Paper Info Reviews Meta-Review Author Feedback Post-rebuttal Meta-Reviews Back to top
List of Papers By topics Author List
Paper Info Reviews Meta-review Author Feedback Post-Rebuttal Meta-reviews

Authors

Xi Fang, Daeseung Kim, Xuanang Xu, Tianshu Kuang, Nathan Lampen, Jungwook Lee, Hannah H. Deng, Jaime Gateno, Michael A. K. Liebschner, James J. Xia, Pingkun Yan

Abstract

In CMF surgery, the planning of bone movement to achieve a desired facial outcome is a challenging task. Current bone-driven approaches focus on normalizing the bone with the expectation that the facial appearance will be corrected accordingly. However, due to the complex non-linear relationship between bone and the face, such bone-driven methods are insufficient to fully correct facial deformities. Despite efforts to simulate facial changes resulting from bony movement, surgical planning still relies on iterative revisions and educated guesses. To address these issues, we propose a soft-tissue-driven framework that can automatically create and verify surgical plans. Our framework consists of a bony planner network that estimates the bony movements required to achieve the desired facial outcome and a facial simulator that can simulate the possible facial changes resulting from the estimated bony plans. By combining these two models, we can verify and determine the final bony movement required for planning. The proposed framework was evaluated using a clinical dataset, and our experimental results demonstrate that the soft-tissue driven approach greatly improves the accuracy and efficacy of surgical planning when compared to the conventional bone-driven approach.

Link to paper

DOI: https://doi.org/10.1007/978-3-031-43996-4_18

SharedIt: https://rdcu.be/dnwOS

Link to the code repository

N/A

Link to the dataset(s)

N/A

Reviews

Review #4

  • Please describe the contribution of the paper

    This paper proposes a soft-tissue-driven framework that can automatically create and verify surgical plan to address the issues of bone movement planning in CMF surgery.

  • 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.
    1. A soft tissue-driven CMF surgical planning method is proposed, which can greatly reduce planning time by eliminating the need for repeated guessing of bone movements.
    2. A deep learning-based bony planner network is developed, which can estimate the underlying bony movement needed for changing a facial appearance into a targeted one.
    3. The developed facial simulator module can qualitatively assess the effect of surgical plans on facial appearance, for virtual validation.
  • 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.
    1. The paper lacks more comparison with other methods.
    2. The paper lacks comparison with existing clinical methods in time and effect.
  • 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

    The proposed method should be reproducible, but the method is more complex and open source code is recommended.

  • 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/2023/en/REVIEWER-GUIDELINES.html
    1. It is recommended that the authors add more comparisons with other methods, especially in terms of time and efficacy with existing clinical methods.
    2. The proposed method only improves the accuracy by a few tenths of a millimeter. Is this meaningful for clinical applications?
    3. How is the desired facial surface produced?
    4. Only 4096 points are used for the facial and bony surfaces in the paper, and only 1024 points are used for each bony segment. Will too much detail be lost?
    5. Bony planner and facial simulator are two independently trained networks, shouldn’t joint training get better results?
  • 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

    6

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

    I think it is an interesting planning research for craniofacial surgery, which has certain clinical application prospects.

  • Reviewer confidence

    Very confident

  • [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed

    N/A

  • [Post rebuttal] Please justify your decision

    N/A



Review #2

  • Please describe the contribution of the paper
    • This paper aims to address a challenge in craniomaxillofacial surgical planning, namely the planning of bone movement to achieve a desired facial outcome.
    • The authors proposed a soft-tissue-driven framework (deep learning) capable of generating surgical plans automatically, allowing for the prediction of bony movement based on the preoperative and desired facial surfaces.
    • The proposed method was verified using a clinical dataset consisting of 34 sets.
  • 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 was well-structured, and the background and challenges were well explained.
    • The “soft-tissue driven” concept is intriguing, as most previous studies relied on simulators to adjust bone movements until the desired facial surface was achieved. The “soft-tissue driven” idea explores the complex relationship between facial changes and bone movements, which is more straigtforward and has important clinical value.
    • The authors were able to validate the proposed method using clinical data.
    • The authors provided clear figures and visualizations in their results, presenting their findings in an easy-to-understand manner.
  • 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 training process has not been clearly explained. It seems that only the “Bony planner” network needs to be trained. However, it is unclear how the augmentation of data was done on both the facial and bone.

    • The evaluation details are not clear. It is not explained how five-fold cross-validation was performed on the 34 sets of patient data. The final results are presented as statistics in general, without providing the five-fold results.

    • The training data of 34 may not be sufficient to capture the complex relationship between facial changes and bone movement.

    • The perturbation in the proposed method is confusing. The authors have used the bony planner, which relates facial changes to bone movements, to generate the plan. It is not clear why perturbation was performed before the bony planner. Additionally, the physical meaning of perturbation is not easy to understand. It is unclear why the perturbation was not directly added to the generated bony plan.

    • The authors used ACMT-Net (published in MICCAI 2021 and trained with 40 patients) for self-verification in the method. However, ACMT-Net itself needs to be further verified since only a small number of data were used in the ACMT-Net paper.

