Polish researchers have developed one of the most advanced AI models for chest imaging analysis. 

Specialists from the Warsaw University of Technology, in cooperation with the Polish Lung Cancer Group and the Children’s Clinical Hospital, have developed a system based on the world’s largest database of computed tomography (CT) images of the chest. The tool is designed to support doctors in diagnosing diseases in the chest, including lung cancer. 

The xLungs system utilizes 40,000 lung CT images developed by the MI2.AI research team at Warsaw University of Technology, in cooperation with the Polish Lung Cancer Group (PGRP) and the Children’s Clinical Hospital in Warsaw. Approximately 40,000 CDs from CT scans of Polish patients from 2010-2018, among other lung cancer screening tests, were used for AI training. 

As a result, an artificial intelligence model – CTSegMate – was created based on the world’s largest (40 TB) database of chest images. The MI2.AI team worked on it for three years, and the processors involved in the task required more than 180,000 hours of computation. The implementation of the  project was possible thanks to funding from the National Research and Development Center under the INFOSTRATEG I competition.