Summary
A five-biomarker blood test combining cfDNA levels and gene methylation patterns detected lung cancer with 84% accuracy, offering a promising non-invasive early screening tool.
Lung cancer is the deadliest cancer in the world, killing 1.8 million people in 2022 alone. One big reason for this is that most cases are not found until the cancer has already spread. When lung cancer is caught early, survival rates jump dramatically — from around 20% to as high as 90%. That’s why scientists are working hard to find better, earlier detection tools.
A new study tested a blood-based method that looks at five biomarkers — measurable signals in the blood that can indicate disease. Researchers collected blood plasma from 179 lung cancer patients and 82 healthy people. They measured the amount of free-floating DNA in the blood, called cell-free DNA (cfDNA), and checked for chemical changes called methylation on four specific genes: SHOX2, PTGER4, RASSF1A, and H4C6.
Think of methylation like tiny switches on your DNA. In cancer, these switches get flipped in unusual ways. For example, the study found that the SHOX2 gene had too much methylation in cancer patients, while PTGER4 had too little. Cancer patients also had higher overall levels of cfDNA in their blood compared to healthy individuals.
Using a computer-based prediction model that combined all five biomarkers, researchers achieved strong accuracy in identifying lung cancer. The model scored an AUC of 0.84 in testing — meaning it correctly distinguished cancer from non-cancer about 84% of the time, with a sensitivity of 77% and specificity of 92%. Importantly, no single biomarker performed nearly as well on its own, showing the power of combining multiple signals.
While the study had limitations — including a small sample size and no comparison group with non-cancerous lung diseases — the results are encouraging. This simple blood test could one day work alongside CT scans to catch lung cancer sooner, when it’s most treatable.
