- Proof-of-concept reveals that AI can massively reduce reading burden in the underwriting process, which require the review of 30,000 chest x-rays annually
SEOUL, South Korea, Oct. 14, 2020 /PRNewswire/ -- A recent pilot project conducted by a South Korean medical AI company, Lunit, and one of the largest Taiwanese insurance companies, Cathay Life Insurance, reveals that AI can help the current underwriting process to be more accurate and efficient.
The insurance company, which reviews 30,000 chest x-ray images accepted annually as part of the requirements of new insurance contracts, deployed Lunit INSIGHT CXR to assess the impact of AI in the review process. Lunit INSIGHT CXR is an AI software developed by Lunit which detects abnormal findings via chest x-ray images.
For the assessment, nearly 200 chest x-ray images were used as samples to verify the performance of AI. These cases were internally reviewed by Cathay Life's doctors to validate the accuracy of the algorithm. The results showed that the AI's detection rate for positive cases (sensitivity) was 83% and for negative cases (specificity) was 92%, with an overall accuracy of 90%.
According to the doctors who participated in the validation, the AI provided great value in reducing the time spent on the overall reading of chest x-rays. In the current workflow of doctors, chest x-rays to be reviewed arrive in CDs, which the doctors manually insert to read the medical image, then again, manually write an interpretation report that is packaged along with the CD.
"The introduction of AI and digital interpretation system can greatly reduce our working time," said Tzu-Ling, Ke, Senior VP of Underwriting Dept., Cathay Life Insurance. "The entire process of interpretation－accessing the data, the reading, the writing of the report－usually takes several minutes, and we can save 90% of reading time with AI and the digital system."
"We are delighted to hear that our AI can be of help in not only hospital settings, but to improve the services of insurance companies," said Brandon Suh, CEO of Lunit. "People normally go to hospitals when they are sick, which usually means that there have been symptoms and the disease is already in progress. If we can find more patients in earlier stages, without symptoms, the chance for survival dramatically increases."
The project was part of the "Cathay Financial Innovation Lab Program" launched by Cathay Financial Holdings last year, of which nine international startups were recruited to cooperate with Cathay Financial Holdings' subsidiaries for a proof-of-concept of six innovative financial services.
"The project showed that Lunit INSIGHT CXR, the AI software for the analysis of chest x-ray images, detected abnormal lung diseases with an accuracy of 90% and also successfully detected small nodules and pulmonary fibrosis that are not easy to distinguish," said Tzu-Ling, Ke, Senior VP of Underwriting Dept., Cathay Life Insurance. "This helps Cathay Life Insurance accurately assess the medical condition of the insured, with enhanced diagnosis of the potential illness prior to underwriting."
Cathay Life Insurance pointed out that it will actively evaluate the adoption of AI technology to assist diagnosis, which is expected to save the reading time and improve underwriting effectiveness through data-driven and innovative technologies.
Yunyoung Noh, Communications Manager, Lunit