SS 711 - Brain tumours: lesion characterisation and treatment evaluation
SS 711 - Brain tumours: lesion characterisation and treatment evaluationThursday, March 2, 14:00 - 15:30 Room: E2 Session Type: Scientific Session Topics: Oncologic Imaging, Neuro Moderators: M. A. Lucic (Sremska Kamenica/RS), S. Thust (London/UK) Add session to my schedule In your schedule (remove)
Purpose: Vascularity correlates with shortened survival. We aimed to determine whether the tumoural vascular pattern on MRI (TVP-MRI) of newly diagnosed glioblastomas is useful in predicting survival, and to correlate TVP-MRI with dynamic susceptibility contrast (DSC) perfusion and diffusion parameters for contrast-enhancing lesion (CEL) and surrounding non-CEL.
Methods and Materials: Ninety-seven patients (59 men; mean age, 61 years) with histologically proven glioblastoma underwent 1.5T-MRI including anatomical sequences, first-pass DSC images and post-contrast T1-weighted SE images after gadobutrol (Gadovist; Bayer Schering Pharma, Berlin, Germany) at 0.1 mmol/kg) with a 1 mm isometric voxel. Volumes of interest for CEL, non-CEL, and contralateral tissue were obtained for relative cerebral blood volume (rCBV), relative cerebral blood flow (rCBF), delay time (DT), mean temporal maximal intensity projection (tMIP), and apparent diffusion coefficient (ADC) using Olea Sphere V.3.0 software (Olea Medical, La Ciotat, France). Tumours over 5 vessels were classified as hypervascular (hyper-TVP-MRI) on post-contrast 1 mm-T1SE images. Prognostic factors were evaluated by Kaplan-Meier survival and Cox proportional hazards analysis.
Results: Fifty-five (56.7%) glioblastomas were hyper-TVP-MRI. Patients with hyper-TVP-MRI, mean age, volume-CEL, ADC-CEL, DT-CEL, rCBV-CEL, rCBV-nonCEL and tMIP-CEL were higher and DT-nonCEL lower. Mean survival for hypo- and hyper-TVP-MRI glioblastomas treated with surgery and with radiotherapy plus chemotherapy was 12.9±7.7 and 8.3±6.9 months, respectively. TVP-MRI was the best survival predictor for glioblastoma at 1 year (AUC 0.84, 88.5% sensitivity, 77.9% specificity, 82.3% positive predictive value, 87.8% negative predictive value).
Conclusion: TVP-MRI might be a new and promising potential biomarker for predicting survival in newly diagnosed glioblastoma.
Medulloblastoma in adults: identifying imaging biomarkers of genetic status in a prospective multi-centre study
Purpose: Medulloblastomas are rare in adults. Histopathological and molecular profiles are known to influence the course of disease. This study is a first radiogenomic approach to identify MR imaging biomarkers of molecular subtypes in adults with medulloblastomas, which may facilitate pre-operative tumor assessment and elucidate differences between adult and pediatric tumors. This is a subanalysis of the multi-center NOA-07 study.
Methods and Materials: Between 2008 and 2014, 18 neuro-oncological centers recruited 28 patients above 21 years with confirmed and molecularly sub-grouped medulloblastomas and full pre-operative MRI datasets (T1w, T2w, FLAIR and contrast-enhanced T1w sequences). Morphological characteristics (predefined as: enhancement pattern, T1 and T2 signal, hemorrhage) and further tumor and edema volumes, location, Chang criteria and hydrocephalus were evaluated by 3 experienced neuroradiologists. These MR findings were matched to histological and molecular subtypes (Fischer and ANOVA testing). Results were compared to findings from the only radiogenomic medulloblastoma study in children (S Perreault et al., AJNR 2014).
Results: Relation to the 4th ventricle and lower rhombic limb, edema volume, CNS metastases, non-enhancing tumor volume (Chang larger 2) and hydrocephalus differed significantly between molecular subtypes (P=0.01 to 0.04 respectively). None of these criteria matched with corresponding findings in pediatric medulloblastomas. Perifocal edema was larger in the desmoplastic subtype (P=0.01). Except for hemorrhage, none of the predefined MR-morphological characteristics could differentiate between the molecular or histopathological subtypes.
Conclusion: There are indicators that prognostically different molecular subtypes of adult medulloblastoma can be preoperatively identified based on MRI imaging. These seem distinct from findings in children.
Apparent diffusion coefficient and permeability parameters from dynamic contrast-enhanced perfusion MR imaging: preliminary correlation study with glioma genetic profiles
Purpose: Preliminary evaluation of the efficacy of diffusion and DCE-MR imaging in predicting major genetic alterations in glioma.
