The best evaluation for the treatment of tumors is based on patient survival. In those with recurrent disease, progression-free survival (PFS) through radiologic assessment serves as a reliable marker. Any recorded change in tumor size as a result of treatment typically corresponds with duration of survival. With the emergence of anticancer therapies, standardization of radiographic assessment became essential to reliably monitor therapeutic response.

In 1979, the World Health Organization (WHO) published the first major set of criteria to measure the response rate that was subsequently followed by the Response Evaluation Criteria in Solid Tumors (RECIST) in 2000 and RECIST 1.1 in 2009 [20]. In general, the response rate is based on how the tumor shrinks anatomically. However, this requires serial measurements of tumor burden that can be both costly and tedious for both patient and investigator. As found in both WHO and RECIST, this change is defined by either having a complete response, partial response, stable disease, or progressive disease. Unfortunately, interpretation of these findings remains challenging because certain therapies may not shrink tumor size, but may be successful in preventing progression associated with good outcome (i.e., patient survival).

Each of the response criteria suffers from limitations. For example, irregularly shaped tumors are difficult to measure. Furthermore, there are no guidelines on how to select lesions or assess multifocal tumors which greatly affects inter-rater reliability. Each criteria published does not assess non-enhancing tumors. There is also a bias to classify disease as progressive because any significant increase in size of the contrast-enhancing component signifies a tumor progression, which will alter therapy. However, contrast enhancement is nonspecific and can be influenced by medications such as corticosteroids and antiangiogenic agents. Additionally, postsurgical changes, radiation effects, and inflammation all increase enhancement but do not signify an increase in tumor burden. The most commonly experienced limitations for standardization of response criteria are discussed in the following sections.

Gliomas account for about half of all primary brain tumors in adults with an annual incidence of about five per 100,000 people [29]. They typically occur supratentorially in adults and infratentorially in children. There exists a simple histological grading system for gliomas based on nuclear atypia, mitoses, endothelial proliferation, and necrosis. High-grade gliomas have at least two of these features. High-grade gliomas are expansive and frequently extend beyond the tumor margins. Additionally, all high-grade gliomas show some amount of anaplasia and are subdivided into WHO grade III anaplastic astrocytomas (AA) and WHO grade IV glioblastoma multiforme (GBM) [29]. While CT can be used to differentiate tumor grade in some cases, MRI is the modality of choice due to better tissue contrast that aids in the assessment and grading of all brain neoplasms.

Anaplastic astrocytomas (AA) are infiltrating lesions in which about 75 % arise from low-grade gliomas while the remainder forms as a new primary neoplasm. Unlike GBMs, AAs lack necrosis and vascular proliferation. Instead, the presence of pleomorphic astrocytes and mitosis are characteristic of most AAs [29]. On imaging, AAs present as poorly defined heterogeneous lesions. On CT scans, typically there exists a hypodense area of peritumoral edema. When enhancement is present, it is nodular or heterogenous. On MRI, AAs can present with a range of findings better reflecting the variability in histology of AAs. T1WI shows a heterogenous isointense to hypointense lesion. High-signal areas on T2WI represent the tumor itself and a peritumoral vasogenic edema. On T1-weighted contrast scans there is usually patchy enhancement [30] (Fig. 10.4a–d).

