Reducing False Positives and False Negatives in Artificial Intelligence-assisted Imaging Screening: A Critical Review of Accuracy, Error Mechanisms and Clinical Translation
Imaging-based screening programmes are constrained by two opposing failures: false-positive results that generate unnecessary recall, anxiety, further imaging and biopsy, and false-negative results that permit disease to progress undetected until the next screening round or until symptoms appear. Artificial intelligenc...