The following is a summary of ‘’International Association for the Study of Lung Cancer Early Lung Imaging Confederation Open-Source Deep Learning and Quantitative Measurement Initiative,” published in the January 2024  issue of Oncology by Lam, et al.

With more and more countries using computed tomography (CT) to check for lung cancer, people are interested in using artificial intelligence (AI) deep learning techniques to improve clinical care. The International Association for the Study of Lung Cancer funded the Early Lung Imaging Confederation (ELIC) resource creation in 2018. 

For a study, researchers sought to allow AI research using an open-source, cloud-based, globally distributed screening CT imaging data set and computing environment that would align with the strictest international privacy laws and protect researchers’ intellectual property. 

The goal was to talk about ELIC’s new features and show how this tool can be used for AI research that is useful in clinical settings. In the second part of the project, information and screening CT scans from two different times were gathered from 100 people screened in seven countries. An automated deep learning AI lung segmentation algorithm, an automated quantitative emphysema measures algorithm, and a quantitative lung tumor volume measurement algorithm were used to look at these scans. It took 1,394 CTs from 697 subjects to make the study. 

The LAV950 quantitative emphysema measure could help tell the difference between lung cancer and noncancerous cases when the total slice width was at least 2.5 mm. When used on solid lung nodules from high-quality CT scans, lung nodule volume change readings were more sensitive and specific at telling the difference between cancerous and noncancerous nodules. The first tests showed that ELIC can handle deep learning AI and quantitative image studies on a wide range of cloud-based data sets spread worldwide.

Source: sciencedirect.com/science/article/abs/pii/S1556086423007360