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

Loizos Siakallis,Laurens Topff,Rolf Jager,Carole H Sudre,Jeremy Rees,Paul Mulholland,Sotirios Bisdas,Naomi Fersht,M Jorge Cardoso,Jasmina Panovska-Griffiths,Steffi Thust

doi:10.1007/s00234-021-02719-6

Abstract

PurposeSurveillance 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).MethodsStudy 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.ResultsSVM 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).ConclusionOur 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).

Highlights

High-grade gliomas (HGGs) are the most common type of malignant primary brain tumours, representing 80% of newly diagnosed cases, the majority of which (>50%) correspond to glioblastoma (GB) [1, 2]
This study comparatively assessed the performance of an support vector machine (SVM) classifier for the differentiation between progressive disease (PD), stable disease (SD) and PSP during post-treatment surveillance of patients with highgrade glioma (HGG)
Our results demonstrate improved SVM classification performance following the application of combined perfusion and structural MRI features and the introduction of longitudinal perfusion time points

Summary

Introduction

High-grade gliomas (HGGs) are the most common type of malignant primary brain tumours, representing 80% of newly diagnosed cases, the majority of which (>50%) correspond to glioblastoma (GB) [1, 2]. Current reference standard treatment includes maximal safe resection, radiation therapy and concurrent temozolomide (TMZ) [2, 3]. Poor prognosis and heterogeneous response to treatment warrant imaging surveillance for these patients [4]. Differentiation of progressive disease (PD) from pseudoprogression (PsP) remains critical for patient management [5]. Structural MRI, even under evolving diagnostic criteria, has been inefficient to reliably differentiate PD from PsP [2, 5,6,7]. Perfusion MRI has been previously shown to improve this classification [8,9,10]. A recent meta-analysis on the differentiation between PD and PsP by DSC MR perfusion indicates a pooled sensitivity and specificity of 90% and 88% respectively [11]

Objectives

Methods

Results

Conclusion