Time trend analysis of pulmonary embolism diagnosis with single-photon emission computed tomography ventilation/perfusion imaging

Objective Following the implementation of single-photon emission computed tomography ventilation/perfusion (SPECT V/Q) in our department, we previously carried out an observational study over the period 2011-2013 that showed the safety of SPECT V/Q to rule out pulmonary embolism (PE). However, PE prevalence (28%) was high compared with some studies. Our hypothesis was that the use of SPECT V/Q may be responsible for an overdiagnosis of PE, especially at the implementation phase of the test. The aim was to carry out a time trend analysis of the evolution over the years of PE diagnosis with SPECT V/Q. Patients and methods To evaluate the time trend of SPECT V/Q results, the conclusions of all SPECT V/Q scans performed for a suspected acute PE in our department from April 2011 to December 2016 were collected. To confirm the safety of a diagnostic management on the basis of SPECT V/Q to rule out PE, a retrospective observational study of the outcome of patients assessed with SPECT V/Q for suspected acute PE was carried out during the period 2014-2016. SPECT V/Q was interpreted using a diagnostic cutoff of one segmental or two subsegmental mismatches. The final diagnostic conclusion was established by the physician responsible for patient care. The safety of SPECT V/Q was assessed by monitoring the risk of thromboembolic events during the 3-month follow-up period in patients left untreated on the basis of a negative SPECT V/Q. Results Over the period 2011-2016, an average 16% (95% confidence interval: 10-22%) annual reduction in the proportion of positive SPECT V/Q results was observed. The percentage of positive SPECT V/Q results was 25.3% over the period 2011-2013 compared with 16.3% in 2014-2016 (P<0.0001). Out of the 696 patients analyzed over the period 2014-2016, the SPECT V/Q was positive, indeterminate, and negative in 118, 3, and 575 patients, respectively. The 3-month thromboembolic risk in patients in whom PE was deemed absent according to the SPECT V/Q was 3/543=0.55% (95% confidence interval 0.19-1.61). Conclusion A continuous decrease in the annual proportion of positive SPECT V/Q results was mainly observed while maintaining the safety of the test to rule out PE. Explanatory factors include the introduction of a combined SPECT/CT scan, increased experience and familiarity with the test, and an increased awareness of the risks and consequences of overdiagnosing PE.