Adrian Andreou¹, Matthew Orton², David J Collins^1,2, Martin O Leach², and Dow-Mu Koh^1
1Department of Radiology, Royal Marsden Hospital, Sutton, Surrey, United Kingdom, ²CRUK-EPSRC Cancer Imaging Centre, Institute of Cancer Research, Sutton, Surrey, United Kingdom

Applying the principles of intravoxel incoherent motion (IVIM) allows quantitative parameters which reflect tissue microcapillary perfusion and tissue diffusivity to be derived. We evaluated the short term measurement reproducibility of the perfusion fraction (f), pseudo-diffusion (D*) and diffusion coefficients (D) of normal appearing liver and colorectal liver metastases in patients. We found that D showed good to moderate measurement reproducibility in liver and metastases. However, the reproducibility of f and D* was relatively poor but using a Bayesian method for analysis significantly improved the reproducibility of f and D* in metastases and D* in the liver.

April M. Chow^1,2, Darwin S. Gao^1,3, Shu Juan Fan^1,3, Gladys G. Lo⁴, Siu Ki Yu², and Ed X. Wu^1,3
1Laboratory of Biomedical Imaging and Signal Processing, The University of Hong Kong, Pokfulam, Hong Kong SAR, China, People's Republic of, ²Medical Physics & Research Department, Hong Kong Sanatorium & Hospital, Happy Valley, Hong Kong SAR, China, People's Republic of, ³Department of Electrical and Electronic Engineering, The University of Hong Kong, Pokfulam, Hong Kong SAR, China, People's Republic of, ⁴Department of Diagnostic and Interventional Radiology, Hong Kong Sanatorium & Hospital, Happy Valley, Hong Kong SAR, China, People's Republic of

Early diagnosis of liver fibrosis could facilitate early interventions and treatments, preventing its progression to cirrhosis. Recently, diffusing imaging has been shown to be promising in characterization of liver fibrosis. In this study, using intravoxel incoherent motion (IVIM) analysis and a mouse model of CCl₄-induced fibrosis, we showed that both molecular water diffusion and blood microcirculation contribute to the alteration in apparent diffusion changes in liver fibrosis. Among all parameters studied, blood pseudodiffusion coefficient (D*) showed largest percentage change at 2 weeks after CCl₄ insult. IVIM analysis can be valuable in characterizing liver fibrosis at early phase and monitoring its progression.

Andrew B Rosenkrantz¹, Eric E Sigmund¹, Benjamin E Niver¹, Glyn R Morgan², Bradley Spieler¹, Anthony G Gilet¹, and Cristina H Hajdu^3
1Radiology, NYU Langone Medical Center, New York, New York, United States, ²Transplant Surgery, NYU Langone Medical Center, New York, New York, United States,³Pathology, NYU Langone Medical Center, New York, NY, United States

Eight fresh liver explants in patients with hepatocellular carcinoma (HCC) underwent ex-vivo 1.5T MRI, including diffusion kurtosis imaging (DKI). ADC using mono-exponential (ADC) and non-Gaussian (D) fit, as well as mean kurtosis (K), were calculated. There was significant correlation between ADC and D, but not between ADC or D with K. Relative contrast between liver and HCC was greater for K than for ADC or D. ADC and D were lowest in non-necrotic HCC, while K was not different between viable and necrotic HCC. Alternate diffusion protocols may have potential for DKI of HCC following chemo-embolization in a clinical setting.

Miriam Klauss¹, Matthias Gaida², Andreas Lemke³, Katharina Grünberg³, Dirk Simon³, Stefan Delorme³, Hans-Ulrich Kauczor², Lars Grenacher², and Bram Stieltjes^3
1Diagnostic Radiology, University of Heidelberg, Heidelberg, Baden-Württemberg, Germany, ²University of Heidelberg, ³DKFZ

Fibrosis is the most characteristic histopathological feature of pancreatic carcinoma and chronic pancreatitis. It was shown that the IVIM-derived parameter D (true diffusion constant) did not show differences between these two entities. The aim of this study was to compare differences in D with the histopathological grade of fibrosis in pancreatic diseases. We compared the histopathological grade of fibrosis with the true diffusion constant D calculated from diffusion data in 24 patients. We found significant differences between the D-values for moderate and severe fibrosis. The pure diffusion constant could serve as parameter to monitor tumor response in pancreatic carcinoma.