Methods and Materials: 30 gliomas (male: 19, female: 11; mean age: 37.3 y) were retrospectively included and they all underwent DWI and DCE-MRI. ADC and permeability parameters (Ktrans, Ve, Kep and iAUC) maps were generated. 5~10 ROIs were placed on tumour parenchyma. The max, min, mean value of ADC and the median, mean value of permeability parameters were recorded. Four major glioma genetic alterations (IDH-1; n=27), (MGMT; n=26), (ATRX; n=27) and Ki-67 (n=29) were included and their associations with ADC and permeability parameters were analysed.
Results: The max and mean value of ADC and mean value of Ve and iAUC in IDH-1 mutated group were significantly higher than those of the IDH-1 negative group (P<0.040) and the max value of ADC showed the highest diagnostic value; The mean value of ADC and Ve demonstrated the highest sensitivity (96.8%) and specificity (94.8%), respectively. Compared with ATRX-negative group, the max and mean value of ADC also significantly higher (P<0.001) and it demonstrated the highest diagnostic value. Further, in the analysis of MGMT-positive group, the mean/median value of Ktrans, Kep, ADC and the max/mean value of ADC were significantly higher; in addition, we found that only Kep and ADC values were negatively correlated with Ki-67 (P<0.022).
Conclusion: Tumours with mutated IDH-1, ATRX and MGMT prone to have a higher permeability and ADC value. Thus, DWI and DCE-MRI, especially DWI, might be helpful in predicting genetic alterations.
Purpose: The objective of this study was to assess the diagnostic value of integrated 11C-methionine positron emission tomography/magnetic resonance imaging (methionine PET/MRI) and MRI with magnetic resonance spectroscopy (MRS) for suspected primary brain tumour.
Methods and Materials: Thirty-four consecutive patients with suspected primary brain tumour were prospectively enrolled for an integrated 11C-methionine PET-MRI with MRS. 5 Spectra were non diagnostic (25.6%); therefore, 29 spectroscopic data sets were available for definitive evaluation (85.3%). Reference standard was the histopathological report in 21 patients, whereas in the remaining 8 patients follow-up imaging was used. Two radiologists evaluated the integrated PET/MRI data sets and the MRI datasets with MRS regarding the most likely diagnosis and diagnostic confidence on a 5-point Likert-scale.
Results: In 29 patients 30 lesions were detected and classified by the reference standard. Twenty-seven lesions were primary brain tumour: 15 low-grade glioma (LGG), 12 high-grade glioma (HGG). Three lesions were classified as non-oncologic lesions. MRI with MRS correctly identified 27 of the 30 lesions (90%), whereas integrated PET/MRI was correct in 26 of the 30 lesions (Fig 1). Diagnostic confidence was better for MRI with MRS vs. integrated PET/MRI (4.3±0.8 vs. 4.1±1.1), but did not reach significance (p < 0.17).
Conclusion: This is the first study comparing the clinical impact of integrated PET/MRI and MRI with MRS. MRI with MRS was more sensitive and reached a higher diagnostic confidence than integrated PET/MRI, even though it was not significant.
Purpose: Current limitations in brain imaging, especially the lack of tissue discrimination at the cellular level, lead to the inability to detect microscopic spread of cancerous tissue and to verify the survival of cancer cells following surgical resection, chemo- or radio-therapy. We used high sensitivity of X-rays Phase Contrast micro-computed tomography (PCI-CT) to non-invasively study the effects of a novel spatially fractionated single-stage high-dose radiotherapy technique, Microbeam Radiation Therapy (MRT), administered in-vivo to both healthy and cancerous rat brains.
Methods and Materials: 9 Fisher rats implanted with 9L gliosarcoma were treated with MRT (75 microns-thick micro-beams reaching up to 600Gy peak doses) and then imaged by PCI-CT (with 32 keV X-rays and an imaging CCD camera with voxel size of 8³ micron³).
Results: PCI-CT allows recognition and differentiation of brain anatomical details down to cellular level, and identification of single cancerous cells, microscopic cancerous cell-clusters far from the main lesion, tissue necrosis, neuronal degeneration, tumour oedema, high-density calcifications as well as micrometric MRT-transections. Furthermore, the technique permits the segmentation of full brain vessel networks down to blood micro-capillaries.
Conclusion: The ability of PCI-CT to replicate the accuracy of histology and to provide far more anatomical detail than MRI opens a fascinating novel avenue for brain imaging. This technique appears to be well suited for post-mortem studies of brain cancer dissemination and radiotherapy effects. In the future, PCI-CT may provide image-guidance during radiotherapy, e.g. by directing focused X-ray irradiation in the ablation of inlets of cancerous cells.
