Encrypted login | home

Program Information

Evaluation of Dynamic Fractal Signature Dissimilarity as a Novel Biomarker in Therapeutic Response Assessment Using Dynamic Contrast-Enhanced MRI

C Wang

C Wang*, E Subashi , F Yin , Z Chang , Duke University Medical Center, Durham, NC


TH-CD-204-7 (Thursday, July 16, 2015) 10:00 AM - 12:00 PM Room: 204

Purpose:To evaluate dynamic fractal signature dissimilarity (FSD) analysis in the therapeutic response assessment using longitudinal DCE-MRI.

Methods:A novel texture analysis method of dynamic FSD was proposed to quantify the heterogeneity of contrast agent (CA) uptake during a DCE-MRI exam. 16 nu/nu mice with LS-174T implanted were randomly assigned into treatment/control groups (n=8/group) and received bevacizumab/saline three times (Day1/Day4/Day8). All mice received one pre- (Day0) and two post-treatment (Day2/Day9) scans. For each scan, pharmacokinetic (PK) analysis was performed using the extended Tofts model, and a dynamic FSD curve was generated using the DCE images in the first two minutes after CA injection. Two heuristic features of a dynamic FSD curve, Area Under Curve (AUCFSD) and Maximum Enhancement (MEFSD), were compared against the tumor volume V and the tumor permeability rate constant Ktrans. For each post-treatment scan, the relative values (in reference to Day0 values) of AUCFSD, MEFSD, V and Ktrans were recorded. The Mann-Whitney U-test was used to assess the differences of the relative values between treatment/control groups. Experiments using support vector machine in a leave-one-out approach were performed to validate the use of relative AUCFSD, MEFSD and Ktrans as biomarkers in treatment/control group classification.

Results: At Day2, the relative AUCFSD (p=0.012) and MEFSD (p=0.036) in the treatment group were significantly larger than those in the control group, while no significantly differences were observed for the relative Ktrans (p=0.916) and V (p=0.093). At Day9, significant differences between two groups were observed for the relative AUCFSD (p=0.012), MEFSD (p=0.009) and V (p=0.001) but not for relative trans (p=0.462). The classification accuracy using two post-treatment relative values of AUCFSD, MEFSD and Ktrans was 87.5%, 87.5% and 31.3%, respectively.

Conclusion:The proposed dynamic FSD may serve as a reliable DCE-MRI analysis method and is promising for the early assessment of therapeutic response.

Contact Email: