Abstract

Accurate assessment of vascular pathologies, such as aneurysms, requires reliable visualization of the cerebral vasculature. Time-of-Flight MR angiography (TOF-MRA) is a standard MRI technique for non-contrast, radiation-free imaging of the intracranial vasculature. However, anisotropic resolution and orientation-dependent sensitivity limit depiction of small or in-plane vessels, especially with slow flow. While TOF-MRA is primarily structural, 4D Flow MRI provides flow measurements, though long, gated, low-resolution acquisitions restrict brain coverage and limit routine clinical use in neuroimaging. In this work, we propose TOFlow, a neural reconstruction framework that fuses multi-orientation TOF-MRA acquisitions to address anisotropy and orientation-dependent signal effects and enable flow-related information estimation. The method integrates multi-view scans and incorporates a streamline-consistency constraint to resolve directional ambiguity and enforce coherent flow patterns. Experimental results demonstrate that TOFlow produces high-quality isotropic reconstructions while preserving signal-to-noise characteristics and enabling estimation of relative flow patterns and directionality. These findings are supported by quantitative image quality metrics and qualitative radiologists’ evaluation. Voxel-wise flow-direction and relative-intensity maps demonstrate improved directional coherence relative to 4D Flow MRI, as quantified by vessel-alignment and velocity-profile metrics. These results support TOFlow as a TOF-MRA-based framework for isotropic reconstruction and flow mapping.

Links to Paper and Supplementary Materials

Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026/paper/2261_paper.pdf

SharedIt Link: Not yet available

SpringerLink (DOI): Not yet available

Supplementary Material: Not Submitted

Link to the Code Repository

N/A

Link to the Dataset(s)

N/A

BibTex

@InProceedings{HasDar_TOFlow_MICCAI2026,
        author = { Hasin, Daria AND Bermano, Amit H. AND Sadeh Gonik, Udi AND Aizenstein, Orna AND Ben Bashat, Dafna},
        title = { { TOFlow: Learning-Based Isotropic Reconstruction and Flow Mapping from Multi-orientation TOF-MRA } },
        booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026},
        year = {2026},
        publisher = {Springer Nature Switzerland},
        volume = {LNCS 16888},
        month = {September},
        page = {pending}
}


Reviews

Review #1

  • Please describe the contribution of the paper

    This paper proposes an INR based framework for isotropic reconstruction by fusing multi-orientation TOF-MRA acquisitions. It tackles anisotropic resolution and orientation-dependent issues commonly seen in TOF-MRA. The authors employ a streamline-consistency constraint strategy, which improves directional coherence and physiologically consistent flow patterns. Quantitative comparison demonstrated enhanced reconstruction performance as to direct isotropic acquisitions.

  • Please list the major strengths of the paper: you should highlight a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.

    1.The authors address flow-dependent anisotropic signal effects issue and ease the estimation of flow-related information in TOF imaging. 2.The model is capable of learning vector-valued representations of flow magnitude and direction to account for the orientation dependent TOF signal.

  • Please list the major weaknesses of the paper. Please provide details: for instance, if you state that a formulation, way of using data, demonstration of clinical feasibility, or application is not novel, then you must provide specific references to prior work.

    1.The technical contribution seems not well presented. 2.Although the authors mentioned that the quantitative evluation by neuroradiologists also showed coherent and clinically interpretable flow patterns, but the specific results were not presented in the paper.

  • Please rate the clarity and organization of this paper

    Good

  • Please comment on the reproducibility of the paper. Please be aware that providing code and data is a plus, but not a requirement for acceptance.

    The submission does not mention open access to source code or data, but provides a clear and detailed description of the algorithm to ensure reproducibility.

  • Based on your review and your understanding of the MICCAI Scientific Code of Ethics, do you believe this submission may involve a potential ethics concern or violation?

    N/A

  • Optional: If you have any additional comments to share with the authors, please provide them here. Please also refer to our Reviewer’s guide on what makes a good review and pay specific attention to the different assessment criteria for the different paper categories: https://conferences.miccai.org/2026/en/REVIEWER-GUIDELINES.html

    N/A

  • Rate the paper on a scale of 1-6, 6 being the strongest (6-4: accept; 3-1: reject). Please use the entire range of the distribution. Spreading the score helps create a distribution for decision-making.

