argument

Image fusion between free-breathing GRASP MRI and 4D-CT phases, quantified by the maximum cross-correlation coefficient (MCC) as a measure of motion representation quality.

If this is right

GRASP MRI fusion quality is highest at the 50% phase and statistically similar at the 30%, 40%, and 60% phases.
Fusion quality declines significantly at the 0%, 10%, 20%, 80%, and 90% phases.
Blinded radiation oncologist scores remain above 4 for phases 30%-70% but fall below 4 outside that window.
GRASP MRI must be paired with 4D-CT or dynamic imaging for complete motion assessment in liver SBRT.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The phase restriction suggests GRASP MRI may work for mid-cycle gated radiotherapy but not for full-cycle 4D planning without extra data.
Differences in how each modality samples breathing patterns could explain why only mid phases align well.
Testing other free-breathing MRI sequences might reveal whether the limitation is specific to GRASP or common to the approach.

Load-bearing premise

The in-house registration program used for image fusion produces reliable and unbiased maximum cross-correlation coefficients that reflect true motion representation.

What would settle it

Repeating the fusions with a publicly available registration algorithm and finding that the peak match at the 50% phase disappears or shifts to other phases.

Figures

Figures reproduced from arXiv: 2606.08065 by Fei Liu, Guohua Wu, Jianrong Dai, Jiayun Chen, Junchao Li, Shengqi Chen.

read the original abstract

Objective: To determine whether free-breathing golden-angle radial sparse parallel (GRASP) magnetic resonance imaging (MRI) can represent respiratory-induced organ motion in patients with liver malignancies undergoing stereotactic body radiation therapy (SBRT). Methods: A retrospective analysis of 54 patients undergoing liver SBRT was conducted. Four-dimensional computed tomography (4D-CT), the gold standard for motion assessment, was used to characterize liver tumor motion. Image fusion was performed between free-breathing GRASP MRI and each respiratory phase of 4D-CT using an in-house registration program, with fusion quality quantified by maximum cross-correlation coefficient (MCC). Validation involved two blinded radiation oncologists: one repeated image fusion using the Eclipse-built-in module, while the other evaluated clinical relevance on a five-point scale. Results: The 50% respiratory phase of 4D-CT achieved the highest fusion quality with GRASP MRI, showing no significant differences compared to the 30% (P = 0.106), 40% (P = 0.632), and 60% (P = 0.792) phases. In contrast, fusion quality declined significantly beyond the mid-respiratory window (30%-60%), with poor fusion at the 0%, 10%, 20%, 80%, and 90% phases (P < 0.001). Validation by radiation oncologists corroborated these findings, with the 50% phase achieving the highest score. Subjective scores remained above 4 for phases 30%-70%, while scores for the remaining phases fell below 4. Conclusion: Free-breathing GRASP MRI cannot independently represent organ motion across all respiratory phases; it accurately characterizes motion only within the mid-respiratory phases (30%-60%), with optimal performance at the 50% phase. When used as a delineation standard in liver SBRT, GRASP MRI should be combined with 4D-CT or dynamic imaging modalities to ensure comprehensive motion assessment and accurate target volume definition.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

GRASP MRI matches 4D-CT phases reliably only in the mid-respiratory window per this 54-patient study, but the MCC numbers rest on an undocumented in-house registration tool.

read the letter

The main takeaway is straightforward: free-breathing GRASP MRI aligns well with 4D-CT only around the 30-60% phases, best at 50%, and falls off sharply outside that range. The authors reached this by fusing GRASP images to each 4D-CT phase in 54 liver SBRT patients and tracking maximum cross-correlation coefficients, then having two radiation oncologists run a separate Eclipse fusion and score clinical usefulness on a five-point scale.

The work does a few things cleanly. The patient number is reasonable for a retrospective imaging comparison, the statistical tests are reported with P-values, and the clinician scores provide an independent check that lines up with the MCC pattern. That combination gives the phase-specific claim some grounding in real clinical data rather than just theory.

