Longitudinal structural and perfusion MRI enhanced by machine learning outperforms standalone modalities and radiological expertise in high-grade glioma surveillance

Abstract

Purpose Surveillance of patients with high-grade glioma (HGG) and identification of disease progression remain a major challenge in neurooncology. This study aimed to develop a support vector machine (SVM) classifier, employing combined longitudinal structural and perfusion MRI studies, to classify between stable disease, pseudoprogression and progressive disease (3-class problem). Methods Study participants were separated into two groups: group I (total cohort: 64 patients) with a single DSC time point and group II (19 patients) with longitudinal DSC time points (2-3). We retrospectively analysed 269 structural MRI and 92 dynamic susceptibility contrast perfusion (DSC) MRI scans. The SVM classifier was trained using all available MRI studies for each group. Classification accuracy was assessed for different feature dataset and time point combinations and compared to radiologists’ classifications. Results SVM classification based on combined perfusion and structural features outperformed radiologists’ classification across all groups. For the identification of progressive disease, use of combined features and longitudinal DSC time points improved classification performance (lowest error rate 1.6%). Optimal performance was observed in group II (multiple time points) with SVM sensitivity/specificity/accuracy of 100/91.67/94.7% (first time point analysis) and 85.71/100/94.7% (longitudinal analysis), compared to 60/78/68% and 70/90/84.2% for the respective radiologist classifications. In group I (single time point), the SVM classifier also outperformed radiologists’ classifications with sensitivity/specificity/accuracy of 86.49/75.00/81.53% (SVM) compared to 75.7/68.9/73.84% (radiologists). Conclusion Our results indicate that utilisation of a machine learning (SVM) classifier based on analysis of longitudinal perfusion time points and combined structural and perfusion features significantly enhances classification outcome (p value= 0.0001).

Hemodynamic Imaging in Cerebral Diffuse Glioma—Part B: Molecular Correlates, Treatment Effect Monitoring, Prognosis, and Future Directions

“…This rapidly growing area of research has also adapted to using the data obtained from hemodynamic imaging modalities such as DSC [ 213 ], DCE [ 214 ], and ASL [ 215 ]. The use of perfusion imaging-derived radiomic features has shown promising feasibility in predicting MGMT promoter methylation [ 34 ] and IDH mutation status [ 216 , 217 , 218 , 219 ] and for improving the differential diagnosis of gliomas [ 220 ], the diagnostic performance of tumor grading [ 215 , 221 ] as well as pseudoprogression [ 222 , 223 , 224 ], but also prognostication [ 225 ]. Perfusional tumor heterogeneity can also be used to extract the radiomic features needed to train deep learning models for the prediction of glioblastoma recurrence patterns [ 226 ].…”

Treatment Response and Prognosis Evaluation in High‐Grade Glioma: An Imaging Review Based on MRI

“…Other studies have also proved that based on the combination of structure and perfusion MRI multiparameters, the SVM classifier can be used to analyze multiple time points of longitudinal perfusion to classify stable disease, PsP, and disease progression. The sensitivity/specificity/accuracy of SVM were 100%/91.67%/94.7% (first time point analysis) and 85.71%/100%/94.7% (longitudinal analysis), respectively; both are higher than the classification efficiency of radiologists 62 . In addition to the commonly used SVM classifiers, some studies used unsupervised learning methods, volume‐weighted voxel‐based multiparametric clustering (VVMC) methods, to distinguish between TP and PsP within 3 months of radiotherapy and chemotherapy, and to enhance radiation necrosis and lesion recurrence after 3 months.…”

“…The sensitivity/specificity/accuracy of SVM were 100%/91.67%/94.7% (first time point analysis) and 85.71%/100%/94.7% (longitudinal analysis), respectively; both are higher than the classification efficiency of radiologists. 62 In addition to the commonly used SVM classifiers, some studies used unsupervised learning methods, volume-weighted voxel-based multiparametric clustering (VVMC) methods, to distinguish between TP and PsP within 3 months of radiotherapy and chemotherapy, and to enhance radiation necrosis and lesion recurrence after 3 months. Compared with single-parameter (ADC, nCBV) measurement methods, VVMC is a more reproducible imaging biomarker, and VVMC had the highest reader consistency (within-class correlation coefficient, 0.87-0.89).…”

Application of Artificial Intelligence in Diagnosis of Craniopharyngioma

“…Aiming to solve the non-standardization of MRI intensity, Chen et al ( 43 ) used bias correction and Z-score normalization which were implemented in Statistical Parametric Mapping (SPM12) for pre-processing. Siakallis et al ( 44 ) performed log-transformation, normalization, bias field correction and intensity matching on skull stripping images. Kandemirli et al ( 45 ) conducted pre-processing operations including gray-level normalization and discretization.…”