The algorithm was trained to find nodules, pneumothorax, effusions, and interstitial opacities that are commonly seen on chest x-ray images. The algorithm classifies lesions as normal or abnormal and highlights suspected abnormal regions.
Results from clinical trials of the algorithm indicate it reduced the average reading time of medical staff by 50% while improving lesion detection performance by 5.8%. Sensitivity, specificity, and accuracy were all improved with Vuno Med Chest X-ray, and the probability of false positives in normal areas was reduced by 50%.
Vuno is also developing algorithms for bone age and brain analysis.
Copyright © 2019 AuntMinnie.com