基于本地小型语言模型的产科超声报告隐私保护、低标注妊娠状态分类研究
Privacy-preserving, low-annotation pregnancy status classification from obstetric ultrasound reports using on-premises small language models
本研究采用部署于本地的轻量级语言模型,从产科超声报告中以隐私保护方式完成妊娠状态分类,仅需少量标注即取得良好性能。
Acknowledgements
Computational resources were provided by the Faculty of Science, Saint Mary’s University. We thank the annotators for their careful work in labeling and reviewing the dataset.
Funding
This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grants RGPIN-2018-05363 (P.L.) and RGPIN-2026-06675 (V.M.), and by Faculty of Health Dean’s startup funding (V.M.). The funders had no role in study design, data collection, analysis, interpretation, or manuscript preparation.
Author information
Authors and Affiliations
Department of Mathematics and Computing Science, Saint Mary’s University, Halifax, NS, Canada
Nikita Neveditsin & Pawan Lingras
Dhanwantari Hospital, Pune, Maharashtra, India
Salil Patil & Swarup Patil
School of Health Policy & Management, York University, Toronto, ON, Canada
Vijay Mago
Authors
- Nikita Neveditsin
- Salil Patil
- Swarup Patil
- Pawan Lingras
- Vijay Mago
Corresponding authors
Correspondence to Nikita Neveditsin or Vijay Mago.
Ethics declarations
Competing interests
The authors declare no competing interests.
Ethics declarations
This study was performed in accordance with the Declaration of Helsinki and with all relevant institutional guidelines and regulations. Access to the MIMIC-IV corpus was obtained through PhysioNet credentialing and a signed Data Use Agreement. Use of MIMIC-IV in this study complied with institutional requirements for de-identified credentialed research data. Study procedures involving the Indian OCR cohort were reviewed and cleared by the Research Ethics Board at Saint Mary’s University (REB file #25-152), which determined the project to be in accordance with the Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans (TCPS 2) and Saint Mary’s University policies. The study involved secondary analysis of existing, anonymized clinical documents. Consent from individual patients was therefore not sought, consistent with the provisions of TCPS 2 governing secondary use of anonymous information. Institutional authorization for the use of anonymized obstetric reports originating from Dhanwantari Hospital, Nigdi, Pune, India, was granted by the Dhanwantari Medical Foundation on 18 October 2024. The reports were supplied to the researchers in anonymized form, without direct patient identifiers. Individual reports from the Indian OCR cohort and identifying patient information are not published. All reports from the Indian OCR cohort were de-identified prior to analysis. This study was not registered in a public registry. There was no patient or public involvement in the design, conduct, reporting, or dissemination of this study.
Use of generative AI
Generative AI tools were used during the final preparation of the manuscript for language polishing, including improvements to grammar, style, and clarity, and to assist with documenting and packaging the accompanying code for public release. AI assistance with the code was limited to documentation, packaging metadata, and the removal of unused code and environment-specific file paths. All AI-assisted outputs were reviewed and, where necessary, revised by the authors. The authors take full responsibility for the final manuscript and associated code.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Neveditsin, N., Patil, S., Patil, S. et al. Privacy-preserving, low-annotation pregnancy status classification from obstetric ultrasound reports using on-premises small language models. Sci Rep (2026). https://doi.org/10.1038/s41598-026-79573-0
Received: 03 May 2026
Accepted: 02 November 2026
Published: 17 November 2026
DOI: https://doi.org/10.1038/s41598-026-79573-0
Keywords
来源:Nature 健康护理主题 · nature.com