Stress-only myocardial perfusion scintigraphy: a prospective study on the accuracy and observer agreement with quantitative coronary angiography as the gold standard.

OBJECTIVE: Patients with normal stress perfusion have an excellent prognosis. Prospective studies on the diagnostic accuracy of stress-only scans with contemporary, independent examinations as gold standards are lacking. PATIENTS AND METHODS: A total of 109 patients with typical angina and no previous coronary artery disease underwent a 2-day stress (exercise)/rest, gated, and attenuation-corrected (AC), 99m-technetium-sestamibi perfusion study, followed by invasive coronary angiography. The stress datasets were evaluated twice by four physicians with two different training levels (expert and novice): familiar and unfamiliar with AC. The two experts also made a consensus reading of the integrated stress-rest datasets. The consensus reading and quantitative data from the invasive coronary angiography were applied as reference methods.
RESULTS: The sensitivity/specificity were 0.92-1.00/0.73-0.90 (reference: expert consensus reading), 0.93-0.96/0.63-0.82 (reference: ≥1 stenosis>70%), and 0.75-0.88/0.70-0.88 (reference: ≥1 stenosis>50%). The four readers showed a high and fairly equal sensitivity independent of their familiarity with AC. The expert familiar with AC had the highest specificity independent of the reference method. The intraobserver and interobserver agreements on the stress-only readings were good (readers without AC experience) to excellent (readers with AC experience).
CONCLUSION: AC stress-only images yielded a high sensitivity independent of the training level and experience with AC of the nuclear physician, whereas the specificity correlated positively with both. Interobserver and intraobserver agreements tended to be the best for physicians with AC experience.