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: Dec 25, 2025

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.3K

A physics-guided modular deep-learning based automated framework for tumor segmentation in PET.

Kevin H Leung1,2, Wael Marashdeh3, Rick Wray4

  • 1Department of Biomedical Engineering, Johns Hopkins University, Baltimore, MD, United States of America.

Physics in Medicine and Biology
|April 3, 2020
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

Cycle-dependent variation of tumor absorbed dose rates in 177Lu-DOTATATE therapies.

Biomedical physics & engineering express·2026
Same author

Beyond the tumor: Recurrence-prone radiomics for prognostication in negative PSMA PET/CT scans of prostate cancer.

Biomedical physics & engineering express·2026
Same author

Kernel-based Maximum likelihood reconstruction of attenuation and activity (MLAA) in SPECT imaging for improved attenuation correction and activity quantification: Simulation, phantom and patient validation studies.

Physics in medicine and biology·2026
Same author

Microenvironment at a Distance: Multi-Endocrine-Organ Radiomics to Identify Systemic Signatures in PSMA-Negative Prostate Cancer.

Cancers·2026
Same author

Radiological and biological dictionary of radiomics features: addressing understandable AI issues in personalized breast cancer; dictionary version BM1.0.

Physics in medicine and biology·2026
Same author

Enhancing PRRT Outcome Prediction in Neuroendocrine Tumors: Aggregated Multi-Lesion PET Radiomics Incorporating Inter-Tumor Heterogeneity.

Cancers·2025
Same journal

Effective contrast-enhanced preprocessing for intracranial artery segmentation in digital subtraction angiography.

Physics in medicine and biology·2026
Same journal

Improving Plan Quality in Adaptive Proton Therapy Using an Interactive Dose Modification Tool.

Physics in medicine and biology·2026
Same journal

Technical Note: Real-Time MLC Control and Latency Measurement Optimization with External Verification.

Physics in medicine and biology·2026
Same journal

Fetus-Specific Hematopoietic Stem Cell Dosimetry Framework for Leukemia-Relevant Target Cells During Prenatal Development.

Physics in medicine and biology·2026
Same journal

Deep learning-based dose prediction to enhance planning efficiency in cervical brachytherapy with hybrid applicators.

Physics in medicine and biology·2026
Same journal

Corrigendum: Referenceless MR thermometry-a comparison of five methods (2017<i>Phys. Med. Biol</i>.<b>62</b>1-16).

Physics in medicine and biology·2026
See all related articles
This summary is machine-generated.

This study introduces an automated deep learning framework for positron emission tomography (PET) tumor segmentation. The physics-guided approach enhances accuracy for lung cancer treatment planning, even with limited data.

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Oncology

Background:

  • Reliable positron emission tomography (PET) tumor segmentation is crucial for radiation therapy planning and quantitative feature analysis.
  • Challenges include limited spatial resolution and a scarcity of clinical training data with precise tumor boundaries.
  • Existing methods often struggle with accuracy and generalizability across different imaging systems.

Purpose of the Study:

  • To develop an automated, physics-guided deep learning framework for accurate PET tumor segmentation.
  • To address limitations in training data availability and improve segmentation performance in challenging PET images.
  • To evaluate the framework's efficacy in segmenting primary lung tumors in 18F-fluorodeoxyglucose (FDG)-PET scans.

Main Methods:

More Related Videos

Author Spotlight: Standardizing Mouse In Vivo PET Imaging with Body Conforming Molds and Automated Analysis
07:45

Author Spotlight: Standardizing Mouse In Vivo PET Imaging with Body Conforming Molds and Automated Analysis

Published on: October 25, 2024

689

Related Experiment Videos

Last Updated: Dec 25, 2025

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.3K
Author Spotlight: Standardizing Mouse In Vivo PET Imaging with Body Conforming Molds and Automated Analysis
07:45

Author Spotlight: Standardizing Mouse In Vivo PET Imaging with Body Conforming Molds and Automated Analysis

Published on: October 25, 2024

689
  • A three-module framework was proposed: 1) A physics-based approach to generate realistic PET images with ground-truth tumors, addressing data scarcity.
  • 2) A modified U-net trained on simulated data for initial tumor segmentation.
  • 3) Fine-tuning with a small clinical dataset using radiologist delineations as surrogate ground-truth.

Main Results:

  • The framework achieved reliable performance on simulated (Dice Similarity Coefficient [DSC]: 0.87) and patient images (DSC: 0.73).
  • It outperformed several semi-automated methods and accurately segmented small tumors (down to 1.83 cm²).
  • The approach demonstrated generalizability across five PET scanners (DSC: 0.74) and robustness to partial volume effects, requiring minimal training data (DSC: 0.70 with 30 patients).

Conclusions:

  • The automated physics-guided deep learning framework provides a reliable solution for tumor delineation in FDG-PET images.
  • This method effectively overcomes challenges related to limited data and scanner variability in PET tumor segmentation.
  • The framework holds significant potential for improving PET-based cancer diagnosis and treatment planning.