Ventilation-perfusion scintigraphy in suspected pulmonary embolism: correlation with pulmonary angiography and refinement of criteria for interpretation.

Retrospective analysis of 139 ventilation-perfusion (V/Q) lung scans obtained for suspected pulmonary embolism (PE) was undertaken, using the Biello criteria for interpretation. All scans were correlated with chest radiographs obtained within 24 hours and with pulmonary angiograms obtained within 72 hours of the V/Q study. The prevalence of PE in the high-, intermediate-, and low-probability groups was 89.5%, 49.3%, and 3.6%, respectively. Using minor modifications of these criteria, V/Q scans interpreted as high or low probability had a sensitivity of 97.1%, specificity of 94.3%, and accuracy of 95.7%. Perfusion defects corresponding to radiographic abnormalities that were present for at least 10 days were less likely to be associated with PE. Separate analysis of 27 perfusion-only lung scans that met all other criteria for inclusion in the study confirmed the nonspecificity of perfusion-only lung imaging and indicated the need for ventilation imaging with Tc-99m-DTPA aerosol or Kr-81m gas for portable studies.