Pith. sign in

REVIEW 1 major objections 19 references

This paper releases the first public dataset of 1,111 transvaginal ultrasound images and 16 videos with pixel-level annotations for cesarean scar defect segmentation.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.3

2026-06-29 18:26 UTC pith:QO2F4LWQ

load-bearing objection First public dataset for CSD segmentation in transvaginal ultrasound, but no numbers on annotation consistency. the 1 major comments →

arxiv 2605.26774 v2 pith:QO2F4LWQ submitted 2026-05-26 cs.CV

Cesarean Scar Defect Segmentation in Transvaginal Ultrasound Images: a Dataset and Benchmark

Yuan Tian , Yue Li , Wei Xia , Tianyu Xu , Jian Zhang , Liye Shi , Jing Liu , Yang Wang
show 5 more authors
Ming Liu Qing Xu Yixuan Zhang Maggie M. He Xiangjian He
This is my paper
classification cs.CV
keywords cesarean scar defecttransvaginal ultrasoundimage segmentationmedical imaging datasetbenchmarkcesarean delivery complicationsultrasound annotation
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The authors establish that no public dataset has existed for training or benchmarking algorithms to segment cesarean scar defects in transvaginal ultrasound. They supply 501 confirmed positive cases with precise manual outlines created under clinical guidelines. A sympathetic reader would care because CSD is common after cesarean delivery yet often missed due to small size and image challenges, and better automated tools could improve diagnosis and treatment decisions.

Core claim

No public dataset exists for transvaginal ultrasound CSD segmentation. The authors address the gap by releasing a dataset of 1,111 images and 16 videos that contains 501 positive samples with confirmed CSD and pixel-level manual annotations performed by experienced sonographers and trained PhD students following standardized clinical guidelines.

What carries the argument

The CSD dataset itself, consisting of transvaginal ultrasound images and videos with pixel-level annotations for defect boundaries.

Load-bearing premise

The manual annotations accurately capture true CSD boundaries according to standardized clinical guidelines.

What would settle it

An independent set of expert sonographers re-annotating a random subset of the images and producing boundary outlines that differ substantially from the released labels.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Segmentation algorithms trained on the dataset can produce consistent CSD outlines and dimensions for clinical review.
  • The benchmark enables direct comparison of new medical image segmentation methods on this task.
  • Wider availability of the data can increase clinical awareness of CSD in settings with limited specialist expertise.
  • Improved detection supports more timely treatment decisions that affect reproductive-age women.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The dataset could serve as a template for creating similar annotated collections for other ultrasound-detected gynecological conditions.
  • Models trained here might be tested for transfer to related scar or defect segmentation tasks in different imaging modalities.
  • Routine use of such benchmarks could accelerate adoption of automated measurement tools in routine transvaginal screening protocols.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

1 major / 0 minor

Summary. The paper claims to introduce the first public dataset for Cesarean Scar Defect (CSD) segmentation in transvaginal ultrasound, consisting of 1,111 images and 16 videos that yield 501 positive samples with confirmed CSD and pixel-level manual annotations performed by experienced sonographers collaborating with trained PhD students following standardized clinical guidelines. It positions the release as a benchmark resource to advance medical image segmentation algorithms and clinical practice for CSD diagnosis.

Significance. If the ground-truth annotations are shown to be reliable, the dataset would address a documented absence of public resources for this clinically relevant task, enabling reproducible development and comparison of segmentation methods for a condition whose small size and irregular morphology make it prone to oversight. The core empirical contribution of data collection and expert annotation is directly supported by the description.

major comments (1)
  1. [Abstract] Abstract: the assertion of 'precise pixel-level manual annotations' and 'high-quality benchmark resources' is not supported by any reported quantitative validation of annotation quality, such as inter-annotator agreement (Dice/IoU), disagreement-resolution protocol, or per-image quality metrics. Given the abstract's own description of CSDs as small and irregular, this omission leaves the reliability of the ground truth untested and directly affects the central claim that the dataset supplies usable benchmark data.

Simulated Author's Rebuttal

1 responses · 1 unresolved

We thank the referee for their constructive comments on our manuscript. We address the single major comment below and agree that revisions are warranted to better support our claims regarding the dataset.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the assertion of 'precise pixel-level manual annotations' and 'high-quality benchmark resources' is not supported by any reported quantitative validation of annotation quality, such as inter-annotator agreement (Dice/IoU), disagreement-resolution protocol, or per-image quality metrics. Given the abstract's own description of CSDs as small and irregular, this omission leaves the reliability of the ground truth untested and directly affects the central claim that the dataset supplies usable benchmark data.

