T₂^*-weighted DSC MRI, also referred to as bolus-tracking MRI, is the most commonly used method to assess perfusion in brain tumors with MRI (see Chapters 11 to 14 and 24 to 28 and 32). The relative cerebral blood volume (rCBV) calculated with this technique is the most robust and widely used metric obtained from DSC MRI.²⁵ Relative, rather than absolute values of CBV, are usually derived because of the technical difficulties involved with derivation of the arterial input function (AIF). The rCBV is typically normalized to contralateral normal appearing white matter.²⁵ In DSC MRI, rapid repeated imaging is performed that depicts reduction of signal intensity during the first-pass of bolus administration of an exogenous, nondiffusible, paramagnetic gadolinium-based contrast agent (GBCA). Within each voxel, a signal intensity time curve is generated, and in the absence of significant contrast agent leakage or recirculation, the transverse relaxivity on the spin echo (R₂) or gradient echo (R₂^*) is expressed as:

where SI_t is signal intensity at time t, SI₀ is the baseline signal prior to contrast arrival, and TE denotes echo time. The ΔR₂^* is assumed to be linearly proportional to the contrast agent concentration, however, this relationship may not always be valid and could lead to errors in cerebral blood flow (CBF) and CBV determination.^26,27,28,29 The area under the contrast agent concentration–time curve is proportional to CBV.^30,31,32

DSC MRI most often uses single-shot echo-planar imaging (EPI) with multislice gradient echo (GRE) sequences to rapidly acquire images. Spin-echo (SE) EPI methods have also been used, however, GRE sequences provide a higher signal-to-noise ratio (SNR) than SE. Although GRE sequences appear to be sensitive to both capillaries and larger vessels, SE appears to be more sensitive to capillaries.³³ In addition, there are fewer susceptibility artifacts associated with SE compared with GRE sequences. GRE techniques appear to have a better correlation with glioma grade compared with SE, likely because the angiogenesis of these tumors results in enlarged microvessels that do not resemble capillaries.^34,35,36,37

T₁-weighted dynamic contrast enhancement (DCE) MRI, sometimes referred to as permeability MRI, also measures changes in tissue signal intensity over time following administration of a GBCA (see Chapters 23 and 29). Nevertheless, it usually employs more complex pharmacokinetic models (PKMs) than DSC MRI. In general, the T₁ relaxation rate is measured before, during, and after the intravenous administration of a GBCA for several minutes by repeatedly imaging a region of interest using a T₁-weighted GRE or spoiled gradient echo type. The latter is less sensitive to T₂ effects that degrade the T₁ signal intensity, but with inferior SNR compared with GRE. Nevertheless, improvement of SNR can be accomplished with the use of three-dimensional spoiled gradient echo or equivalent techniques.³⁸ There are often conflicting demands on high temporal resolution, high spatial resolution, SNR, and anatomical coverage.³⁹

Acquisition of T₁-weighted DCE data usually involves the use of precontrast T₁ mapping techniques and dynamic three-dimensional acquisition to obtain images that provide an estimated GBCA concentration at each location and time point. Then an estimate of the AIF and a PKM are incorporated to derive metrics, such as the volume transfer constant between the capillary plasma space and extravascular-extracellular space (EES), (K^trans), initial area under the contrast agent concentration–time curve and the fractional volume of the extravascular-extracellular space (v_e). PKMs, which also include the contribution of the intravascular tracer to the MR signal, will also provide estimates of the fractional plasma volume (v_p). Note that V_e = v_e·V_t, with V_t being the total tissue volume and V_p = v_p·V_t.⁴⁰

The changes of signal intensity with time will be converted into GBCA concentration–time curves. Unlike the case with computed tomography perfusion, the relationship between MR signal intensity and GBCA concentration is not always linear.⁴¹ Accurate determination of GBCA tissue concentration from T₁ mapping can be facilitated by numerous techniques such as the variable flip angle approach, in which baseline precontrast T₁-weighted images with multiple flip angles are acquired just prior to the DCE acquisition. Pathologic tissues enhance rapidly, and greater temporal resolution allows for more sophisticated PKM, which can determine physiologic markers such as K^trans with superior specificity.⁴²

Rather than relying on an exogenous contrast agent, ASL is an MR perfusion technique that uses magnetically labeled arterial water protons as an endogenous tracer (see Chapters 15 to 22 and 30).²⁷ It is capable of providing an estimate of absolute CBF, although for brain tumor evaluation, relative values appear to be sufficient because absolute CBF must be corrected for age- and patient-dependent mean perfusion.⁴³ Absolute CBF can, however, allow comparison of values in a given individual patient throughout the course of treatment.

