Multiparametric MRI in Localization of Prostate Cancer with Patient-specific MRI-based Prostate Mold Validation

Project Collaborators (sorted by Institute): 
Center for Interventional Oncology, CC
Laboratory of Pathology, Center for Cancer Research, NCI
Urology Oncology Branch, Center for Cancer Research, NCI
Molecular Imaging Program, NCI
Molecular Imaging Program, Center for Cancer Research, NCI
Urologic Oncology Branch, Center for Cancer Research, NCI
Project Brief: 

The NCI Molecular Imaging Program develops innovative methods of localizing prostate cancer based on in vivo multiparametric MRI.  SPIS contributes to this effort in two areas.  (1) Assisted in the design, implementation, and validation of individualized MR-based molds to facilitate the correlation of the multiparametric MRI imaging with downstream histopathology.  Achieving sufficient MR/histopathology image correlation (i.e., 3D alignment) is critical when evaluating the efficacy of new prostate imaging protocols as potential cancer diagnostics.  During the initial mold development phase, SPIS 3D-printed ~130 patient-specific molds.  Recently, SPIS has resumed the mold printing (2-5 surgeries/week), and is now responsible for automating the process of creating the mold 3D model files from the MRI imaging files.  (2) Developing an unsupervised multi-characteristic framework for prostate cancer localization using diffusion-weighted magnetic resonance imaging (DW-MRI).

More specifically, the potential single imaging modality localization method consists of combining a number of unique parameters from different models that characterize underlying tissue structure estimated from DW-MRI signal attenuation. The process of combining unique parameters from multiple models allows for the reduction of false positive and false negative voxels compared to using a single parameter, such as only apparent diffusion coefficient or kurtosis. This is due to the fact that theoretically a ‘positive-if-all-positive’ selection process favors higher specificity.

Process of MR imaging to patient-specific prostate mold/slicer

Process of making the patient-specific sectioning-mold from the MRI imaging: (a) prostate MRI images, (b) segmentation, (c) patient-specific prostate model, and (d) 3D-printed mold with knife slots for 6mm sectioning.

Positive-if-all-Positive approach for combining three unique DW-MRI parameters

Estimated characteristics: apparent diffusion coefficient, D (a); pseudo-diffusion fraction, f, (b); kurtosis, K, (c); corresponding threshold maps (d-f); unsupervised framework map with cancerous voxels in red and tumor suspicious voxels in green (g).

Unsupervised framework map, weighted MR image, and histology slide

Unsupervised framework map with cancerous voxels in red and tumor suspicious voxels in green (a); axial T2-weighted MR image confirming large regions high level of risk for prostate carcinoma (b); corresponding histology slide confirms a large focus of Gleason 4+4 cancer (c).

Awards: 
  • 2012 NIH Director’s Award: Design and implementation of a method for correlating in vivo prostate MRI and histopathology using individualized MR-based molds resulting in  a standard-of-care at NIH