A Tumour-Aware Multimodal Framework for Explainable Prediction of IDH Mutation in Gliomas from Multi-Sequence MRI
Authors: Tahsien Al-Quraishi, Waseem Alwan Zaboon, O. S. Albahri, Raza Nowrozy
Organizations: School of Business, Melbourne Institute of Technology, Melbourne, 3000, Victoria, Australia · Al-Ayen Iraqi University, Nasiriyah, Iraq · Mechanical Engineering Department, College of Engineering, Wasit University, Al Kut, Iraq · Federation University Australia, Ballarat, Victoria, Australia · College of Art, Business, Law, Education and IT (CoABLEIT), Victoria University, Melbourne, 3000, Victoria, Australia
Accurate prediction of isocitrate dehydrogenase (IDH) mutation status is essential for glioma diagnosis, prognosis, and treatment planning. Although recent deep learning approaches have demonstrated promising performance, many rely on whole-brain feature extraction and provide limited interpretability. This study proposes a tumour-aware multimodal framework for explainable, non-invasive prediction of IDH mutation status from multiparametric magnetic resonance imaging (MRI). The proposed framework integrates four MRI sequences (T1, T1CE, T2, and FLAIR), tumour subregion segmentation, quantitative radiomic descriptors, and clinical information within a unified deep learning architecture. An ablation study was conducted to evaluate the contribution of tumour-aware learning, radiomic features, and clinical variables. Model performance was assessed using multiple complementary evaluation metrics, including accuracy, precision, recall, F1-score, balanced accuracy, receiver operating characteristic analysis, confusion matrix evaluation, and Grad-CAM visualisation. Experimental results demonstrated that the proposed multimodal framework consistently outperformed the baseline MRI-only model, with the Clinical Fusion model achieving the strongest overall predictive performance. The integration of tumour-aware learning, radiomic descriptors, and clinical information progressively improved classification performance, while Grad-CAM visualisations indicated that the model predominantly focused on biologically relevant tumour regions. These findings suggest that tumour-aware multimodal feature integration enables accurate, interpretable, and non-invasive molecular profiling of gliomas and provides a promising foundation for future clinical decision support in precision neuro-oncology.
Figures & tables
Study
Year
DL
TA
Rad
Clin
XAI
EV
Key Innovation
Choi et al. [ 25 ]
2021
✓
✓
✓
×
×
✓
Fully automated CNN–radiomics framework
Cluceru et al. [ 26 ]
2022
✓
×
×
×
×
×
Deep learning for WHO 2021 molecular glioma classification
Truong et al. [ 37 ]
2024
△
✓
✓
×
×
✓
Two-stage multicontrast MRI radiomics
Lost et al. [ 38 ]
2024
×
×
✓
×
×
✓
Externally validated radiomics model
Siddiqui et al. [ 31 ]
2025
✓
×
△
×
×
×
Hybrid whole-brain MRI framework
Niu et al. [ 28 ]
2025
✓
×
✓
×
×
✓
Transformer–radiomics fusion
Table 1: Comparison of MRI-based IDH mutation prediction methods
Characteristic
Value
Original cohort size
625
Excluded (missing IDH status)
3
Final cohort size
622
Age (years)
55.0 ± 15.5 (18–85)
Male
369 (59.3%)
Female
253 (40.7%)
Table 2: Characteristics of the study cohort
Figure 1: Distribution of IDH mutation status in the study cohort
Figure 2: Age distribution of patients included in the study cohort.
Figure 3: Distribution of tumour grades in the study cohort.
Figure 4: Sex distribution of patients included in the study cohort.
Figure 5: Distribution of IDH mutation status across tumour grades.
Figure 6: Percentage of IDH mutation status within each tumour grade.
Figure 7: Representative tumour segmentation mask
Figure 8: Representative T1CE MRI and tumour segmentation mask
Figure 9: Overview of the proposed tumour-aware multimodal framework.
Model
Accuracy (%)
Precision (%)
Recall (%)
F1-score (%)
AUC
Baseline CNN
64.89
40.63
48.15
44.07
0.684
No-Grade Ensemble
80.85
62.86
81.48
70.97
0.895
Radiomics Fusion
80.85
62.16
85.19
71.88
0.917
Clinical Fusion
86.17
70.59
88.89
78.69
0.944
Table 3: Performance comparison of the evaluated models for IDH mutation prediction.
Configuration
Components Included
Baseline CNN
MRI only
Tumour-aware Ensemble
MRI + Tumour-aware learning
Radiomics Fusion
MRI + Tumour-aware learning + Radiomic features
Clinical Fusion (Proposed)
MRI + Tumour-aware learning + Radiomic features + Clinical variables
Table 4: Ablation configurations of the proposed multimodal framework.
