Triage of patients for attenuation-corrected stress-first Tc-99m SPECT MPI using a simplified clinical pre-test...

Stress-only Tc-99m SPECT MPI saves time and reduces radiation exposure while a normal study has a benign prognosis. However, no guidelines exist as to which patients should undergo stress-first MPI. The purpose of this study was to validate a previously published pre-test prediction scoring model and refine the stress-first triage process further if possible.

We retrospectively reviewed all patients who underwent an attenuation-corrected Tc-99m SPECT MPI over a 39-month period. Based on 17-segment model semi-quantitative scoring, a successful stress-first MPI was defined as a summed stress attenuation-corrected score ≤ 1. Based on results from multivariate analysis, the previously published prediction score (comprised eight clinical and demographic variables) was compared to triage based on coronary artery disease (CAD) status alone and with the addition of other highly associated variables. Logistic regression and Chi-squared analyses were used to determine the magnitude of variable effect and to compare model results.

A total of 2,277 patients were included, and the prediction score successfully stratified patients into low-risk (91.1% successful stress-first), intermediate-risk (79.4%), and high-risk (50.7%) groups. Comparing the use of the prediction score to the use of a history of CAD as the only triage factor, 69.0% of patients would be accurately triaged using the prediction score with a cutoff of 7 (maximized sensitivity and specificity), while 78.6% were correctly triaged with CAD status alone (P < .0001). The addition of variables highly associated with a successful stress-first protocol (congestive heart failure [OR 3.4] and an abnormal resting ECG [OR 2.1]) to CAD status further enhanced triage accuracy to 81% (P < .0001).

While the previously described prediction score effectively identifies patients who can successfully undergo stress-first MPI, it is cumbersome. Triaging based solely on CAD status and with the addition of other key variables is practical and provides improved predictive accuracy for successful stress-first MPI. Utilizing this simplified pre-test scoring model may allow for wider adoption of stress-first imaging protocols which have clear advantages over traditional rest-stress protocols.