Mayo Clinic, Stanford researchers develop blood test to predict cancer immunotherapy response

USA—Researchers at Mayo Clinic and Stanford Medicine have developed a blood test that can map the complex environment surrounding cancer cells, potentially providing a more accurate way to predict which patients will respond to immunotherapy.

The findings, published in Nature, mark an advance in precision oncology by allowing clinicians to assess the tumor microenvironment through a blood sample rather than an invasive tumor biopsy.

Mapping the tumor environment

Immunotherapy has transformed cancer treatment, but only a proportion of patients benefit.

Existing biomarkers, including the number of DNA mutations in a tumor and the levels of certain proteins on cancer cells, provide limited information about how a tumor will respond.

“Until now, liquid biopsies or blood tests have focused almost entirely on tumor cells,” said Aadel Chaudhuri, M.D., Ph.D., professor of radiation oncology at Mayo Clinic and co-senior author of the study.

“For the first time, we can use a simple blood test to understand the tumor’s microenvironment.”

To develop the test, the researchers used spatial transcriptomics, a technology that shows how different cells are organized and interact within tumors.

Their analysis identified nine distinct cellular neighborhoods, known as spatial ecotypes that reflect different combinations of immune and stromal cells surrounding tumors.

The team found these ecotypes across all 17 cancer types examined.

Some occurred predominantly at the boundary between tumors and healthy tissue, while others appeared deeper within tumors.

Importantly, several ecotypes were associated with patient survival and responses to immunotherapy.

Turning tumor mapping into a blood test

The researchers then collaborated with Aaron Newman, Ph.D., associate professor of biomedical data science at Stanford Medicine, whose team developed methods to identify the cellular neighborhoods in tumor samples and an artificial intelligence (AI) framework for detecting them in blood.

The resulting liquid biopsy analyzes methylation, or chemical modifications of DNA, in cell-free DNA released by tumors into the bloodstream.

 This approach allows the test to capture information about the tumor microenvironment in addition to cancer cells.

The researchers evaluated the approach in more than 1,300 patients with several cancers, including melanoma, lung, bladder and gastric cancers.

Specific spatial ecotypes showed strong associations with treatment outcomes, with some linked to better immunotherapy responses and others associated with treatment resistance and poorer survival.

The researchers said the new approach outperformed standard biomarkers in predicting treatment outcomes.

Potential for treatment monitoring

The test could also help clinicians identify patients who are unlikely to benefit from immunotherapy before treatment begins, potentially allowing them to consider alternative therapies sooner.

Because the test requires only a blood sample, researchers also believe it could enable repeated monitoring of changes in the tumor microenvironment during treatment.

Early findings suggest that shifts in spatial ecotypes may indicate treatment response or resistance months before conventional imaging detects changes.

Although the published study focused primarily on melanoma, the researchers are investigating the approach in other cancers, including lung and bladder cancer.

They are also evaluating whether spatial ecotypes can predict responses to antibody-drug conjugate combination therapies.

Further studies will assess the test in larger patient populations and determine whether it can be incorporated into routine clinical practice.

Chaudhuri and Newman have disclosed patent filings related to liquid biopsy and cancer biomarkers, while both researchers have reported ownership or consulting interests in companies in the field.