Image-based Survival Analysis for Lung Cancer Patients using CNNs
Traditional survival models such as the Cox proportional hazards model are typically based on scalar or categorical clinical features. With the advent of increasingly large image datasets, it has become feasible to incorporate quantitative image features into survival prediction. So far, this kind of analysis is mostly based on radiomics features, i.e. a fixed set of features that is mathematically defined a priori. In order to capture highly abstract information, it is desirable to learn the feature extraction using convolutional neural networks. However, for tomographic medical images, model training is difficult because on one hand, only few samples of 3D image data fit into one batch at once and on the other hand, survival loss functions are essentially ordering measures that require large batch sizes. In this work, we show that by simplifying survival analysis to median survival classification, convolutional neural networks can be trained with small batch sizes and learn features that predict survival equally well as end-to-end hazard prediction networks. Furthermore, we demonstrate that adding features from a fine-tuned convolutional neural network improves the predictive accuracy of Cox models that otherwise only rely on radiomics features. Moreover, we propose survival label noise as a means of data augmentation for deep image based survival analysis.
READ FULL TEXT