Ina Nora Kompan^1,2, Benjamin Richard Knowles³, and Matthias Guenther^1,2
1Fraunhofer MEVIS, Bremen, Bremen, Germany, ²mediri GmbH, Heidelberg, Baden-Württemberg, Germany, ³Universitätsklinikum Freiburg, Freiburg, Baden-Württemberg, Germany

Pharmacokinetic modeling is used in dynamic contrast-enhanced MRI to quantify tissue physiology. High temporal and spatial resolutions are both needed, but are not compatible. In this work, a resolution adaption sequence is presented which in real-time adapts temporal resolutions to the measured contrast-enhanced signal changes. The sequence is validated using a perfusion phantom and is compared to an only low spatial and only high spatial resolution sequence. For the phantom, the adaptive sequence provides comparable fitting accuracy to the low temporal resolution sequence and high spatial resolution images at the signal peak.

James H Holmes¹, Kang Wang¹, Courtney K Morrison², Frank R Korosec³, Ersin Bayram⁴, Roberta M Strigel³, Diego Hernando³, Scott B Reeder³, Edward F Jackson², and Ryan J Bosca^2
1Global MR Applications and Workflow, GE Healthcare, Madison, WI, United States, ²Medical Physics, University of Wisconsin-Madison, WI, United States,³Radiology, University of Wisconsin-Madison, WI, United States, ⁴Global MR Applications and Workflow, GE Healthcare, Houston, WI, United States

Robust fat suppression in breast dynamic contrast enhanced (DCE)-MRI has been primarily viewed as a critical need for qualitative evaluation. However, current pharmacokinetic models used to quantitate high-resolution acquisitions fail to account for the complicated signals in voxels arising from voxels containing fatty tissue and enhancing water compartments. Such combinations of fat and water are found commonly in, for example, non-mass enhancing lesions of the breast. In this work we demonstrate the influences of fat signal on DCE parameters and evaluate fat suppression and separation methods including a 2 point Dixon method for high-spatiotemporal quantitative DCE acquisitions.

Jin Zhang^1,2, Kerryanne Winters^1,2, and Sungheon Gene Kim^1,2
1Center for Advanced Imaging Innovation and Research (CAI2R), Dept. Radiology, NYU School of Medicine, New York, NY, United States, ²Bernard and Irene Schwartz Center for Biomedical Imaging, Dept. Radiology, NYU School of Medicine, New York, NY, United States

In our previous studies, we introduced Active-Contrast Encoding MRI (ACE-MRI) which measures both B₁ and T₁ values along with kinetic parameters from a single DCE-MRI data. The model free approach of ACE-MRI we proposed separates estimation of B₁/T₁ from estimation of contrast kinetic parameters, and consequently improves parameter estimation accuracy and precision. This in vivo cross-validation study was conducted to compare the contrast kinetic parameters estimated from ACE-MRI data with those estimated from conventional DCE-MRI experiments with separate measurements of B₁ and T₁.

Frank FJ Simonis¹, Alessandro Sbrizzi², Ellis Beld¹, Jan JW Lagendijk¹, and Cornelis AT van den Berg^1
1Radiotherapy, UMC Utrecht, Utrecht, Utrecht, Netherlands, ²Radiology, UMC Utrecht, Utrecht, Netherlands

Acquiring an accurate arterial input function (AIF) is essential in dynamic contrast-enhanced (DCE) MRI analysis. Determining AIFs using MR magnitude data faces challenges due to experimental difficulties such as inflow, B1 non-uniformity and saturation effects. However phase-based AIFs have problems estimating the baseline and the tail of the AIF. Here we demonstrate that fitting the enhancement data in the complex plane can be used in DCE AIF estimation to mitigate noise and bias that arise from solely using phase or magnitude data. The technique is applied to 3T DCE-MRI data of prostate cancer patients

Jen Moroz¹, Andrew Yung¹, Piotr Kozlowski^2,3, and Stefan Reinsberg^1
1Physics and Astronomy, UBC, Vancouver, BC, Canada, ²Radiology, UBC, Vancouver, BC, Canada, ³MRI Research Centre, UBC, Vancouver, BC, Canada

This study investigates the potential for interleaving a DCE-MRI experiment with the acquisition of a high temporal resolution AIF using a radial projection-based approach. A mouse tail phantom was scanned with two slice packs: one for the AIF (radial, single slice) and one for the DCE experiment (Cartesian, multi-slice). By interleaving these acquisitions, we maintain a high temporal resolution for the AIF, and sufficient spatial and temporal resolutions for the DCE experiment. This technique is expected to improve the accuracy of fitted model parameters as the AIF is specific to the individual study.

