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 Experiment Video

Updated: Jan 15, 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

489

Deep Learning Models to Screen Electronic Health Records for Breast and Colorectal Cancer Progression: Performance

Pascal Lambert1,2, Rayyan Khan1,3, Marshall Pitz1,4,5,6

  • 1Paul Albrechtsen Research Institute CancerCare Manitoba, Winnipeg, MB, Canada.

JMIR AI
|October 13, 2025
PubMed
Summary

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

Diagnostic yield of the one-year surveillance colonoscopy after curative colorectal cancer resection.

International journal of colorectal disease·2026
Same author

Should we screen for colorectal cancer with biennial FIT beginning at age 45 in Canada?

Journal of the Canadian Association of Gastroenterology·2026
Same author

The impact of the COVID-19 pandemic on breast cancer screening, treatment, and overall survival.

Breast cancer research and treatment·2026
Same author

A Population-Based Cohort Study of Fracture Risk Among 122,255 Adults Before and After Cancer Diagnosis Compared with Matched Controls.

Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research·2026
Same author

Surgically Treated Cervical Cancer in Manitoba: A Retrospective Study of the Impact of Geography on Care.

Current oncology (Toronto, Ont.)·2026
Same author

Medico-legal cases involving gastroenterologists in Canada between 2017 and 2021.

Journal of the Canadian Association of Gastroenterology·2026
This summary is machine-generated.

Deep learning language models effectively identify cancer progression in electronic health records (EHRs), significantly reducing the need for manual chart reviews. Clinical-BigBird and Clinical-Longformer showed superior performance in breast and colorectal cancer progression detection.

Area of Science:

  • Computational linguistics
  • Oncology
  • Health informatics

Background:

  • Cancer progression is a critical outcome in research, often buried in unstructured electronic health record (EHR) text.
  • Manual chart reviews for extracting this data are time-consuming and expensive.

Purpose of the Study:

  • To evaluate the effectiveness of three deep learning language models in identifying breast and colorectal cancer progression within EHRs.
  • To compare the performance of Bio+ClinicalBERT, Clinical-BigBird, and Clinical-Longformer models.

Main Methods:

  • Retrospective analysis of EHRs for stage 4 breast and colorectal cancer patients (2004-2020).
  • Utilized pretrained deep learning models (Bio+ClinicalBERT, Clinical-BigBird, Clinical-Longformer) for data analysis.
  • Evaluated model performance using sensitivity, positive predictive value, AUC, and scaled Brier scores; identified influential tokens.
Keywords:
Bio+ClinicalBERTClinical-BigBirdClinical-Longformerbreast cancercancer progressioncolorectal cancernatural language processingretrospective chart review

More Related Videos

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
07:15

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Published on: August 16, 2020

7.4K

Related Experiment Videos

Last Updated: Jan 15, 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

489
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
07:15

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Published on: August 16, 2020

7.4K

Main Results:

  • Clinical-BigBird and Clinical-Longformer demonstrated higher accuracy and sensitivity than Bio+ClinicalBERT for both cancer types.
  • All models successfully reduced the chart review workload by over 84%.
  • The term 'progression' was identified as a highly influential token in model predictions.

Conclusions:

  • Deep learning models can automate the identification of cancer progression in EHRs, substantially decreasing manual review efforts.
  • Model performance can be enhanced by larger training datasets and sentence-level EHR analysis.
  • Understanding influential tokens is key to refining model interpretability and accuracy.