Perfusion is physiologically defined as the steady state delivery of blood to an element of tissue. The term "perfusion" is also used to emphasize contact with the tissue, or in other words capillary blood flow. Perfusion is variably used for different physiologic parameters that also affect the MR signal, e.g. blood volume, blood velocity and blood oxygenation.
Perfusion is defined as millilitres of blood per minute and 100g of tissue which is the standard definition of the cerebral blood flow (CBF).
Cerebral blood volume (CBV) is defined as the fraction of the total tissue volume within a voxel occupied by blood (arteries, capillaries and veins. In the normal functioning brain perfusion and blood volume are often strongly correlated.
Mean transit time (MTT) is an other important parameter which describes the transport kinetics of an agent. The unit used is seconds.
Because perfusion and blood volume is disturbed in many disease processes, monitoring of this key physiological parameter can often provide insight into disease. Consequently, the measurement of perfusion for medical purposes has been performed in almost all organs using different technical approaches.
During the last decade several methods have been described to non-invasively measure perfusion with magnetic resonance imaging. Most effort in this context have been made in the perfusion imaging of the brain.
Today there are several approaches to assess perfusion with dynamic contrast enhanced MRI techniques. The two major approaches to measure cerebral perfusion with MRI are:
The basic technique is to inject a gadolinium chelate and acquire images rapidly as the bolus of contrast agent passes through the blood vessels in the brain . The contrast agent causes a signal change; this signal change over time can then be analysed to measure cerebral hemodynamics. The passage of a rapid bolus of paramagnetic contrast agent through the cerebrovascular system is monitored by acquiring a series of T2*-weighted images. The signal-time curves are converted into concentration-time curves from which blood volumes within regions of interest (ROI) can be calculated. A relative determination of regional cerebral blood volume (rCBV) values, which is possible with the standard methodology, proved to be sufficient for most of the current clinical applications such as the early detection of ischemia. However, for oncological applications including follow-up studies, absolute quantifications of the rCBV and regional cerebral blood flow (rCBF) are required. These necessitate a knowledge of the arterial input function (AIF) in order to calculate the tissue response to the bolus ± based concentration-time curve. There are different potential methods for determination of the AIF based on intracerebral or brain feeding arteries.
In the earlier years a double-slice technique, which allowed the simultaneous measurement of the AIF in the brain feeding arteries and tissue concentration-time curves within the brain was used. This enabled an absolute quantification of the blood volume but only in one slice. From each of the two sections, 55 T2*-weighted FLASH images (TR/TE1/TE2/flip = 31/15/25/10°) were acquired after bolus application of gadolinium. Post processing was performed on a separate workstation.
In recent years EPI based sequences which allow to cover nearly the whole brain have replaced the double slice techniques in most fields. Most of the currently used EPI sequences (Gradient-Echo or Spin-Echo) are not able to allow an absolute quantification of blood flow and volume, newer studies presented sequence types with shorter echo times or dual echo EPI techniques which might allow the absolute quantification.
Dynamic contrast enhanced magnetic resonance imaging (DCE-MRP) is the acquisition of serial images before, during and after the administration of extracellular low-molecular weighted MR contrast media. The resulting signal intensity measurements of the tumor reflect a composite of tumor perfusion, vessel permeability, and the extravascular-extracellular space [literature?].
DCE-MRI has been investigated for a range of clinical oncologic applications including cancer detection, diagnosis, staging and assessment of treatment response. The DCE-MRP allows to measure permeability and its aberrations, while microvascular density (MVD) measures only the histopathologically partial picture of the tissue microvasculature. Furthermore, MVD is also heterogeneous property of tumors and is limited by histophatologic sampling and are generally hotspot values. Tumor microvascular measurements by DCE-MRI have been found to correlate with prognostic factors such as tumor grade, microvessel density (MVD), and vascular endothelial growth factor expression (VEGF) and with recurrence and survival outcomes.
