Contrast-enhanced thin-sliced MRI imaging using gadolinium remains the gold standard imaging modality for target delineation for gliomas. The problems are that following surgery or radiotherapy, the signal change based on gadolinium enhancement can be non-specific making accurate target definition a challenge. Most gliomas recur within 4 cm from the margin of the original lesion, and this is often where non-specific signal change due to surgery and prior treatment are also apparent (Gaspar et al. 1992; Chan et al. 2002).

To optimise the outcome of radiotherapy treatment, improved target definition is of paramount importance. In this context, alternative biological imaging may improve the definition of the relevant target, for example, amino acid PET (SPECT)/CT/MRI image fusion to determine the gross tumour volume is currently under investigation (GLIAA-NCT01252459).

Typically, the gross tumour volume is delineated as the new or progressive contrast-enhancing lesion. For stereotactic radiosurgery or fractionated stereotactic techniques, typically no margin is added; however, based on physician preference or other treatment parameters, a small 1–2 mm margin may be added. Recently, several studies have assessed the value of adding a lower dose to the surrounding T2/FLAIR volume due to the possibility that this may encompass low-grade or transforming tumour (Clark et al. 2014).

There is no gold standard imaging modality that can distinguish true tissue necrosis from tumour recurrence associated with recurrence (Ullrich et al. 2008). Conventional MRI techniques have limitations when discriminating tumour recurrence and treatment-induced injury. This may become possible with advances in imaging technology such as perfusion-weighted, diffusion-weighted and magnetic resonance spectroscopy (Bobek-Billewicz et al. 2010). Early studies using ¹⁸F-FDG PET reported sensitivities of 81–86 % and specificities of 40–94 % when distinguishing between radiation necrosis and recurrent tumour (Langleben and Segall 2000). Further studies are awaited but co-registration of MRI and amino acid tracers may improve the diagnostic accuracy (Chen 2007; Götz and Grosu 2013).

For many years, the Macdonald criteria developed in 1990 was the gold standard in assessing response to treatment in high-grade gliomas. These criteria were based on two-dimensional tumour measurements and clinical assessments (Macdonald et al. 1990). These criteria focused on the contrast-enhancing lesions only and have limited use to differentiate pseudoprogression from true progression. The new standardised response criteria for glioma clinical trials are the updated Response Assessment Criteria in Neuro-Oncology (RANO) criteria developed in 2010 (Wen et al. 2010). These updated criteria differentiate between measurable and non-measurable lesions on contrast-enhanced CT, MRI T1 post-contrast and T2 imaging. A measurable lesion is defined as a contrast-enhancing lesion with clearly demarcated margins, seen on two or more axial slices with a maximum diameter of at least 10 mm excluding a cystic cavity. Non-measurable lesions are those with cystic or necrotic components or a surgical cavity. For these lesions, the peripheral nodular component is considered measurable. Criteria for progression of disease based on RANO criteria include a 25 % or more increase in enhancing lesions despite increasing steroid dose; a significant increase in non-enhancing T2/FLAIR, not attributable to other causes; any new lesions; and clinical deterioration.

The diagnosis of true tumour progression can be difficult as approximately 20–30 % of patients may develop pseudoprogression after concurrent chemoradiotherapy (Brandsma et al. 2008). Pseudoprogression is a manifestation of treatment-related effects and is more likely in MGMT-methylated glioblastoma patients. Newer imaging techniques have been developed which may also aid in the differentiation between early tumour progression and pseudoprogression. Perfusion MRI imaging measuring cerebral blood volume changes may predict early tumour progression as true tumour progression manifests as an increase in relative cerebral blood volume (rCBV), and radionecrosis more likely manifests with a relative decrease in rCBV (Surapaneni et al. 2015). Additionally, diffusion-weighted MRI imaging and apparent diffusion coefficient maps may also predict early tumour progression as true progression is more likely correlated with low ADC values (Chu et al. 2013).