Google develops AI to predict risk of lung cancer


Cancer cells

Google has developed a new imaging technology to assist medical staff in their diagnosis. The first results are already very encouraging, with a higher detection rate than doctors.

The engineers of Google introduced a new use of theartificial intelligence in the field of medicine. The latter, in association with several medical centers, developed a technology to anticipate the risk of lung cancer in an individual.

AI compares tests to accurately detect the risk of cancer in a patient

The system is based on machine learning. It consists of a database containing thousands of radio images and the analysis tool uses them as a point of comparison for to detect the appearance of lung tissue and nodules in a patient, almost invisible to the naked eye.

To track lung cancer, radiologists usually examine hundreds of images from a single CT scan. Google wants to bring them here one technological assistance and the first results obtained are optimistic.

Promising results to be confirmed by other tests

In the first tests, Google asked AI to review 45,856 screens Thoracic computed tomography The company then compared the results to those of six radiologists accredited by the representative bodies of the medical profession.

The system finally detected 5% more cases compared to human analysis and reduces false positives by about 11%.

Google will continue to improve its device and pass many other clinical trials before considering offering it in hospitals. Mountain View expects to increase screening for patients, particularly in the United States, where only 2-4% will pass the test during their lifetime.

Source: Engadget


Source link