|Renal Parenchymal & Perfusion MR Imaging|
A Simple Two-Compartment Model That Describes Dynamic
Contrast-Enhanced MRI Signal in the Kidney
Paul S. Tofts1, Iosif A. Mendichovszky2, David L. Buckley3, Kenneth A. Miles1, A Michael Peters1, Isky Gordon2
1Brighton and Sussex Medical School, Brighton, UK; 2University College, London, UK; 3Manchester Medical School, UK
A two-compartment model, based on widely-accepted work in tumours, is able to fit the complexity of DCE renal data with 2.5s temporal resolution. The bolus delay and dispersion from aorta to kidney are accounted for in the model. The estimated parameter K-trans is the GFR per unit volume of kidney. In 10 normals controls, the mean value was 0.48 min-1 (sd=0.09), close to established values. Repeated measurements showed differences of 30% (1sd). The model was implemented on a simple spreadsheet. The model overcomes problems with the established Patlak approach. Significant Gd efflux from parenchymal voxels was found even after 30s.
|DCE MRI Based Perfusion Parameters in Renal
Transplants: Influence of the Choice of the Arterial Input Function
Martine Isabella Dujardin1, Steven Sourbron2, Rob Luypaert1, Yves Fierens1, Jacques Sennesael3, Tadeusz Stadnik1, Johan de Mey1
1UZ Brussel, Brussels, Belgium; 2Klinikum Grosshadern, Munich, Germany; 3UZ brussel, Brussels, Belgium
This study tries to estimate the influence of dispersion and partial volume by comparing quantitative DCE-based perfusion results in renal transplant data deconvolved with two AIFs. RVDlob (small proximate lobar artery AIF) values are higher than RVDil (more distant iliac artery AIF), consistent with AIF underestimations due to partial volume effects. In MTT the trend in the data (MTTlob smaller compared to MTTil) can be attributed to bolus dispersion. RBF is influenced by both effects, but the results show that partial volume effects dominate. We conclude that for quantitative DCE-based perfusion parameters in transplants the choice of the AIF in a larger more distant artery is more accurate.
Assessment of Renal Function Using MR
Renography Without Aortic Input Information
Jeff Lei Zhang1, Henry Rusinek1, Qun Chen1, Pippa Storey1, Louisa Bokacheva1, Ting Song1, 2, Vivian S. Lee1
1New York University, New York, New York, USA; 2Columbia University, New York, New York, USA
Most current models for analyzing MR renographic data require an aortic input curve, which suffers from several measurement errors. In this study, we explored a new approach that estimates GFR but obviates input function measurements. Patient study showed that the root mean square errors for fitting medullary curve averaged only 0.016 ▒ 0.006 mM, and there was excellent correlation (R = 0.95) between the proposed parameter ├ and GFR measured by scintigraphy. The Monte Carlo simulation indicated that parameter ├ could be estimated with high precision and accuracy under typical levels of noise.
Noninvasive Monitoring of Acute Tubular Necrosis Induced
by Ischemia/Reperfusion in a Rat Model Using 23Na MRI
Bharath Atthe1, Andriy Babsky1, Navin Bansal1
1Indiana University School of Medicine, Indianapolis, USA
23Na MRI was applied to study the early alteration in renal sodium distribution in the rat kidney with acute tubular necrosis (ATN) using the ischemia-reperfusion injury model. The images showed ~80% higher 23Na MRI SI in the medulla compared to the cortex. The medulla to cortex Na+ gradient decreased progressively after 10, 30 and 50 min of ischemia. No recovery was observed after 1 hour of reperfusion in the 30 and 50 min ischemia groups, suggesting irreversible tubular injury. 23Na MRI should prove useful as a noninvasive technique for diagnose of evolving ATN.
Measurement Precision of Gadolinium Enhanced Magnetic
Resonance Nephro-Urography Using Rapid 3D Imaging and a
Multi-Compartment Kinetic Model
Khalil Nabeel Salman1, Puneet Sharma1, Bobby Kalb1, John Votaw1, Diego R. Martin1
1Emory University, Atlanta, Georgia, USA
It is currently known that an estimation of renal perfusion kinetics can be determined non-invasively using MRI perfusion techniques. Apart from accuracy, the utility of such techniques depend highly on the precision of the acquisition and kinetic modeling steps, to ensure the physiological parameters are comparable and applicable clinically. The purpose of this investigation was to determine precision of kinetic parameters of the kidney in normal subjects, using an accelerated volumetric acquisition technique and semi-automated kinetic modeling steps. From the results, the standard deviations for all measurement variables were low, ensuring measurement reproducibility.
DCE-MRI Demonstrates Immediate Post-Perfusion
Microvascular Hyperpermeability in the Mouse Renal Cortex Following
Ischemia Induced by Renal Artery Clamping
Negar Geula Knowles1, Li-wen Lai1, Suzu Igarashi1, Christine M. Howison1, Y. Howard Lien1, Natarajan Raghunand1
1University of Arizona, Tucson, Arizona , USA
DCE-MRI imaging of the mouse kidney following surgically induced ischemia and reperfusion was performed using an intravascular contrast agent. Enhancement curves centered over the renal cortices in the reperfused and control kidneys demonstrate a period of slow uptake of contrast in the reperfused renal cortex following the initial steep slope of contrast uptake. This pattern, not seen in the control kidney, is likely due to capillary leak with interstitial retention of contrast following reperfusion. T1-weighted contrast-enhanced images demonstrate enlargement of the reperfused kidney and increased corticomedullary differentiation, attributed to increased capillary permeability in the reperfused cortex.
