List of Papers By topics Author List
| Paper Info | Reviews | Meta-review | Author Feedback | Post-Rebuttal Meta-reviews |
Authors
Meng Song, Shi-Qi Liu, Xiao-Liang Xie, Xiao-Hu Zhou, Zeng-Guang Hou, Yan-Jie Zhou, Xi-Yao Ma
Abstract
In endovascular interventional therapy, automatic common
iliac artery morphological analysis can help physicians plan surgical procedures
and assist in the selection of appropriate stents to improve surgical
safety. However, different people have distinct blood vessel shapes,
and many patients have severe malformations of iliac artery due to hemangiomas.
Besides, the uneven distribution of contrast media make
it difficult to make an accurate morphological analysis of the common
iliac artery. In this paper, a novel fusion network, combining CNN and
Transformer is proposed to address the above issues. The proposed FTU-Net
consists of a parallel encoder and a Cross-Fusion module to capture
and fuse global context information and local representation. Besides, a
hybrid decoder module is designed to better adapt the fused features.
Extensive experiments have demonstrated that our proposed method
significantly outperforms the best previously published results for this
task and achieves the state-of-the-art results on the common iliac artery
dataset built by us and two other public medical image datasets. To the
best of our knowledge, this is the first approach capable of common iliac
artery segmentation.
Link to paper
DOI: https://link.springer.com/chapter/10.1007/978-3-031-16449-1_6
SharedIt: https://rdcu.be/cVRUM
Link to the code repository
https://github.com/SongCASIA/FTU_Net
Link to the dataset(s)
N/A
Reviews
Review #2
- Please describe the contribution of the paper
This paper proposed a new method for morphological analysis of the common iliac artery. This paper provided a new framework that combines CNN and Transformer, addressing the challenging task of CIA morphological analysis. This paper is the first automatic approach to morphological analysis of the common iliac artery.
- Please list the main strengths of the paper; you should write about 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 application is novel.
- Morphological analysis is a sound idea.
- It is good to see that the authors attempted to use the ablative analysis to back up the claims regarding the proposed network modules.
- Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.
- I feel that there is little methodology advance from a technical point of view.
- Although I found the work interesting and potentially very useful, my major concern is on whether analysis morphological algorithm can “obtain access diameter, the minimum inner diameter, and the maximum curvature of CIA with arbitrary shapes.” Otherwise, I think the experiment is far from validating this strong claim.
- Please rate the clarity and organization of this paper
Satisfactory
- Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance
Paper is clear that an expert could reproduce
- Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://conferences.miccai.org/2022/en/REVIEWER-GUIDELINES.html
- Limited innovation in methodology. The U-shaped and Cross-fused is actually not new. They are simple modifications and applications of [12] and [3], respectively.
- Insufficient results for validation. Could you add more experiments to show the accuracy of morphological? The paper is the lacks evaluation related to comparing the quantitative results between anatomical information and the ground truth.
- The author claims “the shape and size of the iliac leg are important indicators”, which require some references.
- The method is complicated and has included many modules. However, there are modules, such as Hybrid Decoder, that are not well validated.
- There are some misspellings in the manuscript. e.g.” the maximum value of the remains represent the largest tortuosity of the vessel”.
- Rate the paper on a scale of 1-8, 8 being the strongest (8-5: accept; 4-1: reject). Spreading the score helps create a distribution for decision-making
3
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
Although the clinical hypothesis is meaningful, the paper lacks technical innovation. More importantly, the paper is the lacks evaluation related to comparing the quantitative results between anatomical information and the ground truth.
- Number of papers in your stack
5
- What is the ranking of this paper in your review stack?
3
- Reviewer confidence
Confident but not absolutely certain
- [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed
5
- [Post rebuttal] Please justify your decision
According to the challenges of the specific clinical task, the design of the Cross Fusion Block is reasonable. And more specific experimental results will be added in the final version. So I think the author addresses the main shortcomings of this paper.
Review #3
- Please describe the contribution of the paper
The paper proposed a new module Cross-Fusion block and Transformer structure. And the paper designed to provide a new solution for common iliac artery segmentation and morphological analysis tasks.
- Please list the main strengths of the paper; you should write about 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.
Clear description: The structure diagram and formula are clearly described, and the method part is clearly organized. Clear motivation: Fluent writing and clear clinical meaning.
- Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.
The comparative experiment is not sufficient and lacks the comparison and analysis of the number of parameters and computation (e,g, FLOPS) There were some minor mistakes.
- Please rate the clarity and organization of this paper
Very Good
- Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance
No source code is given. But, the method section of the paper is introduced so that readers can hopefully repeat or refer to it.
- Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://conferences.miccai.org/2022/en/REVIEWER-GUIDELINES.html
- Is there any comparison between the method in the paper and criss-fusion Block[15] or other attention methods, such as spatial attention, CBAM(Convolutional Block Attention Module), etc.