    Authors: We agree with this observation. The current abstract uses terms such as 'precise' and 'high-quality' without accompanying quantitative evidence of annotation reliability. In the manuscript body we describe a collaborative annotation workflow between experienced sonographers and trained PhD students that follows standardized clinical guidelines, but we do not report inter-annotator agreement statistics or a formal disagreement-resolution protocol. Because annotations were produced collaboratively rather than by independent raters, Dice or IoU agreement metrics between multiple annotators are not available from the existing data. We will revise the abstract to remove or qualify these strong claims, add a dedicated subsection in the Methods that details the annotation protocol and consensus process, and explicitly note the absence of quantitative inter-rater metrics as a limitation of the released resource. revision: yes

standing simulated objections not resolved
  • Quantitative inter-annotator agreement (Dice/IoU) cannot be reported because the annotations were generated collaboratively rather than by independent annotators.

Circularity Check

0 steps flagged

No circularity: empirical dataset release with no derivations

full rationale

The paper presents a new CSD ultrasound dataset (1,111 images, 16 videos, 501 positive samples) with pixel-level annotations performed per clinical guidelines. No equations, fitted parameters, predictions, or derivation chains exist in the manuscript. The contribution is purely empirical data collection and annotation; no self-citation load-bearing steps, ansatzes, or renamings of results are present. The central claim reduces only to the act of releasing the data itself, which is self-contained and externally verifiable by inspection of the released files.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Dataset paper; no mathematical model, free parameters, or postulated entities are introduced. The contribution rests on empirical collection and manual labeling rather than derivation.

pith-pipeline@v0.9.1-grok · 5742 in / 1037 out tokens · 28620 ms · 2026-06-29T18:26:42.388540+00:00 · methodology

0 comments
read the original abstract

Cesarean Scar Defect (CSD) is one of the most prevalent complications following cesarean delivery. Transvaginal ultrasonography is widely used for primary CSD screening. Accurate determination of CSD outline and dimensions is crucial for treatment. However, CSDs are frequently overlooked by sonographers due to small size and irregular morphology, suboptimal image quality, and limited clinical awareness in resource-constrained settings. Despite artificial intelligence advances in medical imaging, no public dataset exists for transvaginal ultrasound CSD segmentation. To address this gap, we present a comprehensive CSD dataset comprising 1,111 images and 16 videos, yielding 501 positive samples with confirmed CSD and precise pixel-level manual annotations. Annotations are performed following standardized clinical guidelines through collaboration between experienced sonographers and trained PhD students. This work provides high-quality benchmark resources for advancing medical image segmentation algorithms and promoting clinical innovation. Ultimately, improved CSD diagnosis and subsequent treatment strategies can enhance the quality of life in women of reproductive age, representing significant value for both medical research and clinical practice.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

19 extracted references · 1 canonical work pages

  1. [1]

    Acta Obstetricia et Gynecologica Scandinavica 98(4), 413–422 (2019)

    Pan, H., Zeng, M., Xu, T., Li, D., Mol, B.W., Sun, J., Zhang, J.: The prevalence and risk predictors of cesarean scar defect at 6 weeks postpartum in shanghai, china: A prospective cohort study. Acta Obstetricia et Gynecologica Scandinavica 98(4), 413–422 (2019)

    2019

  2. [2]

    JAMA network open6(3), 235321–235321 (2023)

    Meuleman, S.J.K., Murji, A., Bosch, T., Donnez, O., Grimbizis, G., Saridogan, E., Chantraine, F., Bourne, T., Timmerman, D., Huirne, J.A.,et al.: Definition and criteria for diagnosing cesarean scar disorder. JAMA network open6(3), 235321–235321 (2023)

    2023

  3. [3]

    American Journal of Obstetrics and Gynecology228(6), 712–1 (2023)

    Zhang, J., Zhu, C., Yan, L., Wang, Y., Zhu, Q., He, C., He, X., Ji, S., Tian, Y., Xie, L.,et al.: Comparing levonorgestrel intrauterine system with hysteroscopic niche resection in women with postmenstrual spotting related to a niche in the uterine 9 cesarean scar: a randomized, open-label, controlled trial. American Journal of Obstetrics and Gynecology22...

    2023

  4. [4]

    Ultrasound in Obstetrics & Gynecology51(2), 169–175 (2018)

    Cal` ı, G., Timor-Tritsch, I., Palacios-Jaraquemada, J., Monteaugudo, A., Buca, D., Forlani, F., Familiari, A., Scambia, G., Acharya, G., D’Antonio, F.: Outcome of cesarean scar pregnancy managed expectantly: systematic review and meta- analysis. Ultrasound in Obstetrics & Gynecology51(2), 169–175 (2018)

    2018

  5. [5]

    European Journal of Obstetrics & Gynecology and Reproductive Biology284, 136–142 (2023)

    He, C., Xia, W., Yan, L., Wang, Y., Tian, Y., Mol, B.W., Zhang, J., Huirne, J.: Fertility outcomes after hysteroscopic niche resection compared with expec- tant management in women with a niche in the uterine cesarean scar. European Journal of Obstetrics & Gynecology and Reproductive Biology284, 136–142 (2023)

    2023

  6. [6]

    Ginekologia Polska92(10), 726–730 (2021)

