Background-We have advocated the use of a 5-category "normal," "probably normal," "equivocal," "probably abnormal," and "definitely abnormal" approach to final interpretation of myocardial perfusion single-photon emission computed tomography (SPECT). The prognostic value of expressing levels of certainty compared with a dichotomous normal/abnormal classification or categories for summed stress scores is unclear. Methods and Results-Myocardial perfusion SPECT (MPS) was visually assessed using a standard semiquantitative approach, yielding summed scores that were used for preliminary interpretation using 5 levels of certainty. The interpreter was permitted to then shift the level of certainty in the final interpretation by 1 degree, based on nonperfusion MPS variables and available clinical information. To examine the prognostic value of expressing levels of clinical certainty, we evaluated 20 740 unique consecutive patients who underwent rest Tl-201/stress Tc-99m sestamibi MPS (34.3% vasodilator stress), of whom 845 (4.4%) were lost to follow-up and 1695 were excluded from prognostic analysis due to an early revascularization (Ͻ60 days after MPS). The remaining 18 200 patients (59.1% men; age, 65Ϯ13 years) were followed up for cardiac death for a mean of 2.7Ϯ1.7 years. During the follow-up, a total of 591 cardiac death events occurred. By univariable analysis, there were substantial differences in the distribution of follow-up cardiac death by the category of clinical MPS certainty. The clinical certainty was found to be an independent multivariable predictor of cardiac death in the study population and better identified patients at increased risk of cardiac death than the approaches based solely on the standard categories of summed perfusion scores or based solely on categories of segmental perfusion scores.
Conclusions-The