A Fully Automated Method for Bladder Segmentation in PSMA PET/CT Scans
Image Segmentation Radiomics PyTorch
We developed a fully automated method using the 3D U-Net architecture for bladder segmentation in PSMA PET/CT scans. The proposed approach is a critical step towards automating prostate lesion detection and improving standardization of clinical reporting. This work was done in collaboration with the Quantitative Radiomolecular Imaging and Therapy (Qurit) lab which is led by UBC Professor Arman Rahmim. We presented our results at the EANM'21 conference.