- A blood test that uses machine learning to search for cancer DNA could reveal lung cancer in patients who would never have been examined otherwise.
- The system is accurate in about two-thirds of patients, but that may be enough to save 1,200 additional lives per year, according to the scientists who developed it.
- Visit the BGR home page for more stories.
A team of scientists in the United States has developed a blood test that uses machine learning to look for telltale signs of lung cancer. The system, still in an early stage of development, could eventually replace CT scans as a first-line screening measure for suspected lung cancer patients.
The test looks for tumor DNA that circulates in a person's blood and is much less expensive than CT scans, which are generally used to diagnose lung cancer. It is not yet ready to be widely used or trusted in real settings, but the research is incredibly promising and could be a powerful weapon against one of the deadliest cancers.
Medical researchers and doctors already know that cancer DNA circulating in a patient's bloodstream could serve as a tool to diagnose the disease. However, it is unrealistic to commission medical technicians to look for those tiny spots of cancer that flow through the blood. Artificial intelligence, on the other hand, can do that job much easier than a human.
In their first tests, the researchers discovered that their new system is capable of identifying stage 1 lung cancer in 63% of patients who had it. That means he still missed cancers in about a third of those who were screened. That may sound unacceptable, and CT scans are often riddled with false positives, but they are generally more accurate. However, that might not be a deciding factor for this type of new form of testing.
Due to the high cost of CT scans, many potential lung cancer patients do not get screened at all, leaving them at increased risk of being positively diagnosed with even more advanced stages of cancer later. A blood test that can detect cancer DNA in the blood, even if it's only 63% of the time, could be an easy screening option that saves a lot of lives.
Scientists who developed the system believe that using the system in this way could extend the lives of up to 1,200 people per year. In addition, the system may also be able to be modified to look for other types of cancer. It is a potentially revolutionary development in cancer screening, but it may be a while before we see these options in hospitals and clinics across the country or around the world.