Related Experiment Video
Updated: May 5, 2026
09:25
Lineage Tracing and Clonal Analysis in Developing Cerebral Cortex Using Mosaic Analysis with Double Markers MADM
Published on: May 8, 2020
10.6K
Cross-Modality Image Translation From Brain 18 F-FDG PET/CT Images to Fluid-Attenuated Inversion Recovery Images
Sangwon Lee1, Jin Ho Jung1, Yong Choi1
1From the Department of Electronic Engineering, Sogang University, Seoul, Republic of Korea.
Clinical Nuclear Medicine
|September 26, 2024
Summary
A novel deep learning framework, CypixGAN, generates high-quality synthetic FLAIR images from PET/CT scans. This approach enhances dementia diagnosis cost-effectiveness when MRI is unavailable.
Area of Science:
- Artificial Intelligence in Medical Imaging
- Deep Learning for Neuroimaging
- Generative Adversarial Networks (GANs)
Background:
- Positron Emission Tomography/Computed Tomography (PET/CT) and Magnetic Resonance Imaging (MRI) are crucial for dementia diagnosis but present cost and accessibility challenges.
- Developing advanced imaging techniques is essential to improve diagnostic accuracy and patient convenience.
Purpose of the Study:
- To develop a generative adversarial network (GAN)-based framework, CypixGAN, for synthesizing Fluid-Attenuated Inversion Recovery (FLAIR) images from 18F-FDG PET and CT scans.
- To evaluate the performance of CypixGAN in generating high-fidelity synthetic FLAIR images compared to existing methods.
Main Methods:
- Utilized a dataset of 143 patients undergoing PET/CT and MRI for training, validation, and testing of deep learning models.
- Employed pix2pix, CycleGAN, and the proposed CypixGAN (combining CycleGAN with pix2pix's L1 loss) to generate synthetic FLAIR images.
- Segmented white matter hyperintensities (WMHs) from generated images and compared quantitative metrics (PSNR, SSIM) and visual quality.
Main Results:
- CypixGAN demonstrated superior performance in generating synthetic FLAIR images with significantly higher Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM) compared to pix2pix and CycleGAN.
- Quantitative metrics for CypixGAN were (mean ± SD): PSNR 20.23 ± 1.31, SSIM 0.80 ± 0.02, outperforming pix2pix and CycleGAN.
- Synthetic FLAIR images from CypixGAN accurately represented WMHs, showing minimal volume differences and high Dice similarity coefficients compared to ground-truth images.
Conclusions:
- The CypixGAN framework successfully generated high-quality synthetic FLAIR images, preserving spatial information even with unpaired data.
- This GAN-based approach offers a promising solution for improving the diagnostic performance and cost-effectiveness of PET/CT in dementia assessment, particularly when MRI is not accessible.
Related Concept Videos
01:25Imaging Studies II: Positron Emission Tomography and Scintigraphy
853
Positron Emission Tomography (PET) is a medical imaging technique that provides crucial insights into the body's physiological functions at a molecular level. It is an indispensable resource for diagnosing, staging, and monitoring various illnesses, notably cancer, neurological disorders, and cardiovascular conditions.
Fundamental Principles of PET
Fundamental Principles of PET
853
01:27Imaging Studies III: Computed Tomography
893
DefinitionComputed Tomography (CT) of the genitourinary (GU) tract is a non-invasive imaging modality that utilizes X-rays and computer processing to generate detailed cross-sectional images of the urinary system, encompassing the kidneys, ureters, bladder, and adjacent structures such as the adrenal glands.PurposeCT scans of the GU tract serve several diagnostic and therapeutic purposes, including:Diagnosis of Urinary Tract Diseases: Detects kidney stones, tumors, cysts, and congenital...
893