In laboratories at Stanford and the Mayo Clinic, researchers have found that a simple blood draw may reveal the quiet communities of healthy cells that surround a tumor, a hidden neighborhood that has long shaped how patients fare despite sharing the same diagnosis. For decades, medicine has largely studied the tumor itself, while the cells living beside it went unread unless someone cut into the body to look. This work, published in Nature, suggests that those neighbors can be read from fragments of DNA drifting in the bloodstream, offering a gentler way to see how the ground beneath a diseas
Blood Test Maps Tumor 'Neighborhoods' That May Predict Immunotherapy Response
Cancer therapy has been like whack-a-mole
Mark
So what is actually new here? Blood tests for cancer already exist.
Mimi
What's new is what the blood is asked to reveal. Not just that a tumor is there, but how the healthy cells around it are arranged and behaving.
Mark
And that arrangement matters?
Mimi
Newman's argument is that it does. Patients with similar tumors fare very differently, and the surrounding cells may explain part of that.
Luke
I'd keep that as an argument. The correlation with immunotherapy response is reported for several ecotypes, but the piece doesn't say how strong it is or in how many patients.
Mark
What did they check the blood against?
Mimi
Predicted ecotypes from blood were compared with those from biopsy and surgical samples from the same patients, across several experiments.
Luke
Which is validation, not proof it changes outcomes. Nothing here shows doctors steering treatment with it and patients doing better.
Mark
The sample was over 100 specimens across 10 cancer types.
Luke
Right, so about ten per cancer type on average. Broad, but not deep. That's my inference from the numbers given, not a stated limitation.
Mimi
The authors themselves say more studies are needed before routine clinical use.
Mark
And the hope is repeat testing?
Mimi
Yes, choosing a first therapy, then watching the microenvironment change to decide when to switch.
Der Puls
- Cancer cells recruit their healthy neighbors into building blood supply lines and shielding the tumor from immune attack, and this hidden cooperation has been largely invisible to doctors.
- Biopsies have been the only window into these tissue neighborhoods, yet they are invasive, rarely repeated, and offer only a single frozen moment in a changing disease.
- Researchers identified nine recurring cell neighborhoods across ten cancer types, several of which track whether patients respond to immunotherapy.
- A new AI tool called Liquid EcoTyper reads methylation patterns in blood DNA and reconstructs those neighborhoods, matching results from surgical tissue in the same patients.
- The test is not yet cleared for routine clinical use, but teams envision repeated blood draws that could signal when a patient should switch therapies.
In laboratories at Stanford and the Mayo Clinic, researchers have found that a simple blood draw may reveal the quiet communities of healthy cells that surround a tumor, a hidden neighborhood that has long shaped how patients fare despite sharing the same diagnosis. For decades, medicine has largely studied the tumor itself, while the cells living beside it went unread unless someone cut into the body to look. This work, published in Nature, suggests that those neighbors can be read from fragments of DNA drifting in the bloodstream, offering a gentler way to see how the ground beneath a disease is changing. Its promise lies in the possibility of watching that terrain over time, so that treatment might follow the shifting landscape rather than a single snapshot taken long ago.
Researchers at Stanford Medicine and the Mayo Clinic, working with colleagues at Washington University in St. Louis, report a development that could change how doctors look at cancer. Their blood test can reveal how healthy cells are arranged around a tumor, the surrounding ecosystem that helps explain why patients with similar cancers can have very different outcomes. The findings appeared in Nature.
For years, the field has leaned on targeted drugs aimed at specific mutations, a approach Aaron Newman, a Stanford biomedical data scientist, describes as a game of whack-a-mole. Many patients with matching tumors still respond differently, and his team argues the missing piece is the community of nonmalignant cells surrounding the tumor. Cancer exploits those neighbors, coaxing them to build blood vessels, quieting immune cells, and laying down collagen that shields the growth. Studying these relationships has required biopsies, which are invasive and rarely repeated, so a single sample offers little insight into a disease that keeps changing.
Using machine learning tools from Newman's lab, the team examined more than 100 tumor samples across ten cancer types, focusing on carcinomas and melanomas. They identified nine recurring cellular neighborhoods, some sitting at the border between tumor and healthy tissue and others deeper within. Several of these patterns correlated with immunotherapy response, suggesting they could guide treatment choices.
