Deep Learning versus Classical Regression for Brain Tumor Patient Survival Prediction

TL;DR

This study compares deep learning and classical regression methods for brain tumor survival prediction, finding that traditional approaches outperform CNNs in small datasets and emphasizing the importance of clinical data.

Contribution

It provides a comparative analysis of CNNs and classical regression with radiomic features for survival prediction, highlighting the need for more data and clinical info for deep learning models.

Findings

Support Vector Classifier outperformed CNNs in accuracy.

Ensemble of SVCs achieved 72.2% cross-validated accuracy on BraTS 2018.

Deep learning models showed unstable results and require more data.

Abstract

Deep learning for regression tasks on medical imaging data has shown promising results. However, compared to other approaches, their power is strongly linked to the dataset size. In this study, we evaluate 3D-convolutional neural networks (CNNs) and classical regression methods with hand-crafted features for survival time regression of patients with high grade brain tumors. The tested CNNs for regression showed promising but unstable results. The best performing deep learning approach reached an accuracy of 51.5% on held-out samples of the training set. All tested deep learning experiments were outperformed by a Support Vector Classifier (SVC) using 30 radiomic features. The investigated features included intensity, shape, location and deep features. The submitted method to the BraTS 2018 survival prediction challenge is an ensemble of SVCs, which reached a cross-validated accuracy of 72.2% on the BraTS 2018 training set, 57.1% on the validation set, and 42.9% on the testing set. The results suggest that more training data is necessary for a stable performance of a CNN model for direct regression from magnetic resonance images, and that non-imaging clinical patient information is crucial along with imaging information.