Eric Sigmund PhD, Presenter: Nothing to Disclose

Jason Ostenson, Abstract Co-Author: Nothing to Disclose

Linda Moy MD, Abstract Co-Author: Nothing to Disclose

Sungheon Kim, Abstract Co-Author: Nothing to Disclose

Gene Cho, Abstract Co-Author: Officer, RF Test Labs Shareholder, RF Test Labs

Komal Jhaveri MD, Abstract Co-Author: Nothing to Disclose

Christopher Glielmi PhD, Abstract Co-Author: Employee, Siemens AG

Thorsten Feiweier DIPLPHYS, PhD, Abstract Co-Author: Employee, Siemens AG Stockholder, Siemens AG Patent holder, Siemens AG

Kimberly Jackson, Abstract Co-Author: Nothing to Disclose

Amy Noel Melsaether MD, Abstract Co-Author: Nothing to Disclose

Diffusion-weighted MRI (DWI) and FDG-PET have strong track records in the setting of breast cancer, but their intratumoral interrelationship is not completely understood. In a simultaneous MR/PET system, we derive intra-voxel incoherent motion (IVIM) diffusion and perfusion metrics and compare them with spatially correlated FDG SUV values in breast cancer patients.

In this IRB approved, HIPAA compliant study, we imaged 17 women with histologically proven invasive cancers on a 3T Siemens Biograph mMR MR/PET system immediately following the patients’ whole body PET/CT without additional FDG injection. Concurrent with FDG-PET, we acquired DWI using a prototype fat-suppressed, twice-refocused spin echo EPI sequence with eddy-current distortion correction and 10 b-values (0 to 800 s/mm2), followed by Gd-DTPA contrast agent and T1-weighted imaging. DWI and PET volumetric images were independently registered to the post-contrast T1 and resampled to its matrix and resolution. DWI were fit voxelwise to a bi-exponential IVIM model to derive tissue diffusion (Dt) and perfusion fraction (fp). Pearson correlation coefficients for SUV vs. Dt and SUV vs. fp were derived for each lesion. SUV and diffusion/perfusion relationships were visualized using correlated 2D histogram analysis.

The relationships between SUV vs. Dt/fp were individually distinct, but most fell into two categories: Type A (N=6, SUV vs. Dt negative correlation, SUV vs. fp positive or no correlation) and Type B (N=6, SUV vs. Dt showing a Dt limit, and SUV vs. fp a negative correlation). The remaining 5 cases were too small or heterogeneous to be classified. The Pearson’s r values reflect these visual correlations. Type B cases were more likely than Type A’s to be metastatic (80% vs. 50%) and Type B’s had a higher Ki-67 expression (55% vs. 36%).

Co-registration of DWI and PET imaging in breast cancer has revealed diverse relationships between diffusion, perfusion, and glucose uptake on an intra-lesion level which can be visualized using a correlated 2D histogram and each lesion broadly typed. Further correlation with clinical data may reveal more connections between 2D histograms and prognostic factors or clinical endpoints.

The spatial correlation of tumor metabolism with microstructural markers via PET and DWI betters understanding of lesion heterogeneity and how it relates to disease aggression and prognosis.

Sigmund, E, Ostenson, J, Moy, L, Kim, S, Cho, G, Jhaveri, K, Glielmi, C, Feiweier, T, Jackson, K, Melsaether, A, Voxel-Wise Correlations of Glucose Uptake and Intra-voxel Incoherent Motion Parameters from Simultaneous MR/PET in Locally Advanced Breast Cancer.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14013442.html