A simple blood test that warns of cancer years before symptoms show up is no longer just a hope. New research in Cell Genomics shows that genome-wide methylation patterns in cell-free DNA (cfDNA) from blood can sort people by their risk for prostate and breast cancer long before doctors can spot the disease—sometimes as early as eight years ahead.
According to Stanford Medicine, methylation-based blood tests are already used in commercial multi-cancer screening products such as Galleri and Cancerguard, which detect epigenetic signals rather than DNA mutations.
The team used cfDNA immunoprecipitation sequencing (cfMeDIP-seq) to map methylation across the genome. They then turned to machine learning to find differentially methylated regions (DMRs) that could separate people who would go on to develop cancer from those who would not. The results for prostate cancer stood out. People flagged as high risk by cfDNA silencer methylation were 3.55 times more likely to be diagnosed during follow-up. These risk signals could be seen up to eight years before clinical detection. For breast cancer, the story was less clear. Enhancer methylation signals changed by disease subtype and stage. The high-risk group had an estimated 2.3 times higher risk, but this did not reach statistical significance.
These findings matter because early-stage cancers are much easier to treat than advanced ones. Spotting high-risk people years before symptoms could change how we screen and intervene. The study’s models, trained on 70% of age-matched participants, reached area under the receiver operating characteristic (AUROC) scores of 0.78 for pre-diagnostic prostate cancer and 0.62 for pre-diagnostic breast cancer in the discovery set. Scores were a bit lower in the test sets. The prostate cancer signature also told apart localized from metastatic disease. The breast cancer signature worked best for late-stage cases.
Stanford Medicine experts emphasize that methylation-based blood tests are not yet ready for routine screening of healthy individuals. The sensitivity is currently insufficient for reliably detecting the small amounts of tumor-derived cfDNA present at early cancer stages, and there is no evidence yet that such testing reduces cancer mortality.
The researchers dug deeper and found that DMRs in cfDNA often landed in regulatory regions—enhancers, promoters, silencers—and repetitive sequences. There was strong enrichment in pathways linked to inflammation and cancer biology, including TNF-α, IL-2-STAT5, mTOR, MAPK, KRAS, and p53. These methylation patterns matched those seen in cancer tissues and immune cells. Some signals also appeared in normal tissues, suggesting a mix of tissue-of-origin and immune system effects.
There are limits to the study. Sample sizes for rare cancer subtypes were small. Follow-up among controls was not always the same. The risk models have not yet been validated by outside groups. The breast cancer classifier, in particular, only showed modest power for early disease and did not beat standard screening like mammography. The authors say cfDNA methylation testing for breast cancer should be seen as a possible add-on, not a replacement, for current screening.
Still, the potential is clear. If larger, independent studies confirm these results, plasma cfDNA methylation profiling could become a strong tool for early cancer risk screening—especially for prostate cancer, where current options are limited. Finding molecular warning signs years before symptoms could change how we prevent and catch cancer early. For now, the evidence shows that the molecular clock for cancer may start ticking long before the first symptom, and blood-based methylation signatures are a promising way to hear it.
Reference: Cheng, N., Ouellette, T., Skead, K. et al. (2026). Pre-diagnosis plasma cell-free DNA reveals early signatures of prostate and breast cancer risk up to eight years prior to clinical detection. Cell Genomics, DOI: 10.1016/j.xgen.2026.101364