Ashley M Stokes¹ and C. Chad Quarles^1
1Institute of Imaging Science, Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States

Contrast agent (CA) leakage in tumors leads to inaccurate spin echo (SE) cerebral blood volume (CBV) and mean vessel diameter (mVD) measures in DSC-MRI. Here, we propose a new strategy to correct for T₂ leakage effects on T₁-insensitive ΔR₂ timecourses. This simple and efficient approach was validated in rat tumor models by comparing uncorrected and corrected SE CBV and mVD against reference values obtained with an intravascular CA. Correction of both T₁ and T₂ leakage effects improved the reliability of SE CBV and mVD. This highlights the need for comprehensive correction techniques to more accurately represent non-leaky tumor hemodynamics.

Sajan Goud Lingala¹, Yi Guo¹, Yinghua Zhu¹, Samuel Barnes², R. Marc Lebel³, and Krishna S Nayak^1
1Electrical Engineering, University of Southern California, Los Angeles, California, United States, ²Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, California, United States, ³GE Healthcare, Calgary, Canada

DCE-MRI of the brain is a powerful technique to assess the blood brain barrier permeability, and other neuro-vascular parameters. Current clinical DCE-MRI protocols have restricted spatial resolution, and slice coverage due to the slow MRI encoding process. In this work, we propose a novel pharmaco-kinetic dictionary approach to constrain the recovery of concentration profiles from under-sampled DCE -MRI acquisitions. Using the Patlak pharmaco-kinetic model, we construct a dictionary of temporal bases that characterize all possible time-intensity curves. The dictionary bases are extremely tolerant of noise and incoherent under-sampling artifacts as these are poorly described in the dictionary. Our approach enabled faithful reconstruction of upto reduction factor of 20 fold. We demonstrate superior fidelity in recovering Pharmaco-kinetic parameter maps in comparison to image reconstructions that are based on constraints blind to pharmaco-kinetic modeling (such as the temporal total variation constraint).

Ruogu Fang^1
1School of Computing and Information Sciences, Florida International University, Miami, FL, United States

A 4-D Tensor Total Variation (TTV) deconvolution approach is adapted and evaluated for dynamic susceptibility contrast magnetic resonance imaging (DSC-MRI) in the brain. TTV exploits the temporal continuity of the contrast agent concentration and the spatial correlation of the non-random structure of the microvasculature. The algorithm is guaranteed to converge to global optima and outperforms the baseline methods (singular value decomposition-based and Tikhonov regularization) on both synthetic and real data of subjects with glioblastomas, in terms of improving accuracy of residue functions, quantification of perfusion maps, and localization of tumor tissue.

Hatef Mehrabian^1,2, Anne L Martel^1,2, Johann Le Floc'h¹, Hany Soliman^1,3, Arjun Sahgal^1,4, and Greg J Stanisz^1,2
1Physical Sciences, Sunnybrook Research Institute, Toronto, Ontario, Canada, ²Medical Biophysics, University of Toronto, Toronto, Ontario, Canada, ³Odette Cancer Centre, Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada, ⁴Radiation Oncology, Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada

Tumor response to therapy could be assessed using imaging biomarker derived from DCE-MRI. We have previously developed an independent component analysis (ICA)-based algorithm to split DCE-MRI data into its intravascular and extravascular compartments. The objective of current study is to employ these intravascular/extravascular signals in a two-compartment relaxation model of MRI signal and calculate the water exchange rates between these two compartments, as well as the compartment sizes. Our results show that the model can robustly estimate exchange parameters, and that these parameters are more sensitive to changes in the tumor due to radiotherapy compared to conventional pharmacokinetic modeling.

Zi-Min Lei¹, Cheng-He Li¹, Sheng-Min Huang¹, Chin-Tien Lu¹, Kung-Chu Ho², and Fu-Nien Wang^1
1Biomedical Engineering and Environmental Sciences, National Tsing Hua University, HsinChu, Taiwan, ²Nuclear Medicine, Chang Gung Memorial Hospital, Taoyuan, Taiwan

In this study of rodent brain imaging, using a new strategy of monitoring heavy water tracer by the attenuation of 1H signal, we further re-investigate the heavy water dynamics with a multi-compartmental analysis method, and investigated the spatial distribution of fast and slow compartments of heavy water perfusion.