Download presentation
Presentation is loading. Please wait.
Published byTrevion Reilly Modified over 9 years ago
1
ISMRM Hawaii 2009 slide 1 P.S.Tofts 1,2, M. Cutajar 1,2, I.Gordon 2 1 = Brighton and Sussex Medical School, UK 2 = Institute of Neurology, University College London, UK 3 = Institute of Child Health, University College London, UK Estimating GFR from early (uptake) portion of DCE MRI renal data, using a 3-compartment model, improves reproducibility and may eliminate need for cortical segmentation P.S.Tofts 1,2, M. Cutajar 1,2, I.A.Mendichovszky 2, I.Gordon 2 1 = Brighton and Sussex Medical School, UK 2 = Institute of Neurology, University College London, UK 3 = Institute of Child Health, University College London, UK Declaration of Conflict of Interest or Relationship Speaker Name: Paul Tofts I have no conflicts of interest to disclose with regard to the subject matter of this presentation.
2
2ISMRM Hawaii 2009 Introduction 2008 fit full curve with 3 compartment model (6 free parameters) spreadsheet; fit curve in 3s reproducibility in normals CV=20% innovations 2009: 1. fit uptake phase only (forbid efflux) 2: use over-inclusive ROI to get total GFR in a slice 3: fit perfusion phase only to improve vascular parameters
3
3ISMRM Hawaii 2009 MRI Acquisition Siemens Avanto 1.5 T scanner Abdominal TIM coil Gradient-echo 3D-FLASH pulse-sequence TR =1.63 ms; TE = 0.53 ms Flip angle = 17 ° Strong fat saturation; PAT factor = 2 (GRAPPA) FOV = 400 x 325 mm 2 18 x 7.5 mm slices covering entire kidney (no gap) Voxel size = 3.1 × 3.1 × 7.5 mm 3 Frames acquired every 2.5 s Gd dose = 0.05 mmole/kg Cortical ROITime of peak Arterial ROIArterial Input Function AIF
4
glomerular plasma g(t) = Glomerular Input Response Function tubular tracer C d nb can forbid efflux F 2 =0 total tracer C t Kidney model K trans = transfer constant from plasma to kidney (=GFR per unit volume of tissue)
5
5ISMRM Hawaii 2009 Fit cortical ROI with efflux turned off blood and cortical signals blood and cortical signals Fit up to 60s Fit up to 60s ‘uptake phase’ ‘uptake phase’ residuals residuals efflux at 56s efflux at 56s model separates Gd: model separates Gd: IV glomerular IV glomerular EV tubular EV tubular For GFR For GFR Tubular Gd is of interest Tubular Gd is of interest Glomerular Gd is incidental Glomerular Gd is incidental Use parenchymal ROI to 90s Use parenchymal ROI to 90s
6
6ISMRM Hawaii 2009 Normal values 10 subjects; repeated measurements 10 subjects; repeated measurements parenchymal ROI to 90s parenchymal ROI to 90s Bland-Altmann-analysis gives sd of single measurement Bland-Altmann-analysis gives sd of single measurement In 2008 (#454), allowing efflux and fitting data for 240s, we reported: In 2008 (#454), allowing efflux and fitting data for 240s, we reported: larger values of K trans (0.48 min-1) probably because of model degeneracy during efflux phase larger values of K trans (0.48 min-1) probably because of model degeneracy during efflux phase Larger values of single-measurement sd (21%) (more degrees of freedom in model?) Larger values of single-measurement sd (21%) (more degrees of freedom in model?) meansd Single measurement sd K trans GFR/unit vol tissue 0.25 min -1 0.04 min -1 0.03 min -1 (14%)* VbVbVbVb fractional blood volume 0.42%0.10% 0.07% (15%)* art-del arterial delay 2.7 s 1.05 s 0.5s (18%) * * 2009 abstract is wrong (1.4 times larger)
7
7ISMRM Hawaii 2009 Total GFR in a slice K trans = GFR/unit volume of tissue K trans = GFR/unit volume of tissue an intensive quantity (like blood flow or density) an intensive quantity (like blood flow or density) K trans x volume of ROI = GFR in that ROI K trans x volume of ROI = GFR in that ROI Defining cortical ROI is boring and subjective Defining cortical ROI is boring and subjective Could we measure total uptake? Could we measure total uptake? Removal of Gd from blood, i.e. total GFR Removal of Gd from blood, i.e. total GFR without spatial information without spatial information Use over-inclusive ROI (s); GFR _ROI should plateau when ROI > kidney Use over-inclusive ROI (s); GFR _ROI should plateau when ROI > kidney Kidney GFR by adding each slice Kidney GFR by adding each slice Precedent: Object Strength in presence of strong partial volume effect. Precedent: Object Strength in presence of strong partial volume effect. Tofts Magma 2005:18:162 Tofts Magma 2005:18:162 7 ROI’s (cortex ~ 280; parenchyma 500) ‘cortical’ ROI missing some uptake K trans values decrease with increasing ROI size
8
8ISMRM Hawaii 2009 Estimating vascular parameters use vascular phase only! blood volume v b, arterial delay, blood flow? blood volume v b, arterial delay, blood flow? Use vascular phase (first pass) ~20 s Use vascular phase (first pass) ~20 s Cortical ROI Cortical ROI Minimal contribution from tubular uptake Minimal contribution from tubular uptake Might improve estimation Might improve estimation Accuracy, reproducibility Accuracy, reproducibility
9
9ISMRM Hawaii 2009 Estimation of Blood Flow F from cortical Glomerular Impulse Response Function work in progress Gaussian girf Fits better than exponential peak = 0.36 s -1 ; vb=0.36 falling exponential girf (interpolated to 0.4s) peak = 0.39 s -1 ; v b =0.42 Blood flow = (blood volume) * peak F = v b g(0) F = 982 ml blood (100ml tissue) -1 min -1 normal value ~ 400 (whole kidney) vbvb FF Cortical ROI Ensemble of compartments with different arrival times Central-limit theorem: gaussian convolution of exponentials, reduced peak height Smaller ROI’s? arterial delay (~3s) Could ? give whole kidney blood flow (with renal artery blood volume)
10
10ISMRM Hawaii 2009 summary Fitting to different time- portions and ROI’s Fitting to different time- portions and ROI’s Uptake phase for K trans (=GFR/ml tissue) Uptake phase for K trans (=GFR/ml tissue) Cortical perfusion phase for blood volume Cortical perfusion phase for blood volume Glomerular Impulse Response Function may give blood flow Glomerular Impulse Response Function may give blood flow Total GFR is accessible with over-inclusive ROIs Total GFR is accessible with over-inclusive ROIs e-poster #4135 Characteristics and reproducibility of the arterial input function (AIF) derived from dynamic contrast- enhanced magnetic resonance imaging (DCE-MRI) studies and its effect on renal functional parameters. M. Cutajar, P. S. Tofts, I. A. Mendichovszky, and I. Gordon.
Similar presentations
© 2024 SlidePlayer.com Inc.
All rights reserved.