Scientists Develop AI System to Predict Cancer Spread Using Genetic Patterns

Researchers have developed a new artificial intelligence system capable of predicting the likelihood of cancer spreading to other parts of the body. The AI system, named **“Mangrove GS”**, analyzes complex patterns of gene activity to estimate the metastatic potential of tumors.

The study, published in the scientific journal **Cell Reports**, could help doctors assess the risk of metastasis in patients more accurately in the future.

Understanding why some tumors spread rapidly while others remain localized has long been one of the biggest scientific challenges in oncology. The risk of death increases significantly when cancers of the **colon**, **breast**, or **lungs** metastasize to other organs.

According to **Ariel Ruiz i Altaba**, it is incorrect to view cancer solely as a result of random cellular mutations. Instead, it can be seen as a set of biological programs that are normally active during early development but become reactivated at the wrong time or in the wrong location.

To understand this process, researchers studied cloned cells derived from colon tumors, analyzing the activity of hundreds of genes to determine how different cells respond. During the research, scientists identified specific gene activities strongly associated with a cell’s ability to spread cancer.

Based on these findings, the team developed **“Mangrove Gene Signatures”**, an AI-based system that simultaneously analyzes dozens or even hundreds of genetic signatures to predict the likelihood of cancer metastasis.

This technology could mark a significant step forward in personalized cancer treatment, enabling earlier interventions and better management of high-risk tumors.