Figure 10: ROC curves of the evaluated models for IDH mutation prediction.
Figure 11: Confusion matrix of the proposed Clinical Fusion model
Figure 12: Tumour-focused saliency analysis of the proposed Clinical Fusion model.
Abbreviation
Definition
AI
Artificial Intelligence
AUC
Area Under the Receiver Operating Characteristic Curve
BraTS
Brain Tumor Segmentation Challenge
CNN
Convolutional Neural Network
FLAIR
Fluid-Attenuated Inversion Recovery
F1-score
Harmonic Mean of Precision and Recall
Table 5: List of abbreviations used in this study.
Multimodal deep learning has improved prognostic accuracy for brain tumours by integrating histopathology and genomic data, yet the contribution of volumetric MRI within unified survival frameworks remains unexplored. This pilot study extends a bimodal framework by incorporating Fluid Attenuated Inversion Recovery (FLAIR) MRI from BraTS2021 as a third modality. Using the TCGA-GBMLGG cohort (664 patients), we evaluate three unimodal models, nine bimodal configurations, and three trimodal configurations across early, late, and joint fusion strategies. In this small cohort setting, trimodal early fusion achieves an exploratory Composite Score (CS = 0.854), with a controlled ΔCS of +0.011 over the bimodal baseline on identical patients, though this difference is not statistically significant (p = 0.250, permutation test). MRI achieves reasonable unimodal discrimination (CS = 0.755) but does not substantially improve bimodal pairs, while providing measurable uplift in the three-way combination. All MRI containing experiments are constrained to 19 test patients, yielding wide bootstrap confidence intervals (e.g. [0.400,1.000]) that preclude definitive conclusions. These findings provide preliminary evidence that a third imaging modality may add prognostic value even with limited sample sizes, and that additional modalities require sufficient multimodal context to contribute effectively.
Iain Swift, JingHua Ye
Department of Computer Science, Munster Technological University, Cork, Ireland
Clinicians diagnose brain tumors by synthesizing patient symptoms, medical history, and quantitative imaging data from modalities such as MRI and CT scans into a unified clinical judgement. However, most deep learning models rely on MRI/CT images alone, failing to replicate the clinicians multimodal reasoning. We explore a two-branch multimodal network combining raw MRI scans with 91 extracted radiomic features (intensity, texture, shape, and boundary descriptors) to classify brain tumors into glioma, meningioma, pituitary, and no-tumor. A pre-trained CNN backbone encodes the image stream, whereas a dedicated MLP encodes the radiomic stream. Both streams are fused via concatenation, gated, or bidirectional cross-modal attention strategies. Across nine experimental runs on a balanced 7,200 image dataset, all multimodal configurations outperform unimodal baselines with gated fusion achieving the best accuracy of 96.13%.
Wajih ul Islam, Muhammad Yaqoob, Javed Ali Khan +1
School of Physics, Engineering and Computer Science, University of Hertfordshire, UK
Precise molecular subtyping of gliomas, including isocitrate dehydrogenase (IDH) mutation and 1p/19q codeletion, directly guides surgical and therapeutic decisions, yet currently relies on invasive tissue sampling. Deep learning on structural MRI has emerged as a non-invasive alternative, but anatomy-only approaches cannot capture the hemodynamic signatures that distinguish molecular subtypes. Radiogenomics based on dynamic susceptibility contrast (DSC) MRI holds immense potential for non-invasively characterizing glioma molecular subtypes, yet clinical deployment has been hindered by inter-site variability and the limitations of voxel-wise analysis. We introduce HiPerfGNN, a framework that first learns discrete hemodynamic representations from raw time-intensity curves using a vector-quantized variational autoencoder (VQ-VAE). These quantized perfusion codes define coarse-level graph nodes representing functional tumor habitats, each of which is hierarchically subdivided into fine-level subregions guided by structural MRI. A hierarchical graph neural network then propagates information across scales for molecular prediction. On an internal cohort (n=475), the model achieved AUCs of 0.96 (IDH), 0.89 (1p/19q), and 0.84 (WHO grade), and maintained robust IDH performance (AUC 0.89) on an independent external cohort (n=397) without recalibration. Gradient-based saliency analysis confirms biologically grounded attention patterns aligned with known glioma pathophysiology. Our results demonstrate the added value of integrating perfusion dynamics into radiogenomic pipelines for glioma molecular subtyping. Code is available at https://github.com/janghana/HiPerfGNN.
Han Jang, Junhyeok Lee, Heeseong Eum +4
Interdisciplinary Program in Bioengineering, Seoul National University, Seoul, South Korea · Interdisciplinary Program in Cancer Biology, Seoul National University College of Medicine, Seoul, South Korea · Dept. of Biomedical Sciences, Seoul National University, Seoul, South Korea +4