Ventilation-perfusion scintigraphy in the PIOPED study. Part II. Evaluation of the scintigraphic criteria and interpretations.

This article presents an evaluation of the criteria used for categorical interpretation of the ventilation-perfusion (V/Q) scans performed in the PIOPED study. In addition, the correlation of percent probability estimates with the actual frequency of pulmonary embolism (PE) is presented. Cases which met the PIOPED criteria for various diagnostic categories were selected by computerized search of the detailed scan descriptions that had been done as part of the study. The process by which the scans were described was detailed in Part I of this report. Most of the criteria appropriately categorized V/Q scans which satisfied them. However, we recommend that three criteria should be reconsidered: 1. A single moderate perfusion defect is appropriately categorized as intermediate, rather than as low probability. 2. Extensive matched V/Q abnormalities are appropriate for low probability, provided that the chest radiograph is clear. On the other hand, single-matched defects may be better categorized as intermediate probability. Although due to the small number of cases with this finding, no definite, statistically founded recommendation can be made. 3. Two segmental mismatches may not be the optimum threshold for high probability, and in some cases should be considered for intermediate probability. However, due to the small number of cases with this finding, no definite, statistically founded recommendation can be made. We suggest that the revised criteria resulting from these adjustments should now be used for the interpretation of V/Q scans.