Paper Info Reviews Author Feedback Meta-Review Back to top
List of Papers Browse by Subject Areas Author List

Abstract

Accurate identification of euploid embryos is essential for improving outcomes in in vitro fertilization. While time-lapse imaging and deep learning enable non-invasive embryo assessment, most existing approaches model developmental sequences within a unified representation space, overlooking substantial inter-individual variability in morphology, developmental rhythm, and global developmental deviation. To address this limitation, we propose a multi-level individualized modeling framework that explicitly characterizes embryo variability from spatial, temporal, and global perspectives. First, at the morphological level, we characterize embryo-specific spatial heterogeneity by adaptively composing shared morphological components, enabling the model to preserve canonical structures while capturing individualized structural variations. Second, at the temporal level, we model individualized developmental dynamics by adaptively reweighting time points, allowing the framework to emphasize embryo-specific critical stages rather than enforcing uniform temporal importance. Third, at the global level, we incorporate reliability-aware modeling to explicitly characterize uncertainty and developmental deviation, balancing population-level regularities with individual-specificity. Extensive experiments demonstrate that the proposed method achieves state-of-the-art performance, highlighting the importance of explicitly modeling embryo individuality for euploidy prediction. The GitHub link is https://github.com/Renard131/MIEP.

Links to Paper and Supplementary Materials

Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026/paper/0713_paper.pdf

SharedIt Link: Not yet available

SpringerLink (DOI): Not yet available

Supplementary Material: Not Submitted

Link to the Code Repository

https://github.com/Renard131/MIEP

Link to the Dataset(s)

N/A

BibTex

@InProceedings{TanShu_Born_MICCAI2026,
        author = { Tan, Shuangyi AND Song, Xinxian AND Mai, Qingyun AND Li, Zhen AND Yang, Sibei AND Li, Guanbin},
        title = { { Born Different: A Multi-level Individualized Modeling Framework for Embryo Euploidy Prediction } },
        booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026},
        year = {2026},
        publisher = {Springer Nature Switzerland},
        volume = {LNCS 16896},
        month = {September},
        page = {pending}
}

Reviews

Review #1

  • Please describe the contribution of the paper

    The paper proposes a novel deep learning system for a problem that has been explored by a number of papers: the prediction of embryo ploidy. The system takes into account three streams of information: morphological, kinetic (temporal) and clinical parameters of the patient. The paper introduces custom modules for each of these streams and makes use of a novel loss function.

  • 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 makes a number of contributions over the literature including its GDRE and diversity loss.
    • The paper demonstrates the utility of each of these components through an ablation study.
    • The paper is, for the most part, well-written and easy to follow.
  • 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.
    • (Introduction) The authors should justify why different embryos differ in “stage importance”.
    • (Introduction) “Systematically review” implies a systematic review (which isn’t present). Consider rewording for clarity.
    • (Introduction) The paper would benefit from a more explicit definition of “global developmental reliability” in the introduction.
    • (Data) It might be worth stating the name of the hospital / department.
    • (Data) It would be worth stating which PGT-P platform was used, as different techniques can vary in reliability.
    • (Experiment) The paper would benefit from an explanation of the difference between the 868 and 869 Embryoscopes.
    • (Experiment) Tables 1 and 2 would benefit from some indicator of variability in performance (e. g. standard deviation)
    • (Experiment) I think the claims that the system “generalizes well across different time-lapse incubators” is a bit overstated. It would be better-supported if, say, you demonstrated that there is no performance degradation in a model trained on the 868 when tested on a test set from the 869 (and vice versa).
  • 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 paper makes a decent contribution over the existing literature. However, the performance improvement results could be more credible with some indicator of variability. Moreover, given the current evidence, I think claims of generalisability are overstated. That said, the paper is solid and I’d be willing to go for a full accept if these issues are addressed.

