In a report published in the journal Nature, the team described their work, including testing machine learning systems and the ability to find DNA (DNA is a molecule carrying genetic information, showing how to create it). proteins important for the growth and development of all organisms) circulating tumor (ctDNA) in blood samples.

Lung cancer is one of the most dangerous cancers. Like many other cancers detected early, the higher the chance of survival of patients. Unfortunately, the only way to detect it now is through a CT scan, which is not only expensive but also has a high false positive rate.

For that reason, medical researchers have considered the possibility of developing a blood test that can detect lung cancer at an early stage. Such blood tests will involve scanning blood samples for traces of ctDNA. And machine learning in previous studies has shown it to be useful in identifying early cancers so lung cancer cannot be ignored.

The researchers have trained a machine learning model on related data. After training, the model was programmed to make estimates of lung cancer in a given patient.

During the trial, the system detected 63% of tumors for patients with stage 1 lung cancer, not as good as a CT scan, but it could be good enough to serve as a baseline for those considered to have high risk of lung cancer.

The researchers noted that many of these patients are currently not screened at all. If a patient receives a positive result, they will be advised to undergo more sophisticated testing. It is also important to note that such screening can prolong the lives of about 600 to 1,200 people each year. Therefore, the researchers also believe that the method could also be used to check for other types of cancer.