    (4) Weak Accept — marginally above the acceptance threshold, but would not mind if rejected, dependent on rebuttal

  • Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

    Although this study is not particularly innovative, it is well-suited to modeling flow-aware and orientation-dependent scenario typically existing in TOF-MRA acquisition.

  • Reviewer confidence

    Not confident (1)

  • [Post rebuttal] After reading the authors’ rebuttal, please state your final opinion of the paper.

    Accept

  • [Post rebuttal] Please justify your final decision from above.

    The methodological contribution is sound. A more structured radiological qualitative assessment is preferred in the camera-ready version. Given that the small cohort reflects a genuine field-level constraint, not a study design flaw, I recommend accept.



Review #2

  • Please describe the contribution of the paper

    This paper introduces TOFlow, a neural reconstruction framework for isotropic super-resolution and flow mapping from multi-orientation TOF-MRA. The key idea is to exploit the orientation-dependent signal properties of TOF-MRA by combining multiple acquisitions (axial, rotated axial, coronal, sagittal) and learning a vector-valued implicit neural representation (INR). The model extends NeSVoR to predict both magnitude and direction components of flow and introduces a streamline-consistency loss to enforce coherent flow patterns and resolve directional ambiguity.

    Experiments on a small cohort (6 subjects) show improved isotropic reconstruction quality compared to SVRTK and NeSVoR, with better preservation of vascular structures. The method also demonstrates the ability to estimate relative flow direction, with qualitative agreement and some quantitative improvement over 4D Flow MRI in a single subject. Radiologist evaluation suggests improved visualization of small vessels and clinically interpretable outputs.

  • Please list the major strengths of the paper: you should highlight a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.

    1.A major strength is the idea of leveraging multi-orientation TOF-MRA to infer both isotropic structure and flow-related information. This is conceptually novel and clinically relevant. 2.The adaptation of implicit neural representations to model vector-valued outputs (magnitude + direction) is a meaningful technical contribution. 3.The proposed streamline-consistency loss is interesting. It introduces a physically inspired constraint that improves directional coherence (as shown by improved FJA). 4.The inclusion of radiologist evaluation is a strong point.

  • Please list the major weaknesses of the paper. Please provide details: for instance, if you state that a formulation, way of using data, demonstration of clinical feasibility, or application is not novel, then you must provide specific references to prior work.

    1.The most critical limitation is the very small sample size. Only 3 subjects with ground truth isotropic scans and only 1 subject with 4D Flow MRI for validation of the flow. 2.The flow estimation is insufficiently validated. Comparison with 4D Flow MRI is done on only one subject, and the true velocity accuracy measured by the proposed method was not evaluated. I understand that flow mapping results are preliminary, but the paper somewhat overstates its significance in flow estimation. 3.The method requires multiple acquisitions (including rotated scans and orthogonal 2D TOF) which will require a long acquisition time. This may limit clinical adoption unless acquisition protocols and runtime are further optimized.

  • Please rate the clarity and organization of this paper

    Good

  • Please comment on the reproducibility of the paper. Please be aware that providing code and data is a plus, but not a requirement for acceptance.

    The submission does not mention open access to source code or data, but provides a clear and detailed description of the algorithm to ensure reproducibility.

  • Based on your review and your understanding of the MICCAI Scientific Code of Ethics, do you believe this submission may involve a potential ethics concern or violation?

    N/A

  • Optional: If you have any additional comments to share with the authors, please provide them here. Please also refer to our Reviewer’s guide on what makes a good review and pay specific attention to the different assessment criteria for the different paper categories: https://conferences.miccai.org/2026/en/REVIEWER-GUIDELINES.html

    N/A

  • Rate the paper on a scale of 1-6, 6 being the strongest (6-4: accept; 3-1: reject). Please use the entire range of the distribution. Spreading the score helps create a distribution for decision-making.

    (3) Weak Reject — marginally below the acceptance threshold, but would not mind if accepted, dependent on rebuttal

  • Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

    While the preliminary results are promising, the proposed technique is relying on modifications of the clinical protocol, long acquisition time, and is only validated on a small subject number.

  • Reviewer confidence

    Confident but not absolutely certain (3)

  • [Post rebuttal] After reading the authors’ rebuttal, please state your final opinion of the paper.

    Reject

  • [Post rebuttal] Please justify your final decision from above.