- It is recommended to add comparison of speed or calculation amount with Unet or other networks.
- Minor problems: Page 4: Ψ∈ R^(H+W−1)×(WW) , I think it should be HW
- Is there a quantitative analysis for calculating vessel diameter and curvature? For example, comparison with manual measurement through radiologist, or, clinical significance and guidance.
- Rate the paper on a scale of 1-8, 8 being the strongest (8-5: accept; 4-1: reject). Spreading the score helps create a distribution for decision-making
5
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
This paper has a clear motive and structure, and the network design is also innovative. There are details that can be improved and modified in the experimental section.
- Number of papers in your stack
4
- What is the ranking of this paper in your review stack?
4
- Reviewer confidence
Confident but not absolutely certain
- [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed
Not Answered
- [Post rebuttal] Please justify your decision
Not Answered
Review #4
- Please describe the contribution of the paper
The authors propose a novel fusion network, combining CNN and Transformer is proposed to address the morphological analysis of common iliac artery issues. The experiment demonstrates that their method outperforms the best previously published results for this task.
- Please list the main strengths of the paper; you should write about 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 designs an interesting way that combines CNN and Swin Transformer for the ends of the common iliac artery and the edge pixels segmentation. The authors propose a morphological analysis algorithm to obtain anatomical information on the common iliac arteries (the minimum inner diameter, access diameter, and tortuosity).
- Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.
The design of the method mainly focuses on the combination of CNN and Swin Transformer, which is weakly related to morphological analysis of the common iliac artery.
- Please rate the clarity and organization of this paper
Good
- Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance
The description of the proposed method is clear and the paper is reproducible.
- Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://conferences.miccai.org/2022/en/REVIEWER-GUIDELINES.html
The proposed morphological analysis algorithm isn’t outlined in Abstract. “make it difficult” is changed to “makes it difficult” in Abstract. The chaotic topological order of modules in Fig. 1 makes the figure unclear at a glance. The relationship between the method and task can be better explained. “represent the largest tortuosity” is changed to “represents the largest tortuosity” In Diagnosis Algorithm
- Rate the paper on a scale of 1-8, 8 being the strongest (8-5: accept; 4-1: reject). Spreading the score helps create a distribution for decision-making
5
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
The proposed network that combines CNN and Swin Transformer is appealing. However, the task relevance of the method can be better designed.
- Number of papers in your stack
5
- What is the ranking of this paper in your review stack?
1
- Reviewer confidence
Very confident
- [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed
Not Answered
- [Post rebuttal] Please justify your decision
Not Answered
Primary Meta-Review
- Please provide your assessment of this work, taking into account all reviews. Summarize the key strengths and weaknesses of the paper and justify your recommendation. 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.
This paper presents a novel fusion network for morphological analysis of common iliac artery
Strength: 1) well-written 2) sound methodology 3) better performance than five other methods.
Weakness: 1) experiments could be improved. 2) the novelty of the methodology should be further explained.
Three reviewers have given the following comments: 1) R2 mainly considers the limited novelty and insufficient experiments. 2) R3 mainly considers clear description and motivation, but the insufficient experiments. 3) R4 mainly considers interesting idea but weakly related to morphological analysis of the common iliac artery.
- What is the ranking of this paper in your stack? Use a number between 1 (best paper in your stack) and n (worst paper in your stack of n papers). If this paper is among the bottom 30% of your stack, feel free to use NR (not ranked).
3
Author Feedback
Thanks for all your constructive feedback and valuable suggestions. We apologize for the minor mistakes in the paper and they will be carefully corrected in the final version.
We’d like to emphasize the significance and novelty of our work, which is the first automatic approach for segmentation and morphology analysis of the common iliac artery(CIA). The size and curvature of the CIA determine the choice of surgical instruments and the planning of the surgical procedure. Improper device selection may lead to postoperative thrombosis and internal leakage. So, we built an expert-annotated dataset and proposed this method, which achieved state-of-the-art segmentation performance and automatic acquisition of anatomical information.
“Limited innovation in methodology” Firstly, due to the shape variability, uneven contrast agent, and background noise of DSA images, we need to combine local details and global contextual information to address these issues. The fusion method in TransFuse[13] is a modification of spatial attention and channel attention, which cannot generate dense contextual information and different features from Transformer and CNN obtain the same contextual dependencies. Therefore, we propose the Cross Fusion block to incorporate dense and feature-wise contextual information. In this block, a dense cross-connection with dual input channels is designed to model the interaction between two different features. Secondly, past works mostly adopt the ASPP structure as a decoder, which can lead to the loss of information in the up-sampling process. So, the Hybrid Decoder is designed to sort the different parts of the fusion features that are relevant to CNN and Transformers so that the features can restore the segmentation image better. We verified this module in the ablation experiment. When the decoder part was replaced by ASPP with Hybrid Decoder, the performance has been improved and better segmentation results have been achieved. Thirdly, the U-Shaped structure can increase the receptive field and integrate multi-scare features through down-sampling and skip-connection, so it is the most commonly used structure in medical image segmentation. So, we followed this structure.
