Blood test maps the tumor microenvironment with cfDNA

The tumor microenvironment determines how cancer behaves. The mix of immune cells, fibroblasts, macrophages, and lymphocytes surrounding the tumor dictates whether the immune system can attack it, whether it will spread, and whether immunotherapy will work. Until recently, mapping that microenvironment required a tissue biopsy: an invasive procedure that captures a single snapshot of a single tumor region.

A new approach is changing that. Researchers have developed a blood test that uses the methylation pattern of circulating tumor DNA (cfDNA) to classify the tumor microenvironment with the help of artificial intelligence. The tool, called Liquid Ecotyper, can do from a blood sample what previously required surgery.

What is cfDNA methylation?

When cancer cells die, they release fragments of their DNA into the bloodstream. These fragments carry epigenetic marks, particularly DNA methylation patterns, that reflect the cell type from which they came. Methylation is a chemical modification that does not change the DNA sequence but controls gene expression: a methylated gene is typically silenced, while an unmethylated gene can be expressed.

These methylation marks are cell-type-specific. A T lymphocyte has a different methylation pattern than a macrophage or a fibroblast. Liquid Ecotyper exploits this specificity to deconvolve the cellular composition of the tumor microenvironment from a blood sample.

How Liquid Ecotyper works

Liquid Ecotyper is a machine learning model trained on paired tissue biopsies and blood samples. It learns the methylation signatures associated with each cell type in the tumor microenvironment and then uses those signatures to classify new blood samples.

The model operates in two steps. First, it deconvolves the cfDNA signal to estimate the proportion of each cell type present in the tumor microenvironment. Second, it classifies those proportions into ecotypes: recurring cellular patterns that predict clinical behavior and treatment response.

Three published studies underpin the tool. The first demonstrated high correlation between cfDNA methylation-derived microenvironment profiles and those obtained from tissue biopsies in the same patients. The second validated the approach across multiple cancer types using a pan-cancer cfDNA methylation atlas. The third showed that Liquid Ecotyper could predict which patients would respond to immunotherapy before treatment started, based on blood alone.

The tumor microenvironment and immunotherapy

Not all tumors are equally accessible to the immune system. Some are immunologically hot: they are infiltrated by CD8+ T cells that recognize tumor antigens, and they respond well to checkpoint inhibitors. Others are immunologically cold: they exclude immune cells through physical barriers, suppressive signals, or the absence of recognizable antigens.

Liquid Ecotyper classifies tumors along this spectrum. A tumor identified as immunologically cold is unlikely to respond to a PD-1 or CTLA-4 inhibitor alone. That information can guide the oncologist toward alternative strategies: combination therapies, cancer vaccines, or approaches designed to convert cold tumors into hot ones.

The microenvironment also reveals the dominant suppressive mechanism. A tumor rich in M2 macrophages (immunosuppressive) behaves differently than one dominated by regulatory T cells or by cancer-associated fibroblasts. Each of those profiles may require a different therapeutic approach.

Clinical applications

The most immediate application is treatment selection. Rather than waiting for a patient to fail one immunotherapy regimen and then trying another, a blood test taken at diagnosis or at a key decision point can guide which approach to try first.

The second application is monitoring. Because cfDNA is shed continuously into the bloodstream, repeated blood tests can track how the tumor microenvironment changes over time: whether it becomes more or less immunologically accessible, whether treatment is working, or whether resistance mechanisms are emerging.

This is particularly valuable because tumor heterogeneity means that a single tissue biopsy may not represent the full picture. A blood test integrates signals from all tumor lesions simultaneously, providing a more comprehensive view than a single biopsy site can offer.

Limitations and what remains to be validated

Liquid Ecotyper is a promising tool at an early clinical stage. The studies published to date are largely retrospective: they analyze stored samples from patients whose outcomes are already known. Prospective trials, where the test result actively changes treatment decisions in real time, are needed to confirm clinical benefit.

The signal in blood is also diluted. In patients with low tumor burden, cfDNA may be sparse, reducing the reliability of the methylation-based classification. Technical standardization across laboratories remains an active area of development.

Conclusion

Liquid Ecotyper represents a shift in how oncology thinks about tumor characterization. Rather than a one-time biopsy that captures a single location at a single moment, liquid biopsy methylation profiling enables repeated, non-invasive snapshots of the full tumor ecosystem. If prospective trials confirm its predictive value, it could become a routine tool for treatment selection and monitoring in oncology.