{
  "abstract": "Background: Thyroid fine-needle aspiration cytology is widely used to evaluate thyroid nodules before surgery. Recent advances in deep learning have increased the potential for artificial intelligence-assisted cytological diagnosis. However, developing accurate models usually requires large numbers of images labelled by experts, making dataset preparation time-consuming and costly. We investigate whether self-supervised learning can improve thyroid cytology image classification while reducing the need for labelled data.\n\nMethods: We retrospectively analysed cytology images from 393 thyroid nodules and extracted 142,158 image patches. Three self-supervised learning methods were used to pre-train vision transformer models on unlabelled images, followed by classification training with different amounts of labelled data. Performance in classifying eight lesion types was assessed using five-fold cross-validation, precision–recall area under the curve, and macro-averaged F1 scores. Model performance was compared with ImageNet-pretrained models and four cytotechnologists.\n\nResults: Here we show that self-supervised learning maintains high classification performance when labelled training data are limited and consistently outperforms ImageNet pre-training. With 1% of the labelled data, the best self-supervised model achieves a precision–recall area under the curve of 0.899, compared with 0.707 for ImageNet pre-training. The diagnostic pipeline achieves an average class-specific accuracy of 0.94, exceeding that of the four participating cytotechnologists. Self-supervised image representations also form lesion-specific clusters without diagnostic labels.\n\nConclusions: Self-supervised learning enables accurate classification of thyroid cytology images while substantially reducing the need for expert annotation. This approach may facilitate the development of more efficient artificial intelligence-assisted cytology systems, although validation using data from multiple institutions is needed.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": null,
  "author_total": 11,
  "authors": [
    {
      "name": "Masatoshi Abe",
      "orcid": "0000-0002-4597-6109"
    },
    {
      "name": "Ayana Suzuki"
    },
    {
      "name": "Kumiko Kamada"
    },
    {
      "name": "Toshiya Kanno"
    },
    {
      "name": "Mayu Shimasaki"
    },
    {
      "name": "Toshinori Mori"
    },
    {
      "name": "Hajime Nagahara"
    },
    {
      "name": "Akira Miyauchi"
    },
    {
      "name": "Takashi Akamizu",
      "orcid": "0000-0003-4960-8271"
    },
    {
      "name": "Mitsuyoshi Hirokawa"
    },
    {
      "name": "Hirohiko Niioka",
      "orcid": "0000-0003-4542-2883"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/7vdjkdrmnmmwe.bib",
  "cited_by": 0,
  "doi": "10.1038/s43856-026-01874-2",
  "fulltext_url": null,
  "id": "7vdjkdrmnmmwe",
  "issns": [
    "2730-664X"
  ],
  "landing_url": "https://www.nature.com/articles/s43856-026-01874-2",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0",
  "metadata_licence": null,
  "native_subjects": [
    [
      "openalex",
      "1702|AI in cancer detection"
    ]
  ],
  "openalex": "W7220712899",
  "pdf_url": "https://www.nature.com/articles/s43856-026-01874-2_reference.pdf",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-06",
  "publisher": "Springer Science and Business Media LLC",
  "ris": "https://research.draconitic.com/paper/7vdjkdrmnmmwe.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:17:53Z",
      "native_id": "10.1038/s43856-026-01874-2",
      "source": "crossref",
      "url": "https://www.nature.com/articles/s43856-026-01874-2"
    }
  ],
  "subjects": [
    {
      "field": "medicine",
      "origin": "issn:2730-664X",
      "subfield": "general"
    }
  ],
  "title": "Automated acquisition of thyroid cytology image features using self-supervised learning and application to image classification",
  "title_text": "Automated acquisition of thyroid cytology image features using self-supervised learning and application to image classification",
  "type": "peer-reviewed",
  "updated_at": "2026-10-30T12:17:53Z",
  "url": "https://research.draconitic.com/paper/7vdjkdrmnmmwe",
  "venue": "Communications Medicine",
  "versions": []
}