Medical AI company Lunit announced on the 28th that it will present seven latest research achievements utilizing its AI imaging analysis solution, Lunit INSIGHT, at the 2024 European Congress of Radiology (ECR 2024) held in Vienna, Austria, from the 28th to the 3rd of next month (local time).
Following its continuous demonstration of the clinical effectiveness of chest X-ray and mammography image analysis technology, Lunit plans to share innovative diagnostic technology achievements through four oral presentations and three poster presentations, which are major studies of the congress.
Among these, the most significant study is an oral presentation analyzing 16,996 chest X-ray images, showing how AI identifies lung diseases and assists physicians' diagnostic decisions. This study used Lunit INSIGHT CXR, a chest X-ray AI imaging analysis solution, to interpret chest X-ray images and precisely calculate the sensitivity and specificity for various lung diseases. Sensitivity refers to the probability of diagnosing actual patients with the disease as positive, while specificity refers to the probability of diagnosing patients without the disease as negative.
As a result, the sensitivity for each disease reached up to 94%, and specificity up to 99%. Lunit explained that this study "suggests that AI can be particularly useful for rapid interpretation in emergency situations and accurate interpretation of relatively low-risk diseases," and "provides insights into optimal AI threshold settings for each disease."
Additionally, Lunit will present an oral presentation comparing the lung nodule detection capabilities of commercial AI solutions. This study, first published in the international radiology journal Radiology in January, analyzed data from 386 patients using seven AI solutions, including Lunit INSIGHT CXR and AI solutions released in Germany, France, Australia, and China. Among all AI solutions analyzed, Lunit INSIGHT CXR showed the best performance with an area under the curve (AUC) of 0.93, a performance evaluation metric for AI models. AUC is an indicator where values closer to 1 indicate better performance, and models with AUC above 0.8 are classified as high-performance models.
Furthermore, a study exploring the role of AI in the European mammography screening process, which is based on double reading by two radiologists, will also be presented. Using 249,402 mammography cases collected from Danish medical institutions, this study analyzed three scenarios in which the mammography AI imaging analysis solution Lunit INSIGHT MMG could replace the roles of the first and second reading radiologists. As a result, AI showed statistically no significant difference in performance compared to the existing double reading in almost all cases. Notably, in the second scenario, it demonstrated clinical effectiveness by increasing the positive predictive value (PPV) by 4.7%.
Seobum Seok, CEO of Lunit, said, “Continuously presenting research results at globally prestigious congresses like ECR is one of Lunit’s core competitive advantages,” adding, “We will continue to expand our influence in the European market, where interest in technological advancement is particularly high.”
Meanwhile, Lunit will participate in ECR 2024 together with Volpara Health Technologies, a New Zealand-based multinational company currently undergoing acquisition. Lunit has signed a distribution agreement with Volpara for Lunit INSIGHT MMG, and at this congress, Volpara will introduce Lunit INSIGHT MMG as part of its product lineup to overseas customers at its booth.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
