Close
Longitude with pathology challenge
Longitude intra-subject registration with pathology Paper references: Prastawa M, Bullitt E, Gerig G, “Simulation of Brain Tumors in MR Images for Evaluation of Segmentation Efficacy.” Medical Image Analysis (MedIA), Vol 13, No 2, Apri More »

Longitude intra-subject registration with pathology

Rendered View

Paper references:

Prastawa M, Bullitt E, Gerig G, “Simulation of Brain Tumors in MR Images for Evaluation of Segmentation Efficacy.” Medical Image Analysis (MedIA), Vol 13, No 2, April 2009, Pages 297-311. [http://www.sci.utah.edu/~prastawa/papers/MedIA2009_Prastawa_TumorSim.pdf ](http://www.sci.utah.edu/~prastawa/papers/MedIA2009_Prastawa_TumorSim.pdf )

Niethammer M, Hart G, Pace D, Vespa P, Irimia A, Van Horn J, Aylward S, “Geometric Metamorphosis.” Lecture Notes in Computer Science, Vol. 6893, Medical Image Computing and Computer Assisted Intervention (MICCAI) 2011. Pages 639-646 http://wwwx.cs.unc.edu/~mn/sites/default/files/niethammer2011_geometric_metamorphosis_miccai.pdf

Description:

<p> This collection of synthetic data is intended for the development and evaluation of methods for intra-subject registration in the presence of changing pathologies. Specifically, we developed this data to evaluate our Geometric Metamorphosis registration method for the quantification of tumor progression (e.g., estimating its infiltrating and displacing components). This method is also being applied to the quantification and prediction of chronic blood perfusion changes after Traumatic Brain Injuries.
</p>

<p> In this data collection three variables are explored: <ol> <li>Tumor location: near edge of brain versus in the center of brain.</li> <li>Tumor size: small versus large.</li> <li>Tumor type: highly infiltrative, mix of infiltrative and displacive, and highly displacive.</li> </ol> </p>

InfiltrativeTumor Figure 1. An example of a large, ring-enhancing, highly infiltrative tumor near the edge of the brain.

License:

This data is released under the Creative-Commons Attribution (CC-BY) license. Please cite the above papers if you use this data in your research. For more details, see http://wiki.creativecommons.org/Creative_Commons_Attribution

Data:

The image data were generated using the TumorSim 1.2 software developed at the University of Utah by Marcel Prastawa. That software is available at http://www.nitrc.org/projects/tumorsim

The software combines physical and statistical modeling to generate synthetic multi-modal 3D brain MRI with tumor and edema, along with the underlying anatomical ground truth. Emphasis is placed on simulating the major effects of tumors, particularly as captured in MRI, such as contrast enhancement, local distortion of healthy tissue, infiltrating edema adjacent to tumors, destruction and deformation of fiber tracts, and multi-modal MRI contrast of healthy tissue and pathology.

Simulation was initiated using a healthy MRI study from the BrainWeb database: <br /> http://www.bic.mni.mcgill.ca/brainweb/anatomic_normal_20.html

For each simulation, the results as well as the "Sim_[XXX].xml" parameters file used to generate those results are available.

This data generation effort was supported in-part by: <ol> <li>NIH/NIBIB grant "National Alliance of Medical Image Computing" (NA-MIC, PI: Kikinis, 1U54EB005149)</li> <li>NIH/NCI grant "Image Registration for Ultrasound-Based Neurosurgical Navigation" (TubeTK, PI: Wells, Aylward, 1R01CA138419)</li> <li>NIH/NCI grant “A Needle Guidance System for Hepatic Tumor Ablation that Fuses Real-Time Ultrasound” (InnerOptic, PI: Razzaque, R1R44CA143234-02A1)</li> </ol>

Name Size Modified
Public 5 years ago

Stats

  • 2 members

Actions

Checked

Info