USYD & Nepean Hospital

Created in March 01, 2022

2022

Deep Learning for Prostate Cancer Detection and Diagnosis Using Multimodal MRI, Sydney, Australia

  • Proposed a Zonal-aware Self-supervised Mesh Network (Z-SSMNet) for clinically significant prostate cancer (csPCa) detection and diagnosis using bi-parametric MRI (bpMRI).
  • The proposed network adaptively fused multiple 2D/2.5D/3D convolutional neural networks (CNNs) to capture a balanced representation of dense intra-plane information and sparse inter-plane information of bpMRI.
  • Introduced a self-supervised learning technique to pre-train the network using large-scale unlabeled data.
  • Proposed a network constraint technique to enhance csPCa detection by focusing on zonal anatomical regions.