Difference between revisions of "2013 Project Week:PkModeling for prostate DCE MRI"

From NAMIC Wiki
Jump to: navigation, search
 
Line 25: Line 25:
 
* The main source of the inconsistency was due to different models used by the two tools being compared: 2-parameter (Ktrans+ve) vs 3-parameter (Ktrans+ve+fpv). After fixing this, the results were very similar (difference <1% vs 1000%).
 
* The main source of the inconsistency was due to different models used by the two tools being compared: 2-parameter (Ktrans+ve) vs 3-parameter (Ktrans+ve+fpv). After fixing this, the results were very similar (difference <1% vs 1000%).
 
* Bolus arrival time was not estimated in the Slicer PkModeling tool.
 
* Bolus arrival time was not estimated in the Slicer PkModeling tool.
* We may need to re-engineer the module to accommodate multiple models. Further discussions are needed.
+
* Work in progress:
 +
** Andrey: update the code to propagate model selection through various layers/filters of the module
 +
** Jim: look into the issue of synchronizing bolus arrival time for the plasma/tissue curves
 
</div>
 
</div>
 
</div>
 
</div>

Latest revision as of 00:16, 11 January 2013

Home < 2013 Project Week:PkModeling for prostate DCE MRI

Key Investigators

  • BWH: Andrey Fedorov
  • GE: Jim Miller

Project Description

Objective

  • debug and test the pharmacokinetic modeling extension on prostate DCE MRI data

Approach, Plan

  • previously validated non-open-source implementation of the PK modeling will be used as reference for comparison with the results produced by PkModeling extension
  • prostate DCE MRI data will be used for testing using identically defined AIF
  • success will be achieved if the results produced by PkModeling and the reference software agree

Progress

  • The main source of the inconsistency was due to different models used by the two tools being compared: 2-parameter (Ktrans+ve) vs 3-parameter (Ktrans+ve+fpv). After fixing this, the results were very similar (difference <1% vs 1000%).
  • Bolus arrival time was not estimated in the Slicer PkModeling tool.
  • Work in progress:
    • Andrey: update the code to propagate model selection through various layers/filters of the module
    • Jim: look into the issue of synchronizing bolus arrival time for the plasma/tissue curves

References