COMPARISON OF GENERATIVE ADVERSARIAL NETWORKS IN MEDICAL IMAGING APPLICATIONS - MR to CT image synthesis

Cancer is one of the leading causes of death worldwide with about half of all cancer patients undergoing radiation therapy, either as a standalone treatmentor in combination with other methods such as chemotherapy. \u91The dose planning of radiation therapy is based on medical images such as CT and MR images. A recent trend is to move towards “MR-only” work\u83ows, and previous work have shown good results when synthesizing CT images from MR with machine learning methods. Eliminating the need for CT scans in the dose planning procedure removes a potentially carcinogenic part of the procedure, saves clinical resources, and could shorten the time until treatment can begin. Th\u91is thesis builds upon earlier work, where a Cycle-Consistent Adversarial Network (CycleGAN) was used successfully to synthesize CT images from MR Iimages. We compare the CycleGAN architecture to the original Generative Adversarial Network (GAN), and a variant called Wasserstein GAN (WGAN). Th\u91e UNetwas used as a generative sub-network for all models.Th\u91e CycleGAN utilized a PatchGAN discriminative sub-network while the other models used a custom Convolutional Neural Network (CNN). Th\u91e GAN architecture was tested both with paired and unpaired data. Th\u91e best results were obtained by the GAN using unpaired data, with an MAE of 30:4 HU and PSNR of 26:7 dB. \u91The CycleGAN model failed to produce images what could pass as suitable synthetic CT images. We could not reproduce the results of earlier work in this regard while either using the hyperp arameters of previous studies or other con\u80gurations. But we conclude that it is possible to produce synthesized CT images using GANs with both paired and unpaired data