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: Mar 13, 2026

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

1.2K

Developing a segmentation cascade deep learning network based on automated prompts.

Yuhe Yao1, Shiran Sun1, Xuena Yan1

  • 1National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.

Quantitative Imaging in Medicine and Surgery
|March 12, 2026
PubMed
Summary
This summary is machine-generated.

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

Induction chemotherapy for locally advanced OPSCC: real-world evidence and HPV-dependent prognostic value of treatment response.

Radiation oncology (London, England)·2026
Same author

Dynamic changes of physicochemical, microbial, and flavor characteristics in Xinjiang rawish cereal vinegar fermentation.

Food chemistry: X·2026
Same author

Neoadjuvant immunochemotherapy followed by definitive radiotherapy in locally advanced head and neck squamous cell carcinoma: two-year survival outcomes.

Cancer immunology, immunotherapy : CII·2026
Same author

Evaluation of Prognostic Factors in Adult Sinonasal Rhabdomyosarcoma With Consideration of Resection Status.

Otolaryngology--head and neck surgery : official journal of American Academy of Otolaryngology-Head and Neck Surgery·2026
Same author

Optimizing Cardiac Dose Prediction in Left-Sided Breast Cancer Radiation Therapy: Clinical Strategies for Identifying Beneficiaries of the Deep Inspiration Breath-Hold Technique.

Advances in radiation oncology·2026
Same author

Dosimetric Impact and Correction of Lipiodol Deposition on Photon-Beam Radiotherapy for Hepatocellular Carcinoma Using a Novel Simulation Model.

Journal of hepatocellular carcinoma·2026

Accurate auto-segmentation of nasopharyngeal carcinoma gross tumor volume (GTVnx) is improved by a novel two-stage deep learning framework. This cascaded approach enhances precision and efficiency for GTVnx segmentation in clinical settings.

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Oncology

Background:

  • Accurate auto-segmentation of nasopharyngeal carcinoma gross tumor volume (GTVnx) presents a significant clinical challenge.
  • Existing single-stage deep learning models often face limitations in precision for GTVnx segmentation.

Purpose of the Study:

  • To introduce a novel two-stage deep learning cascade framework for enhanced GTVnx auto-segmentation.
  • To improve the accuracy and computational efficiency of GTVnx segmentation compared to single-stage models.

Main Methods:

  • Developed an end-to-end model with a localization prompt generation unit (PGU) and a fine segmentation unit (FSU), linked by an attention mechanism.
  • Evaluated three PGU strategies (prompt-mask, prompt-box, dual-prompt) on 276 nasopharyngeal carcinoma (NPC) patient datasets.
Keywords:
Auto-segmentationpromptstwo-stage

More Related Videos

Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

11.1K

Related Experiment Videos

Last Updated: Mar 13, 2026

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

1.2K
Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

11.1K
  • Assessed performance using Dice similarity coefficient (DSC), 95% Hausdorff distance (HD95), and mean distance to agreement (MDA), with statistical analysis via paired t-tests.

    • Main Results:

    • The dual-prompt strategy significantly outperformed the baseline single-stage model, increasing DSC by 9.7% (0.8219 vs. 0.7489, P<0.001).
    • HD95 decreased by 28.2% (9.22 vs. 12.84 mm, P<0.001) and MDA improved by 31.4% (1.53 vs. 2.23 mm, P<0.001) with the dual-prompt approach.
    • Qualitative analysis confirmed superior anatomical fidelity and statistically significant improvements over single-prompt variants (P<0.01).

    Conclusions:

    • The proposed two-stage deep learning-based auto-segmentation (DLAS) framework offers a computationally efficient solution for GTVnx segmentation.
    • This approach significantly enhances the accuracy and reliability of GTVnx segmentation, featuring a lightweight architecture and high scalability.
    • The framework represents a promising pathway for clinical integration in radiation oncology for nasopharyngeal carcinoma treatment planning.