2597

CT-based Synthetic pelvic T1-weighted MR image generation using a deep convolutional neural network (CNN)

Reza Kalantar¹, Jessica M Winfield^1,2, Christina Messiou^1,3, Dow-Mu Koh^1,3, and Matthew D Blackledge¹
¹Division of Radiotherapy and Imaging, The Institute of Cancer Research, London, United Kingdom, ²The Royal Marsden Hospital, London, United Kingdom, ³Department of Radiology, The Royal Marsden Hospital, London, United Kingdom

With the advent of the MR-Linac automatic pelvic MR segmentation is becoming essential to plan radiotherapy treatments efficiently. We developed a deep learning-based model to generate synthetic T₁W MR images from pelvic CT to enhance datasets for automated MR segmentation development.

Figure 3: Synthetic MR images generated from the validation set along with their corresponding input CT and T₁-weighted MR images. Good visual correspondence was observed between the synthetic image generated from the model and the reference MR images.

Figure 4: Synthetic T₁-weighted MR images generated from CT scans of independent patients (not included in the training/validation set) of the proposed model. Promising sMR images suggest their potential future benefits towards disease segmentation on MRI.