Purpose: For high dose rate prostate (HDR) brachytherapy at our institution, both CT and MRI are acquired to identify catheters and delineate the prostate, respectively. We propose to build a deep-learning model to generate synthetic magnetic resonance imaging (sMRI) with enhanced soft-tissue contrast from computed tomography (CT) scans. sMRI would assist physicians to accurately delineate the prostate without needing to acquire additional planning MRI.
Methods: 58 paired post-implant CT and T2-weighted MRI sets acquired on the same day for HDR prostate patients were retrospectively curated to train and validate the conditional Generative Adversarial Networks (cGAN) algorithm Pix2pix to generate sMRI. CT and MRI images were fused using a mutual information rigid registration algorithm in the Eclipse treatment planning system. The CT images for each patient were resampled to match to the dimensions of the MRI scan. Pix2pix was then trained to generate sMRI from the CT images using the MRI images as the ground truth. The performance of the algorithm was quantitatively evaluated using the metrics of mean absolute error (MAE), mean squared error (MSE) and cosine similarity (CS) over 400 images. MAE and MSE are 0 and CS is 1 when sMRI and MRI images are identical.
Results: sMRI images with enhanced soft-tissue contrast and catheter visualization were generated using Pix2pix for HDR prostate patients. The following performance metrics were obtained: MAE = 0.062±0.008, MSE = 0.008±0.002, CS = 0.987±0.006. Qualitatively, sMRI closely resembled the patient MRI in both soft tissues and bony anatomies in the region of near the prostate.
Conclusion: It is feasible to generate sMRI from CT scans clearly visualizing soft tissues, the boundary of the prostate, and catheters inside the prostate. To our knowledge, this is the first time both the boundary of the prostate and catheters have been visualized on sMRI.
Funding Support, Disclosures, and Conflict of Interest: Funding Support: Cancer Translational Research Development Grants at Loyola University Chicago