    The authors acknowledged the limitation with extremely small sample size (only 3), insufficient validation with flow measurement, and longer scan time than standard TOF MRA, but did not convince me why this work is valuable despite the limitations.



Review #3

  • Please describe the contribution of the paper

    This work presents a neural reconstruction framework for joint super-resolution and blood-flow mapping in cerebral TOF-MRA, integrating anatomical priors with orientation-dependent TOF signal variations obtained from multi-directional acquisitions to encode flow directionality.

  • Please list the major strengths of the paper: you should highlight a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.
    • The proposed method introduces a novel joint approach for super-resolution and blood-flow mapping, addressing two tasks simultaneously
    • The initial results are promising and demonstrate the feasibility of this integrated strategy
    • The evaluation combines standard quantitative reconstruction metrics with expert radiological assessment, providing both technical and clinical perspectives on performance
    • The study includes comparisons with current state-of-the-art SVR methods, allowing the proposed approach to be appropriately contextualised
    • Despite the relatively small dataset, the inclusion of a pathological case is valuable - enabling preliminary assessment in a more clinically relevant scenario
  • Please list the major weaknesses of the paper. Please provide details: for instance, if you state that a formulation, way of using data, demonstration of clinical feasibility, or application is not novel, then you must provide specific references to prior work.
    • While the quantitative evaluation against state-of-the-art methods is extensive, the radiological assessment lacks depth and specificity: “reconstruction-related artifacts in some subjects, including pixelated or artificial appearance” - this statement is quite vague and requires more detailed characterization; given the potential clinical applicability and the role of radiologists as end users, a more thorough and structured radiological qualitative analysis would significantly strengthen the study
    • A joint assessment of both reconstruction quality and the derived blood-flow maps would be particularly valuable
    • The dataset is heterogeneous in terms of available reference standards - while this is understandable given acquisition constraints, the variability limits the strength of comparisons with gold-standard modalities; for a small cohort, a more consistent dataset would enable more robust and interpretable evaluation, e.g., the comparison with 4D Flow MRI is only demonstrated in a single case - although promising, this is insufficient to support strong conclusions.
    • An ablation study would be beneficial for isolating the contribution of key components of the framework, e.g. streamline-consistency loss, methodological choices, etc.
  • Please rate the clarity and organization of this paper

    Satisfactory

  • Please comment on the reproducibility of the paper. Please be aware that providing code and data is a plus, but not a requirement for acceptance.

    The submission does not provide sufficient information for reproducibility.

  • Based on your review and your understanding of the MICCAI Scientific Code of Ethics, do you believe this submission may involve a potential ethics concern or violation?

    N/A

  • Optional: If you have any additional comments to share with the authors, please provide them here. Please also refer to our Reviewer’s guide on what makes a good review and pay specific attention to the different assessment criteria for the different paper categories: https://conferences.miccai.org/2026/en/REVIEWER-GUIDELINES.html

    N/A

  • Rate the paper on a scale of 1-6, 6 being the strongest (6-4: accept; 3-1: reject). Please use the entire range of the distribution. Spreading the score helps create a distribution for decision-making.

    (4) Weak Accept — marginally above the acceptance threshold, but would not mind if rejected, dependent on rebuttal

  • Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

    The proposed work is novel and has promising potential for both research and clinical impact.

  • Reviewer confidence

    Confident but not absolutely certain (3)

  • [Post rebuttal] After reading the authors’ rebuttal, please state your final opinion of the paper.

    N/A

  • [Post rebuttal] Please justify your final decision from above.

    N/A



Author Feedback

We thank the reviewers and meta-reviewer for their feedback. We appreciate the constructive comments and the recognition of the clinical relevance and potential of TOFlow.

To clarify the technical contribution, the novelty of TOFlow lies in leveraging multiple TOF-MRA acquisition orientations for isotropic reconstruction and voxel-wise flow-direction estimation. In addition, TOFlow extends implicit neural reconstruction to predict both flow magnitude and direction components and introduces a streamline-consistency loss to encourage directional consistency along vessels.

We agree that this work represents an initial feasibility study and that larger-scale validation, particularly with 4D Flow MRI, is important future work. The current study focuses on demonstrating the feasibility of estimating flow-direction information from multi-orientation TOF-MRA rather than quantitative flow measurement.

