AI Finds Early Signs of Lung Cancer that Doctors Can't Detect

Cancer detection as early as possible is very important in an effort to increase patient survival rates, especially in preventing the spread of cancer throughout the body. However, what if the doctor cannot see the early signs of cancer that is already in the patient.

Now, that concern can be slightly pushed aside. Recently, researchers in Massachusetts developed artificial intelligence (AI) that can detect signs of lung cancer, even years before doctors can make a perfect diagnosis.

This AI tool is named Sybil and it works by scanning the results of a computerized tomography (CT) scan of the patient's lung organs. Next, Sybil analyzes 3D images of the patient's lungs and looks for growths and patterns that are very small beyond the reach of the human eye.

Then, the results of these findings will be assessed whether a patient has the potential to have lung cancer in the next one to six years or not. In a series of studies using thousands of volunteer CT data, Sybil accuracy rates were found to range between 75% and 94%.

The location of the lungs themselves are deep in the body so they cannot be easily seen or felt. Conventional early detection methods are less effective in detecting lung cancer, so the only way to see the internal organs of the lungs is through a CT scan.

However, sometimes the CT scan results are too small or fine to be captured by the naked eye. This means lung cancer cells are not detected until the patient has symptoms such as chest pain or shortness of breath. Unfortunately when it gets to this point, cancer is much more difficult to treat. Sybil is able to identify subtle patterns that are difficult for the human eye to identify and make an analysis of whether a patient has the potential to have cancer or not.

Further reading: ascopubs.org