List of Papers Browse by Subject Areas Author List
Abstract
The precise segmentation of intracranial aneurysms and their parent vessels (IA-Vessel) is a critical step for hemodynamic analyses, which primarily depend on Computational Fluid Dynamics (CFD). However, current segmentation methods predominantly focus on image-based evaluation metrics, often neglecting their practical effectiveness in subsequent CFD applications. Metrics such as the Dice similarity coefficient are insensitive to geometric topological abnormalities including vessel adhesion and surface irregularities which frequently cause CFD validation failures due to mesh generation errors or flow field distortions. To address these deficiencies, we construct an evaluation benchmark, the Intracranial Aneurysm Vessel Segmentation (IAVS) dataset, which is the first comprehensive, multi-center collection comprising 641 3D MRA images with 587 expert-verified annotations of aneurysms and IA-Vessels. In addition to image-mask pairs, the IAVS dataset provides standardized geometric files including STL models, vascular centerlines, and mesh grids alongside detailed hemodynamic analysis outcomes. Furthermore, we establish a standardized CFD applicability evaluation system that enables the automated and consistent conversion of segmentation masks into simulation-ready CFD models via a CFD conversion pipeline. Based on this system, we introduce a novel application-oriented metric, the CFD-Applicability Score (CFD-AS), to facilitate a comprehensive assessment of segmentation results focused on their clinical utility. The IAVS dataset and evaluation framework offer a robust foundation for bridging the gap between medical image segmentation and clinical hemodynamic research. The dataset, code, and model weights will be released after the paper is accepted.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026/paper/6086_paper.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to the Code Repository
https://github.com/AbsoluteResonance/IAVS
Link to the Dataset(s)
N/A
BibTex
@InProceedings{XiaFei_IAVS_MICCAI2026,
author = { Xiao, Feiyang AND Zhang, Yichi AND Li, Xigui AND Zhou, Yuanye AND Jiang, Chen AND Guo, Xin AND Han, Limei AND Li, Yuxin AND Zhu, Fengping AND Cheng, Yuan},
title = { { IAVS: A Multi-center Dataset and Applicability Evaluation System for Computational Fluid Dynamics-Oriented Intracranial Aneurysm Segmentation } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 16895},
month = {September},
page = {pending}
}
Reviews
Review #1
- Please describe the contribution of the paper
The main contribution of this paper is to reframe intracranial aneurysm segmentation evaluation from purely image-overlap quality to downstream CFD usability. Concretely, the paper introduces IAVS, a multi-center dataset for intracranial aneurysm and parent-vessel segmentation that includes not only MRA images and masks, but also CFD-oriented assets such as STL models, centerlines, mesh files, and hemodynamic analysis results. In addition, the paper proposes a standardized CFD applicability evaluation pipeline and a new metric, the CFD-Applicability Score (CFD-AS), to measure whether segmentation outputs are truly suitable for simulation rather than only achieving high Dice scores. The benchmark results support the paper’s key message that strong conventional segmentation metrics do not necessarily translate into simulation-ready results.
- 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 paper addresses an important and underexplored problem. The key message of the paper is highly relevant: good voxel-wise segmentation metrics do not necessarily imply good downstream CFD applicability. This is a meaningful and potentially impactful perspective for the medical image analysis community. The proposed dataset is valuable and application-oriented. IAVS is more comprehensive than a standard segmentation dataset because it includes segmentation masks, STL models, centerlines, mesh files, and CFD analysis outputs. This makes the dataset potentially useful for both segmentation and computational biomechanics research. The evaluation framework is novel and practically motivated. The proposed CFD applicability pipeline and CFD-AS metric go beyond conventional overlap-based evaluation and explicitly assess whether predicted segmentations can support downstream simulation. This is the most innovative part of the paper. The benchmark results support the central claim. The paper shows that even the best-performing baseline only achieves limited applicability scores, despite improvements in conventional segmentation metrics, highlighting a real gap between image-level performance and downstream clinical/simulation feasibility. The work has good long-term potential. Although the paper is not centered on a novel segmentation architecture, benchmark and dataset contributions of this kind can have substantial community impact if adopted broadly.
- 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.
The dataset description could be clearer in several places. The manuscript states that the dataset contains 641 3D MRA images and 587 expert-verified annotations, but the relationship between these numbers is not fully transparent. It would also help to clarify whether all seven types of data are available for all cases or only for a subset. The interpretation and role of CFD-AS could be discussed more explicitly. The metric is interesting and useful, but the paper would benefit from a clearer discussion of how CFD-AS should be used alongside Dice, HD95, clDice, and related metrics, rather than being interpreted in isolation. The failure analysis could be slightly more detailed. The paper demonstrates that many cases fail to become CFD-applicable, but it would be helpful to summarize the dominant causes of failure, such as localization errors, topological defects, or mesh-generation issues. The manuscript needs language polishing. There are noticeable grammar and style issues, including article usage, sentence construction, notation inconsistency, and mixed spelling conventions. These are mostly presentation issues, but they should be corrected in revision.
- 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 authors claimed to release the source code and/or dataset upon acceptance of the submission.
- 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?