    Budny-Winska, J., Pomorski, M.: Uterine niche after cesarean section: a review of diagnostic methods. Ginekologia Polska92(10), 726–730 (2021)

    2021

  7. [7]

    Ultrasound in Obstetrics & Gynecology43(4), 372–382 (2014)

    Vaate, A., Voet, L., Naji, O., Witmer, M., Veersema, S., Br¨ olmann, H., Bourne, T., Huirne, J.: Prevalence, potential risk factors for development and symptoms related to the presence of uterine niches following cesarean section: systematic review. Ultrasound in Obstetrics & Gynecology43(4), 372–382 (2014)

    2014

  8. [8]

    Journal of minimally invasive gynecology20(3), 386–391 (2013)

    Marotta, M.-L., Donnez, J., Squifflet, J., Jadoul, P., Darii, N., Donnez, O.: Laparoscopic repair of post-cesarean section uterine scar defects diagnosed in nonpregnant women. Journal of minimally invasive gynecology20(3), 386–391 (2013)

    2013

  9. [9]

    Ultrasound in obstetrics & gynecology37(1), 93–99 (2011)

    Vaate, A., Br¨ olmann, H., Van Der Voet, L., Van Der Slikke, J., Veersema, S., Huirne, J.: Ultrasound evaluation of the cesarean scar: relation between a niche and postmenstrual spotting. Ultrasound in obstetrics & gynecology37(1), 93–99 (2011)

    2011

  10. [10]

    Ultrasound in Obstetrics & Gynecology47(4), 499–505 (2016)

    Baranov, A., Gunnarsson, G., Salvesen, K., Isberg, P.-E., Vikhareva, O.: Assessment of cesarean hysterotomy scar in non-pregnant women: reliability of transvaginal sonography with and without contrast enhancement. Ultrasound in Obstetrics & Gynecology47(4), 499–505 (2016)

    2016

  11. [11]

    Uniultra: Interactive parameter-efficient sam2 for universal ultrasound segmentation,

    Li, Y., Xu, Q., Zhang, Y., He, X., Zhang, Q., Yao, Y., Tesem, F.B., Chen, X., Wang, R., Chen, Z., et al.: Uniultra: Interactive parameter-efficient sam2 for universal ultrasound segmentation. arXiv preprint arXiv:2511.15771 (2025)

    work page arXiv 2025

  12. [12]

    Nature communications12(1), 5645 (2021)

    Shen, Y., Shamout, F.E., Oliver, J.R., Witowski, J., Kannan, K., Park, J., Wu, N., Huddleston, C., Wolfson, S., Millet, A.,et al.: Artificial intelligence system reduces false-positive findings in the interpretation of breast ultrasound exams. Nature communications12(1), 5645 (2021)

    2021

  13. [13]

    Medical image analysis 61, 101665 (2020)

    Wang, L., Zhang, L., Zhu, M., Qi, X., Yi, Z.: Automatic diagnosis for thyroid 10 nodules in ultrasound images by deep neural networks. Medical image analysis 61, 101665 (2020)

    2020

  14. [14]

    The Lancet Digital Health4(3), 179–187 (2022)

    Gao, Y., Zeng, S., Xu, X., Li, H., Yao, S., Song, K., Li, X., Chen, L., Tang, J., Xing, H.,et al.: Deep learning-enabled pelvic ultrasound images for accurate diagnosis of ovarian cancer in china: a retrospective, multicentre, diagnostic study. The Lancet Digital Health4(3), 179–187 (2022)

    2022

  15. [15]

    Journal of Digital Imaging35(4), 983–992 (2022)

    Jin, J., Zhu, H., Teng, Y., Ai, Y., Xie, C., Jin, X.: The accuracy and radiomics fea- ture effects of multiple u-net-based automatic segmentation models for transvagi- nal ultrasound images of cervical cancer. Journal of Digital Imaging35(4), 983–992 (2022)

    2022

  16. [16]

    In: Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18, pp

    Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomed- ical image segmentation. In: Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18, pp. 234–241 (2015). Springer

    2015

  17. [17]

    In: Proceed- ings of the European Conference on Computer Vision (ECCV), pp

    Chen, L.-C., Zhu, Y., Papandreou, G., Schroff, F., Adam, H.: Encoder-decoder with atrous separable convolution for semantic image segmentation. In: Proceed- ings of the European Conference on Computer Vision (ECCV), pp. 801–818 (2018)

    2018

  18. [18]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops, pp

    Cao, Y., Xu, J., Lin, S., Wei, F., Hu, H.: Gcnet: Non-local networks meet squeeze- excitation networks and beyond. In: Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops, pp. 0–0 (2019)

    2019

  19. [19]

    In: European Conference on Computer Vision, pp

    Cao, H., Wang, Y., Chen, J., Jiang, D., Zhang, X., Tian, Q., Wang, M.: Swin- unet: Unet-like pure transformer for medical image segmentation. In: European Conference on Computer Vision, pp. 205–218 (2022). Springer 11

    2022