very low probability lung scan findings: a need for change

this study was conducted to determine the prevalence of pe in patients with v/q scans interpreted as representing a very low probability (vlp) of pe (1-3 small segmental perfusion defects with a normal chest radiograph). an analysis of the entire data set obtained during the pioped study was performed. of the total of 1,359 patients in whom the presence or absence of pe was confirmed, 80 (6%) patients had the vlp pattern. only 2 of 80 patients (2.5%) had pe, both patients had partially occlusive thrombus within a single lower lobe segment. both patients were also referred from surgicl wards, had a history of previous surgery within 3 months and a history of immobilization within 3 days prior to scan. both patients with pe and vlp patterns had intermediate pretest odds. however, 36 patients with similar lung scan findings and no evidence of pe also had intermediate pretest odds. the remaining 40 patients had low pretest odds (two patients did not have pretest odds entered into the data base). we conclude that the very law probability interpretation criterion is a valid diagnostic category and should be a separate part of any schema for interpreting v/q lung scans. the pretest clinical likelihood of pe did not change the post test probability of pe for very low probability lung scan interpretation as they did for low, intermediate or high probability lung scan interpretations.