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 Videos

Deep Learning-Based Auto-Segmentation for Liver Yttrium-90 Selective Internal Radiation Therapy.

Jun Li1, Wookjin Choi1, Rani Anne1

  • 1Department of Radiation Oncology, Thomas Jefferson University, Philadelphia, PA, USA.

Technology in Cancer Research & Treatment
|March 28, 2025
PubMed
Summary

A novel deep learning (DL) model accurately segments livers for Y-90 selective internal radiation therapy (SIRT). This advanced auto-segmentation method outperforms traditional atlas-based approaches, improving treatment planning reliability.

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

Cross-Institutional Validation of a novel LLM-Based Cardiac Event Extraction framework from Electronic Health Records.

International journal of radiation oncology, biology, physics·2026
Same author

Contour and dosimetric evaluation of an in-room mobile CBCT scanner for cylinder based brachytherapy.

Brachytherapy·2026
Same author

Process-first relief of emergency department overcrowding: a before-after analysis of time-to-decision and boarding in a regional center.

Clinical and experimental emergency medicine·2026
Same author

Weekly trends in emergency department visits, hospitalizations, and mortality in patients with end-stage kidney disease: nationwide registry analysis.

PloS one·2026
Same author

Development and validation of an AI-based lung lobe auto-contouring tool using radiation therapy planning free-breathing images.

Reports of practical oncology and radiotherapy : journal of Greatpoland Cancer Center in Poznan and Polish Society of Radiation Oncology·2026
Same author

Impact of health insurance status on hospitalization and mortality from emergency department admission in patients with end-stage kidney disease: a Korean nationwide registry analysis.

Kidney research and clinical practice·2025

Area of Science:

  • Medical Imaging
  • Artificial Intelligence in Medicine
  • Radiotherapy

Background:

  • Accurate liver delineation is crucial for Y-90 selective internal radiation therapy (SIRT) planning.
  • Manual segmentation is time-consuming and subject to inter-observer variability.
  • Existing automated methods, like atlas-based segmentation, have limitations in accuracy.

Purpose of the Study:

  • To evaluate a deep learning (DL)-based auto-segmentation method for liver delineation in Y-90 SIRT.
  • To compare the performance of the DL model against physician manual delineations and an atlas-based method.

Main Methods:

  • A U-Net3D deep learning architecture was developed for liver segmentation.
  • The DL model was tested on CT images from SIRT patients.
Keywords:
atlasauto-segmentationdeep learningliver delineationresin yttrium-90

Related Experiment Videos

  • Performance was assessed using Dice Similarity Coefficient (DSC), Mean Distance to Agreement (MDA), Volume Ratio (RV), and Activity Ratio (RA).

    • Main Results:

    • The DL model achieved higher agreement with manual delineations compared to the atlas-based method (DSC: 0.94 vs. 0.83; MDA: 1.8 mm vs. 7.1 mm).
    • Volume and activity ratios (RV: 0.99, RA: 1.00) indicated excellent agreement between DL-based and manual segmentations.
    • The DL model demonstrated reliable liver identification and segmentation in CT images.

    Conclusions:

    • The DL-based auto-segmentation method provides accurate and reliable liver delineation for Y-90 SIRT.
    • This approach surpasses traditional atlas-based methods in performance.
    • The developed DL model is suitable for clinical application in SIRT procedures.