AI solution to assist in diagnosing respiratory diseases(Korea AI Technology)
Problem Statement
Chest radiography is one of the most basic and fundamental diagnostic tests used in medicine, accounting for 25% of the annual total numbers of diagnostic imaging procedures. It has been shown that radiologic information changed clinical practice in more than 60% of those who received chest radiography. Unfortunately, miss rates for proper interpretation of chest radiographs go as high as 30% even for experts, leading to increased mortality from treatable diseases. Moreover, the interpretive performance of chest radiographs differ significantly between specialists and non-specialists, up to 30%. Additionally, 10% of chest radiographs are reported to be held back for 30 days until the final report is issued, and only 60% of radiographs are reported by radiologists due to overflowing number of cases to interpret. Improvement in the radiology workflow and efficiency can greatly alleviate the burden.
Key Value Proposition
Prevent difficult cases of chest abnormalities from being missed upon reading chest radiographs.
Help physicians make early diagnosis of chest abnormalities in chest radiographs.
Increases workflow efficiency in interpretation through decreasing reading time by 34%.
Trained to individually detect and locate 10 different radiologic findings
The user can customize detectable findings and its visualization method according to user clinical environment
Automatically generates case report which includes analysis of each radiologic findings and its location information
Provides TB screening AI score to identify tuberculosis on the chest radiograph
Training & Validation/ Xây dựng và đánh giá mô hình
Trained with a large-scale (>200,000 cases), high-quality (clinically/CT-proven cases) training set.
Demonstrated to perform at a standalone accuracy of 97-99% in ROC AUC.19
Certified with CE Mark and approved by Korea MFDS.
Currently in preparation for regulatory approval in various markets worldwide, including FDA.