Realize’s CT lung nodule detection application, now available for investigational use, boasts a nodule miss rate of just 4% at a specificity of 4 false positives per scan. For comparison, a leading FDA-approved competitor has a 21% miss rate at the same specificity, and radiologist miss rates as high as 56% have been documented. Solutions for detecting lung nodules in chest X-ray, pulmonary embolism in CT angiography, and intracranial hemorrhage in head CT are currently in development.
Realize’s CT lung nodule detection application, available soon for investigational use, provides a sensitivity rate of 96% at 4 false positives per scan.
Realize is developing broad capability detecting abnormalities in chest X-ray, pulmonary embolism in CT angiography, and intracranial hemorrhage in head CT.
We are developing systems to prioritize scans with algorithmically-suspected pulmonary embolism, and intracranial hemorrhage in emergency work queues, shaving critical minutes off of treatment response times.
Early detection can lead to vastly improved outcomes for patients with lung cancer. However, lung nodules pose a challenge to clinicians, varying greatly in size, texture, density, etc. Our CT lung solution, trained on thousands of real nodules across hundreds of scans, greatly outperforms competing products which have been clinically proven to increase the accuracy of radiologists. We anticipate similar performance for our forthcoming chest X-ray, pulmonary embolism, and intracranial hemorrhage solutions. In addition, we are currently developing systems to prioritize scans with algorithmically-suspected pulmonary embolism or intracranial hemorrhage in emergency work queues, shaving critical minutes off of treatment response times.
CAUTION–Investigational device. Limited by Federal (or United States) law to investigational use. We are currently working towards FDA approval for our detection products, but welcome anyone to try or pilot our technology on retrospective data. Our prioritization products do not require FDA approval, provided that clinicians are not informed of the decisions made by the algorithm, and may be used commercially once development is complete.