Radiological phenotyping of IDH mutation status in gliomas using dynamic susceptibility contrast perfusion-weighted MRI

AbstractBackgroundMachine learning assisted MRI radiomics, which combines MRI techniques with machine learning methodology, is rapidly gaining attention as a promising method for staging of brain gliomas. This study assesses the diagnostic value of such framework applied to dynamic susceptibility contrast (DSC)-MRI in classifying treatment-naïve gliomas from a multi-center patient pool into WHO grades II-IV and across their isocitrate dehydrogenase (IDH) mutation status.Methods333 patients from 6 tertiary centres, diagnosed histologically and molecularly with primary gliomas (IDH-mutant=151 or IDH-wildtype=182) were retrospectively identified. Raw DSC-MRI data was post-processed for normalised leakage-corrected relative cerebral blood volume (rCBV) maps. Shape, intensity distribution (histogram) and rotational invariant Haralick texture features over the tumour mask were extracted. Differences in extracted features between IDH-wildtype and IDH-mutant gliomas and across three glioma grades were tested using the Wilcoxon two-sample test. A random forest algorithm was employed (2-fold cross-validation, 250 repeats) to predict grades or mutation status using the extracted features.ResultsFeatures from all types (shape, distribution, texture) showed significant differences across mutation status. WHO grade II-III differentiation was mostly driven by shape features while texture and intensity feature were more relevant for the III-IV separation. Increased number of features became significant when differentiating grades further apart from one another. Gliomas were correctly stratified by IDH mutation status in 71% of the cases and by grade in 53% of the cases. In addition, 87% of the gliomas grades predicted with an error distance up to 1.ConclusionDespite large heterogeneity in the multi-center dataset, machine learning assisted DSC-MRI radiomics hold potential to address the inherent variability and presents a promising approach for non-invasive glioma molecular subtyping and grading.Key points-On highly heterogenous, multi-centre data, machine learning on DSC-MRI features can correctly predict glioma IDH subtyping in 71% of cases and glioma grade II-IV in 53% of the cases (87% <1 grade difference)-Shape features distinguish best grade II from grade III gliomas.-Texture and distribution features distinguish best grade III from grade IV tumours.Importance of studyThis work illustrates the diagnostic value of combining machine learning and dynamic susceptibility contrast-enhanced MRI (DSC-MRI) radiomics in classifying gliomas into WHO grades II-IV as well as across their isocitrate dehydrogenase (IDH) mutation status. Despite the data heterogeneity inherent to the multi-centre design of the studied cohort (333 subjects, 6 centres) that greatly increases the theoretical challenges of machine learning frameworks, good classification performance (accuracy of 53% across grades (87% <1 grade difference) and 71% across mutation status) was obtained. Therefore, our results provide a proof-of-concept for this emerging precision medicine field that has good generalisability and scalability properties. Introspection on the classification errors highlighted mostly borderline cases and helped underline the challenges of a categorical classification in a pathological continuum.With its strong generalisability property, its ability to further incorporate participating centres and its possible use to identify borderline cases, the proposed machine learning framework has the potential to contribute to the clinical translation of machine-learning assisted diagnostic tools in neuro-oncology.

Download Full-text

P04.11 Voxelwise correlation between vascular MRI parameters obtained with ASL and DSC is decreased in IDH-wt non-enhancing glioma

Neuro-Oncology ◽

10.1093/neuonc/noz126.106 ◽

2019 ◽

Vol 21 (Supplement_3) ◽

pp. iii31-iii31

Author(s):

E A H Warnert ◽

F Incekara ◽

A J P E Vincent ◽

J W Schouten ◽

M J van den Bent ◽

...

Keyword(s):

Cerebral Blood Volume ◽

Correlation Coefficients ◽

Region Of Interest ◽

Dynamic Susceptibility ◽

Idh Mutation ◽

Dynamic Susceptibility Contrast ◽

Mutation Status ◽

Susceptibility Contrast ◽

Perfusion Measurements ◽

Future Work

Abstract BACKGROUND Comparative studies of dynamic susceptibility contrast (DSC) based measurement of cerebral blood volume (CBV) or cerebral blood flow (CBF) and arterial spin labelling (ASL) based measurement of CBF have previously shown good correlation of these parameters in human glioma. However, these studies were mostly done before inclusion of the mutation status of the isocitrate dehydrogenase (IDH) encoding gene in brain tumour classification. In light of the call for gadolinium-free imaging, here we investigate the effect of IDH-mutation status on the correlation between ASL and DSC-based perfusion measurements in non-enhancing glioma. MATERIAL AND METHODS Twenty-two patients with non-enhancing glioma and confirmed IDH-mutation status (next generation sequencing, 6 IDH-wildtype and 16 IDH-mutated) underwent 3T MRI scanning (GE, Milwaukee, WI, USA). Image acquisition included a 3D spiral pseudocontinuous ASL with time-encoded labelling (7 effective label delays from 0.8 to 2 s, reconstruction matrix 128x128x42, resolution 1.9x1.9x3.5 mm3), and 2D DSC imaging (122 TRs, TR/TE 1500m/18.6ms, 15 slices, voxel size: 1.88x1.88x4 mm3) in which a bolus of 7.5ml of gadolinium-based contrast agent (Gadovist, Bayer, Leverkussen, GE) was injected. A pre-load bolus of equal size was given 5 minutes prior to DSC imaging. DSC and ASL images were motion corrected and linearly registered to high resolution FLAIR images (FSL, version 5.0.9, Oxford, UK). DSC-relative CBV (rCBV), DSC-relative CBF (rCBF), and ASL-CBF maps were calculated via previously described methods. The glioma region of interest (ROI) was determined via manual segmentation on the FLAIR images. Voxel-wise Pearson’s linear correlation coefficients (ρ) within this ROI were calculated between ASL-CBF and DSC-rCBV, and between ASL-CBF and DSC-rCBF. RESULTS Normalised histograms indicate that IDH-wt glioma has higher values for ASL-CBF, DSC-rCBV, and DSC-rCBF than IDH-mutated glioma. IDH-wildtype glioma has a significantly lower ρ ASL-CBF vs DSC-rCBV and ρ ASL-CBF vs DSC-rCBF than IDH-mutated glioma (two-sample t-tests p < 0.05). CONCLUSION IDH-mutation status of non-enhancing glioma potentially affects the correlation between ASL-CBF and DSC-rCBF/rCBV and should be taken into account when moving towards ASL-only imaging. The decreased correlation between ASL and DSC-based vascular parameters in IDH-wt gliomas may be due to more aggressive vasculature in subtypes of IDH-wt tumours. Future work includes expansion of the current patient cohort (part of the ongoing iGENE study).

Download Full-text