  • 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
    • The novel framework proposed in this study explicitly models individual variation across three key dimensions—morphology, temporal progression, and developmental stage—to predict embryonic euploidy.
    • A comprehensive dataset comprising 1,034 embryos was constructed, integrating time-lapse imaging sequences with structured clinical information (7 slices per time point), enabling evaluation grounded in clinical context.
    • Extensive experimental validation demonstrated superior predictive accuracy compared to existing methods, quantitatively demonstrating the importance of accounting for individual variation.
  • 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.
    • Implements highly innovative modeling approaches (AMRM, ITIM, GDRE)
    • AMRM: Extracts embryo-specific structural variation features by adaptively combining shared morphological patterns conditioned on clinical information
    • ITIM: Features an adaptive weighting mechanism that emphasizes critical developmental stages specific to each embryo
    • GDRE: Utilizes heteroscedastic loss to automatically account for the “uncertainty” of embryos that deviate significantly from standard developmental trajectories, thereby enhancing training stability
    • Rather than simply combining features, it employs TabGLM to extract relationships between clinical factors and incorporates these insights into the image feature encoding
    • The team has independently collected clinical data to construct the dataset. Notably, they have gone as far as collecting data specific to different incubator models.
  • 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.
    • In the ablation study, removing clinical information resulted in a significant AUC drop from 0.826 to 0.728, indicating that performance may be severely limited in environments with missing clinical data.
    • While it’s commendable that the paper examines bias differences across incubators in detail, the specific breakdown of each dataset and any actual biases observed remain unclear. Notably, the F1 score shows substantial variation between incubator types, raising questions about the underlying reasons for this discrepancy.
    • This study divides embryo development into 16 temporal segments for training and evaluation, but different developmental stages likely require varying resolution for accurate euploidy prediction. The authors suggest addressing this through foundation model utilization and temporal attention strategies, but the paper lacks a detailed analysis of which specific time points contribute most significantly to prediction accuracy.
  • 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

    I consider this to be a truly remarkable study. While I fully understand that the dataset contains highly sensitive and difficult-to-handle data, given its exceptional value for IVF research, I hope you’ll consider making it publicly available. This would significantly advance the field and I’d appreciate your positive consideration.

  • 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?

    This study effectively implements a clear conceptual framework for modeling inter-individual variability across three key dimensions - spatial distribution, temporal dynamics, and uncertainty - to address a critical clinical challenge in IVF embryo evaluation. The research demonstrates particularly high technical innovation. Notably, its design effectively utilizes clinical metadata not merely as supplementary information but as the basis for determining module weights. Furthermore, the construction of the dataset represents highly significant work. The comprehensive evaluation approach, including ablation studies, makes the overall structure exceptionally well-balanced. While some aspects require refinement in the dataset details and result interpretation, these can be addressed through rebuttals. Therefore, I assign a “weak accept” rating.

  • 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 #3

  • Please describe the contribution of the paper

    The method fuses multi-focus embryo time-lapse images, encodes them with a pretrained model, and predicts euploidy by modeling embryo individuality at three levels: frame morphology through the Adaptive Morphological Representation Module (AMRM), temporal stage importance through Individualized Temporal Importance Modeling (ITIM), and whole-trajectory reliability through the Global Developmental Reliability Estimator (GDRE). Clinical information guides all three parts, and the final loss combines classification, clinical-image alignment, and temporal diversity terms.

  • 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.

    Novel in using three-level individualized modeling framework: AMRM, ITIM, GDRE. Combination is reasonably distinctive, not just images with clinical data.

  • 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.

    No explanation on how morphology experts are implemented. From Fig 1, are they layers of FFNs together or a collection of individual FFNs? What exactly are “canonical morphological patterns shared across embryos”? cell shapes, fragmentation, structure? it says the router is a function f_router([pt; ec]) followed by softmax, but it does not specify whether this is a linear layer, MLP or anything more complex.

  • 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 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?

    A well-designed synthesis of known ingredients arranged into a three-level personalized framework.

  • Reviewer confidence

    Somewhat confident (2)

  • [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 consistently recognize the paper contribution with only minor concerns



back to top