Blood Test May Predict Cancer Risk Years Before Diagnosis
Key Summary:
- Researchers analysed 491 blood samples collected before cancer diagnosis or from cancer-free controls.
- DNA methylation signatures were associated with prostate cancer risk up to eight years before diagnosis.
- The breast cancer model showed weaker performance and requires further validation.
CHANGES in circulating DNA could help identify people at increased risk of developing prostate or breast cancer years before clinical diagnosis, according to new research.
Researchers identified DNA methylation patterns in blood samples collected before cancer diagnosis, with certain molecular signatures associated with prostate cancer risk up to eight years before the disease was detected.
Investigating Early Cancer Signals in Blood
Cell-free DNA (cfDNA) consists of fragments of genetic material released into the bloodstream by cells throughout the body.
Changes in DNA methylation, a chemical modification that regulates gene activity, may provide information about biological processes associated with cancer development.
Researchers investigated whether these changes could be detected before a clinical cancer diagnosis and used to identify individuals at increased risk.
The study analysed 491 plasma samples from participants in the Ontario Health Study, including 171 samples from individuals who subsequently developed breast cancer, 93 from those who developed prostate cancer, and 227 from cancer-free controls.
Samples from participants who later developed cancer had been collected between two weeks and nine years before diagnosis.
Researchers used genome-wide DNA methylation profiling and machine-learning models to identify molecular signatures associated with subsequent cancer development.
Prostate Cancer Risk Identified Years Earlier
The researchers found that methylation changes in DNA regions known as silencers, which regulate gene expression, were associated with future prostate cancer risk.
Participants classified as high risk using these molecular signatures had a 3.55-times higher adjusted hazard of prostate cancer diagnosis during follow-up compared with those classified as low risk.
The prostate cancer model achieved an area under the receiver operating characteristic curve of 0.78 in the discovery dataset and 0.73 in the held-out test dataset, indicating moderate discriminatory performance.
Researchers also identified methylation changes associated with breast cancer, particularly in regulatory regions known as enhancers.
However, the breast cancer model demonstrated weaker performance before diagnosis, with an area under the curve of 0.58 in the test dataset.
Although the estimated hazard of subsequent breast cancer was 2.3-times higher in the group classified as high risk, this association did not reach conventional statistical significance.
Could Blood Tests Improve Cancer Screening?
The findings suggest that circulating DNA methylation patterns may contain information about cancer risk several years before clinical diagnosis.
Such biomarkers could potentially complement existing screening strategies by identifying individuals who may benefit from closer monitoring or further investigation.
However, the researchers emphasise that the findings require validation in larger, independent prospective cohorts.
The relatively small number of cancer cases and differences in model performance between prostate and breast cancer also limit the immediate clinical applicability of the findings.
Importantly, the study investigated risk stratification rather than demonstrating that a blood test can reliably diagnose cancer years before symptoms develop.
The researchers conclude that cfDNA methylation profiling warrants further investigation as a potential tool for identifying individuals at increased cancer risk and supporting earlier detection strategies.
Reference
Cheng N et al. Pre-diagnosis plasma cell-free DNA reveals early signatures of prostate and breast cancer risk up to eight years prior to clinical detection. Cell Genom. 2026;101364. doi:10.1016/j.xgen.2026.101364.
Featured image: motortion on AdobeStockAuthor:
Each article is made available under the terms of the Creative Commons Attribution-Non Commercial 4.0 License.
Share:
