AI Assistance Significantly Reduces Missed Rate of IPE, Shortens Time to Dx in Oncology

Background: Pulmonary embolism (PE) may be symptomatic or incidental. Incidental pulmonary embolism (IPE) is actually common in oncology patients and places the patient in a high-risk category for recurrence. IPEs can be seen on about 3% of routine contrast-enhanced CTs in oncology patients, and are associated with poor outcomes (disease progression and reduced overall survival). Timely diagnosis and proper management of IPEs is essential. With increasing work volume in radiology departments, increasing turnaround times (TATs) are common and pose risks for delay in diagnosis of critical findings. Artificial intelligence (AI) programs can automatically detect critical findings and can be used to prioritize scans on the reading worklist, and have been used with varying results to detect other critical imaging findings. IPE detection presents additional challenges because filling defects may be more difficult to detect if there is suboptimal contrast enhancement of the pulmonary arterie more...