Published in Frontiers in Oncology (2022).

Work by Fabian Sinzinger, Mehdi Astaraki, Örjan Smedby, and Rodrigo Moreno.

Abstract

Survival-rate prediction (SRP) is a valuable tool to assist in the clinical diagnosis and treatment planning of lung-cancer patients. In recent years, deep-learning (DL) based methods have shown great potential in medical image processing in general and SRP in particular.

This study proposes a fully automated method for SRP from computed tomography (CT) images that combines automatic segmentation of the tumour and a DL-based method for extracting rotation-invariant features. The methodology uses a variational autoencoder to provide information to a U-Net segmentation model, and the 3D volumetric image of the tumour is projected onto 2D spherical maps that serve as input to a spherical convolutional neural network. The approach exemplifies the application of geometric deep-learning methods in oncology, specifically addressing survival prediction through spherical CNNs.