Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
Artificial intelligence (AI) is revolutionising science and healthcare, and cancer research is no exception. Multiple AI approaches have been used for cancer diagnosis and prognosis, drug target prediction, and the analysis of tumour composition from multimodal data, amongst many other tasks.
With this cross-journal Collection, the editors at Nature Communications, npj Digital Medicine, npj Precision Oncology, Communications Biology, Communications Medicine, and Scientific Reports invite submissions with a focus on AI in cancer. We welcome papers covering recent advances in the development and application of AI techniques – including machine learning and deep learning methods – with the purpose of deciphering cancer biology, improving diagnosis, prognosis and treatment, and leveraging the vast amounts of available datasets for the benefit of cancer patients.
This Collection supports and amplifies research related to SDG 3: Good Health and Well-Being.
Supervised deep learning models hold promise for the interpretation of histology images, but are limited by cost and quality of training datasets. Here, the authors develop a self-supervised deep learning method that can automatically discover features in cancer histology images that are associated with diagnosis, survival, and molecular phenotypes.
Sahlsten et al. systematically evaluate two Bayesian deep learning methods and eight uncertainty measures for the segmentation of oropharyngeal cancer primary gross tumor volume with a multi-institute PET/CT dataset. The uncertainty-aware approach can accurately predict the segmentation quality that enables automatic segmentation quality control.