Arterial Input Function (AIF)
What is the arterial input function (AIF) and why do you need to measure it?
|
|
|
In the ideal DSC study we would like gadolinium to be delivered to tissue in a short, compact arterial bolus. In practice, however, the initially intravenous bolus becomes delayed and dispersed by passage through the heart and lungs. The shape of the tissue concentration curve measured in a DSC experiment is highly dependent on the shape of the arriving arterial input function (AIF), a(t). Without knowledge of the AIF, valuable information about intrinsic tissue perfusion dynamics may be lost.
|
Shape of the tissue contrast concentration curve, ctissue(t), is highly dependent on the shape of the arterial input function
|
The simplest method to obtain an AIF is to place a region of interest over an easily identifiable artery. In abdominal imaging this might be the aorta or renal artery; in the head it often a major named cerebral artery (ACA, MCA, etc) or one of its proximal branches. If an AIF obtained in this fashion is then applied to all voxels within a slice, it is called a global AIF. However, an appreciable distance may exist between the site of AIF measurement and voxel of interest. Bolus delay and dispersion of contrast can then produce significant errors in blood flow quantification. A potentially better method, the subject of considerable research over the last decade, is to identify local AIFs arising from small arteries located very close to each tissue voxels being imaged.
Voxel-by-voxel interrogation to determine the best AIF in a 1 cm² region near the middle cerebral artery. The "optimal" AIF is generally one with a flat baseline and sharp peak.
|
Modern perfusion software packages now provide automated or semi-automated methods for identifying optimal local AIFs. An example is shown left where an array of possible AIFs from a small region of brain can be interrogated to find the best candidate. The voxel(s) selected for the AIF should ideally have a flat baseline, sharp peak, and compact shape. In this example, a voxel in the 2nd (or 3rd) column would likely be the best choice.
|
Once an appropriate AIF has been selected, software is used to generate pixel-by-pixel maps of important perfusion parameters, including blood volume, blood flow, and mean transit time. Exactly how this is accomplished is the subject of the next Q&A.
References
Bleeker EJW, van Buchem MA, van Osch MJP. Optimal location for arterial input function measurements near the middle cerebral artery in first-pass perfusion MRI. J Cereb Blood Flow Metab 2009; 29:840-852. (Best location for the AIF may be just outside the MCA, not within it).
Calamante F, Mørup M, Hansen LK. Defining a local arterial input function for perfusion MRI using independent component analysis. Magn Reson Med 2004; 52:789-797.
Carroll TJ, Rowley HA, Haughton VM. Automatic calculation of the arterial input function for cerebral perfusion imaging with MR imaging. Radiology 2003; 227:593-600.
Mlynash M, Eyngorn I, Bammer R, et al. Automated method for generating the arterial input function on perfusion-weighted MR imaging: validation in patients with stroke. AJNR Am J Neuroradiol 2005; 26:1479-1486. (Computers generally due a more consistent/reliable job of finding the best AIF than people!)
Bleeker EJW, van Buchem MA, van Osch MJP. Optimal location for arterial input function measurements near the middle cerebral artery in first-pass perfusion MRI. J Cereb Blood Flow Metab 2009; 29:840-852. (Best location for the AIF may be just outside the MCA, not within it).
Calamante F, Mørup M, Hansen LK. Defining a local arterial input function for perfusion MRI using independent component analysis. Magn Reson Med 2004; 52:789-797.
Carroll TJ, Rowley HA, Haughton VM. Automatic calculation of the arterial input function for cerebral perfusion imaging with MR imaging. Radiology 2003; 227:593-600.
Mlynash M, Eyngorn I, Bammer R, et al. Automated method for generating the arterial input function on perfusion-weighted MR imaging: validation in patients with stroke. AJNR Am J Neuroradiol 2005; 26:1479-1486. (Computers generally due a more consistent/reliable job of finding the best AIF than people!)
Related Questions
How is the arterial input function used to extract more quantitative flow information from the DSC data?
How is the arterial input function used to extract more quantitative flow information from the DSC data?