Glioma surveillance imaging: current strategies, shortcomings, challenges and outlook

Abstract

Inaccurate assessment of surveillance imaging to assess response to glioma therapy may have life-changing consequences. Varied management plans including chemotherapy, radiotherapy or immunotherapy may all contribute to heterogeneous post-treatment appearances and the overlap between the morphological features of pseudoprogression, pseudoresponse and radiation necrosis can make their discrimination very challenging. Therefore, there has been a drive to develop objective strategies for post-treatment assessment of brain gliomas. This review discusses the most important of these approaches such as the RANO "Response Assessment in Neuro-Oncology", iRANO "Immunotherapy Response Assessment in Neuro-Oncology" and RAPNO "Response Assessment in Paediatric Neuro-Oncology" models. In addition to these systematic approaches for glioma surveillance, the relatively limited information provided by conventional imaging modalities alone has motivated the development of novel advanced magnetic resonance (MR) and metabolic imaging methods for further discrimination between viable tumour and treatment induced changes. Multiple clinical trials and meta-analyses have investigated the diagnostic performance of these novel techniques in the follow up of brain gliomas, including both single modality descriptive studies and comparative imaging assessment. In this manuscript, we review the literature and discuss the promises and pitfalls of frequently studied modalities in glioma surveillance imaging, including MR perfusion, MR diffusion and MR spectroscopy. In addition, we evaluate other promising MR techniques such as chemical exchange saturation transfer as well as fludeoxyglucose and non-FDG positron emission tomography techniques.

Figure 1.

Flow chart for glioma treatment response including pseudoprogression. CR, complete response; PD, progressive disease; PR, partial response; SD, stable disease.

Figure 2.

True progression. Treatment surveillance after surgical resection chemoradiation in a patient with GBM. (A, B and C) T₁WI post-gadolinium imaging show progressive increase in the volume of the heterogeneously enhanced residual tumour tissue. (D) Perfusion imaging demonstrates some focally increased cerebral blood volume mostly on the lateral parts of the heterogeneous mass lesion. T₁WI, T₁ weighted imaging.

Figure 3.

Patient with right occipital glioblastoma underwent diagnostic biopsy and concurrent adjuvant chemoradiation with temozolomide. 6 months after the baseline MRI (A–C), which showed well-circumscribed oedema (arrow in A) with rather solid enhancing (arrow in B) tumour and strong perfusion features in the DSC-derived rCBV map (arrow in C), there is further increase in the oedematous area (arrowhead in D) with expansile partly necrotic tumour mass (arrowhead in E). According to the mRANO criteria, the lesion is classified as preliminary progressive disease but the obvious decrease in the tumour vascularity in the rCBV map (arrowhead in F) demonstrates clearly the pseudoprogression effect. The enhancing lesion was stable in the next MRI follow-up and started resolving 3 months later. DSC, dynamic susceptibility contrast; RANO, Response Assessment in Neuro-Oncology; rCBV, relative cerebral blood volume.

Figure 4.

Early progression of a histopathologically verified GBM (IDH wild type). (A) T₂WI displays diffuse abnormal high signal in the left periinsular region. (B) T₁WI with gadolinium shows no significant enhancement within the diffuse glioma. (C) Baseline perfusion image shows foci of high CBV (yellow arrow), eventually suggesting anaplastic foci within diffuse glioma in line with the histopathological diagnosis. (D) T₁WI with gadolinium after 2 months surveillance shows two distinct enhancing lesions. (E) Perfusion image demonstrates remarkably high vascularity in the enhancing lesions with an rCBV ratio of ~9. T₁WI, T₁ weighted imaging; rCBV, relative cerebral blood volume.

Figure 5.

Serial images for a patient with a resected GBM. (A, B) Serial T₁WI post-gadolinium (with a 1-month interval) show increasing irregularly enhanced tissue outlining the resection cavity. (C) Perfusion image demonstrates normal to slightly elevated focal Ktrans values (~0.1) predominantly in line with treatment-related changes (pseudoprogression). T₁WI, T₁ weighted imaging.

Figure 6.

Post-contrast T₁WI (arrow in A) in a patient with recurrent glioblastoma on the left frontal lobe shows the enhancing tumour relapse. The patient underwent combined chemotherapy with temozolomide and bevacizumab showing partial response after 6 months in the post-contrast T₁WI (arrow in B) and in the DCE-derived Ktrans map (arrow in C) due to the anti-neoangiogenic effect of the administered bevacizumab. The MR spectroscopy indicates that the conventional and perfusion MR appearances are largely misleading and show pseudoresponse as the choline map still shows a very high tumour burden (arrows in D). DCE, dynamic contrast enhancement; T₁WI, T₁ weighted imaging.

Figure 7.