“Insufficient results for validation of morphology algorithm” Manually calibrated diameter and curvature by experts using Siemens commercial software are collected. To demonstrate the accuracy of the morphological algorithm, mean error and mean square error (MSE) between manual calibration and our results are calculated. For the minimized diameter, the mean error is 2.04mm and the MSE is 7.88. For the access diameter, the mean error is 2.11mm and the MSE is 9.69. For the curvature of vessels, the mean error is 0.2. through the retrospective analysis of actual surgical cases, it is found that the results of the automatic algorithm can accurately help physicians to make surgical procedures and select medical instruments, so the measurement indexes are considered to meet clinical needs.
“Comparison between the method in the paper and other attention methods” To demonstrate the progressiveness of the Cross Fusion block, we compare it with spatial attention, CBAM, and criss-cross attention[15]. The mean Dice of these methods are 0.8579, 0.8532, and 0.8646, respectively, none of which is as good as our method’s result of 0.9281. Also, we calculate the FPS, GFLOPs, and params of our proposed network, TransUNet and TransFuse. There is no significant increase in the three values. But in the same order of magnitude of params, our results improved by 3.0% and 1.1% over the other two methods, respectively. The detailed tables of the above experimental results will be presented in the final version.
Thank you again for your careful reading and valuable comments.
Post-rebuttal Meta-Reviews
Meta-review # 1 (Primary)
- Please provide your assessment of the paper taking all information into account, including rebuttal. Highlight the key strengths and weaknesses of the paper, clarify how you reconciled contrasting review comments and scores, indicate if concerns were successfully addressed in the rebuttal, and provide a clear justification of your decision. If you disagree with some of the (meta)reviewer statements, you can indicate so in your meta-review. Please make sure that the authors, program chairs, and the public can understand the reason for your decision.
The rebuttal addresses my concerns. 1)Although the comparison indicators in the experimental part are too few, only the mean error and mean square error (MSE), experimental design is good . I hope that you can add more results to explain the superiority of the method in final version. 2)The method mainly depends on the combination of the network part.
- After you have reviewed the rebuttal, please provide your final rating based on all reviews and the authors’ rebuttal.
Accept
- What is the rank of this paper among all your rebuttal papers? Use a number between 1/n (best paper in your stack) and n/n (worst paper in your stack of n papers). If this paper is among the bottom 30% of your stack, feel free to use NR (not ranked).
8
Meta-review #2
- Please provide your assessment of the paper taking all information into account, including rebuttal. Highlight the key strengths and weaknesses of the paper, clarify how you reconciled contrasting review comments and scores, indicate if concerns were successfully addressed in the rebuttal, and provide a clear justification of your decision. If you disagree with some of the (meta)reviewer statements, you can indicate so in your meta-review. Please make sure that the authors, program chairs, and the public can understand the reason for your decision.
The rebuttal addressed the conerns on contribution and limited experiments, as reflected by the acceptance by all reviewers.
- After you have reviewed the rebuttal, please provide your final rating based on all reviews and the authors’ rebuttal.
Accept
- What is the rank of this paper among all your rebuttal papers? Use a number between 1/n (best paper in your stack) and n/n (worst paper in your stack of n papers). If this paper is among the bottom 30% of your stack, feel free to use NR (not ranked).
4
Meta-review #3
- Please provide your assessment of the paper taking all information into account, including rebuttal. Highlight the key strengths and weaknesses of the paper, clarify how you reconciled contrasting review comments and scores, indicate if concerns were successfully addressed in the rebuttal, and provide a clear justification of your decision. If you disagree with some of the (meta)reviewer statements, you can indicate so in your meta-review. Please make sure that the authors, program chairs, and the public can understand the reason for your decision.
The key strength of this work is a network model for analyzing the morphology of the common iliac artery. Reviewers were concerned regarding the originality of the method, however, authors were convincingly demonstrating their novelty. In the context of the relevant downstream clinical application, i.e. selection of appropriate surgical instruments based on the morphology, the work is of interest for the MICCAI community.
A number of clarification regarding the experimental setup were given in the rebuttal, as a response to further reviewer comments. After rebuttal phase, all reviewers agreed on acceptance. This is another piece of evidence that the work is likely to be of interest for MICCAI.
- After you have reviewed the rebuttal, please provide your final rating based on all reviews and the authors’ rebuttal.
Accept
- What is the rank of this paper among all your rebuttal papers? Use a number between 1/n (best paper in your stack) and n/n (worst paper in your stack of n papers). If this paper is among the bottom 30% of your stack, feel free to use NR (not ranked).
9