Multicenter DSC–MRI-Based Radiomics Predict IDH Mutation in Gliomas
by
, , ,
Georgios C. Manikis
1,*
,
Georgios S. Ioannidis
1
,
Loizos Siakallis
2,
Katerina Nikiforaki
1,
Michael Iv
3,
Diana Vozlic
4,5,
Katarina Surlan-Popovic
4,5,
Max Wintermark
3,
Sotirios Bisdas
2,6,† and
Kostas Marias
1,7,†
1
Computational BioMedicine Laboratory (CBML), Foundation for Research and Technology—Hellas (FORTH), 70013 Heraklion, Greece
2
Department of Brain Repair and Rehabilitation, Queen Square Institute of Neurology, UCL, London WC1N 3BG, UK
3
Division of Neuroimaging and Neurointervention, Department of Radiology, Stanford University, Stanford, CA 94305, USA
4
Department of Radiology, Faculty of Medicine, University of Ljubljana, 1000 Ljubljana, Slovenia
5
Department of Neuroradiology, University Medical Centre, 1000 Ljubljana, Slovenia
6
Department of Neuroradiology, The National Hospital for Neurology and Neurosurgery, University College London NHS Foundation Trust, London WC1N 3BG, UK
7
Department of Electrical & Computer Engineering, Hellenic Mediterranean University, 71410 Heraklion, Greece
*
Author to whom correspondence should be addressed.
†
These authors contributed equally as last authors.
Cancers 2021, 13(16), 3965; https://doi.org/10.3390/cancers13163965
Submission received: 24 June 2021
/
Revised: 25 July 2021
/
Accepted: 31 July 2021
/
Published: 5 August 2021
(This article belongs to the Special Issue Radiomics in Brain Tumor Imaging)
Simple Summary
Significant efforts have been put toward developing MRI-based radiogenomics for IDH status subtyping predictions; however, in the vast majority of these approaches, the external validation sets are absent. Another limitation in current studies is the lack of explainability in radiomics models, which hampers clinical trust and translation. Motivated by these challenges, we proposed a multicenter DSC–MRI-based radiomics study based on an independent exploratory set, which was externally validated on two independent cohorts, for IDH mutation status prediction. Our results demonstrated that DSC–MRI radiogenomics in gliomas, coupled with dynamic-based image standardization techniques, hold the potential to provide (a) increased predictive performance by offering models that generalize well, (b) reasoning behind the IDH mutation status predictions, and (c) interpretability of the radiomics features’ impacts in model performance.
Abstract
To address the current lack of dynamic susceptibility contrast magnetic resonance imaging (DSC–MRI)-based radiomics to predict isocitrate dehydrogenase (IDH) mutations in gliomas, we present a multicenter study that featured an independent exploratory set for radiomics model development and external validation using two independent cohorts. The maximum performance of the IDH mutation status prediction on the validation set had an accuracy of 0.544 (Cohen’s kappa: 0.145, F1-score: 0.415, area under the curve-AUC: 0.639, sensitivity: 0.733, specificity: 0.491), which significantly improved to an accuracy of 0.706 (Cohen’s kappa: 0.282, F1-score: 0.474, AUC: 0.667, sensitivity: 0.6, specificity: 0.736) when dynamic-based standardization of the images was performed prior to the radiomics. Model explainability using local interpretable model-agnostic explanations (LIME) and Shapley additive explanations (SHAP) revealed potential intuitive correlations between the IDH–wildtype increased heterogeneity and the texture complexity. These results strengthened our hypothesis that DSC–MRI radiogenomics in gliomas hold the potential to provide increased predictive performance from models that generalize well and provide understandable patterns between IDH mutation status and the extracted features toward enabling the clinical translation of radiogenomics in neuro-oncology.
