From
Send to

New Study by Lunit and MGH Validates Effectiveness of AI as Aid in Chest X-Ray Interpretation - Published in JAMA Network Open

Sept. 13, 2022 - 23:40 By
  • Joint study with Massachusetts General Hospital proves that Lunit's AI-powered chest x-ray analysis software enhances the overall performance and efficiency of radiologists

SEOUL, South Korea, Sept. 13, 2022 /PRNewswire/ -- A recent study published in JAMA Network Open demonstrated that Lunit INSIGHT CXR, Lunit's AI-powered chest x-ray analysis solution, can significantly improve the accuracy and efficiency of radiologists diagnosing major chest findings.

The study was conducted by Lunit (KRX: 328130.KQ) and Dr. Mannudeep K. Kalra of Massachusetts General Hospital (MGH) from April to November 2021.

With a total of 497 chest x-rays, researchers compared Lunit INSIGHT CXR's standalone performance, radiologists' assessments with AI and without AI, diagnosing pneumonia, pneumothorax, pleural effusion, and pulmonary nodules. Reviewers included two radiologists, two thoracic radiology fellows, and two residents.

According to the study, Lunit INSIGHT CXR achieved detection rate of 82.5% for pulmonary nodules (94 of 114 findings), 88.7% for pneumonia (173 of 195 findings), 87.2% for pleural effusion (130 of 149 findings), and 100% for pneumothorax (80 of 80 findings). The AI's standalone performance remained high even when non-target findings were present on the chest radiograph.

For all four abnormalities, AI-assisted analysis substantially improved radiologists' sensitivity in comparison to unassisted assessment. Sensitivity increased 17.3%, 6.2% and 4.6% for pneumothorax, pulmonary nodules, and pneumonia respectively.

Moreover, the study also showed that Lunit INSIGHT CXR can improve efficiency. With the AI-assistance, radiologists saved 10% of their reading time compared with the group reporting without AI.

"The use of Lunit's AI product was associated with sensitivity gains for all four targeted chest findings and efficiency improvement across all readers regardless of their experience and training status," said Dr. Kalra.

"This joint study can be a remarkable milestone as it is conducted with the best medical staff from one of the top-ranked hospitals in the US," said Brandon Suh, CEO of Lunit. "The research proved that Lunit INSIGHT CXR can support both radiologists and patients with its exceptional accuracy and efficiency. I strongly believe that it will build a pillar for Lunit as we intend to expand the reach of our product across the US market."

JAMA Network Open is an international peer-reviewed medical journal published by the American Medical Association, with an impact factor of 13.353.