In a lab just north of Munich, a team of physicists, oncologists, and engineers has found a way to eavesdrop on the molecular murmurs of cancer.
Using a machine trained to read the vibrations of molecules,they’ve found a way to detect cancer signatures in blood plasma — with striking success in identifying lung cancer. Their study, published in ACS Central Science, introduces a new kind of blood test that relies on light. This emerging method is called electric-field molecular fingerprinting (EMF) and could usher in a new era in cancer diagnostics.
Molecular Fingerprints Written in Light
To understand what the researchers have accomplished, consider how blood plasma — the clear, cell-free part of blood — acts as a messenger of health status. It carries hormones, proteins, lipids, and chemical signals throughout the body. When someone develops cancer, the mix of molecules in their plasma subtly shifts.
But these shifts are complex, and until now, difficult to interpret without invasive biopsies or highly targeted chemical tests.
Žigman and her team tried something different.
They directed pulses of infrared laser light through plasma samples from more than 2,000 people, including patients with lung, prostate, breast, and bladder cancer, as well as healthy individuals. When the light hit the molecules in the plasma, it triggered tiny vibrations. Each molecule reflected or absorbed energy in a way that produced a unique “infrared molecular fingerprint.”
Those patterns — too intricate for the human eye to decipher — were then fed into a machine learning model trained to distinguish cancer from non-cancer. The computer learned to recognize telltale differences in these molecular signatures.
When tested on a new group of 430 patients, the model was able to correctly identify lung cancer cases up to 81% of the time.
That level of accuracy is promising. It means that in eight out of ten cases, the AI could correctly sort a lung cancer patient from a healthy individual using just a drop of blood.
Promise and Caution
To see if the technology works outside the lab, the researchers launched the Lasers4Life study, a clinical trial that recruited 2,533 participants. Blood samples were collected from individuals newly diagnosed with lung, prostate, breast, or bladder cancer—none of whom had begun treatment—as well as from healthy controls. Each sample was analyzed in just three and a half minutes.
The test didn’t perform equally well across all cancer types. While it succeeded in picking up many lung cancer cases, it identified only about half of breast cancer cases in the same trial. The performance for prostate and bladder cancers was also lower.
Still, the researchers believe this is just the beginning and performance can be improved.
The team plans to train the system on more samples, hoping to improve detection across cancer types and stages. Their long-term goal: a general blood test that could screen for many cancers early, accurately, and without the need for biopsies.
For now, the technology is still in its infancy. But it could complement — and someday even rival — existing approaches.
More importantly, even if it’s imperfect it could lead to earlier detection, which remains the single biggest factor in cancer survival. Many types of cancer are perfectly treatable if caught early, but detection remains challenging. A blood test that flags it at an earlier stage, when it’s more treatable, could save countless lives.
“There’s no silver bullet in cancer diagnostics,” said Kepesidis, co-lead author of the study. “But we think this could be a powerful new tool in the kit.”
