Deep learning in image segmentation for cancer

This article discusses the role of deep learning (DL) in cancer imaging, focusing on its applications for automatic image segmentation. It highlights two studies that demonstrate how U-Net- and convolutional neural networks-based architectures have improved the speed and accuracy of body composition analysis in CT scans and rectal tumour segmentation in MRI images. While the results are promising, the article stresses the need for further research to address issues like image quality variability across different imaging systems.