Both CT and MRI images are acquired for HDR prostate brachytherapy patients at our institution. CT is used to identify catheters and MRI is used to segment the prostate. To address scenarios of limited MRI access, we developed a novel Generative Adversarial Network (GAN) to generate synthetic MRI (sMRI) from CT with sufficient soft-tissue contrast to provide accurate prostate segmentation without MRI (rMRI). 
Approach: Our hybrid GAN, PxCGAN was trained utilizing 58 paired CT-MRI datasets from our HDR prostate patients. Using 20 independent CT-MRI datasets, the image quality of sMRI was tested using MAE, MSE, PSNR, and SSIM. These metrics were compared with the metrics of sMRI generated using Pix2Pix and CycleGAN. The accuracy of prostate segmentation on sMRI was evaluated using the Dice similarity coefficient (DSC), Hausdorff distance (HD), and mean surface distance (MSD) on the prostate delineated by three radiation oncologists (RO) on sMRI vs. rMRI. To estimate inter-observer variability (IOV), these metrics between prostate contours delineated by each RO on rMRI and the prostate delineated by treating RO on rMRI ("gold standard") were calculated.
Main results: Qualitatively, sMRI images show enhanced soft-tissue contrast at the prostate boundary compared with CT scans. For MAE and MSE, PxCGAN and CycleGAN have similar results, while the MAE of PxCGAN is smaller than that of Pix2Pix. PSNR and SSIM of PxCGAN are significantly higher than Pix2Pix and CycleGAN (p<0.01). The DSC for sMRI vs. rMRI is within the range of the IOV, while HD for sMRI vs. rMRI is smaller than the HD for IOV for all ROs (p≤0.03).
Significance: PxCGAN generates sMRI images from the treatment planning CT scans which depicts enhanced soft-tissue contrast at the prostate boundary. The accuracy of prostate segmentation on sMRI compared to rMRI is within the segmentation variation on rMRI between different radiation oncologists.
Physics in medicine and biology. 2023 Jul 11 [Epub ahead of print]
Hyejoo Kang, Alexander R Podgorsak, Bhanu Prasad Venkatesulu, Anjali L Saripalli, Brian Chou, Abhishek A Solanki, Matthew M Harkenrider, Steven M Shea, John C Roeske, Mohammed Abuhamad
Radiation Oncology, Stritch School of Medicine, Cardinal Bernadin Cancer Center, Loyola University Chicago, 2160 S 1st Ave, Chicago, Illinois, 60611-2001, UNITED STATES., Radiation Oncology, Stritch School of Medicine, Cardinal Bernadin Cancer Center, Loyola University Chicago, 2160 SOUTH 1ST AVE, MAYWOOD, Chicago, Illinois, 60611-2001, UNITED STATES., Radiology, Loyola University Chicago Stritch School of Medicine, 2160 SOUTH 1ST AVE, MAYWOOD, MAYWOOD, Illinois, 60153, UNITED STATES., Radiation Oncology Department, Stritch School of Medicine, Cardinal Bernadin Cancer Center, Loyola University Chicago, 2160 S. First Ave, Maywood, IL 60153, USA, Chicago, Illinois, 60611-2001, UNITED STATES., Computer Science, Loyola University Chicago, 1032 W. Sheridan Rd, Chicago, Illinois, 60611-2001, UNITED STATES.