I recommended acceptance because I find the paper novel, meaningful, and potentially impactful. Its main contribution is not a new segmentation architecture, but a strong reframing of the problem: good Dice scores do not necessarily imply good downstream CFD applicability. The proposed IAVS dataset, CFD applicability evaluation pipeline, and CFD-AS metric are valuable resources for the community, and the benchmark results support the paper’s central claim. The main weaknesses are in dataset clarity, discussion of CFD-AS, failure-case analysis, and language/presentation quality, rather than in the core idea itself. Since these issues appear addressable through revision, I believe minor revision is appropriate.
- 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
Review #2
- Please describe the contribution of the paper
This paper introduces a comprehensive, application-oriented dataset for IA-Vessel segmentation, designed to facilitate evaluation in the context of CFD. The dataset encompasses multi-level annotations, including segmentation masks, surface meshes, centerlines, and derived geometric representations, enabling assessment of downstream CFD applicability. Furthermore, the authors propose a novel metric to quantify whether segmentation results are directly usable for CFD simulation, and conduct extensive experiments to validate the effectiveness of the proposed evaluation framework.
- 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 paper provides a comprehensive dataset that includes image data, segmentation masks, surface meshes, centerlines, and other related annotations, which can be used to evaluate the applicability of segmentation results for downstream CFD tasks; 2.It presents a complete data construction pipeline and implements an automated processing workflow through scripting, enabling standardized and reproducible generation of CFD-ready data; 3.The authors propose the CFD-Applicability Score, a novel metric designed to assess whether segmentation results are suitable for downstream CFD applications. This metric considers multiple aspects, including segmentation quality, mesh conversion feasibility, and CFD simulation success, offering a more comprehensive evaluation than conventional metrics.
- 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.Regarding the automated scripting pipeline, the paper does not clearly specify how surface meshes are extracted from the segmentation results. If standard methods such as marching cubes are used, the resulting meshes are typically low-quality and often require substantial manual post-processing. It is therefore unclear whether a fully automated pipeline can reliably produce high-quality meshes; 2.The methods used for repairing holes and discontinuities in the segmentation are not sufficiently justified. In practice, fully automated repair techniques often have limited success rates and may introduce artifacts, potentially leading to a significant portion of unusable data or degraded geometric fidelity; 3.The process for automatically determining and cutting inlet and outlet boundaries is not clearly described. Given the complexity of vascular structures, it remains unclear how this step can be robustly automated without manual intervention; 4.The paper does not provide sufficient validation of the CFD simulation results. Specifically, it is unclear whether the repaired and processed meshes can produce hemodynamic simulations that are consistent with those derived from ground-truth geometries; 5.It is also unclear whether the dataset has undergone independent expert evaluation. The reliability and clinical validity of the generated meshes require further verification, particularly given the extensive automated processing involved.
- 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 authors claimed to release the source code and/or dataset upon acceptance of the submission.
- 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?
I believe this paper makes a meaningful contribution from an application-oriented perspective. The introduction of the IAVS dataset, along with the focus on evaluating the applicability of CFD, addresses an important gap in the field and has the potential to significantly impact future research. However, due to space limitations, several key components of the methodology are not described in sufficient detail, particularly regarding the automation pipeline and geometric processing steps. These omissions make it difficult to fully assess the robustness and reproducibility of the proposed framework. Overall, I find the core idea and practical relevance of the work compelling. With additional clarification and more detailed explanations, the paper would be significantly strengthened and would be suitable for acceptance.
- Reviewer confidence
Very confident (4)
- [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
Review #3
- Please describe the contribution of the paper
This paper presents a new benchmark named IAVS. It incorporates 3 datasets with 641 3D MRA images and annotations with CFD analysis results. It benchmarks several SOTA methods and does meaningful analysis.
- 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 gap is real. Dice score alone cannot show topological error and mesh generation failure. 2.The dataset scope is meaningful. The authors extend the current dataset with various annotation files to enrich the evaluation.
- 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.Inter-reader agreement should be presented. Given such a large volume of data, annotators may lead to errors, especially for segmentation tasks. How this problem is handled should be discussed. 2.The paper defines VTA, MGA, and BFA as three new metrics. But these are not presented in the results table. The usage of these metrics should be explained. 3.The paper says there are 641 images and 587 annotations. For images without annotations, the authors should clarify.
- 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 authors claimed to release the source code and/or dataset upon acceptance of the submission.
- 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 dataset is meaningful with newly curated images and rich annotations.
- 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
N/A
Meta-Review
Meta-review #1
- Your recommendation
Provisional Accept
- 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.
All three reviewers recognized the significant value and clinical relevance of your work. The consensus is that reframing the evaluation of intracranial aneurysm segmentation from pure voxel-wise overlap (Dice) to downstream computational fluid dynamics (CFD) applicability is a highly impactful contribution to the medical image analysis community. Furthermore, the introduction of the comprehensive IAVS dataset and the novel CFD-Applicability Score (CFD-AS) provides a much-needed benchmark and evaluation framework for future research. However, the reviewers have raised several important points regarding clarity, methodological details, and presentation. You are suggested to address the following issues in your final version.