In addition, changes of DCE-MRP in follow-up studies during therapeutic intervention have been shown to correlate with outcome, suggesting a role for DCE-MRI as a predictive marker.
In contrast to conventional (static post contrast T1-w) enhanced MRI, which simply presents a snapshot of enhancement at one time point, DCE-MRI permits a fuller depiction of the wash-in and wash-out contrast kinetics within tumors, and this provides insight into the nature of the bulk tissue properties on its microvascular level.
With the strong demand in drug development (especially with the introduction of anti-VEGF trials) to identify a biomarker that can assess tumor microvascular properties non-invasive in animal as well in human studies, this technique seem to be most appealing as an possible imaging biomarker.
DCE-MRI also allows further non-invasive characterisation of brain lesion and its microcirculatory properties. This information can be used for improved biopsy and treatment planning. The gained information can also be used for monitoring therapeutic interventions e.g. chemo- and/or radiotherapy. This is supported by the fact that the dynamic information reflects the angiogenic profile and heterogeneity of a tumor.
Today there exicst several approaches to perform and to assess DCE-MRP data. As DCE-MRI for diagnostic and therapeutic monitoring emerges, a bewildering variety of analysis approaches have been described.
Measurements of the contrast agent extravasation have been broadly based on the approaches proposed by Tofts and Brix.
Both approaches ignore intravascular contrast agent in the tissue and assume an either pre-defined or a modelled AIF, which induce an uncertainty in DCE-MRI quantification compared to other novel approaches, not implemented in most of the available analysis software, using a measured AIF.
Newer approaches and software developments may overcome this problem.
Today there are several indications for the different cerebral perfusion MRI techniques. Depending on the type of disease and the presence of enhancing tissue lesions, the one or other technique or a combination ob both techniques using a double injection are used.
The major CNS indications routinely used in increasing numbers of imaging centers include cerebrovascular disease, tumor imaging and recently psychiatric disorders.
All three of the main descriptors of perfusion (CBV, CBF, and MTT) are useful in imaging acute cerebral ischemia. Of these three, CBV maps have been shown to correlate best with final infarct volume, implying that rCBV maps may include flow that comes in via collateral vessels, and thus give a snapshot of cerebrovascular reserve. MTT is an easy-to-interpret format that is homogeneous in normal areas. This is useful because the gray-white differences that are present in CBF and CBV maps can make interpretation difficult. MTT does have a tendency to overestimate infarct size, and generally MTT maps tend towards a binary classification; they show either areas as normal or else uniformly abnormal (with no gradation). This can be helpful for identifying areas of abnormal hemodynamics. Unfortunately, MTT and other timing maps do not appear to do an adequate task in distinguishing between levels of hemodynamic compromise. That is, the feature that makes them easy to interpret - being somewhat all or nothing, normal or abnormal - does not allow for the graduation of abnormalities, the difference between blood flow decreases that are mild to moderate versus moderate to severe. And of course, all of these perfusion techniques (both MRI and non-MRI) appear to be unable to distinguish between acute and chronic hemodynamic compromise.
To make this distinction diffusion weighted MR imaging is added to the protocol.
Other studies have confirmed that perfusion MRI might also be helpful in some non-acute pathophysiological states of cerebrovascular disease such as stenosis or occlusion of brain supplying arteries or venous disease.
Diagnosis of the presence of CNS tumors has been greatly aided by MRI and its outstanding anatomic detail. Characterization of a tumor's malignant potential can be more difficult using conventional techniques, particularly because T1 and T2 relaxation times as well as the integrity of the blood-brain barrier of neoplastic tissue are only moderately specific indicators of malignancy, which seems to be directly associated with tumor angiogenesis. The degree of angiogenesis has been linked to tumor grade in human gliomas, with higher grade lesions showing increased angiogenic factors.