Renal Tracer Kinetics with a Reabsorption
Steven Sourbron1, Henrik J. Michaely2, Stefan O. Schoenberg2, Maximilian F. Reiser, Michael Peller1
1Ludwig-Maximilian-University Munich, Munich, Germany; 2Universitaetsklinikum Mannheim, Mannheim, Germany
A 2- and 3-compartment model are proposed to account for the effect of reabsorption in renal tracer kinetics. Possible applications of the models are assessed analytically, and the methods are evaluated using data from healthy kidneys. Both allow a measurement of the reabsorption fraction, provided additional data are available in the form of a venous outflow curve. Measurements in healthy kidneys produce realistic values for the reabsorption fraction. The 3-compartment model is the most attractive alternative, since it more accurately models the kidney physiology and allows to measure the glomerular filtration rate from a fit to the kidney curve alone.
MR Elastography of the Kidneys: Preliminary Results
Sudhakar Kundapur Venkatesh1, 2, Meng Yin1, Roger C. Grimm1, Phillip J. Rossman1, Jun Chen1, Armando Manduca1, Anthony J. Romano, 13, Richard L. Ehman1
1Mayo Clinic, Rochester, Minnesota, USA; 2NUS, Singapore, Singapore; 3Naval Research Laboratory, Washington, District Of Columbia, USA
A method for applying MR Elastography to quantitatively assess the mechanical stiffness of the human kidney in vivo was developed and tested in a series of volunteers. A large passive pneumatically-activated acoustic driver, placed against the posterior abdominal wall in the supine position was found to provide excellent shear wave illumination of the kidneys. Elastograms were successfully obtained in all three orthogonal planes in 25 seconds of acquisition time for each cross-sectional image. The study results provide a practical protocol for further studies of renal MRE and initial normative values for the stiffness of renal parenchyma in humans in vivo.
Retrospective Respiratory Triggering for 2D Abdominal
Ulrike Attenberger1, Steven Sourbron1, Henrik Michaely2, Mike Notohamiprodjio1, Maximilian Reiser1, Christian Glaser1, Karin Herrmann1
1Ludwig-Maximilians-University, Munich, Germany; 2University of Heidelberg, Mannheim, Germany
Quantification of renal blood flow and glomerular filtration rate is possible with MRI measurements. As acquisition times of several minutes are needed an alternative strategy is required to compensate the effect of respiratory motion. We used a retrospective triggering approach in this study which might be especially promising in critically ill patients.The aim of this study was a first evaluation of this method using patient data.9 consecutive patients (4 women, 5 men; mean age 59) underwent renal perfusion measurements after intravenous injection of 7 ml Gd-BOPTA (Multihance«, Bracco) at 4 ml/s at 3.0T (Magnetom Tim Trio; Siemens Medical Solutions, Erlangen, Germany). Data were acquired with a 2D Saturation-recovery TurboFLASH sequence measuring 5 slices (4 coronal, 1 axial) with a temporal resolution of 0.9s and a pixel size of 2.3 mm (slice thickness 8mm, FA 8 , TI 71ms, TR 177ms, TE 0.93ms, matrix 192x176).A 4-pixel region-of-interest (ROI) was drawn manually in the lumen of the aorta to measure the Arterial Input Function (AIF). For the measurement of the triggering signal, a rectangular triggering ROI was placed at the interface between tissue and air on the axial slice. . In order to test the dependence of the results on the precise choice of the triggering ROI, the analysis was repeated for three widely differing ROIs at the tissue-to-air interface. Our results show that without triggering the image is blurred due to the combined influence of data measured during in- and expiration. After triggering, the contours are sharper and the image has a stronger contrast. The results show that retrospective respiratory triggering is a feasible approach to correct for breathing motion in the quantification of renal functional parameters using a two-compartment model.
Adaptive 2D Centric View Ordering for Fast Abdominal
Ersin Bayram1, Charles R. Michelich1, Anthony T. Vu1, James H. Akao1, Shaorong Chang1, Sanjay M. Joshi1, Reed F. Busse2, Anja C. S. Brau3, Philip J. Beatty3, Lloyd Estkowski1
1GE Healthcare, Waukesha, Wisconsin, USA; 2GE Healthcare, Madison, Wisconsin, USA; 3GE Healthcare, Menlo Park, California , USA
Resolution and coverage in abdominal MRI are largely limited by the breath-hold scan times. 2D parallel imaging can be utilized to address this issue. The potential for breath-hold mismatch between the external calibration scan and the accelerated scan makes auto-calibrated techniques more appealing, at the expense of reduced net acceleration. The scan time penalty becomes even more evident with magnetization-prepared fat suppression approaches in which a segmented view ordering scheme must be employed. Segmentation forces the number of total views along the segmentation direction to be an integer multiple of the segment size and also to remain constant from one kx-ky or kx-kz plane to the next. This restriction forces the acquisition to fall on a separable ky-kz grid as shown in Fig.1A. An adaptive centric view ordering scheme is proposed that supports more efficient sampling patterns including non-separable grids as well as k-space corner removal