eduzhai > Applied Sciences > Engineering >

Confidence-guided Lesion Mask-based Simultaneous Synthesis of Anatomic and Molecular MR Images in Patients with Post-treatment Malignant Gliomas

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 10 pages

Abstract: Data-driven automatic approaches have demonstrated their great potential inresolving various clinical diagnostic dilemmas in neuro-oncology, especiallywith the help of standard anatomic and advanced molecular MR images. However,data quantity and quality remain a key determinant of, and a significant limiton, the potential of such applications. In our previous work, we exploredsynthesis of anatomic and molecular MR image network (SAMR) in patients withpost-treatment malignant glioms. Now, we extend it and propose ConfidenceGuided SAMR (CG-SAMR) that synthesizes data from lesion information tomulti-modal anatomic sequences, including T1-weighted (T1w), gadoliniumenhanced T1w (Gd-T1w), T2-weighted (T2w), and fluid-attenuated inversionrecovery (FLAIR), and the molecular amide proton transfer-weighted (APTw)sequence. We introduce a module which guides the synthesis based on confidencemeasure about the intermediate results. Furthermore, we extend the proposedarchitecture for unsupervised synthesis so that unpaired data can be used fortraining the network. Extensive experiments on real clinical data demonstratethat the proposed model can perform better than the state-of-theart synthesismethods.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×