Dynamic contrast enhanced computed tomography of kidney physiology in small animals

Ganapathy Krishnamurthi, Purdue University

Abstract

Goals. The dissertation explores the feasibility of Dynamic Contrast Enhanced Computed Tomography (DCE-CT) imaging of small animals. The objectives are to (a) study the reproducibility and sensitivity of DCE-CT renal physiologic parameter estimates, (b) validate a renal physiologic parameter, Glomerular Filtration Rate (GFR) against standard clearance techniques of Creatinine clearance and Fluoroscein Iso-ThioCyanate Conjugated Inulin (FITC-Inulin) clearance, (c) investigate renal physiology using DCE-CT along with compartmental models in the pcy mouse model of the Polycystic Kidney Disease (PKD). Experiments. (a) A group of n = 5 mice was used to study the reproducibility of physiologic parameters under various image temporal sampling conditions by varying the contrast injection rates. Three injection rates of 0.5ml/min,1ml/min and 2ml/min were used. A group of n= 3 mice was used to study the impact of X-ray tube current i.e. mAs on the reproducibility of the physiologic parameters. DCE-CT was performed at 4 different tube currents, 80, 60, 25 and 10mAs. The10mAs tube current condition was created by enclosing the mice inside an aluminum cylinder. The cylinder reduces the X-ray photon flux and the image noise equivalent to that of 10 mAs was obtained as verified using a water phantom, (b) The tissue physiologic parameters from the kidney cortex were extracted in a group n = 8 mice, where 3 of them were control CD1 and 5 were late stage pcy mice. The GFR estimated using three different compartmental models of the kidney were compared to FITC-Inulin and Creatinine clearance measurements. The models were used to estimate the Renal Blood Flow (RBF), GFR and fractional vascular volumes in the kidney cortex. (c) DCE-CT was performed on a group of n= 27 mice consisting of control CD1(n = 11), pcy early stage (n = 4) and pcy late stage (n = 12) to test the sensitivity of DCE-CT to changes in organ function decline in terms of Renal Blood Flow (RBF) and GFR. Results: The reproducibility studies under different injection rates showed that the variation of the parameters was within the 20% limit required for following in-vivo physiological changes in a disease or tumor model of small animals. The variation of the physiological parameters as a function of image noise determined by the X-ray tube currents show that the parameter estimates are unstable below the 25mAs tube current. The results are significant because it can be used to reduce the radiation dose to the small animals in the case of longitudinal studies requiring frequent physiological monitoring. The preliminary validation studies for GFR indicated a good correlation (0.91) between the Creatinine clearance and the GFR estimates from the compartmental models. The correlation between the FITC-Inulin clearance and the functional CT (fCT) GFR estimates were poor but both results showed significant difference between the control CD1 mice and the late stage pcy mice where the GFR estimates for the late stage pcy mice were significantly lower. The experiments to monitor kidney physiology in different groups of mice show a statistically significant difference (P < 0.02) between the late stage pcy and control mice while no statistically significant (P = 0.3) difference was observed between the early stage pcy and control mice, in terms of the Flow, GFR and fractional vascular volumes. The results were consistent for all compartmental models. The goodness of fit criterion defined as the residual sum of squares (RSS) were lower for the 2CM models because of the better fits to the renal cortex tissue curves. The Akaike (AIC) and Schwarz (BIC) criterion were also lower for the two compartmental models because the lower values of the goodness of fit described offset the penalty imposed by the AIC and BIC for the increased number of degrees of freedom. Conclusions. DCE-CT can be used for physiologic imaging of small animals in a pre-clinical setting. The pcy mouse model show a decline in renal function in the later stages of the disease and retains kidney function in the early stages despite changes to morphology i.e. there was no statistically significant difference between the control CD1 and the early stage pcy mice. The various compartmental models gave consistent results. The results show that DCE-CT combined with compartmental modeling can be an effective non-invasive tool to monitor functional changes in the kidney and correlate it to disease progression. The use of contrast agent means that a high resolution image of the structure of the kidney can also be obtained. Thus DCE-CT can be used to correlate structural changes to functional changes and vice versa. The models and methods can be transferred to the human studies with sufficient modifications to imaging protocols to reduce dose.

Degree

Ph.D.

Advisors

Liang, Purdue University.

Subject Area

Medical imaging|Physiology

COinS