The soft spot is exactly where the stress-test flagged it. The primary quantitative result depends on an in-house registration program whose algorithm, transformation model, and validation are not described. Without those details it is hard to know whether the MCC drop at the extreme phases reflects true motion mismatch or a registration failure mode that varies with phase. The Eclipse-based oncologist check uses a different tool and therefore does not fully rescue the MCC findings.

This paper is aimed at medical physicists and radiation oncologists who already use or are considering GRASP for liver motion assessment. A reader working on SBRT planning protocols would get practical guidance on when GRASP can stand alone and when it needs 4D-CT backup. It is not a methods paper and does not claim new technique, but the empirical comparison is the sort of incremental validation that clinics actually use.

I would send it to peer review. The data and clinician input are there, the question is clinically relevant, and the registration gap is fixable with added methods detail rather than a fatal flaw.

Referee Report

2 major / 0 minor

Summary. The manuscript reports a retrospective analysis of 54 patients with liver malignancies undergoing SBRT. It compares free-breathing GRASP MRI to each respiratory phase of 4D-CT via image fusion performed with an in-house registration program, quantifying quality by maximum cross-correlation coefficient (MCC). The central finding is that fusion quality is highest at the 50% phase, statistically indistinguishable from the 30%, 40%, and 60% phases (P = 0.106, 0.632, 0.792) but significantly poorer outside the 30-60% mid-window (P < 0.001 for 0%, 10%, 20%, 80%, 90%). Blinded radiation-oncologist validation on a five-point scale corroborates the phase-specific pattern.

Significance. If the quantitative results hold, the work indicates that GRASP MRI cannot serve as a standalone motion surrogate across the full respiratory cycle and should be combined with 4D-CT for target delineation in liver SBRT. The study draws on a moderately sized cohort, reports explicit P-values for phase comparisons, and includes independent clinician scoring, providing direct empirical grounding rather than purely self-referential metrics.

major comments (2)

[Methods (image fusion paragraph)] Methods (image fusion paragraph): The in-house registration program that computes the MCC values is entirely undocumented; no description is supplied of the transformation model, similarity metric, optimization procedure, regularization, or any validation against ground-truth deformations. Because the statistically significant MCC differences outside the 30-60% window rest exclusively on this program, it is impossible to exclude registration artifacts (local minima, intensity mismatch, or phase-dependent bias) as the source of the reported P < 0.001 drops.
[Results (MCC and clinician validation)] Results (MCC and clinician validation): The primary quantitative claim is based on MCC from the undocumented in-house program, while the Eclipse-based fusion performed by one oncologist uses an independent registration module; therefore the clinician scores do not constitute an independent check of the MCC-derived phase ranking that drives the central conclusion.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which help clarify the methodological transparency and validation approach in our study. We respond to each major comment below.

read point-by-point responses

Referee: [Methods (image fusion paragraph)] Methods (image fusion paragraph): The in-house registration program that computes the MCC values is entirely undocumented; no description is supplied of the transformation model, similarity metric, optimization procedure, regularization, or any validation against ground-truth deformations. Because the statistically significant MCC differences outside the 30-60% window rest exclusively on this program, it is impossible to exclude registration artifacts (local minima, intensity mismatch, or phase-dependent bias) as the source of the reported P < 0.001 drops.

Authors: We agree that the in-house registration program was insufficiently documented in the submitted manuscript. In the revised version we will add a complete description of the registration procedure, including the transformation model (rigid plus affine), the similarity metric (maximum cross-correlation), the optimization algorithm, regularization terms if any, and results of validation against synthetic ground-truth deformations. This addition will allow readers to evaluate the risk of phase-dependent artifacts. revision: yes
Referee: [Results (MCC and clinician validation)] Results (MCC and clinician validation): The primary quantitative claim is based on MCC from the undocumented in-house program, while the Eclipse-based fusion performed by one oncologist uses an independent registration module; therefore the clinician scores do not constitute an independent check of the MCC-derived phase ranking that drives the central conclusion.

