Response to "Erroneous data on renal hemodynamics derived from DCE-MRI in rats".

Abstract

We thank Dr. Seeliger and colleagues for their interest in our work and for their thoughtful comments. As mentioned in our article and established studies, application of the tracer-kinetic model to dynamic contrast-enhanced magnetic resonance imaging ((DCE-MRI) renography involves several considerations regarding image acquisition and analysis,1, 2 eg, the injection speed, tracer dose, cardiac function, dispersion, blood recirculation, choice of kinetic model, and the location of regions of interest (ROIs), all of which will influence the shape of the renographic curve and thus influence the quantitation of renal function. Let us note that the denoted glomerular filtration rate (GFR) in this study, ie, the parameter K in Eq. [1],3 is actually an apparent rate constant describing the transport of tracer from the cortex to the outer medulla, which can be considered a good estimate of the renal clearance as indicated by dominant linear correlation with the absolute GFR. Because there is no standard that can be referred to, we compare our GFR derived by modified Toft with an established Kcl reported by Baumann and Rudin4 and Laurent et al,5 who used a cortex-to-medulla model (BR model). The Kcl is reported with a range of 1.09–3.4 min−1 for rat kidney, which also shows considerable variabilities with the similar analysis and animals. After normalized by Hct (0.42) and rat cortical kidney weight (0.8 g measured in our study), the normalized Kcl ranges from 0.79–2.45 mL/min/g in cortical kidney. From Annet et al6 and our previous report,7 the normalized GFR of rabbit measured by a cortical-compartment model ranges from 0.24–0.32 mL/min/g in cortical kidney. Therefore, we believe that our measurements are partially comparable to these established values. From our experience, the modified 2C model is somewhat better than other models, ie, BR, Toft, Patlak, uptake, or filtration model, regarding accuracy and reproducibility. And many factors, as we mentioned above, could make a great influence on the results of model-dependent analysis. The reproducibility needs to be comprehensively validated in further study. After all, because each animal acted as its own control in this study, the measure variability can be avoided maximally. Thus, we believe our observations can partially explain the mechanisms of acute kidney damage induced by the compartment models with different physicochemical properties. These findings now receive more support in the nephrotic literature.8, 9 Indeed, we agree with Dr. Seeliger and colleagues that the renal perfusion by our measurements are hard to reconcile with the established values, and seems to be significantly underestimated. This can be explained as follows: first, in order to satisfy the theory of compartment model that the blood tracer concentrations must be equal at all data sampling regions, in this study a very slow injection rate (1 mL/sec) and a smaller amount (0.025 mmol/kg, about 0.5 ml/rat) of contrast agents was used to make the difference as small as possible. The slow injection and low-dose agents may influence the sharp of arterial input function and thus influence the deconvolution analysis. Second, the renal perfusion measured by deconvolution algorithms can be greatly affected by the choice of time window. Let us note that the blood perfusion can be underestimated, while the blood volume can be overestimated by increasing the time window (Fig. 1). In our study, we chose the 4:30-minute time window for the analysis of renal perfusion and GFR. This time window is advisable for a 2C-based Toft model, while it is too long for the one compartment-based deconvolution algorithm. Underestimation of renal perfusion may be associated with blood recirculation. The influence of time window on renal perfusion based on deconvolution algorithms. The renal blood flow is underestimated (a,b), while the renal blood volume is overestimated (c) with the increased time windows. In the next work, we would like to write a review article, in particular for responding to your comments regarding the pitfalls, optimization, and application of DCE-MRI for renal functional imaging. Contract grant sponsor: National Natural Science Foundation of China; Contract grant sponsor: NSFC; contract grant number: 81301191; Contract grant sponsor: Priority Academic Program Development of Jiangsu Higher Education Institutions; contract grant number: PAPD; JX10231801