Researchers from Stanford Medicine and the Mayo Clinic have developed the first non-invasive blood test capable of revealing the cellular microenvironment of a tumor. This opens new possibilities for predicting how cancer patients will respond to immunotherapy without the need for an invasive tissue biopsy.
Mapping the tumor microenvironment without a biopsy
The study, published in the journal "Nature", describes nine recurrent multicellular ecosystems – so-called spatial ecotypes – common to various types of cancer, and shows that they can be identified from a simple blood sample using artificial intelligence.
The team, led by senior authors Aaron Newman from Stanford University and Aadel Chaudhuri from the Mayo Clinic, integrated over 10 million single-cell and spatial transcriptomic profiles (at the "spots" level) from 132 tumor samples of ten different cancer types to identify nine spatial ecotypes.
Each of these ecotypes represents a unique cellular environment about the size of a human hair, characterized by a specific combination of immune cells, stromal cells, and tumor cells with their own gene expression programs.
"Just as some plants thrive in certain soils and die in others, cancer cells show completely different growth patterns in different cellular environments," explains Newman. "Our goal is to understand on a broad scale how these environments, or spatial ecotypes, influence cancer growth and response to therapy, as well as how they vary between individual patients."
Liquid EcoTyper: AI framework that replaces biopsy with a blood test
After identifying the ecotypes, the team developed "Liquid EcoTyper" – an artificial intelligence framework that reconstructs spatial patterns within the tumor based on signals from cell-free DNA methylation in blood plasma.
This approach eliminates the need for invasive tissue biopsies and builds on earlier computational methods from the Newman lab, including "CytoSPACE" and "Spatial EcoTyper," which map cells to precise locations in the tumor and predict cell-cell interactions according to gene expression data.
In practice, "Liquid EcoTyper" translates the complex spatial organization of the tumor into "signals" in circulating DNA that can be captured by a blood test – thus providing a remote "portrait" of the tumor microenvironment.
Better prognosis than existing biomarkers
Clinical validation involved over 1,300 patients with melanoma, lung, bladder, and stomach cancer and showed that certain spatial ecotypes are reliably associated with response to immunotherapy and survival metrics.
In plasma samples taken before the start of treatment in nearly 100 melanoma patients, signals from so-called "liquid ecotypes" demonstrated an average Area Under the ROC Curve (AUC) of 0.87 for ecotype SE7, associated with a good response – a result that significantly outperforms tumor mutation burden (TMB), "PD-L1" expression, and circulating tumor DNA as predictive markers.
In a multivariate analysis across various solid tumors, signals from liquid spatial ecotypes showed a stronger association with overall survival than both TMB and "PD-L1." This suggests that information about the tumor microenvironment may be more valuable for predicting the success of immunotherapy than classic genetic and protein markers.
From scientific breakthrough to clinical test
The study results are already attracting interest from the industry. The diagnostic company "LiquidCell Dx," co-founded by Newman, announced the development of "LiquidTME" – a clinical assay based on the findings described in the "Nature" article, aimed at selecting optimal therapeutic tactics for different cancer types and treatment approaches.
The concept behind "LiquidTME" is to turn the "Liquid EcoTyper" scientific framework into a practical tool for oncologists – a test to help in decision-making on which patient has the best chance of benefiting from immunotherapy, combination therapy, or alternative approaches.
Newman outlines a future where blood testing becomes a primary tool for selecting initial therapy and monitoring changes in the tumor microenvironment over time. "This method has the potential to provide a significantly more complete and profound picture than any of the existing ways of studying the tumor microenvironment," he says. "The clinical prospects are truly impressive."
If the technology is successfully implemented into routine practice, oncology teams could have a dynamic map of the "ecosystem" around the tumor, obtained from just a few milliliters of blood – and tailor treatment much more precisely to the individual patient.