Acute pulmonary embolism: sensitivity and specificity of ventilation-perfusion scintigraphy in PIOPED II study

Radiology. 2008 Mar;246(3):941-6. doi: 10.1148/radiol.2463070270. Epub 2008 Jan 14.


Purpose: To use Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) II data to retrospectively determine sensitivity and specificity of ventilation-perfusion (V/Q) scintigraphic studies categorized as pulmonary embolism (PE) present or PE absent and the proportion of patients for whom these categories applied.

Materials and methods: The PIOPED II study had institutional review board approval at all participating centers. Patient informed consent was obtained; the study was HIPAA compliant. Approval and consent included those for future retrospective research. Patients in the PIOPED II database of clinical and imaging results were included if they had diagnosis at computed tomographic (CT) angiography, Wells score, and diagnosis at V/Q scanning. V/Q scan central readings were recategorized as PE present (PIOPED II reading = high probability of PE), PE absent (PIOPED II reading = very low probability of PE or normal), or nondiagnostic (PIOPED II reading = low or intermediate probability of PE). A composite reference standard was used: the PIOPED II digital subtraction angiographic (DSA) result, or if there was no definitive DSA result, CT angiographic results that were concordant with the Wells score (ie, positive CT angiographic result and Wells score > 2 or negative CT angiographic result and Wells score < 6). Sensitivity and specificity of recategorized central readings were computed.

Results: With the exclusion of patients with intermediate or low probability, the sensitivity of a high probability (PE present) scan finding was 77.4% (95% confidence interval [CI]: 69.7%, 85.0%), while the specificity of very low probability or normal (PE absent) scan finding was 97.7% (95% CI: 96.4%, 98.9%). The percentage of patients with a PE present or PE absent scan finding was 73.5% (95% CI: 70.7%, 76.4%).

Conclusion: In a population similar to that in PIOPED II, results of V/Q scintigraphy can be diagnostically definitive in a majority of patients; thus, it can be considered an appropriate pulmonary imaging procedure in patients for whom CT angiography may be disadvantageous.

Publication types

  • Multicenter Study
  • Research Support, N.I.H., Extramural

MeSH terms

  • Acute Disease
  • Angiography, Digital Subtraction
  • Female
  • Humans
  • Male
  • Pulmonary Embolism / diagnostic imaging*
  • Radionuclide Imaging
  • Retrospective Studies
  • Sensitivity and Specificity
  • Tomography, X-Ray Computed