Lunit "AI-based Tuberculosis Diagnosis Research Published in Nature Sister Journal"

Medical AI company Lunit announced on the 8th that the research results on tuberculosis diagnosis using the chest X-ray AI image analysis solution 'Lunit Insight CXR' have been published in 'Scientific Reports,' a sister journal of the scientific journal Nature.

Lunit's AI chest X-ray image analysis solution 'Lunit INSIGHT CXR'.
[Photo by Lunit]

According to Lunit, tuberculosis diagnosis is generally conducted through sputum tests after chest X-ray imaging. However, this process takes time, and in urgent medical environments such as emergency rooms, there has been a demand for rapid diagnosis of tuberculosis patients and prompt isolation upon diagnosis.

The study, led by the research team of Professors Kim Ji-hoon, Choi Arom, and Choi Soyeon at Severance Hospital, analyzed data from 8,374 suspected tuberculosis patients aged 18 or older who visited the Severance Hospital emergency room from January 2018 to December 2021. The chest X-ray image reading assistance was performed using 'Lunit Insight CXR.'

The results showed that the sensitivity of sputum smear tests, which directly observe tuberculosis bacteria under a microscope, was on average 41.2% in accurately diagnosing actual tuberculosis cases, while the relatively faster PCR sputum test showed a sensitivity of 22.6%. In contrast, the reading sensitivity using Lunit Insight CXR was 70.6%, which was statistically significantly higher than the sensitivity of existing test methods.

When using a multi-diagnostic model including the Lunit solution, the AI model's performance evaluation metric, AUROC (Area Under the Receiver Operating Characteristic), averaged 0.924. Generally, the closer the AUROC value is to 1, the higher the discriminative power, meaning greater reliability.

Lunit explained that these research results demonstrate that the Lunit AI solution can complement existing tuberculosis diagnostic methods and play a role in early diagnosis.

Seobumseok Seo, CEO of Lunit, said, "This study shows that AI can significantly contribute to rapid diagnosis, isolation, and early treatment of patients in emergency situations by increasing speed and sensitivity in the tuberculosis diagnosis process," adding, "Tuberculosis is a representative disease of developing countries, and the introduction of AI technology in countries with insufficient medical infrastructure will bring innovative changes to tuberculosis management."

This content was produced with the assistance of AI translation services.