ABSTRACTWhen selective magnification pulmonary angiography was done within 48 hours of a Ventilation/Perfusion (VP) scan, disagreement between scan and angiogram findings was common. Correlation was poor in both high and low probability of embolism groups as defined by scan. Analysis of individual scan diagnostic criteria showed them to be as unreliable as all of them taken together.The patients coming to angiography were those in which the VP scan disagreed with clinical impressions about the probability of embolism. These problem cases have biased the results against scanning, but they reflect how pulmonary angiography is used in clinical practice. North American studies are reviewed in detail, to show that poor correlation between VP scan and angiography is neither a local aberration nor confined to retrospective analyses.If clinical and scan probabilities of embolism do not agree, pulmonary angiography can safely decide the issue.