Classification of paediatric brain tumours by diffusion weighted imaging and machine learning

Ján Novák ,Daniel Rodriguez Gutierrez,Andrew Peet,Shivaram Avula,Theodoros N Arvanitis ,Lesley Macpherson,Darren Hargrave,Patrick W Hales ,Adam Oates,Niloufar Zarinabad,Simon Bailey,Rajvir Kaur ,Christopher A Clark ,Dipayan Mitra,Benjamin Pinkey,Tim Jaspan,Richard G Grundy ,Heather Rose,Dorothee P Auer ,Laurence Abernethy,Nigel P Davies

doi:10.1038/s41598-021-82214-3

Abstract

To determine if apparent diffusion coefficients (ADC) can discriminate between posterior fossa brain tumours on a multicentre basis. A total of 124 paediatric patients with posterior fossa tumours (including 55 Medulloblastomas, 36 Pilocytic Astrocytomas and 26 Ependymomas) were scanned using diffusion weighted imaging across 12 different hospitals using a total of 18 different scanners. Apparent diffusion coefficient maps were produced and histogram data was extracted from tumour regions of interest. Total histograms and histogram metrics (mean, variance, skew, kurtosis and 10th, 20th and 50th quantiles) were used as data input for classifiers with accuracy determined by tenfold cross validation. Mean ADC values from the tumour regions of interest differed between tumour types, (ANOVA P < 0.001). A cut off value for mean ADC between Ependymomas and Medulloblastomas was found to be of 0.984 × 10−3 mm2 s−1 with sensitivity 80.8% and specificity 80.0%. Overall classification for the ADC histogram metrics were 85% using Naïve Bayes and 84% for Random Forest classifiers. The most commonly occurring posterior fossa paediatric brain tumours can be classified using Apparent Diffusion Coefficient histogram values to a high accuracy on a multicentre basis.

Highlights

Brain tumours are the most common solid tumours in childhood and the largest cause of death from cancer in this age group
All of the metrics used for the classification were assessed between the raters and no significant differences were detected between metrics derived from the raters’ ROIs
As can be seen in the images, the apparent diffusion coefficients (ADC) maps look distinctly different for the Medulloblastomas and the Pilocytic Astrocytomas with higher ADC observed for the latter

Summary

Introduction

Brain tumours are the most common solid tumours in childhood and the largest cause of death from cancer in this age group. About half of the tumours arise from the posterior fossa with the most common site being the cerebellum making them amenable to surgical resection but with a significant risk of subsequent morbidity. The degree of the resection required is dependent on the type of tumour and so pre-operative diagnosis is desirable as it can aid in surgical planning. Discrimination of the three main types of brain tumours in the posterior fossa (Ependymoma, Medulloblastoma and Pilocytic Astrocytoma) using qualitative assessment of MRI is challenging due to overlapping radiological characteristics but can be improved by the inclusion of diffusion weighted imaging (DWI)[1,2,3].

Methods

Results

Conclusion