Program Information
Testing DSC-MRI Leakage Correction Using a Brain-Like Digital Reference Object
M Jen*, P Hou , R Stafford , H Liu , UT MD Anderson Cancer Center, Houston, TX
Presentations
MO-RPM-GePD-IT-3 (Monday, July 31, 2017) 3:45 PM - 4:15 PM Room: Imaging ePoster Theater
Purpose: To test the post-processing based leakage correction in dynamic susceptibility contrast (DSC) MRI for the quantification of tumor blood volume using a brain-like digital reference object (DRO) in a clinical implemented software.
Methods: A digital brain phantom was constructed using templates including gray matter (GM), white matter (WM), artery and choroid plexus. DSC MRI signal time curves were simulated with tissue-specific physiological parameters. Tumor lesions objects were assigned with distinct level of contrast extravasation (K₂= +0.02, 0, -0.02) and blood volume (CBV = 0.74, 2.27, 7.02 mL/100mL) using the Weisskoff model. The DRO perfusion data was analyzed with the Nordic ICE software version 2.3.14. The software generated cerebral blood volume (CBV) map from the DRO with lesions was compared with those without lesions. Post-processing was performed both with and without the leakage correction. The relative CBV (rCBV) in lesions and GM was obtained by contrast to CBV from WM.
Results: The rCBV in GM was found to be identical in each processed with and without the lesion objects added to the digital phantom. It was unchanged after the leakage correction was applied. As expected, the tumor rCBV was under-estimated with the T1-dominant effect and over-estimated with the T2-dominant effect. After the leakage correction was applied, the errors significantly reduced to a mean of 3.51% and 1.32%, respectively, with the T1-dominant and T2-dominant effect.
Conclusion: The brain-like DRO was successfully applied for the testing of a clinical software for processing of the DSC-MRI. Lesion simulating contrast extravasation did not interfere with the selection of normal tissue time curve required for the leakage correction in this study. Further study will include testing and comparing various clinical software using the DRO.
Contact Email: