untitled design

Sber’s AI service learned to detect cancer at an early stage using lung images

The artificial intelligence-based service of CT of Lungs, which is being developed by SberMedII, has received significant improvements. Now he is able to detect cancer at an early stage by analyzing chest images. This will allow doctors to pay attention to dangerous areas in time and decide on further treatment.

Artificial intelligence finds minimal nodular neoplasms in the lungs and highlights these areas. “CT Lung” can detect and highlight neoplasms larger than 4 mm (threshold) on CT scans of a patient’s lungs that are loaded by the doctor. It also determines the presence of pneumonia (including from COVID-19), indicates the percentage of affected tissues and assigns a degree. The system was trained on data from open sources, which is almost 12 thousand images. To do this, radiologists manually marked them, creating one of the largest databases for such AI models. Currently, the new algorithm is being tested in one of the oncological dispensaries.

According to Yuri Krestinsky, Vice President and Head of the Healthcare Industry of Sberbank, the Lung CT service has shown high efficiency during the COVID-19 pandemic, when the algorithm processed each image within a few minutes, and doctors were primarily provided with images of patients with suspected pathology … Adding the ability to detect cancer at an early stage will make it possible to diagnose a dangerous disease in time and save lives.

The Lung CT service is already being used in 30 clinics in 19 regions of Russia to diagnose viral pneumonia.

You may also like

Get the latest

Stay Informed: Get the Latest Updates and Insights

 

Most popular