Patient with glioblastoma undertreatment shows extensive vasogenic oedema on the right frontal lobe (arrow in A) with fairly well-circumscribed enhancement on the genu of the corpus callosum (arrow in B). The enhancing region demonstrates markedly restricted diffusion (arrows in C, D) without haemorrhagic features (E) suggesting hence hypercellular tumour recurrence.

Figure 8.

Patient with recurrent glioblastoma, treated with bevacizumab as mono-therapy, shows diffuse oedema in the T₂ FLAIR images (arrows in A) surrounding the surgical cavity. The post-contrast T₁WI show patchy, confluent enhancement in the area with abnormal T₂ signal intensity (arrows in B) and a few areas with restricted diffusion (arrows in C). 6 months later, the gadolinium enhancement has decreased (arrows in D) suggesting hence response. The diffusion restriction (arrows in E) appears to be increasing in the areas of the previous patchy enhancement and the DWI retains its high signal 6 months later (arrows in F) confirming partial response to therapy. The contrast enhancement has also stable appearances (G). The diffusion restriction is believed to represent ischaemic tissue with coagulative necrosis. DWI, diffusion-weighted imaging; FLAIR, fluid attenuation inversion recovery; T₁WI, T₁ weighted imaging..

Figure 9.

Patient with left precentral anaplastic (WHO Grade 3) glioma treated with partial resection and adjuvant concurrent chemoradiation. Three months after treatment end, the patient started experiencing progressing hyperintense T₂WI signal changes in the left frontoparietal hemisphere and undergoes multiparametric hybrid MR-PET imaging with methionine. The enhancing lesion shows a tracer ”hot-spot” (arrow in A) with increased vascularity (arrow in B, DSC-derived relative cerebral blood volume map) and increased focal cell proliferation rate denoted by the increased ratio of choline to creatine in the MR spectroscopy (arrow in C). Interestingly, the diffusion weighted imaging shows pathological signal (arrow in D) on the medial part of the lesion, which appears “cold” on the other biomarkers’ maps. The finding probably reflects the different stages of tumour metabolism and growth captured by the different imaging modalities in a complementary way. DSC, dynamic susceptibility contrast; MR-PET, magnetic resonance-positron emmision imaging; T₂WI T₂ weighted imaging.

Figure 10.

Patchy enhancement (notched arrow in A) in the left frontal precentral area in a patient with anaplastic oligodendroglioma 6 months after treatment end. The choline metabolic map (B) from MR spectroscopy (Chemical Shift Imaging) shows significantly high amount of this metabolite (notched arrow and curved arrow) in the tumour area indicating possible recurrence. The DCE-Ktrans map (C) shows that the precentral area has increased permeability values (notched arrow) and the hybrid MR-PET imaging with methionine (notched arrow in the fused PET with T2FLAIR images in D) corroborates the suspected tumour recurrence. Further avid tracer uptake anterior to the enhancing lesion (arrow in C) denotes tumour relapse without disruption of the blood-brain-barrier or increased neo-angiogenesis. DCE, dynamic contrast enhancement; FLAIR, fluid-attenuated inversion recovery; MR-PET, magnetic resonance-positron emission tomography.

Figure 11.

Multiparametric hybrid MR-PET imaging in a patient with IDH-mutant, WHO Grade 3 astrocytoma on the left frontal lobe treated with surgery and adjuvant chemoradiation. Progressive signal changes in the T2FLAIR images (not shown) with some tiny patchy enhancement (A) prompted the comprehensive MR-PET exam. Pathological changes related to neovascularity are visible in the permeability map (B) and intravascular blood volume (C) derived from the DCE with slight incongruence in their spatial distribution zone, reflecting the different pathophysiology that these biomarkers may capture. The intravascular blood volume map is more in line with the DSC-derived rCBV map (D) and the blood flow map from the arterial spin labelling measurement (E). (F) The fused post-contrast T₁WI with the FET-PET image shows avid tracer uptake in the area of hypervascularity. FLAIR, fluid-attenuated inversion recovery; PET, positron emission tomography; T₁WI T₁ weighted imaging.

Figure 12.

Multiparametric MRI in a patient with GBM on the left frontal lobe treated with surgery and adjuvant radiation with concurrent temozolomide. The arrows on the images in the upper row demonstrate the ring-enhancing treated tumour bed (A) with strong diffusion attenuation (C, E) due to presumable necrotic material and blood and punctate increase in the relative cerebral blood volume (G) and permeability-Ktrans (J) maps. The interstitial volume map (L) shows significant, treatment-related increase in the extravascular extracellular space (ve). The adjacent slice (B) shows punctate enhancement (notched arrow) with strong diffusion attenuation and perifocal oedema (D,F). The enhancing focus shows also increased vascularity in the relative cerebral blood volume (H) and permeability- Ktrans (K) maps, though its conspicuity is less on the DSC imaging (notched arrows). The lesion demonstrates also pathological “ve” values (M). The final diagnosis suggests predominantly treatment-related changes with foci of tumour residual/recurrent disease. DSC, dynamic susceptibility contrast.