Cedars-Sinai
Researchers have unveiled a groundbreaking advancement focusing on “spatial” gene expression, which explores how genetic activity differs across various regions within a single tumor.
At Cedars-Sinai, scientists have crafted an innovative artificial intelligence tool designed to streamline the analysis of genetic activity in cancer tumors. This development may pave the way for broader access to personalized cancer treatments.
Dubbed Path2Space, this tool was recently highlighted in the scientific journal Cell. The researchers explain that it leverages AI to anticipate gene expression patterns in tumors by analyzing digital images obtained from biopsy slides.
The emphasis on “spatial” gene expression allows for an understanding of genetic activity fluctuations in diverse tumor regions. This insight could explain the varying responses of different cancers to therapeutic interventions.
Traditional methods for profiling spatial gene expression often span several weeks and can incur costs in the thousands, as noted by the authors of the study. In contrast, the AI-driven approach generates results within minutes, utilizing standard pathology images.
The research team trained the model with breast cancer biopsy data and then assessed its precision using additional patient datasets. They reported that the system could accurately predict the activity of thousands of genes in tumor samples, achieving results comparable to laboratory measurements.
Lead researcher Eytan Ruppin highlighted that this technology could enable scientists to study a significantly larger number of tumor samples than current methods allow.
Moreover, the researchers believe the tool has the potential to identify biomarkers associated with treatment responses or negative outcomes. However, they cautioned that further validation through clinical trials is essential before it can be implemented into routine patient care.