Perfusion MRI, with its sensitivity to the capillary bed, especially when using spin echo techniques, may therefore be ideally suited for evaluating tumor angiogenesis in vivo. In recent years considerable clinical experience has been gained with perfusion MR imaging of tumors. Perfusion MRI can play an important role at the major clinical decision points: diagnosis, intervention, and post-treatment monitoring.
The contrast enhanced methods have been applied in the diagnostic work-up of low-grade astrocytomas since there is minimal blood brain barrier breakdown and thus no contrast enhancement. The differential diagnosis can be ischemic infarcts. Due to tumor-induced angiogenesis, most of these low grade astrocytomas present an increased blood volume compared to the surrounding white matter. Following radiotherapy, there is a reduction in intratumoral blood volume. Patients with high pretherapeutic intratumoural rCBV values have a worse outcome after radiotherapy indicating more aggressive tumor growth correlated to a higher tumor-induced angiogenesis. This DSC technique may also allow a non-invasive functional assessment of delayed radiation injury, which is based on fibrosing and occlusion of small vessels. Patients who have had whole-brain radiotherapy reveal a significantly decreased rCBV in normal brain tissue. In contrast, patients with grade II astrocytoma after conformal radiotherapy have only a relatively moderate decrease in rCBV of normal tissue after therapy. The data demonstrate a measurable sparing of normal tissues with advanced radiotherapy techniques with regard to blood volume. Other studies using pre-therapeutic measurements of rCBF and rCBV could be used to predict the response of brain metastases to radiation therapy. The results demonstrated that high pre-therapeutic rCBV seem to be an indicator for a poor treatment outcome. After radiosurgery, patients with tumor remission and stable disease demonstrated decreased rCBV despite a temporary tumor volume increase. This was in contrast to the increased rCBV seen in patients with tumor progression at 3 months follow-up. No effects of radiosurgery on surrounding normal brain tissue were observed.
There is an increasing interest in imaging studies of psychiatric diseases. The most important diseases are the senile dementia of the Alzheimer type (AD) and schizophrenia. While first volumetric studies have been performed functional MRI studies are applied more frequently.
Dementia is one of the most common diseases and still difficult to diagnose radiographically. The most common subtype, senile dementia of the Alzheimer type (AD) remains particularly challenging. Previous PET studies have shown alterated rCBV and rCBF values in both grey and white matter. Perfusion imaging with fMRI has been applied to this problem as well, with early studies indicating similar results to those of PET and SPECT. These early studies indicate that rCBV mapping may have a sensitivity and specificity similar to that of nuclear medicine approaches for the diagnosis of AD.
Vascular causes of dementia have also been studied with perfusion MRI. These studies have shown that these patients present with a generally decreased cerebral perfusion. The white matter changes are more pronounced than in patients with AD. As in many other areas of clinical application, these perfusion studies are still in progress.
There are only a limited number of studies available which examined the perfusion changes in patients with schizophrenic patients. PET and SPECT examinations have shown that regional perfusion may contribute to the prediction of pharmacological therapy response.
Dynamic susceptibility contrast magnetic resonance imaging was used to evaluate cerebral blood volume in the right and left occipital cortex, basal ganglia, and cerebellum of subjects with schizophrenia and healthy comparison subjects. Cerebral blood volume was found to be greater in the schizophrenic subjects in every region studied suggesting an abnormality of the configuration of cerebral blood vessels in schizophrenia.
Perfusion MRI has been used to study a variety of other conditions. Reports of perfusion changes are available in patients with toxic diseases, e.g. after psychotropic medications, hypercapnia, or cocaine. Preliminary studies are available in infectious diseases e.g. cerebral abscesses, toxoplasmosis or HIV. Perfusion MRI has also been used to study cerebral hemodynamics in patients with intraparenchymal hemorrhage, sickle cell anemia, and in patients undergoing evaluation of cerebrovascular reserve. It may also be that pharmacological or vascular interventions could be monitored with perfusion MRI studies.
Several Studies are available in patients with epilepsia.