Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Impact of doublet post-transplant maintenance on outcomes in multiple myeloma: A propensity score matching analysis.

Bone marrow transplantation·2026
Same author

Transcranial Doppler As Ancillary Testing for Pediatric Brain Death/Death by Neurologic Criteria: Retrospective Study of the Israel Transplant Organ Donor Registry, 2016-2024.

Pediatric critical care medicine : a journal of the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies·2026
Same author

Phase 2, open-label INSPIRE trial to assess the tolerability and effectiveness of transdermal cannabidiol gel in children and adolescents with 22q11.2 deletion syndrome (ZYN2-CL-031).

Journal of neurodevelopmental disorders·2026
Same author

Outcomes in Children With Enterovirus Meningitis in London, England: A Retrospective Multicenter Cohort Study, 2013-2023.

Open forum infectious diseases·2026
Same author

Management of ERCP- and EUS-related duodenal perforations using over-the-scope clips.

iGIE : innovation, investigation and insights·2026
Same author

Efficacy, feasibility, and safety of endoscopic double closure in the GI tract.

iGIE : innovation, investigation and insights·2026

Related Experiment Video

Updated: Dec 6, 2025

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

2.2K

Weakly supervised instance learning for thyroid malignancy prediction from whole slide cytopathology images.

David Dov1, Shahar Z Kovalsky2, Serge Assaad1

  • 1Department of Electrical and Computer Engineering, Duke University, Durham, NC 27708, USA.

Medical Image Analysis
|October 13, 2020
PubMed
Summary

This study introduces a novel machine learning algorithm for thyroid cancer prediction from whole-slide images. The method improves accuracy by analyzing image patches and incorporating multiple diagnostic labels for better malignancy detection.

Keywords:
AIDeep learningHealthcareHuman levelMedical image analysisMultiple instance learningPathologyThyroid

More Related Videos

Author Spotlight: Integrating Ultrasound Imaging with Biochemical Markers for Thyroid Disease Diagnosis
05:41

Author Spotlight: Integrating Ultrasound Imaging with Biochemical Markers for Thyroid Disease Diagnosis

Published on: February 9, 2024

924
Computer-Aided Three-Dimensional Visualization in the Treatment of Locally Advanced Thyroid Cancer
03:55

Computer-Aided Three-Dimensional Visualization in the Treatment of Locally Advanced Thyroid Cancer

Published on: June 9, 2023

828

Related Experiment Videos

Last Updated: Dec 6, 2025

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

2.2K
Author Spotlight: Integrating Ultrasound Imaging with Biochemical Markers for Thyroid Disease Diagnosis
05:41

Author Spotlight: Integrating Ultrasound Imaging with Biochemical Markers for Thyroid Disease Diagnosis

Published on: February 9, 2024

924
Computer-Aided Three-Dimensional Visualization in the Treatment of Locally Advanced Thyroid Cancer
03:55

Computer-Aided Three-Dimensional Visualization in the Treatment of Locally Advanced Thyroid Cancer

Published on: June 9, 2023

828

Area of Science:

  • Computational pathology
  • Machine learning in diagnostics
  • Medical image analysis

Background:

  • Machine learning for thyroid malignancy prediction from whole-slide images (WSI) faces challenges with cytopathology slide structures.
  • Traditional multiple instance learning (MIL) methods struggle with sparsely located, heterogeneous informative instances in cytopathology.
  • Existing approaches do not fully leverage diverse label types (bag-level, instance-level) for improved diagnostic accuracy.

Purpose of the Study:

  • To develop an advanced machine learning framework for accurate thyroid malignancy prediction from cytopathology WSIs.
  • To address limitations of standard MIL by incorporating multiple label types and unique slide structures.
  • To enhance diagnostic performance and potentially augment human expert decisions in thyroid cancer pathology.

Main Methods:

  • Proposed a maximum likelihood estimation (MLE) framework to integrate bag-level malignancy, diagnostic scores, and instance-level informativeness/abnormality labels.
  • Developed a two-stage deep-learning algorithm identifying informative instances and assigning local malignancy scores for global prediction.
  • Derived a lower bound of the MLE for a weakly supervised training strategy and extended the algorithm for simultaneous multi-label prediction.

Main Results:

  • The proposed algorithm achieved competitive performance against existing methods in thyroid malignancy prediction.
  • The approach demonstrated expert human-level performance, indicating high diagnostic accuracy.
  • The algorithm showed potential for augmenting human decisions, aiding pathologists in diagnosis.

Conclusions:

  • The novel MLE-based deep learning algorithm effectively addresses challenges in cytopathology WSI analysis for thyroid malignancy prediction.
  • Integrating multiple label types and a refined training strategy significantly improves diagnostic performance.
  • The developed method offers a powerful tool for enhancing diagnostic accuracy and supporting clinical decision-making in thyroid pathology.