The breakthrough came when the researchers asked whether these neighborhoods could be detected in blood. Dying cells release DNA into circulation, and chemical marks on that DNA reveal which genes were active. Building on this, the team created Liquid EcoTyper, which reconstructs the tumor environment from those marks. In several experiments, the picture drawn from blood matched the one drawn from tissue taken from the same patients.
More studies are needed before the test reaches routine practice. The researchers imagine a future in which a clinician uses the blood test to choose an initial therapy, then repeats it over time to watch the surrounding ecosystem evolve and to signal when a change in treatment may be needed.
Researchers at Stanford Medicine and the Mayo Clinic report that a blood draw can reveal how healthy cells are arranged around a cancerous tumor. It is the first noninvasive way to study what is called the tumor microenvironment, the surrounding ecosystem that helps explain why patients with similar tumors can have very different outcomes after diagnosis and treatment. The work was published in Nature.
The senior authors are Aaron Newman, Ph.D., an associate professor of biomedical data science at Stanford, and Aadel Chaudhuri, MD, Ph.D., a professor of radiation oncology at the Mayo Clinic in Rochester, Minnesota. The lead authors are postdoctoral scholar Wubing Zhang, graduate student Erin Brown, and Abul Usmani, a senior scientist from Washington University in St. Louis.
Newman describes cancer therapy to date as a game of whack-a-mole. Targeted drugs work against tumors with particular mutations, he said, yet many patients with similar tumors respond differently. His argument is that the overlooked factor is the community of nonmalignant cells around the tumor. Cancer cells exploit those neighbors: they coax them into building new blood vessels that feed the tumor, persuade immune cells to ignore them, and push fibroblasts to produce collagen scaffolding that supports growth and blocks immune access. Once such a permissive environment is in place, spread becomes easier. Newman compares it to plants that thrive in some soils and die in others.
Until now, studying who sits next to whom has depended on biopsies, either thin tissue slices fixed on a slide or bulk tissue analyzed for gene activity. Biopsies are invasive and done only occasionally during treatment, if at all. Because cellular relationships change as disease progresses and as therapy is applied, a single biopsy offers limited clinical value. The team likens it to judging a 40-year class reunion from an old yearbook.
The team used machine learning tools from Newman's lab on two major solid cancer groups, carcinomas and melanomas. Carcinomas account for 80% to 90% of human cancers, including breast, lung, colorectal, liver and prostate. One tool, CytoSPACE, designed in 2023, maps cells to precise locations in tumor tissue. Another, Spatial EcoTyper, predicts cell types from the abundance and patterns of genetic messages in a sample, infers what the cells are doing, and identifies which cells they interact with. The output is a set of cellular neighborhoods, called spatial ecotypes.
Applying these tools to more than 100 tumor specimens across 10 cancer types, the researchers mapped gene expression in nine cell types at different tumor locations. They identified nine spatial ecotypes, each roughly the width of a human hair. The pattern held across all the tumors studied. Some ecotypes tended to sit at the border between tumor and healthy tissue, others deeper inside. Several correlated with whether a tumor responded to immunotherapy, suggesting they could inform treatment decisions.
Newman noted that a CD8 T cell may appear in several neighborhoods but express different genes depending on its neighbors. The team also found something it did not expect: each ecotype carries consensus genetic programs shared across cell types, in addition to the unique program of each cell type.
That raised the question of whether these neighborhood signatures could be detected in blood. DNA enters circulation when cancer or healthy cells die, and methyl groups on that DNA indicate which genes were active in the cell of origin. Newman said the team realized methylation patterns could stand in for gene expression, and so for cell state and type in the microenvironment. They built Liquid EcoTyper, an AI tool that reconstructs the tumor microenvironment from methylation patterns. In multiple experiments, ecotypes predicted from blood matched those derived from biopsy and surgical samples from the same patients.
More studies are needed before the test is approved for routine clinical use. The researchers envision clinicians using it to choose a first treatment, then repeating it to track how the microenvironment evolves and signal when to switch therapies.
Bemerkenswerte Zitate
Now we can infer these clinically vital, spatial landscapes in a tumor without having to do any tissue biopsy at all.— Aaron Newman, Stanford Medicine
Like plants that thrive in some soils and die in others, cancer cells exhibit vastly different growth patterns in different cellular environments.— Aaron Newman, Stanford Medicine
Möchten Sie die ganze Geschichte?
Das Original lesen bei Medical Xpress ↗
Health & Medicine