Patch-Based Label Fusion for Automatic Prostate Segmentation in MR Images


An accurate automated method for prostate segmentation in magnetic resonance imaging (MRI).

Key Benefits

  • Yields faster results.
  • Produces less subjective results compared to manual method.

Market Summary

Cancer indications have provided increased utility for MRI machines, since much of the cost associated with cancer treatment is related to diagnostic imaging. Using MRI to decrease the number of random biopsies, now requires more efficient and accurate methods of MRI segmentation. Prostate MRI segmentation is useful for various aspects: to accurately localize prostate boundaries for radiotherapy, perform volume estimation to track disease progression, to initialize multi-modal registration algorithms or to obtain the region of interest for computer-aided detection of prostate cancer. Manual segmentation of the prostate is time consuming and subject to inter- and intra-observer variation. The inventors have developed an automated segmentation method to address this technical challenge.

Technical Summary

Researchers from Emory University have developed an automated prostate MRI segmentation approach based on nonlocal patch-based label fusion. This software is a 3D multi-atlas-based prostate segmentation method for MR images, which is based on patch-based label fusion strategy. The atlases with the most similar appearance are selected to serve as the best subjects in the label fusion. A local patch-based atlas fusion is performed using voxel weighting based on anatomical signature. This will be useful for image-guided interventions in prostate-cancer diagnosis and treatment.

Developmental Stage

Inventors have demonstrated its clinical feasibility, and validated its accuracy with manual segmentations.

Publication: Yang, X et al. Proc. SPIE 9786. Medical Imaging 2016: Image-Guided Procedures, Robotic Interventions, and Modeling. 978621.

Patent Information

Tech ID: 15228
Published: 4/16/2018