News Research Diagnostics & Imaging Thyroid Disease Management Research and Evidence

AI dataset targets nerve injury risk during thyroid surgery 

Annotated surgical images enable AI-based identification of vulnerable anatomy. 

April 14, 2026 By Matthew Solan 3 min read
Share

5 Key Takeaways

  • 1

    A dataset of over 18,000 surgical images demonstrates AI's feasibility in identifying the recurrent laryngeal nerve during thyroid surgery.

  • 2

    The recurrent laryngeal nerve is at risk of injury in 3% to 8% of thyroidectomies, potentially leading to serious complications.

  • 3

    The ThyRLN-PUMCH dataset includes images from 28 patients and reflects various intraoperative conditions encountered in surgery.

  • 4

    DeepLabV3+ and Mask2Former models were tested, achieving 64% and 67% recall, respectively, with Mask2Former showing improved precision.

  • 5

    The study highlights the need for further research to enhance AI's accuracy in detecting nerve structures during thyroid surgeries.

AACE Endocrine AI is published by Conexiant under a license arrangement with the American Association of Clinical Endocrinology, Inc. (AACE®). The ideas and opinions expressed in AACE Endocrine AI do not necessarily reflect those of Conexiant or AACE. For more information, see Policies.

Related Content