Torso Synthetic CT generation by integrating Deep Learning and Segmentation for FDG-PET/MR attenuation correction.
Authors
Affiliations (17)
Affiliations (17)
- Department of Radiology, Case Western Reserve University, 10900 Euclid Ave, Cleveland, 44106-5056, UNITED STATES.
- School of Artificial Intelligence and Computer Science, Jiangnan University, N/A, Wuxi, Jiangsu, 214122, CHINA.
- Department of Radiology, Case Western Reserve University, 10900 Euclid Ave, Cleveland, Ohio, 44106-7078, UNITED STATES.
- Jiangnan University, N/A, Wuxi, Jiangsu, 214122, CHINA.
- School of Digital Media, Jiangnan University, N/A, Wuxi, Jiangsu, 214122, CHINA.
- Medical Physics, University Hospitals Cleveland Medical Center, 11100 Euclid Avenue, Cleveland, Ohio, 44106-1716, UNITED STATES.
- HOPPR, N/A, Chicago, Illinois, N/A, UNITED STATES.
- Department of Radiology, University Hospitals Cleveland Medical Center, 10900 Euclid Avenue, Cleveland, Ohio, 44106-1716, UNITED STATES.
- Department of Public Health Sciences, Penn State College of Medicine, 700 HMC Crescent Road, Hershey, Pennsylvania, 17033-2360, UNITED STATES.
- Department of Radiation Oncology, Penn State University, 700 HMC Crescent Road, Hershey, Pennsylvania, 17033, UNITED STATES.
- Department of Radiation Oncology, University Hospitals Cleveland Medical Center, 10900 Euclid Avenue, Cleveland, Ohio, 44106-1716, UNITED STATES.
- Department of Obstetrics and Gynecology, Allegheny Health Network, Federal North Medical Office Building, 1307 Federal Street, Pittsburgh, Pennsylvania, 15212, UNITED STATES.
- Department of Obstetrics and Gynecology, Cleveland Clinic, 10900 Euclid Avenue, Cleveland, 44195-5243, UNITED STATES.
- Department of Radiation Oncology, Case Western Reserve University, 10900 Euclid Avenue, Cleveland, 44106-7078, UNITED STATES.
- Department of Radiation Oncology, St Luke's University Health Network, 1872 St. Luke's Boulevard, Easton, Pennsylvania, 18045, UNITED STATES.
- Department of Radiation Oncology, Mayo Clinic, 200 First St., Rochester, Minnesota, 55905-0002, UNITED STATES.
- Radiology, Case Western Reserve University, 11100 Euclid Ave, Cleveland, Ohio, 44106, UNITED STATES.
Abstract
Positron Emission Tomography/Magnetic Resonance (PET/MR) offers benefits over PET/CT including simultaneous PET and MR acquisition, intrinsic spatial registration accuracy, MR-based functional information, and superior soft tissue contrast. However, accurate attenuation correction (AC) for PET remains challenging as MR signals do not directly correspond to attenuation. Using deep learning algorithms that learn complex relationships, we generate synthetic CT (sCT) from MR for AC. Our novel method for AC, merges deep learning with threshold-based segmentation, to produce an AC map for the entire torso from Dixon MR images, which heretofore has not been demonstrated.

Twenty-nine prospectively collected, paired FDG-PET/CT and MR datasets were used for training and validation using the U-net Residual Network conditional Generative Adversarial Network integrated with tissue segmentation (URcGANmod) from Dixon MR data. Our application focused on torso (base of the skull to mid-thigh) AC, a common but challenging field of view (FOV). Performance was compared to that of 4 previously published methods.

Using 15 paired datasets for training and 14 independent datasets for testing, the URcGANmod generates an accurate torso sCT with a mean absolute difference of 32±4 HU per voxel. When applied for AC for FDG images, and considering evaluable (SUV ≥ 0.1 g/mL) voxels across all regions of interest, absolute values of the differences were within 4.4% from those determined using the measured CT for AC. Reproducibility was excellent with less than 3.5% standard deviation. The results demonstrate the accuracy and precision of URcGANmod method for torso sCT generation for quantitatively accurate MR-based AC (MRAC), exceeding the comparison methods.

Combining deep learning and segmentation enhances MRAC accuracy in torso FDG-PET/MR, improves SUV accuracy throughout the torso, achieves less than 4.4% SUV error, and outperforms comparison methods. Given the excellent sCT and SUV accuracy and precision, our proposed method warrants further studies for quantitative longitudinal multicenter trials.
.