Artificial neural network for diagnosis of acute pulmonary embolism: effect of case and observer selection.

1,168 patients with suspected pulmonary embolism (1,064 cases from the PIOPED study with definitive angiographic outcome, and 104 patients in whom pulmonary angiography was essential for diagnosis)

Artificial neural network (ANN) predicting pulmonary embolism using findings from ventilation-perfusion lung scans and chest radiographs

Nuclear medicine physicians (two study-specific readers and the original readers)

Diagnostic performance for detecting or excluding pulmonary embolismsurrogate