Thais C Mussi¹, Bradley Spieler¹, and Andrew B Rosenkrantz^1
1Radiology, NYU Langone Medical Center, New York, New York, United States

19 patients with urothelial carcinoma of the bladder who underwent MRI including DWI at 1.5T were included. Among these, 8 had nodal or distant metastatic disease, and 11 had an absence of metastatic disease (mean follow-up 434 days). ADC was measured for each lesion and was significantly lower among those cases with metastatic disease (p=0.0012). Using ADC for prediction of metastatic disease, ROC analysis identified an AUC of 0.920, sensitivity of 87.5%, specificity of 100%, PPV of 100%, and NPV of 91.7%.

Paul Summers¹, Antonio Gregoraci², Giuseppe Petralia¹, Anna Caroli³, Roberto Di Filippi¹, Luca Antiga³, and Massimo Bellomi^1
1Department of Radiology, European Institute of Oncology, Milan, Italy, ²Faculty of Medicine and Surgery, University of Milan, Milan, Italy, ³Mario Negri Institute, Bergamo, Italy

We examined the impact of prandial status, choice of b-values in one or three directions, and ROI definition on repeatability of whole liver ADC measurements in 10 healthy subjects. The inclusion of the b=0 image or limiting the b-values used to those greater than 700 s/mm2 significantly increased both the estimated mean ADC and its variability between subjects. Excluding these choices of b-values, intra- and intersubject standard deviations were similar (< 10%, and 15-22% respectively) for both 1 and 3 encoding directions. Prandial status and exclusion of visible vessels from the ROI had little effect on ADC.

Olaf Dietrich¹, Andreas Biffar¹, Andrea Baur-Melnyk², Gerwin Schmidt², and Maximilian F. Reiser^1,2
1Josef Lissner Laboratory for Biomedical Imaging, Department of Clinical Radiology, Ludwig Maximilian University of Munich, Munich, Germany, ²Department of Clinical Radiology, Ludwig Maximilian University of Munich, Munich, Germany

Diffusion-weighted steady-state free-precession (DW-SSFP) sequences are extremely valuable for the differential diagnosis of benign osteoporotic and malignant neoplastic vertebral compression fractures. The purpose of the present study was to provide a detailed theoretical and experimental analysis of the DW-SSFP signal in the spine. It could be demonstrated that the main reasons for the different lesion contrasts are an opposed-phase readout in combination with the differing fat fractions and T₂^*-values in the fractures as well as in normal-appearing VBM of both entities. The observed signal contrast is therefore rather fat- and T₂^*-weighted than diffusion-weighted.

G. M. Kukuk¹, M. Kaschner¹, F. Träber¹, D. Skowasch², J. Gieseke^1,3, H. H. Schild¹, W. A. Willinek¹, and P. Mürtz^1
1Department of Radiology, University of Bonn, Bonn, NRW, Germany, ²Department of Pneumology, University of Bonn, Bonn, Germany, ³Philips Healthcare, Best, Netherlands

High field MRI has introduced new challenges for DWIBS body imaging with respect to susceptibility artefacts, fat suppression, dielectric shading, and longer acquisition times. Dual-source parallel RF excitation has been shown to allow for a more homogeneous excitation. Therefore, we evaluated the effect of this technique on DWIBS in 40 patients. Shorter acquisition times by 33%, significantly improved signal homogeneity and fat suppression, and reduced dielectric shading without influencing ADC values was found by using parallel RF excitation in DWIBS at 3.0 T.

Lan Lu¹, Bernadette Erowku¹, Gregory R Lee¹, Katherine Dell^2,3, and Chris A Flask^1,4
1Radiology, Case Western Reserve University, Cleveland, OH, United States, ²Pediatrics, Case Western Reserve University, Cleveland, OH, United States, ³MetroHealth Medical Center, Cleveland, OH, United States, ⁴Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States

Diffusion MRI applications typically utilize rapid Echo-Planar Imaging (EPI) acquisition techniques. However, EPI techniques are susceptible to off-resonance distortion and eddy current artifacts, especially for high field MRI applications. We have developed a rapid Diffusion Prepared - Fast Imaging with Steady-State Free Precession (DP-FISP) MRI acquisition for a 7T small animal MRI scanner that combines a diffusion preparation scheme with a rapid FISP imaging readout to provide diffusion-weighted images in less than 500 ms and DTI data (7 directions) in less than 1 minute. Herein, DP-FISP results validated in a water phantom and in vivo polycystic rat kidneys.