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

Magnetic Resonance Imaging01:24

Magnetic Resonance Imaging

5.0K
Magnetic resonance imaging (MRI) is a noninvasive medical imaging technique based on a phenomenon of nuclear physics discovered in the 1930s, in which matter exposed to magnetic fields and radio waves was found to emit radio signals. In 1970, a physician and researcher named Raymond Damadian noticed that malignant (cancerous) tissue gave off different signals than normal body tissue. He applied for a patent for the first MRI scanning device in clinical use by the early 1980s. The early MRI...
5.0K
Positron Emission Tomography01:29

Positron Emission Tomography

4.1K
Positron emission tomography (PET) is a medical imaging technique involving radiopharmaceuticals — substances that emit short-lived radiation. Although the first PET scanner was introduced in 1961, it took 15 more years before radiopharmaceuticals were combined with the technique and revolutionized its potential.
One of the main requirements of a PET scan is a positron-emitting radioisotope, which is produced in a cyclotron and then attached to a substance used by the part of the body...
4.1K
  1. Home
  3. Comparison Of The Accuracy Of A Deep Learning Method For Lesion Detection In Pet/ct And Pet/mri Images.
  1. Home
  3. Comparison Of The Accuracy Of A Deep Learning Method For Lesion Detection In Pet/ct And Pet/mri Images.

Related Experiment Video

MRI and PET in Mouse Models of Myocardial Infarction
10:46

MRI and PET in Mouse Models of Myocardial Infarction

Published on: December 19, 2013

11.7K

Comparison of the Accuracy of a Deep Learning Method for Lesion Detection in PET/CT and PET/MRI Images.

Lifang Pang1,2,3,4, Zheng Zhang5, Guobing Liu1,2,3,4

  • 1Department of Nuclear Medicine, Zhongshan Hospital, Fudan University, No. 180, Fenglin Road, Shanghai, 200032, People's Republic of China.

Molecular Imaging and Biology
|August 14, 2024

View abstract on PubMed

Summary
This summary is machine-generated.

A new deep learning model accurately detects lesions in both PET/CT and PET/MRI scans. While PET/MRI has longer delays and lower signal-to-noise ratio, it identifies more lesions, with minimal impact from image quality on accuracy.

Keywords:
Deep learningLesion detectionPET/CTPET/MRI

More Related Videos

Positron Emission Tomography Imaging for In Vivo Measuring of Myelin Content in the Lysolecithin Rat Model of Multiple Sclerosis
08:40

Positron Emission Tomography Imaging for In Vivo Measuring of Myelin Content in the Lysolecithin Rat Model of Multiple Sclerosis

Published on: February 28, 2021

4.1K
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

6.8K

Related Experiment Videos

MRI and PET in Mouse Models of Myocardial Infarction
10:46

MRI and PET in Mouse Models of Myocardial Infarction

Published on: December 19, 2013

11.7K
Positron Emission Tomography Imaging for In Vivo Measuring of Myelin Content in the Lysolecithin Rat Model of Multiple Sclerosis
08:40

Positron Emission Tomography Imaging for In Vivo Measuring of Myelin Content in the Lysolecithin Rat Model of Multiple Sclerosis

Published on: February 28, 2021

4.1K
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

6.8K

Area of Science:

  • Medical Imaging Analysis
  • Artificial Intelligence in Oncology
  • Radiomics and Quantitative Imaging

Background:

  • Developing universal lesion recognition algorithms for Positron Emission Tomography/Computed Tomography (PET/CT) and PET/Magnetic Resonance Imaging (PET/MRI) is crucial for accurate cancer staging and treatment monitoring.
  • Existing deep learning models often require modality-specific training, limiting their applicability across different imaging techniques.
  • The integration of synthetic CT (sCT) generation from MRI data presents a pathway for creating unified deep learning frameworks.

Purpose of the Study:

  • To develop and validate a universal deep learning algorithm for lesion detection applicable to both PET/CT and PET/MRI.
  • To investigate the performance of the algorithm across different imaging modalities and identify factors influencing its accuracy.
  • To compare quantitative imaging features, such as Total Lesion Glycolysis (TLG) and Metabolic Tumor Volume (MTV), between PET/CT and PET/MRI.

Main Methods:

  • A deep learning lesion detection model was trained using 2D and 3D fractional-residual (F-Res) networks on a large PET/CT dataset (AutoPet Challenge 2022).
  • A synthetic CT (sCT) generation network was developed to adapt the model for PET/MRI, using paired MR/CT data.
  • The universal algorithm was validated on 38 patients' PET/CT and PET/MRI data, with performance assessed by detection accuracy, precision, recall, and Dice coefficients.

Main Results:

  • PET/MRI scans exhibited significantly longer delay times (135 min vs. 61 min) and lower signal-to-noise ratio (SNR) compared to PET/CT.
  • Despite lower SNR, PET/MRI detected more lesions, and key quantitative metrics like TLG and MTV showed no significant differences between modalities.
  • Outlier analysis identified modality-specific false positive regions, with ureters for PET/CT and intestines/testicles for PET/MRI.

Conclusions:

  • The developed deep learning model demonstrates robust performance for universal lesion detection in both PET/CT and PET/MRI.
  • While SNR and reconstruction parameters have minimal impact, post-injection delay time significantly affects recognition accuracy.
  • The findings support the potential of a unified deep learning approach for lesion analysis across hybrid PET imaging modalities.