EUVIMEDEUVIMED entdecken
Crossref

Would artificial intelligence make “Good” decisions about thyroid nodule biopsies?

Abhay Barage, Mia Austria, Emily Vertosick iD, Laura Boucai iD, Azeez Farooki, Stephanie Fish, Samantha K Newman, Armando Patrizio, Lubaina Presswala, Tyler Ryan, Mona Sabra iD, Rebecca Zeuren und 6 weitere

The Journal of Clinical Endocrinology & Metabolism · 28.10.2026

DOI: 10.1210/clinem/dgag433 ↗

Kurz gefasst

Fachartikel in der Zeitschrift The Journal of Clinical Endocrinology & Metabolism, online erschienen am 28.10.2026. Verfasst von Abhay Barage und 17 weiteren Autor:innen. Beteiligte Einrichtung: Memorial Sloan Kettering Cancer Center. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 24 Literaturangaben. Gefördert von: Patient Centered Outcomes Research Institute, National Institutes of Health.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Large language models (LLMs) have shown promise for clinical decision support (CDS), but evaluation under clinical equipoise—scenarios lacking a consensus correct answer—remains rare. We evaluated commercially available LLM performance for thyroid nodule biopsy recommendations under clinical equipoise.

Methods

One hundred clinical vignettes describing patients with incidentally discovered thyroid nodules of varying sizes and risk features were presented to 94 clinicians, recruited by convenience sampling from professional networks, and to multiple LLMs (GPT-5, Claude Sonnet 3.5/4.0/4.5, Llama 3.2-90B, DeepSeek R1, and OpenEvidence). Responses were dichotomized as biopsy vs no biopsy. Agreement was assessed using Krippendorff alpha, and a deterministic algorithm served as a provisional benchmark. Selected LLM responses were reviewed by experts.

Results

Clinicians recommended biopsy in 58% of responses (95% CI 50%, 66%), with unanimous responses in only 50 vignettes. Large language model biopsy rates ranged from 46% (Claude 3.5) to 74% (Llama 3.2-90B). Large language model–clinician agreement was moderate (α = .372-.453). Agreement with the deterministic algorithm ranged from high (α = .954, Detailed Thyroid Imaging and Data Reporting System GPT-5) to low (α = .265, Claude Sonnet 4.5). Llama 3.2-90B recommended biopsy in 8 of 8 cases where clinicians unanimously recommended against it.

Conclusion

Large language models differed meaningfully in biopsy recommendation patterns, with potential population-level consequences. Evaluating and aligning LLMs for CDS under equipoise requires stakeholders and processes analogous to clinical guideline committees. Because LLMs used as CDS could powerfully disseminate different practice patterns, this creates new responsibilities for guideline committees and other stakeholders.

Schlagwörter

Stichwörter der Autor:innen

artificial intelligenceclinical decision supportdiagnosisequipoisesurveythyroid nodule

Autor:innen und Institutionen

  • Memorial Sloan Kettering Cancer Center.
  • Azienda Ospedaliero Universitaria Pisana.
  • Weill Cornell Medicine.
  • Columbia University.
  • Vanderbilt University.
  • National Institutes of Health.

ORCID-Profile: Emily Vertosick, Laura Boucai, Mona Sabra, James H Flory

Förderung

  • Patient Centered Outcomes Research Institute
  • National Institutes of Health

Zitieren

APA

Barage, A., Austria, M., Vertosick, E., Boucai, L., Farooki, A., Fish, S., Newman, S. K., Patrizio, A., Presswala, L., Ryan, T., Sabra, M., Zeuren, R., Vickers, A., Kuperman, G., Ancker, J. S., Kim, S. Y. H., Tuttle, R. M., & Flory, J. H. (2026). Would artificial intelligence make “Good” decisions about thyroid nodule biopsies?. The Journal of Clinical Endocrinology & Metabolism. https://doi.org/10.1210/clinem/dgag433

Vancouver

Barage A, Austria M, Vertosick E, Boucai L, Farooki A, Fish S, et al. Would artificial intelligence make “Good” decisions about thyroid nodule biopsies?. The Journal of Clinical Endocrinology & Metabolism. 2026. doi:10.1210/clinem/dgag433

BibTeX
@article{Barage2026,
  author = {Barage, Abhay and Austria, Mia and Vertosick, Emily and Boucai, Laura and Farooki, Azeez and Fish, Stephanie and Newman, Samantha K and Patrizio, Armando and Presswala, Lubaina and Ryan, Tyler and Sabra, Mona and Zeuren, Rebecca and Vickers, Andrew and Kuperman, Gilad and Ancker, Jessica S and Kim, Scott Y H and Tuttle, R Michael and Flory, James H},
  title = {Would artificial intelligence make “Good” decisions about thyroid nodule biopsies?},
  journal = {The Journal of Clinical Endocrinology & Metabolism},
  year = {2026},
  doi = {10.1210/clinem/dgag433}
}
Kostenlos starten →