Precision cancer medicine aims to obtain treatment-impacting information on individual cancer cases by performing in-depth cancer profiling. Profiling often includes tissue imaging and, more recently, sequencing of cancer genomes. Machine learning methods such as ChatGPT have been shown to be capable of solving a wide range of problems expressed in natural language or as images. Here, we seek to employ similar machine learning approaches in precision cancer medicine to answer questions such as what is the molecular subtype and suitable treatments for particular cancer, or whether a cancer may be aggressive or indolent. To do this, we will develop machine learning models that do not require guidance by experts, but instead learn to automatically recognize similarities and differences between cancers based on histological images and genomics data. These models can then be used to create software tools to assist clinicians to better diagnose and treat patients.

2025

2029