GBMs are primary neoplasms in 60 % of cases, while the remainder arises from AAs. These tumors usually occur in subcortical white matter of the temporal (31 %), parietal (24 %), frontal (23 %), and occipital (16 %) lobes [31]. Brainstem gliomas can also occur, but they are rare and usually found in children secondary to transformation from a low-grade glioma. On imaging, necrosis is the hallmark of GBMs. On CT, GBMs appear as low-density lesions that are heterogenous due to necrosis, hemorrhage, and increasing cellularity. Additionally, edema can be found surrounding the tumor and extending along adjacent white matter tracts causing a mass effect. On contrast-enhanced CT, GBMs appear with heterogeneous ringlike enhancement and an irregular and nodular wall. On T1-weighted MRI, GBMs appear heterogenous and characteristically have a necrotic center, peritumoral edema, and a thick irregular margin. As with CT, post-contrast T1-weighted imaging shows irregular infiltrative enhancement. On T1WI, the necrotic center will appear hypointense, while an enhanced irregular border surrounds the mass. Any peripheral non-enhancing areas most likely represent vasogenic edema. It is important to note that the enhanced ring-like structure is not the true border of the tumor since malignant glioma cells can be found beyond a 2-cm margin. On T2WI, GBMs also appear heterogeneous due to necrosis, hemorrhage, or tumor vascularity. The rim of vasogenic edema is better appreciated on T2WI. Unfortunately, distinguishing tumor from edema is challenging due to similar signal characteristics on T2 [30] (Fig. 10.5a–d).

On DWI, ADC values have been shown to correlate with tumor cellularity and tumor grade. GBMs have high cellularity compared to lower-grade gliomas. The increased cellular density causes restricted diffusion in the extracellular space, which produces low ADC values. As a result, high-grade gliomas present as hyperintense lesions on DWI.

Anatomically, GBMs can spread along white matter tracts and cross hemispheres via the corpus callous (Fig. 10.5a–d). Other common pathways for dissemination include the optic radiations, the internal capsule, the anterior commissure, and posterior commissure. GBMs remain solitary lesions 99 % of the time. Rare multicentric tumors commonly spread via the meningeal-subarachnoid space and appear widespread throughout the brain without evidence of intracranial connections (those with some continuity are multifocal) [32]. They can also disseminate throughout the brain via subependymal, intraventricular, or direct penetration routes through white matter. To support its own tumor growth, GBMs undergo angiogenesis, which is a critical parameter to classify tumor grade.

Determination of the regional vascularity is an important parameter to grade gliomas. While contrast enhancement can indicate a disruption in the blood-brain barrier, it cannot reveal the vascularity of the tumor. Furthermore, as many as 33 % of high-grade gliomas demonstrate no significant contrast enhancement and 20 % of low-grade gliomas demonstrate contrast enhancement. Perfusion-weighted imaging can be used to assess the underlying vascularity of the tumor. Regional cerebral blood volume (rCBV) calculated from perfusion-weighted imaging is a sensitive marker of the microvasculature of tumor (Fig. 10.5a–d) [33]. Numerous studies show that rCBV correlates positively with tumor grade [33, 34]. Low-grade gliomas typically have rCBV value that range between 1.11 and 2.14, while high-grade gliomas range between 3.54 and 7.32 [35]. CBV mapping also has the added benefit of directing stereotactic biopsy sites to highly vascular areas so as to reduce sampling errors in histopathologic diagnosis of gliomas [33].

Studies suggest that MRSI through the calculation of metabolite ratios and levels can be helpful in grading tumors (Fig. 10.5a–d). Choline (Cho) is a cell membrane marker that increases during high rates of cell membrane turnover or proliferation; elevations in Cho commonly represent high-grade tumors. It is important to note that while low levels of Cho are typically associated with low-grade tumors, in GBMs with extensive necrosis, Cho levels may be low. Creatine (Cr), a marker for energy metabolism, is commonly low in high-grade tumors that consume energy due to increased metabolic activity. N-acetylaspartate (NAA), a marker for neuronal integrity, is also commonly low when high-grade tumors are present. In high-grade gliomas, lactate levels are increased after the tumor outgrows the blood supply and begins to utilize anaerobic glycolysis. Myoinositol, a glial cell marker, can also be used to distinguish high-grade and low-grade gliomas because GBMs tend to have low levels of this metabolite. Ratios of these metabolites, namely, Cho/Cr, NAA/Cr, and NAA/Cho, can aid in the grading of tumors [12, 18]. Table 10.2 summarizes MRSI findings useful in grading tumors.