Distinguishing between paediatric brain tumour types using multi-parametric magnetic resonance imaging and machine learning: A multi-site study.

James T Grist,Stephanie Withey,Dipayan Mitra,Stephen Powell,Laurence Abernethy,Shivaram Avula,Andrew C Peet,Richard Grundy,Barry Pizer,Adam Oates,Dorothee P Auer,Simon Bailey,Jan Novak,Lesley Macpherson,Theodoros N Arvanitis

doi:10.1016/j.nicl.2020.102172

Abstract

The imaging and subsequent accurate diagnosis of paediatric brain tumours presents a radiological challenge, with magnetic resonance imaging playing a key role in providing tumour specific imaging information. Diffusion weighted and perfusion imaging are commonly used to aid the non-invasive diagnosis of children's brain tumours, but are usually evaluated by expert qualitative review. Quantitative studies are mainly single centre and single modality. The aim of this work was to combine multi-centre diffusion and perfusion imaging, with machine learning, to develop machine learning based classifiers to discriminate between three common paediatric tumour types. The results show that diffusion and perfusion weighted imaging of both the tumour and whole brain provide significant features which differ between tumour types, and that combining these features gives the optimal machine learning classifier with >80% predictive precision. This work represents a step forward to aid in the non-invasive diagnosis of paediatric brain tumours, using advanced clinical imaging.

Highlights

Brain tumours are the most common solid tumours in children, accounting for approximately 25% of all childhood cancers
The results show that diffusion and perfusion weighted imaging of both the tumour and whole brain provide significant features which differ between tumour types, and that combining these features gives the optimal machine learning classifier with >80% predictive precision
This study has demonstrated that a combination of multiparametric Magnetic resonance imaging (MRI), univariate analysis, and machine learning techniques can be employed to distinguish between both high- and low-grade paediatric brain tumours, as well as enabling tumour type classification with high accuracy

Summary

Introduction

Brain tumours are the most common solid tumours in children, accounting for approximately 25% of all childhood cancers. Magnetic resonance imaging (MRI) is commonly performed for children suspected of having a brain tumour at presentation. Challenges are faced by paediatric radiologists to diagnose paediatric brain tumour type using MRI, especially in tumours which do not enhance with gadolinium contrast agent (a significant fraction in paediatric radiology) (Koob and Girard, 2014). If a combination of imaging methods can be used to quantify tumour cellular microstructure and perfusion, it may be possible to discriminate between low and high

Objectives

Methods

Results

Conclusion