PURPOSES: To provide abdominal contrast-enhanced MR image synthesis, we developed an gradient regularized multi-modal multi-discrimination sparse attention fusion generative adversarial network (GRMM-GAN) to avoid repeated contrast injections to patients and facilitate adaptive monitoring. METHODS: With IRB approval, 165 abdominal MR studies from 61 liver cancer patients were retrospectively solicited from our institutional database. Each study included T2, T1 pre-contrast (T1pre), and T1 contrast-enhanced (T1ce) images. The GRMM-GAN synthesis pipeline consists of a sparse attention fusion network, an image gradient regularizer (GR), and a generative adversarial network with multi-discrimination. The studies were randomly divided into 115 for training, 20 for validation, and 30 for testing. The two pre-contrast MR modalities, T2 and T1pre images, were adopted as inputs in the training phase. The T1ce image at the portal venous phase was used as an output. The synthesized T1ce images were compared with the ground truth T1ce images. The evaluation metrics include peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and mean squared error (MSE). A Turing test and experts contours evaluated the image synthesis quality. RESULTS: The proposed GRMM-GAN model achieved a PSNR of 28.56, an SSIM of 0.869, and an MSE of 83.27. The proposed model showed statistically significant improvements in all metrics tested with p-values < 0.05 over the state-of-the-art model comparisons. The average Turing test score was 52.33%, which is close to random guessing, supporting the models effectiveness for clinical application. In the tumor-specific region analysis, the average tumor contrast-to-noise ratio (CNR) of the synthesized MR images was not statistically significant from the real MR images. The average DICE from real vs. synthetic images was 0.90 compared to the inter-operator DICE of 0.91. CONCLUSION: We demonstrated the function of a novel multi-modal MR image synthesis neural network GRMM-GAN for T1ce MR synthesis based on pre-contrast T1 and T2 MR images. GRMM-GAN shows promise for avoiding repeated contrast injections during radiation therapy treatment.

INTRODUCTION: Inflammatory bowel disease (IBD) is a relative contraindication to external beam radiation therapy (EBRT) for prostate cancer patients due to fear of increased risk of gastrointestinal (GI) toxicity. High-dose-rate (HDR) brachytherapy, capable of minimizing radiation dose to surrounding tissues, is a feasible alternative. Given limited data, this study examined the safety profile of HDR brachytherapy in this setting. MATERIAL AND METHODS: We conducted a retrospective review of patients with localized prostate cancer and IBD treated with HDR brachytherapy at the University of California San Francisco (UCSF), between 2010 and 2022. Eligibility criteria included biopsy-proven prostate cancer, no distant metastases, absence of prior pelvic radiotherapy, IBD diagnosis, and at least one follow-up visit post-treatment. RESULTS: Eleven patients were included, with a median follow-up of 28.7 months. The median dose administered was 2700 cGy (range, 1500-3150 cGy) over 2 fractions (range, 1-3 fractions). Two patients also received EBRT. Rectal spacers (SpaceOAR) were applied in seven patients. All patients experienced acute genitourinary (GU) toxicity, ten of which were grade 1 and one was grade 2. Eight patients experienced late grade 1 GU toxicity, and three patients had late grade 2 GU toxicity. GI toxicities were similarly low-grade, with six grade 1 acute toxicity, no grade 2 or higher acute toxicity, six grade 1 late toxicity, and one late grade 2 GI toxicity. No grade 3 or higher acute or late GI or GU toxicities were reported. CONCLUSIONS: HDR brachytherapy appears to be a safe and tolerable treatment modality for patients with prostate cancer and IBD, with minimal acute and late GI and GU toxicity. These findings warrant multi-institutional validation due to small sample size.