Authors: The Eclipse-based fusion and the separate five-point clinical scoring were performed by two independent radiation oncologists blinded to the MCC results. Although the Eclipse module does not compute MCC, the clinician-assigned scores reproduce the identical phase ranking (highest at 50 %, statistically equivalent in the 30–60 % window, significantly lower outside it). This independent, clinically oriented corroboration supports the MCC-derived conclusion even if it does not numerically replicate the MCC values themselves. revision: no

Circularity Check

0 steps flagged

No circularity: empirical MCC measurements against independent 4D-CT reference

full rationale

The paper reports direct empirical fusion quality (MCC) between free-breathing GRASP MRI and each 4D-CT phase, plus clinician validation scores, against an external gold-standard modality. No derivation, equations, fitted parameters renamed as predictions, or self-citation chains appear in the provided text. The in-house registration program is an undocumented implementation detail that affects reproducibility but does not reduce the reported MCC differences to a definitional identity or self-referential input. The central claim therefore remains grounded in external benchmarks rather than internal construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

No free parameters or invented entities are introduced; the study relies on standard clinical domain assumptions about imaging gold standards and statistical validation.

axioms (1)

domain assumption 4D-CT is the gold standard for motion assessment in liver SBRT
Stated directly in the methods description.

pith-pipeline@v0.9.1-grok · 5925 in / 1170 out tokens · 25028 ms · 2026-06-27T18:59:45.053776+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

5 extracted references

[1]

Methods and Materials 2.1 Participants This retrospective study was approved by the institutional review board (IRB). Patients who underwent SBRT for liver cancer, liver metastasis, or other cancers, with upper abdominal magnetic resonance simulation (MR-Sim) from August 2019 to July 2020, were identified. Patients were included if they underwent upper ab...

2019
[2]

Merlo, J.M. and J.A. Nanzer, Multi-Pass Automotive Synthetic Aperture Radar Image Fusion, in 2023 IEEE Radar Conference (RadarConf23)

2023
[3]

Patnaik, and S

Sarvaiya, J.N., S. Patnaik, and S. Bombaywala, Image Registration by Template Matching 21 Using Normalized Cross-Correlation, in 2009 International Conference on Advances in Computing, Control, and Telecommunication Technologies

2009
[4]

Sirbu, and D

Pavel, S.M., G. Sirbu, and D. Aiordachioaie, Selection of Region of Interest in Thermal Images for the Classification of the Human Emotions, in 2024 International Conference on Development and Application Systems (DAS)

2024
[5]

Marignol, and M

Zhang, B., L. Marignol, and M. Kearney, The volumetric and dosimetric impacts of respiratory motion management in lung SBRT: A systematic review from 2019-2024. J Med Imaging Radiat Sci,

2019

[1] [1]

Methods and Materials 2.1 Participants This retrospective study was approved by the institutional review board (IRB). Patients who underwent SBRT for liver cancer, liver metastasis, or other cancers, with upper abdominal magnetic resonance simulation (MR-Sim) from August 2019 to July 2020, were identified. Patients were included if they underwent upper ab...

2019

[2] [2]

Merlo, J.M. and J.A. Nanzer, Multi-Pass Automotive Synthetic Aperture Radar Image Fusion, in 2023 IEEE Radar Conference (RadarConf23)

2023

[3] [3]

Patnaik, and S

Sarvaiya, J.N., S. Patnaik, and S. Bombaywala, Image Registration by Template Matching 21 Using Normalized Cross-Correlation, in 2009 International Conference on Advances in Computing, Control, and Telecommunication Technologies

2009

[4] [4]

Sirbu, and D

Pavel, S.M., G. Sirbu, and D. Aiordachioaie, Selection of Region of Interest in Thermal Images for the Classification of the Human Emotions, in 2024 International Conference on Development and Application Systems (DAS)

2024

[5] [5]

Marignol, and M

Zhang, B., L. Marignol, and M. Kearney, The volumetric and dosimetric impacts of respiratory motion management in lung SBRT: A systematic review from 2019-2024. J Med Imaging Radiat Sci,

2019