The comparison with 4D Flow MRI in one subject was intended as an initial proof-of-concept reference rather than definitive validation. Acquiring multi-orientation TOF-MRA together with gated 4D Flow MRI substantially increases acquisition complexity and duration, limiting the availability of such datasets in this preliminary study. Despite this limitation, the observed agreement in vessel-aligned directional organization and improved FJA measurements support the feasibility of estimating flow information from multi-orientation TOF-MRA.

We acknowledge that the cohort size is limited. The study required acquisition of multiple TOF orientations, high-resolution isotropic reference scans, and additional 4D Flow MRI in one subject, substantially increasing acquisition complexity. Despite the limited cohort, the method demonstrated consistent reconstruction improvements across subjects using quantitative metrics, radiological evaluation, and comparison against existing reconstruction approaches.

We also agree that the current acquisition protocol is longer than standard clinical TOF-MRA workflows. The goal of this study was to investigate whether complementary orientation-dependent TOF information can be leveraged for joint reconstruction and flow-related estimation. The reconstruction itself is performed offline (~13 minutes per subject per task) and therefore does not affect scan duration or clinical workflow, as the reconstructed volumes would be available to the radiologist together with the final imaging results. Future work will investigate optimized and reduced acquisition configurations, as well as accelerated INR training strategies.

We would also like to clarify that an ablation study evaluating the streamline-consistency loss is already included in Section 3.5.Inclusion of the loss improved the subject-averaged FJA while maintaining similar FD values, demonstrating improved directional coherence without altering the overall velocity distribution.

Finally, the manuscript includes detailed acquisition settings, preprocessing procedures, architecture design, optimization parameters, training details, evaluation metrics, and implementation details.

Overall, we hope the rebuttal clarifies the technical contribution, feasibility-focused scope, and reproducibility of TOFlow.




Meta-Review

Meta-review #1

  • Your recommendation

    Invite for Rebuttal

  • Please justify your decision. In case you deviate from the reviewers’ recommendations, explain in detail the reasons why. In case of an invitation for rebuttal, clarify which points are important to address in the rebuttal.

    Strengths: The work presents a clinically relevant and conceptually novel use of multi-orientation TOF-MRA to jointly improve reconstruction/super-resolution and infer flow-related magnitude and direction information, supported by implicit neural representations, a streamline-consistency loss, comparisons with SVR methods, and some radiological assessment. Weaknesses: The evidence remains preliminary because the sample size and ground-truth flow validation are very limited, the multi-acquisition protocol may be clinically burdensome, the technical contribution and radiological assessment need clearer presentation, and ablations are needed to isolate the value of key components. The authors therefore be invited for the rebuttal to clarify the novelty and reproducibility.

  • After you have reviewed the rebuttal and updated reviews, please provide your recommendation based on all reviews and the authors’ rebuttal.

    Accept

  • Please justify your recommendation.

    The post-rebuttal assessment leans more toward Accept than Reject.



Meta-review #2

  • After you have reviewed the rebuttal and updated reviews, please provide your recommendation based on all reviews and the authors’ rebuttal.

    Accept

  • Please justify your recommendation.

    The authors have answered some comments, esp on methodological novelty and on the ablation study already present. The major limitation if a limited dataset was not (and could not be) addressed in the rebuttal. However, given that the main strength acknowledged universally is a methodological contribution combining blood flow and superresolution in one joint method, merit can be found in the paper and it is likely to lead to significant interest in the community.



Meta-review #3

  • After you have reviewed the rebuttal and updated reviews, please provide your recommendation based on all reviews and the authors’ rebuttal.

    Reject

  • Please justify your recommendation.

    Although the method is interesting, especially the estimation of flow, as the reviewers mentioned, a major disadvantage is that it requires a very long acquisition time. For that reason, the authors just had a few subjects to test it. Thus, there is no clear path towards a translation to the clinics in the future.



Meta-review #4

  • After you have reviewed the rebuttal and updated reviews, please provide your recommendation based on all reviews and the authors’ rebuttal.

    Accept

  • Please justify your recommendation.

    Reviewer 2 remains concerned about the small validation cohort and longer scan times. I agree that the authors’ rebuttal successfully clarified these constraints as inherent to the clinical protocol. Given the strong methodological novelty of the vector-valued INR framework and its compelling radiological utility, I recommend accepting this paper.



back to top