Press "Enter" to skip to content

Google Launches AI to Detect Lung Cancer at Early Stage

Today lung cancer has become one of the deadliest cancers of all. Studies show that this disease has taken more lives than prostate, colorectal, and breast cancer put together. To help reduce the mortality rate, Google introduces an AI to detect lung cancer with using scan feeds. Further, doctors can accurately and consistently detect lung cancer and provide better treatment to the patient using this AI.

According to technical lead of Google, Shravya Shetty (MS), using advanced 3D volumetric modeling along with the datasets from its partners they have developed the lung cancer detecting AI model. The model is capable of identifying malignant tissues in the lungs besides predicting lung cancer.

How Google AI Predicts the Possibility of Lung Cancer in the Patient?

In order to make the AI model accurate, Google trained its model with the help of 45,856 unidentified chest CT scans. While testing, the AI model was able to detect cancer traces in some of the images. Google representatives state that during testing on a single CT scan, model’s performance was better than that of six radiologists. The model was able to detect 5% more cancer cases than human diagnosis during the tests. It also reduces the false-positive count by 11% as compared to unassisted human radiologists.

Google’s new lung cancer detecting AI model is also capable of detecting very minute cancer traces in the patient’s lung. Even if the patient is asymptomatic and has no cancer history, the model can predict whether there are chances of lung cancer or not.

Google believes that these are just stepping stones in proving better assistance to the healthcare industry. With further studies and developments, the true impact of this AI model in clinical practices can be measured.

Let’s see what the future holds for this new Google AI and how it can help patients and the healthcare industry.

VN:F [1.9.22_1171]
Rating: 0.0/5 (0 votes cast)
VN:F [1.9.22_1171]
Rating: 0 (from 0 votes)