South Korean researchers investigated the potential of a commercially available algorithm (Insight CXR, v. 3, Lunit) for spotting abnormalities in specific pediatric age groups. They found that it achieved nearly 97% accuracy when cardiomegaly and children 2 years old and younger were excluded. These results could help kick-start clinical development of AI software in this population, according to the authors led by corresponding author Dr. Eun-Kyung Kim, a radiologist at Yonsei University College of Medicine in Seoul.

Lunit’s Insight CXR received Europe’s CE mark in 2019, and it is…