ADDING A VISUAL LINEAR SCALE PROBABILITY TO THE PIOPED PROBABILITY OFPULMONARY-EMBOLISM

Purpose: Reporting a lung scintigraphy diagnosis as a PIOPED categoric al probability of pulmonary embolism offers the clinician a wide range of interpretation. Therefore the purpose of this study was to analyze the impact on lung scintigraphy reporting of adding a visual linear s cale (VLS) probability assessment to the ordinary PIOPED categorical p robability. Material and Methods: The study material was a re-evaluati on of lung scintigrams from a prospective study of 170 patients. All p atients had been examined by lung scintigraphy and pulmonary angiograp hy. The scintigrams were re-evaluated by 3 raters, and the probability of pulmonary embolism was estimated by the PIOPED categorization and by a VLS probability. The test was repeated after 6 months. Results: T here was no significant difference (p>0.05) in the area under the ROC curve between the PIOPED categorization and the VLS for any of the 3 r aters. Analysis of agreement among raters and for repeatability demons trated low agreement in the mid-range of probabilities. Conclusion: A VLS probability estimate did not significantly improve the overall acc uracy of the diagnosis compared to the categorical PIOPED probability assessment alone. From the data of our present study we cannot recomme nd the addition of a VLS score to the PIOPED categorization.