Looking for cancer in the bloodstream – one patient at a time

What if, instead of searching for the same cancer signal in every patient, we focus on what makes each patient different?

Changes in our bodies often leave traces in the bloodstream, a concept well established for circulating DNA. RNA fragments in blood plasma, known as cell-free RNA (cfRNA), add another layer of information by reflecting gene activity across the body. In our study, we investigated whether cfRNA profiles in blood plasma differ between cancer patients and healthy individuals.

A complex signal hidden in the blood

To explore this, we analyzed more than 600 blood plasma samples from both cancer patients and individuals without cancer, covering a wide range of cancer types. Using high-throughput RNA sequencing, we created detailed snapshots of the RNA molecules present in each sample.

For the blood cancer included in our study, the signal was relatively clear. However, for solid tumors (such as prostate or lung cancer), the situation proved more complex. While we did observe differences between plasma from cancer patients and healthy individuals, these changes varied greatly from one patient to another.

This variability poses a major challenge. Traditional biomarker research typically looks for molecules that behave consistently across many patients. But if each patient shows a slightly different pattern, such “one-size-fits-all” biomarkers become difficult to define.

Turning variability into an advantage

Instead of trying to eliminate this variability, we decided to embrace it.

We developed a patient-centered approach that compares each patient’s RNA profile to that of a reference group of healthy individuals. Rather than asking “Which genes appear different on average?”, we ask: “Which genes are unusually high or low abundant in this specific patient?”

To do this, we measure how much each RNA’s abundance in plasma deviates from the levels typically observed in healthy individuals (see figure above). Think of blood pressure: in a healthy population, most people fall within a certain range, while a few individuals have values that are unusually high or low. These extreme values lie in the “tails” of the distribution. Similarly, we identify RNAs that fall far outside the typical range in healthy controls and refer to them as “tail genes”. In practice, these are the genes whose transcript levels deviate markedly from the norm.

A simple but powerful signal

When we applied this approach, a clear pattern emerged: cancer patients consistently show more tail genes in their blood plasma than healthy individuals.

Simply counting the number of tail genes in a sample allowed us to accurately distinguish cancer patients from controls. This pattern held not only in our initial dataset, but also across independent blood plasma cohorts. While the exact tail genes are not identical across patients, we observe partial overlap among patients, suggesting shared underlying signals alongside individual variation. Importantly, this approach is not limited to blood; we observed similar results when applied to urine cfRNA profiles.

Toward more personalized diagnostics

Our findings highlight an important shift in thinking about biomarkers. Rather than searching for universal signals that apply to everyone, it may be more effective to focus on patient-specific patterns. By looking at individual deviations rather than average differences between groups, we uncover a richer, more personalized view of disease – one that better reflects the inherently heterogeneous nature of cancer.

This approach aligns closely with the goals of personalized medicine, where diagnosis and treatment are tailored to the individual. While further research is needed before clinical application, our results suggest that cfRNA – and especially patient-specific deviations within it – could play an important role in future cancer management strategies.

Find the original published scientific article here.