Oktay K, et al., 2020

Mammograms have long been the standard method for breast cancer screening, but they don’t catch every case. Some cancers are more difficult to detect—especially in younger individuals, those with dense breast tissue, or cases where the tumors grow quickly and aggressively. Because of these limitations, researchers are looking for new tools that can spot early signs of cancer risk in ways mammograms can’t.

One approach gaining traction is the use of blood-based biomarkers—molecules found in the blood that can reflect changes in the body before cancer is clinically visible. These "liquid biopsies" have the potential to flag individuals who may be at greater risk, allowing for earlier intervention and more personalized monitoring.

But before these kinds of blood tests can be used in practice, scientists need to figure out which combinations of molecules are most reliable for telling the difference between someone who is healthy and someone who might be at risk for cancer. This study focuses on identifying that kind of molecular marker.

Full-Text Journal Article PubMed

What did they study?

Researchers analyzed blood samples from women who were cancer-free at the time of donation. Some of these women remained cancer-free, while others were later diagnosed with breast cancer. By comparing the two groups, the team aimed to find markers that might serve as early warning signs.

These samples came from the Komen Tissue Bank, a unique resource that collects and stores healthy breast tissue and blood from volunteers over time. Because participants are followed after donation, researchers could identify individuals who went on to develop breast cancer, making it possible to study early warning signs in those who were initially healthy.

The researchers used computer-based models and statistical methods to search for combinations of blood-based molecules that could help identify who might be at higher risk of developing breast cancer.

What did they find?

1. The researchers identified six different "signatures"—groups of molecules—that could help distinguish between women who remained healthy and those who later developed breast cancer. These signatures varied in their ability to accurately predict risk, with the best-performing group of biomarkers including 13 molecules.
2. One molecule, SCF/KITLG, stood out across all methods used. Levels of SCF/KITLG were consistently lower in individuals who later developed breast cancer.
3. Using computer analysis, the best group of molecules was able to predict future breast cancer risk with 85% accuracy. This suggests that this blood-based approach could effectively predict breast cancer risk in future clinical settings​.

Why is this study important?

Full citation

Oktay K,Santaliz-Casiano A,Patel M,Marino N,Storniolo AMV,Torun H,Acar B,Madak Erdogan Z A Computational Statistics Approach to Evaluate Blood Biomarkers for Breast Cancer Risk Stratification.. Hormones & cancer 2020 Feb; 11 (1). PMID:31858384.