  • 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

    not easy to follow just according to the text. But the authors said they would provide the related code

  • 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/2023/en/REVIEWER-GUIDELINES.html

    The authors have presented an interesting work. I have some suggestions to improve the paper:

    (1) The training process, especially the data augmentation part, needs to be explained in more detail. The authors have only used 34 sets of patient data, and the relationship between facial change and bone movement can be very complex. Therefore, it is important to explain how the authors utilized data augmentation to increase the size of the training dataset.

    (2) More details about the 5-fold cross-validation technique should be added, and it would be better to present the corresponding results.

    (3) The authors mentioned in the paper that previous methods can be time-consuming. Therefore, the time cost of the proposed method needs to be presented.

    (4) It would be helpful if the authors could define the symbols used in Equation (2), such as phi and theta.

  • 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 idea was interesting but the method part is not clear, which may confuse readers

  • 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

    N/A

  • [Post rebuttal] Please justify your decision

    N/A



Review #3

  • Please describe the contribution of the paper

    This paper proposes a novel soft tissue-driven approach for CMF surgical planning, which reduces planning time, developed a deep learning model to estimate the bony movement, and evaluated the effectiveness qualitatively on facial appearance.

  • 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 main strengths of the paper are

    • The proposed framework based on soft-tissue driven approach is novel
    • The cross-validation was performed with comparing with the state-of-the-art bone-driven method
  • 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 main weaknesses of the paper are

    • Anatomical constraints of the bone are not considered well
    • The effectiveness using actual patient data is unclear
  • 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

    Mostly, this work will be reproduced.

  • 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/2023/en/REVIEWER-GUIDELINES.html

    As a whole, this paper is well written. However, the facial surface is derived from the underlying a kind of different mechanisms: skin and fat material properties and thickness, and muscle distribution and connectivity to the skin and bone. In addition, Craniomaxillofacial surgery already has a conventional approach used in clinical procedure. Therefore, the proposed planing should compare with the conventional approach to show its usefulness.

  • 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?

    Major concerns:

    • Anatomical constraints of the bone are not considered well
    • The effectiveness using actual patient data is unclear
    • It is not unknown whether the proposed planing is superior to the conventional approach of clinical Craniomaxillofacial surgery.
  • 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

    N/A

  • [Post rebuttal] Please justify your decision

    N/A



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.

    There is agreement among reviewers that this manuscript describes a novel approach to surgery planning by incorporating soft-tissue motion. The method is evaluated w.r.t. to several baselines and on clinical data.

    The chief weaknesses include some need for clarification and modeling, the overall small dataset size, possibility to further expand on baselines and metrics, and some challenges around modeling.

    As possible, responses to these concerns should be incorporated in an updated version of the paper.



Author Feedback

We appreciate the reviewers and AC for the positive opinions and constructive comments on our work. In the camera-ready version, we will provide further clarity on certain aspects of the data and the model and expand the discussion section to include more potential improvements and clinical insights related to our approach. Some specific responses are included as follows.

In response to the questions from Reviewer#1 regarding the specifics of the model and its implementation, bony planner is the only network to be trained in this work. During the “bony planner” training process, both facial and bone data are subjected to the same random flipping and translation, ensuring that the relative position and scale of facial changes and bony movements remain unchanged. During the “bony planner” inference process, perturbations are applied to the inputs to generate multiple potential post-operative bony surfaces. Since there are multiple alternatives for perturbating potential bony plans, we will explore other methods to create a variety of candidate bony plans in the future. This may include implementing techniques such as network dropout.

In response to Reviewer#1’s inquiry about the specifics of the data used, we clarify that when performing the five-fold cross validation, 34 retrospective clinical data sets are partitioned into five groups: four groups each containing 7 sets, and one group with 6 sets. During each round of validation, four of these groups (folds) are used for training and the remaining group is used for testing. Once all the rounds are completed, the results from all 34 data sets are used for statistical analysis and comparison. We agree that the limited data may restrict the network performance. Thus we have employed data augmentation to aid the network training and to learn the complex relationship between facial changes and bony movement. To demonstrate the learned relationship, in the supplementary file, we visualized the attention between the facial and bony surfaces.

Reviewer#1 and Reviewer#2 had concern regarding the verification of effectiveness of our approach on real patient data. We clarified that all patient data sets are randomly selected from our digital archive of patients who had undergone double-jaw orthognathic surgery. We agree that ACMT-Net may not be enough for self-verification. To further verify the effectiveness of our approach in the future, we plan to utilize biomechanical models, specifically the Finite Element Method (FEM), to validate the efficacy of our approach in the future work.

Thanks Reviewer#2 and Reviewer#3 for pointing out the necessity to demonstrate how our approach improves upon traditional methods. Our methodology is the first soft tissue driven approach that directly predict the surgical plan. The approach creates and selects a bony movement plan in just under a minute. In contrast, current clinical methods rely on the educated guess of a surgeon for an initial plan, followed by iterative refinements with reference to anticipated facial outcome. Each of those iterations could consume a significant amount of time, potentially several hours when facial simulations are performed using the Finite Element Method (FEM). This time-consuming and labor-intensive process is often unfeasible in daily clinical setting. We plan to introduce more bone-driven strategies, for instance, the sparse representation method, for comparative purposes in our future work.



back to top