Lunit said its artificial intelligence (AI)-based imaging analysis solution Lunit Insight CXR's reading accuracy was similar to that of a radiologist with more than 20 years of experience in a recent study.
Lunit conducted the study jointly with the University of Edinburgh's Queen's Medical Research Institute Medical Research Institute (QMRI) and Royal Infirmary of Edinburgh.
The research team compared the results of applying Lunit Insight CXR and readings from two radiologists with 20 years or more experience on 1,960 chest X-ray data from primary care referrals and the emergency department -- 992 and 968 cases --, obtained in 2015 at a U.K. hospital.
The area under the receiver operating characteristic curve (AUROC) was calculated for the normal and 10 common findings -- atelectasis, fibrosis, calcification, consolidation, lung nodules, cardiomegaly, mediastinal widening, pleural effusion, pneumothorax, and pneumoperitoneum.
As a result, Lunit Insight CXR showed high accuracy with an average of 0.940 (0.881-0.999) for predicting abnormal findings in AUROC analysis, which is the performance evaluation index of AI models.
The AUROC for each of the findings between the primary care and emergency department datasets did not differ, except for pleural effusion.
"This study confirmed the possibility that the Lunit AI solution could be used stably in any medical field by showing high accuracy," Professor E. J. R. van Beek at Queen's Medical Research Institute said. "The AI solution is expected to be of great help in resolving the shortage of radiology specialists that Europe, including the UK, is facing."
Lunit CEO Suh Beom-seok also said, "Lunit Insight CXR provides accurate and fast analysis results for lesions such as lung cancer, pneumonia, pneumothorax, and tuberculosis even in environments where the quality of medical images varies."
Through these products, we will help doctors make an accurate and efficient diagnosis, Suh added.
The results of the research were published in Clinical Radiology.