Assessment of intracranial meningioma-associated calcifications using susceptibility-weighted magnetic resonance imaging
Purpose: To determine the diagnostic performance of susceptibility-weighted magnetic resonance imaging (SW-MRI) for the detection of intracranial meningioma-associated calcifications compared to standard MR sequences, using computed tomography (CT) as a standard of reference.
Methods and Materials: MRI scans of 354 patients, who received a clinical brain MR with SW-MRI sequences between January 2014 and July 2016, were retrospectively evaluated and 316 patients were included. 50 patients had positive findings for intracranial meningioma-associated calcifications on CT scans. Calcification diameter was used to assess correlation between imaging modalities. Sensitivity and specificity as well as intra- and interobserver reliability were calculated for SW-MRI and standard MRI sequences.
Results: SW-MRI reached a sensitivity of 94% (95% CI: 83-99%) and a specificity of 95% (95% CI: 92-98%) for the detection of meningioma-associated calcifications, while standard MR yielded a sensitivity of 64% (95% CI: 49-77%) and a specificity of 94% (95% CI: 90-96%). Diameter measurements between SW-MRI and CT showed a close correlation (R2= 0.993, p<0.001) with a slight overestimation of size, which, however, did not reach significance level (SW-MRI: 8.2 mm ± 7.2; CT: 6.8 mm ± 6.4, p>0.05). Interobserver-agreement for size measurements of calcifications was high on SW-MRI (ĸ = 0.85, p < 0.0001) and fair on standard MRI (ĸ = 0.29, p < 0.001).
Conclusion: The combination of SW magnitude and phase images enables an accurate detection of intracranial meningioma-associated calcifications, using CT as standard of reference, and offers a higher sensitivity and specificity than standard MRI sequences.
The role of diffusion-weighted imaging for radiological identification of molecular subgroups of medulloblastoma in children
Purpose: Recent studies on genomic characterisation of medulloblastoma (MB) prompted the development of new classification combining histopathological and molecular approach.The identified subgroups have shown potential for improved risk stratification. The aim is to verify if DWI and ADC maps are able to predict new subgroups of MB.
Methods and Materials: Twenty-six patients with diagnosis of MB, with pretreatment MR imaging, histologic, genomic characterisation after surgery, were retrospectively selected. ADC maps were coregistered with T1w post-contrast and FLAIR images. ROIs were double-blindly selected (ADC): restricted, intermediate, not restricted areas and (post-contrast) absent, inhomogeneous, homogeneous areas. Standard t test was performed to evaluate differences among paired MB subgroups.
Results: No statistical significance was detected. A trend in differences was found for both restricted ADC values between WNT and group 3 (0.61+/- 0.36 vs 1.29+/- 0.10) and ADC values in areas of homogeneous contrast enhancement (0.44 +/- 0.26 vs 1.13 +/- 0.15). Mean values along with standard deviations: for areas of restricted ADC (WNT: 0.67+/- 0.03; SHH: 0.72 +/- 0.06; group 3: 0.81+/- 0.14; group 4: 0.74 +/- 0.11).
Conclusion: The study shows no statistically significant correlation between ADC values and T1 post-contrast between and within different MB subgroups.However, a trend was identified in restricted ADC value and ADC in areas of homogeneous contrast enhancement between WNT and group 3. The study demonstrates that it is not possible to predict the type of medulloblastoma based on ADC/contrast enhancement characteristics. A larger cohort of patients is needed to clarify their real potential to discriminate molecular MB subgroup before surgery as suggested by observed trend.
Purpose: Our purpose was to assess contrast-enhanced ultrasound (CEUS) capability to identify residual tumour mass during glioblastoma (GBM) surgery, in order to increase the extent of resection.
Methods and Materials: We prospectively evaluated 10 patients who underwent surgery for GBM removal with navigated ultrasound guidance. Navigated B-mode and CEUS were performed prior to resection, during resection and after complete tumour resection. Areas suspected for residual tumours on B-mode and CEUS were localised within the surgical field with navigated ultrasound and sent separately for histopathological analysis to confirm tumour presence.
Results: In all cases tumour remnants were visualised as hyperechoic areas on B-mode, highlighted as CEUS positive areas and confirmed as tumoural areas on hystopathological analysis. In 1 case only CEUS partially failed to demonstrate residual tumour because the residual hyperehoic area was devascularised prior to ultrasound contrast agent injection. In all cases CEUS enhanced B-mode findings.
