For the last twenty years therapy response of high-grade glioma (HGG) has been assessed using the Macdonald criteria (1). The criteria are based on 2D measurements of the enhancing component of the cancer on MR/CT imaging scans, in conjunction with clinical assessment and corticosteroid dose. It is assumed that the product of the largest cross-sectional enhancing diameters on a single plane through the mass is a good estimate of the global size of the tumor. Tumor progression is considered when an increase larger than 25% in size is observed. Evaluation of treatment in HGG is determined on the duration of patient survival, or on progression free survival (PFS) based on imaging findings. In the Macdonald Criteria a significant increase of at least 25% in the contrast-enhancing lesion is used as a reliable surrogate marker for tumor progression, and its presence mandates a change in therapy. However, Macdonald criteria have several limitations that became even more obvious with the use of novel therapies such as temozolomide (TMZ), bevacizumab, a monoclonal antibody against vascular endothelial growth factor (VEGF) or cediranib, a VEGF receptor. Radiologists learn very early in their training that contrast enhancement is nonspecific in the evaluation of cancer response. Alteration of the blood-brain barrier (BBB) may be due to recurrent tumor, subacute chemo or radiotherapy effects and delayed radionecrosis. Several agents may affect vessel permeability.

The aim of surgery in glioblastomas (GBM) is maximal safe tumor resection. Progression free survival (PFS) and overall survival (OS) have been correlated with near 100% resection of the enhancing viable component of the mass. A standard therapy protocol includes radiotherapy and concurrent TMZ (2). Ten to 20% of patients will show transient contrast enhancement few months after radio-chemiotherapy. This treatment-related reaction of the cancer leading to an increase in enhancement (flare phenomenon) and/or edema on MR imaging without clinical evidence of increased tumor activity is called Pseudoprogression. It is often detected within the first 12 weeks after radiotherapy and/or chemotherapy with TMZ, gene therapy, immunotherapy or intracavitary chemotherapy. Typically, the diagnosis of pseudoprogression is made in the presence of three criteria: i) absence of clinical signs of cancer progression; ii) lack of additional therapies; iii) decrease in size of the lesion at 6 months MRI follow-up study. Pseudoprogression is associated with local tissue reaction with inflammation, edema and increased abnormal vessel permeability. It usually subsides without further treatment, but in some unfortunate cases it may progress over time into the more severe local tissue reaction with signs of mass effect in addition to disrupted BBB and edema. Delayed radiation necrosis (DRN) is a different entity and it usually occurs 6-12 months after radiotherapy. DRN may progress over time.

Angiogenesis is one of the hallmarks of cancer, including brain tumors. GBM have the highest degree of vascular proliferation among solid tumors. Thus angiogenic pathways represent an attractive target to interfere with tumor growth. Up to date VEGF pathway targeting with specific drugs has yielded interesting therapeutic results. In particular bevacizumab, a monoclonal antibody against VEGF-A, has shown clinical activity in HGG, especially GBM, in terms of a high response rate on MRI and a significant increase in PFS. The main mechanism of action of anti-VEGF agents, including Bev, consists in a transient normalization of the highly abnormal tumor vasculature, leading to a reduction of vasogenic edema (3). The normalization of tumor vessels may theoretically improve the efficacy of chemotherapy by increasing the exposure of tumor cells to cytotoxic drugs and reducing tumor hypoxia, thus improving radioresponsiveness. This vascular normalization is time limited, as the restoration of BBB integrity could lead to ischemia and hypoxia again.

In HGG antiangiogenic agents can produce a marked decrease in contrast enhancement as early as 1 to 2 days after initiation of therapy. It commonly results in high radiological response rate of 25% to 60% (4) and 6 months progression-free survival (PFS-6), but with rather modest effects on overall survival (OS). However these apparent responses may be partly a result of normalization of abnormality permeable tumor vessels and not necessarily indicative of a true anticancer effect.

Pseudoresponse is induced by new antiangiogenic drugs (bevacizumab, cediranib, irinotecan) that modify signal transduction through the VEGF signaling pathways. The first major trial of bevacizumab for GBM reported a 57% response rate and a PFS-6 of 46% (5). The rapid normalization of the BBB within 24 hours, rebound enhancement and edema on drug discontinuation with a rapid "re-response" after restart suggest that a pseudoresponse is responsible for the imaging and clinical response. These imaging changes are so rapid that are unlikely to depend on real tumor shrinkage. Macdonald criteria suggest that radiological responses should persist for at least 4 weeks before they are considered as true responses. Unfortunately, patients treated with anti-VEGF agents may develop progression of the nonenhancing tumor component as shown on FLAIR imaging (6). This unfavorable event may be the result of migration of glioma cells induced by antiangiogenic treatment. This undesirable effect emphasizes that evaluating only the areas of contrast enhancement as a measure of outcome is inaccurate. Unlike the Macdonald criteria, new response assessment have been proposed to consider enlarging areas of nonenhancing tumor as evidence of tumor progression (7). It has been shown that FLAIR imaging may be helpful to identify pseudo-response and progression (8). However, precise quantitation of the area with abnormal signal on FLAIR must be differentiated from other causes such as radiation or other treatment effects, decreased corticosteroid dosing and ischemic injury. The Response Assessment in Neuro-Oncology (RANO) Working Group felt that having an objective measure of progressive nonenhancing recurrent disease similar to the 2D measurement of the contrast-enhancing area was not possible at present given the limitations of current MR methods.

Advanced MR imaging methods such as physiologic imaging with perfusion, permeability and diffusion as well as metabolic imaging with MR spectroscopy and PET are likely going to play an important role to improve response evaluation (9).

Diffusion-weighted images (DWI) has been assessed to differentiate tumor progression from necrosis. ADC values are higher in necrotic tissue than in recurrent tumor. Other investigators with diffusion tensor imaging (DTI) have demonstrated higher fractional anisotropy and lower ADC values in normal appearing white matter adjacent to edema in patients with radiation injury c/w patients with recurrent glioma (10). MR spectroscopy can reveal significant alteration in brain metabolites such as variable changes in choline and NAA loss, signs of anaerobic metabolism with high lactate and of necrosis with abnormal lipid signals. A strong choline signal would favor diagnosis of true disease progression rather than pseudoprogression. Dynamic susceptibility contrast (DSC) perfusion MR has been used to assess brain tumor treatment response with high sensitivity. Percentage of signal intensity recovery is an imaging indicator of microvascular leakiness and it may differentiate recurrent tumor from radiation necrosis (11). Cases with pseudoprogression may show decreased mean rCBV values while cases of true tumor progression an increase in rCBV (12). Preliminary results with permeability DSC seem also very promising. In those cases where advanced methods show that the nonenhancing signal abnormalities represent tumor progression, these patients would also be eligible for enrollment onto clinical trials for recurrent disease, although their tumor will be considered nonmeasurable in size. Although it would be preferable to have an objective measure of progressive nonenhancing recurrent disease, the RANO Working Group felt that this was not possible at present given the limitation of current technology.