Lunit Demonstrates Tuberculosis Screening Performance... "Meets WHO Standards"

Lunit's AI chest X-ray image analysis solution 'Lunit INSIGHT CXR' (Photo by Lunit)

[Asia Economy Reporter Chunhee Lee] Medical artificial intelligence (AI) company Lunit announced on the 15th that its chest X-ray AI image analysis solution, 'Lunit Insight CXR,' applied to prison inmates, met the World Health Organization (WHO) tuberculosis screening performance standards and demonstrated superior accuracy compared to competitors.

This study was conducted by an international research team led by Dr. Julio Croda of Mato Grosso do Sul Federal University in Brazil. It was published in 'The Lancet Regional Health-Americas,' an international academic journal issued by The Lancet group.

The research team applied chest X-ray AI image analysis solutions from three medical AI companies, including Lunit, to a total of 2,075 tuberculosis cases collected from three male prisons in Brazil between October 2017 and December 2019, comparing and analyzing their performance.

As a result, Lunit Insight CXR showed the highest tuberculosis detection ability among the three solutions used in the study, with an Area Under the ROC Curve (AUROC) of 0.91, a performance evaluation metric for AI models. The AUROC score is considered more reliable the closer it is to 1. This figure meets the WHO's tuberculosis patient classification sensitivity (positive detection rate) standard of 90%. Lunit Insight CXR also demonstrated the best performance in detecting tuberculosis under various conditions such as age, smoking status, and presence of tuberculosis symptoms.

Dr. Croda stated, "Due to the closed environment characteristic of prisons, the tuberculosis incidence rate is about ten times higher than in general environments," adding, "Research and investment in tuberculosis screening in such environments are insufficient, so AI-based image analysis solutions are expected to greatly help improve the situation." Lunit CEO Beomseok Seo also said, "Lunit Insight CXR has proven meaningful performance not only in general medical environments but also in settings such as prisons."