Conclusion: As already shown in other neoplastic lesions in other organs CEUS is extremely specific in the identification of residual tumour. CEUS distinction between tumour and artefacts/normal brain on B-mode is based on its capability to show the vascularisation degree and not the echogenicity of the tissues. Therefore, CEUS can play a decisive role in the process of maximising GBM surgical resection.
Texture analysis on diffusion-tensor imaging: discriminating glioblastomas from single brain metastasis
Purpose: Texture analysis has been used to stage, differentiate and predict prognosis in many oncologic tumours. It has been used on CT, MRI and PET. The purpose of this study was to determine the diagnostic accuracy of discriminating glioblastoma (GBM) from single brain metastasis (MET) by assessing the heterogeneity of both the solid tumour and the peritumoural oedema with MRI texture analysis (MRTA).
Methods and Materials: Preoperative MRI examinations done on a 3T scanner of 44 patients were included, 23 GBM and 21 MET. MRTA was performed on the DTI in a representative ROI. The MRTA was assessed using a commercially available research software program (TexRAD) which applies a filtration histogram technique for characterising tumour and peritumoural heterogeneity. The filtration step selectively filters and extracts texture features at different anatomical scales varying from 2mm (fine) to 6mm (coarse). Heterogeneity quantification was obtained by the statistical parameter entropy. A threshold value to differentiate GBM from MET with sensitivity and specificity was calculated by receiver operating characteristics (ROC) analysis.
Results: Quantifying the heterogeneity of the solid part of the tumour showed no significant difference between GBM and MET. However the heterogeneity of the GBMs peritumoural oedema was significantly higher than the oedema surrounding MET, differentiating them with a sensitivity of 90% and specificity of 80%.
Conclusion: Assessing the peritumoural heterogeneity can increase the radiological diagnostic accuracy when discriminating GBM and MET. This will facilitate the medical staging and optimise the planning for surgical resection of the tumour and postoperative management.
Purpose: Definite detection of glioblastoma (GB) progression in follow-up MRIs is essential, both for clinical trials and everyday clinical routine. 3D-volumetry provides quantification of tumour extent, and might therefore facilitate unbiased disease assessment. Aim of the present study was to evaluate the utility of absolute changes in volume (delta) or regional, segmentation-based subtractions for detecting disease progression in longitudinal MRI follow-ups.
Methods and Materials: 165 high-resolution, 3-Tesla MRIs of 30 GB patients (23m, mean age 60y) were retrospectively included. Contrast enhancement (CV) and tumour-associated signal alterations in FLAIR images (FV) were semi-automatically segmented. Delta volumes (dCV, dFV) as well as regional subtractions (sCV, sFV) were assessed. For every follow-up, disease progression was classified according to the histopathologic results of resected specimen, decisions of the local multidisciplinary CNS tumor board and a consensus rating of the radiologic report.
Results: A generalized logistic mixed model for disease progression (yes/no) and the input variables dCV, dFV, sCV and sFV showed that only dCV was significantly associated with prediction of disease progression (P = .005). Delta volumes had a better accuracy than regional, segmentation-based subtractions (79 % versus 72 %) and a higher area under the curve in ROC analysis (.83 vs .75).
Conclusion: Absolute volume changes of the contrast enhancing tumour part were the most accurate volumetric determinant to detect progressive disease in GB follow-ups and outweighed FLAIR changes as well as regional image subtractions. This volumetric parameter might be valuable in upcoming objective response criteria for glioblastoma
Purpose: Neuroimaging plays an essential role in brain tumours resection which is a major prognostic factor in order to increase the survival rate. Contrast enhanced ultrasound (CEUS) using intravascular contrast media enhances image quality and allows visualization of tumour’s perfusion. Perfusion quantification is one of the newest application of CEUS and it may give a clue about tumour’s biological features and vascularisation. Even though this technique is already well established in other fields, it has never been applied to Neurosurgery before. The aim of the study is to evaluate quantitatively different histological types of brain tumours in order to detect specific perfusion patterns which may correlate with molecular parameters and with antiangiogenetic treatment response.
Methods and Materials: We evaluated 102 DICOM clips recorded during brain tumour excision with the perfusion quantification software Vuebox©.
Results: Only 27 video clips were amenable for quantitative analysis. However we found that some parameters, particularly PE, mTTI and FT, seem to be related with tumour’s grade and histology.
Conclusion: The use of Vuebox© underlined some conditions which may increase data variability or make the images inadequate for the quantitative evaluation. This allowed us to define a precise protocol for video clip acquisition which should make data loss less likely. Thanks to quantitative analysis we detected some differences in perfusion patterns between tumours with different histology.