New models reveal tumors as shifting systems, not fixed states

Cancer has long been studied as a collection of static cell types, but recent advances suggest tumors behave more like dynamic systems. Scientists at the UCL Genetics Institute have created models that integrate spatial, temporal, and treatment data to map how cancer cells shift between states. These transitions are influenced not only by genetic changes but also by interactions with surrounding tissues, including stromal cells that can affect tumor behavior across significant distances. The findings challenge traditional views of cancer as fixed entities, instead portraying it as an evolving process where microenvironmental factors often play a dominant role in cell behavior. By identifying regions within tumors where cells exhibit high adaptability—termed 'plastic niches'—these models could help clinicians pinpoint areas most likely to resist treatment or drive recurrence. While current AI tools struggle to fully capture this complexity, particularly in rapidly changing states like epithelial-mesenchymal transitions, hybrid approaches show promise for improving accuracy. The integration of longitudinal data may also enable earlier detection of progression patterns, from pre-cancerous stages to advanced disease. These insights could eventually inform more targeted therapies, though widespread clinical use of such spatial plasticity models remains dependent on further validation and integration with routine diagnostic tools.