Summary
A bioRxiv preprint reports that tumour-cell states, mutations, tumour architecture and immune signals were associated with differences in ctDNA release in ten patients with non-metastatic HR+/HER2− breast cancer. The findings offer a biological framework for understanding why tumour DNA detected in tissue may differ from what appears in blood.
A bioRxiv preprint reports that tumour-cell state and location are associated with how much circulating tumour DNA (ctDNA) is released in HR+/HER2− breast cancer. The researchers combined single-cell and spatial analyses with DNA sequencing from tumour tissue and matched plasma samples from ten patients with non-metastatic disease.
ctDNA consists of DNA fragments shed by tumours into the bloodstream. Because a blood test measures this released material rather than sampling tumour cells directly, a mutation found in tissue may not always be detectable in plasma. The study investigates biological factors that could contribute to that difference. It is a preprint, rather than a peer-reviewed clinical study.
Tumour states and locations differed in release potential
The team developed scMASTER, a framework integrating single-cell gene-expression and mutation data, spatial transcriptomics, and matched tumour and plasma DNA sequencing. In the ten-patient cohort, the researchers identified eight recurring malignant epithelial cell states with different ctDNA-release potential.
Antigen-presentation, stress, mitotic and mesenchymal-like states were associated with higher release. The authors also linked PIK3CA and ESR1 mutations to distinct functional states and changes in ctDNA detectability. These findings suggest that the presence of a tumour mutation alone may not determine whether it is captured in a plasma sample; the state of the cells carrying it may also matter.
Spatial analysis highlighted three tumour architectures associated with high release: regions enriched for epithelial-to-mesenchymal transition (EMT), intraductal luminal progenitor regions, and avascular hypoxic zones. The abstract also reports contrasting immune associations: anti-tumour immunity correlated with lower ctDNA release, while immunomodulatory components correlated with higher release.
Implications for interpreting liquid biopsies
The authors report that a risk-stratification model based on these features identified high-risk, immunosuppressed patients with poor prognosis, with validation across external cohorts. The abstract does not provide the sizes or detailed performance estimates for those cohorts, so the result is best understood as a reported research finding rather than a measure of clinical predictive accuracy.
Taken together, the work proposes a biological explanation for tissue–plasma discordance: ctDNA detection may vary with the tumour cells’ functional state, their position and local immune environment. That framework could inform research into how ctDNA dynamics are interpreted in HR+/HER2− breast cancer. The initial analysis involved ten patients, and the findings remain preclinical in the sense that they describe tumour biology and risk stratification rather than a tested change to patient care.