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

Association between glymphatic dysfunction and glucose hypometabolism in chronic disorders of consciousness: a multimodal PET/MRI study.

Brain communications·2026
Same author

Abnormal Activation and Connection in Middle Frontal Gyrus: A Potential Imaging Feature for Facial Synkinesis Comorbid Depression.

Depression and anxiety·2026
Same author

Stratifying amyloid burden in early Alzheimer's disease using cascaded attention-guided vision transformer using [¹⁸F]Florbetapir PET.

European radiology·2026
Same author

A robust computational framework for methylation age and disease-risk prediction based on pairwise learning.

Nature computational science·2026
Same author

Enhancing interpretability of AI with radiomics-based deep neural network: proof of concept in the classification of Parkinsonian syndromes with <sup>18</sup>F-FDG PET imaging.

European journal of nuclear medicine and molecular imaging·2025
Same author

Evaluating <sup>18</sup>F-Florzolotau tau PET for Alzheimer's disease diagnosis with <sup>18</sup>F-Flortaucipir as reference.

Journal of neurology·2025
Same journal

Clinician Perspectives on Ambient AI Scribes in the Intensive Care Unit: Qualitative Interview Study.

JMIR medical informatics·2026
Same journal

IdeaDistiller-AI Support for Idea Synthesis in Concept Mapping: Algorithm Development and Validation Study.

JMIR medical informatics·2026
Same journal

Pregnancy-Related Clinical Codes in Unlikely Populations in Primary Care.

JMIR medical informatics·2026
Same journal

Selecting, Scaling, and Measuring the Value of Ambient AI in a Nonacademic Health System: Multiphase Pilot Study.

JMIR medical informatics·2026
Same journal

Prediction of Early Hospital Admission (≤24 Hours) After Stroke Using Machine Learning and Deep Learning: Multicenter Study From China.

JMIR medical informatics·2026
Same journal

Assessing the Feasibility and Acceptability of Implementing a Preclinic Vital Signs Assessment in Primary Care: Cross-Sectional Pilot Study.

JMIR medical informatics·2026
See all related articles

Related Experiment Video

Updated: Jan 14, 2026

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
07:13

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model

Published on: April 18, 2025

485

Automated Esophageal Cancer Staging From Free-Text Radiology Reports: Large Language Model Evaluation Study.

Yao Yao1, Xingxing Cen1, Lu Gan2

  • 1Information Center, Shanghai Chest Hospital, Shanghai Jiao Tong University School of Medicine, No 241, West Huaihai Road, Shanghai, 200030, China, 86 22200000.

JMIR Medical Informatics
|October 17, 2025
PubMed
Summary
This summary is machine-generated.

Large language models (LLMs) show promise in esophageal cancer staging. The INF-72B LLM with interpretable reasoning achieved superior accuracy compared to clinicians in staging cancer from radiology reports.

Keywords:
TNMcancer stagingesophageal cancerlarge language modelradiology reporttumor–node metastasis

More Related Videos

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

1.0K
Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
04:09

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma

Published on: October 10, 2018

8.6K

Related Experiment Videos

Last Updated: Jan 14, 2026

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
07:13

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model

Published on: April 18, 2025

485
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

1.0K
Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
04:09

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma

Published on: October 10, 2018

8.6K

Area of Science:

  • Oncology
  • Medical Imaging
  • Artificial Intelligence

Background:

  • Accurate esophageal cancer staging is vital for patient prognosis and treatment planning.
  • Manual interpretation of radiology reports by clinicians presents challenges in accuracy and consistency.
  • The application of large language models (LLMs) in esophageal cancer staging is an emerging area of research.

Purpose of the Study:

  • To evaluate and compare the performance of three locally deployed LLMs against clinicians in preoperative esophageal cancer staging.
  • To assess the utility of different prompting strategies, including an interpretable reasoning (IR) method, for LLM-based cancer staging.

Main Methods:

  • A retrospective study of 200 patients with esophageal cancer from Shanghai Chest Hospital.
  • Analysis of 1134 Chinese free-text radiology reports, with postoperative pathological staging serving as the reference standard.
  • Three LLMs (INF-72B, Qwen2.5-72B, LLaMA3.1-70B) were employed with zero-shot, chain-of-thought, and IR prompting strategies to determine tumor, node, and overall staging.

Main Results:

  • INF-72B combined with the IR strategy achieved significantly higher overall staging accuracy (61.5%) and F1-score (0.60) compared to clinicians (39.5% accuracy, 0.39 F1-score).
  • Qwen2.5-72B with IR also outperformed clinicians, showing 46% accuracy and a 0.51 F1-score.
  • LLaMA3.1-70B did not demonstrate a statistically significant difference in staging performance compared to clinicians.

Conclusions:

  • LLMs, particularly INF-72B with the proposed IR strategy, can accurately stage esophageal cancer from radiology reports.
  • This AI-driven approach offers a transparent and verifiable reasoning process, enhancing diagnostic reliability.
  • LLMs show potential as valuable decision-support tools to augment clinical expertise in esophageal cancer staging.