To screen for lung cancer, radiologists typically view hundreds of images from a single scanner. With this new model of AI, Google can generate an overall prediction of lung cancer malignancy and identify subtle malignant tissue, or pulmonary nodules, that are often difficult to see. The AI also takes into account previous exams, which can help reveal the growth rate of suspicious tissue.
To test the model, Google asked his computer scientist to examine 45,856 chest CT screens. He compared the results with six certified radiologists. In these early studies, Google's AI detected five percent more cancer cases than radiologists. He also reduced false positive tests by more than 11%.
The model requires additional clinical research and testing before it can be deployed, but Google says the first results are encouraging. The company also notes that only two to four percent of eligible patients in the United States are screened for lung cancer. As Google hopes that it will make early detection more accessible, its other tools for detecting diseases based on AI.