Upon completion of this lecture, participants should be able to:

· Identify appropriate applications for perfusion imaging methods;

· Differentiate between different contrast based perfusion imaging methods, including dynamic contrast-enhanced (DCE) and dynamic susceptibility contrast (DSC), and be able to use the properties of each variant to select appropriate application specific methods.

1. Contrast agents and their kinetics

2. Data analysis

3. Applications

· Gadolinium-based contrast agents (GBCA) were originally developed to enhance contrast-to-noise in anatomic images following intravenous administration.

· Their principal mode of action is to increase the relaxation rates (1/T1, 1/T2, 1/T2*) of nearby water molecules thereby modifying local signal intensity.

· Extracellular GBCA (e.g. Gd-DTPA, Gd-DOTA, Gd-HP-DO3A, Gd-BT-DO3A) accumulate in the interstitium of many tissues (e.g. muscle) but do not cross the healthy blood-brain barrier. A number of extracellular agents behave very similarly. The examples given here are simply used for the sake of convenience and are not intended as an endorsement of any agents specifically.

· Functional information may be obtained by treating a GBCA as a tracer and following its kinetics (as opposed to the conventional ‘static’ approach, inject - wait – image once).

· After intravenous administration (normally in the form of a bolus) GBCA transits the venous and arterial system leaking from capillary beds wherever the endothelial cell junctions allow. They are excreted principally via the kidneys (e.g. [1]).

· The choice of imaging method is crucially dependent on what is to be measured and the rate of GBCA delivery and uptake.

· When imaging tissues with leaky capillaries (e.g. tumors), measurements are typically made using rapid T1-weighted spoiled gradient echo sequences. This is known as DCE-MRI.

o GBCA in the plasma has a relatively small influence on the signal except where the vascular volume is large (e.g. in the kidney, heart or certain tumors).

o GBCA in the interstitial space has a larger effect since this often represents a significant fraction of the tissue volume.

· When imaging the brain with an intact blood-brain barrier, measurements are typically made using T2- or T2*-weighted echo-planar sequences. This is known as DSC-MRI.

o GBCA restricted to the blood plasma produces an extensive susceptibility effect resulting in a large signal drop during bolus passage.

· The temporal sampling requirements vary enormously. For perfusion studies a sampling interval of 1 to 3 s is required to measure the rapid transit of the GBCA bolus. For so-called permeability studies the uptake of GBCA may take several minutes and sampling rates can be reduced to tens of seconds or even minutes.

· It is possible to analyze DCE or DSC data visually. When temporal series are viewed as a movie, regions in the images of hyper- or hypo-perfusion may stand out very clearly. However, some form of quantitative analysis is normally undertaken.

· The plotting of signal-time curves is a common way of interpreting the data. These may be compared in different tissue regions or curve shapes may be compared against tabulated standards [2].

· To extract maximum information, tracer kinetic analysis is performed where transport of GBCA can be described using a number of parameters:

o Tissue blood flow (perfusion), F (or plasma perfusion, Fp); capillary permeability-surface area product, PS (a measure of vessel leakiness); plasma volume, vp and interstitial volume, ve.

· When put together in a physiologic model (e.g. Fig. 1) and formulated in mathematical terms the tracer kinetic parameters provide a powerful tool for data analysis.

· One way of describing the tracer kinetics is to calculate the convolution of the arterial input function (AIF, the concentration of GBCA in the plasma of a feeding artery) with the tissue impulse residue function (IRF), the theoretical response in the tissue to an infinitely tight unit bolus [4,5].

$$Ct(t) = Fp.AIF*R(t)$$

where Ct(t) is the tissue concentration of GBCA, Fp is plasma perfusion, R(t) is the IRF and * is the convolution operator.

· The shape of the IRF can be determined in two ways:

o No assumption about R(t). Numerical deconvolution may be performed using Ct and the AIF and this results in an estimate of R(t) directly.

o Model assumed. A mathematical description of R(t) may be derived and then data fitting (usually non-linear regression) to the Ct and AIF data is used to estimate the model parameters.

· T2* DSC-MRI is considered to be the standard approach for perfusion imaging of the brain [6].

· T1 DCE-MRI has largely superseded T2* DSC-MRI for perfusion imaging outside the brain. It is particularly popular in oncological imaging for differential diagnosis, staging and treatment follow-up. Many tumor types have been imaged in such a manner. More recently these imaging techniques have found a niche in the field of drug development [7].

· For certain applications more complex analysis is required:

o Measurement of liver perfusion is complicated by the organ’s dual blood supply (hepatic artery and portal vein). However, techniques have been developed to model this system [8,9].

o Another functional application of tracer kinetic modeling is the assessment of glomerular filtration rate (GFR) [10,11]. GBCA is filtered in the same way as many radioisotope markers of GFR and can be studied as part of a comprehensive renal exam [12].

Book:

A. Jackson, D.L. Buckley, G.J.M. Parker, Editors. Dynamic Contrast-Enhanced Magnetic Resonance Imaging in Oncology Springer-Verlag – Medical Radiology Series (2005).

Review articles:

A.M. Peters, Fundamentals of tracer kinetics for radiologists Brit J Radiol 71:1116-1129 (1998)

S.P. Sourbron, D.L. Buckley, Tracer kinetic modelling in MRI: estimating perfusion and capillary permeability. Phys Med Biol 57:R1-R33 (2012).