There are relatively fewer ASL MRI brain tumor perfusion studies compared with those for perfusion MRI
techniques that employ exogenous GBCAs. ASL does offer some advantages compared with DSC or DCE MRI. ASL can be considered completely noninvasive because it does not require the use of GBCAs. This would allow repeated measurements and also is notable given the recognition of nephrogenic systemic fibrosis (NSF) associated with GBCAs.^44,45 For some pediatric patients, ASL would be advantageous where intravenous access might be difficult, although a contrast-enhanced MRI is generally desired in brain tumor patients.⁴⁶ Also, because ASL uses a diffusible tracer, it appears to be relatively insensitive to permeability, a major obstacle in DSC CBF determination.⁴⁷

Some disadvantages of ASL include intrinsically low SNR, longer acquisition times, and overall technical complexity.^46,48,49 Also, prolonged arterial transit times in severe ischemia can result in relaxation of the spin label and produce underestimation of CBF, particularly in white matter or regions of otherwise prolonged label arrival times.⁵⁰

ASL typically uses a fast acquisition method like EPI or spiral imaging.⁴⁶ Arterial blood water is magnetically “labeled” using a radiofrequency pulse. Because this label decays with T₁ relaxation, only a small amount of ASL water accumulates in the brain.⁵⁰ In order to allow flow of magnetically labeled blood water into the microvasculature and tissue, a postlabeling delay is necessary.⁵¹ Images where there is no magnetic labeling of arterial blood water but where magnetization transfer effects are accounted for, referred to as “control” images, are also obtained. Labeled acquisitions are subtracted from control experiments without magnetization to determine CBF in milliliters per 100 g per minute.⁵⁰ A 0.5% to 1.5% signal difference is seen between inverted and control spins, so experiments must be repeated to improve this poor SNR.⁴⁹

The grading of gliomas is critical to determine prognosis and therapeutic management.²⁵ The clinical management for LGGs may vary from conservative follow up to surgical resection, while HGGs will usually require surgery and adjuvant chemoradiation therapy.⁴⁹ Most perfusion MRI studies in the literature have focused on prediction of tumor grade. In practice, most suspicious tumors are biopsied, and while histopathologic diagnosis remains the gold standard, these studies can potentially validate these microvascular biomarkers.⁵³ Reliance on conventional contrast-enhanced MRI to predict glioma grade can be problematic because, although HGGs more commonly display contrast enhancement, it is not uncommon to be seen in LGGs as well.⁵⁴ Brain tumor biopsies have traditionally relied on contrast-enhanced computed tomography or MRI.^55,56 Nevertheless, because the most malignant part of a tumor might not enhance, up to 25% of tumors appear to be undergraded.^1,57

Other limitations to histopathologic grading include a lack of consensus regarding multiple approaches in grading systems^58,59,60,61 and relatively low reproducibility of interpretation.⁶² Incorporation of rCBV parametric maps has been used in some centers to improve biopsy targeting of the most malignant portions of a tumor (Fig. 58.1).³³ Because metrics such as rCBV might be able to predict clinical outcome, microvascular imaging biomarkers could potentially serve as an adjunct to histopathologic grading of gliomas.⁶³

Elevation of microvascular density, a histopathologic marker of malignancy, is directly related to rCBV.⁶⁴ Multiple studies have demonstrated a strong correlation between glioma grade and rCBV (Figs. 58.2 and 58.3).^{9,54,65,66,67} It should be kept in mind, however, that glioma grading with rCBV should be limited to fibrillary astrocytomas, because other gliomas, especially oligodendrogliomas, could have elevated rCBV regardless of histologic grade.²⁵

Oligodendrogliomas possess what has been called “chicken-wire” vasculature, delicate neoangiogenic vasculature that can result in even low-grade tumors demonstrating quite elevated rCBV, even comparable to that of glioblastoma.⁶⁸ Another DSC MRI metric that is strongly correlated with rCBV and thought to reflect tumor vasculature is peak height, which is calculated as the maximum difference between the precontrast T₂^*-weighted signal intensity and the greatest drop in signal intensity during the first passage of contrast bolus through the investigated tissue.⁶⁹ Barajas et al.⁶⁴ demonstrated that recurrent glioblastoma multiforme (GBM) showed both increased rCBV and peak height compared with delayed radiation necrosis (DRN) using DSC MRI.

Without including tumors with oligodendroglial components, Bisdas et al.⁷⁰ found that astrocytomas with an rCBV_max greater than 4.2 was predictive of recurrence and rCBV_max of 3.8 or less was predictive of 1-year survival. These values appear to be higher than those reported previously by Lev et al.⁷¹ (1.5) and Law et al.⁷² (1.75). These differences could be a reflection in methodologic differences in these studies, including variations in image acquisition, derivation of rCBV values using regions of interest, and methods of assignment of threshold rCBV values. Lev et al.⁷¹ used am SE EPI and found that an rCBV less than 1.5 was predictive of prolonged survival for patients with astrocytomas. Law et al.⁷² examined 189 LGG and HGG patients and found that an rCBV threshold of 1.75 was able to predict median time to progression independent of histopathology (Fig. 58.4). Patients with high initial relative rCBV appeared to have more rapid progression than those with low rCBV. Using a standardized imaging and postprocessing protocol at two separate institutions, Caseiras et al.⁶³ found that LGG patients with an adverse event exhibited a significantly higher baseline rCBV than those without. A longer time to progression was seen in patients with an rCBV less than 1.75 than those greater than 1.75, implying that LGGs with higher rCBV are more likely to behave like HGGs. Danchaivijitr et al.⁷³ demonstrated an increase in rCBV up to 12 months prior to the appearance of contrast enhancement on conventional MRI in LGGs undergoing malignant transformation (Fig. 58.5).