Researchers from Mayo Clinic and Stanford Medicine have developed a blood-based liquid biopsy designed to map the tumor microenvironment and predict patient response to immunotherapy.
The findings, published in the journal Nature, describe what investigators say is the first noninvasive test capable of identifying “spatial ecotypes,” or cellular neighborhoods, within tumors using cell-free DNA from a blood sample.
The researchers say the approach could improve precision oncology by helping clinicians determine which patients are more likely to benefit from immunotherapy and which may require alternative treatment strategies.
“This is a complete paradigm shift,” said Aadel Chaudhuri, MD, PhD, professor of radiation oncology at Mayo Clinic and co-senior author of the study, in a recent press release. “Until now, liquid biopsies or blood tests have focused almost entirely on tumor cells. For the first time, we can use a simple blood test to understand the tumor's microenvironment, which is critical for determining how patients respond to modern cancer therapies.”
Mapping tumor “neighborhoods”
Current biomarkers used to guide immunotherapy decisions, including tumor mutational burden and protein expression markers, have shown limited predictive accuracy in some cancers. According to the investigators, these methods do not adequately capture the complexity of the tumor microenvironment, which includes immune cells, stromal cells, and surrounding tissue structures.
To address this limitation, the team used spatial transcriptomics to analyze tumor samples and identify nine distinct spatial ecotypes shared across 17 cancer types. These ecotypes represented different patterns of immune and stromal cell organization within tumors.
“Almost like geographic mapping, we were able to map where in the tumor microenvironment these neighborhoods of co-associated cells live,” Chaudhuri said.
The researchers then partnered with Aaron Newman, associate professor of biomedical data science at Stanford Medicine and co-senior author of the study, to develop an artificial intelligence framework capable of detecting these ecotypes in blood samples.
Using methylation signatures on circulating cell-free DNA, the AI-based assay was able to infer tumor microenvironment patterns without requiring tissue biopsy.
“This is the first time we've been able to noninvasively profile the tumor microenvironment at this level,” Chaudhuri said.
The study included more than 1,300 patients with melanoma, lung, bladder, and gastric cancers. Specific spatial ecotypes were associated with immunotherapy response, treatment resistance and overall survival outcomes. Investigators reported that standard biomarkers demonstrated weaker predictive performance.
Potential role in treatment monitoring
Because the assay relies on blood samples, researchers believe it could also support longitudinal monitoring during treatment. Early findings suggest shifts in spatial ecotypes may indicate treatment response or emerging resistance months before conventional imaging detects changes.
“This gives us a window into how the tumor microenvironment is changing over time,” Chaudhuri said. “We've never been able to see that before in a practical way.”
The researchers are conducting additional validation studies and exploring whether the platform could predict response to therapies beyond immunotherapy, including antibody-drug conjugate combinations. They also suggest the technology could eventually have applications outside oncology in diseases involving complex tissue environments.
Note: This news summary was generated by AI based on a published press release, followed by a review from human editors.
