Multislice computed tomography coronary angiography for triaging patients with positive radionuclide myocardial perfusion imaging.

Sir, Myocardial perfusion imaging using single-photon emission computed tomography (SPECT) is the most commonly applied diagnostic imaging modality in patients with suspected coronary artery disease in the US, with more than 9 million procedures performed annually. The positive predictive value of SPECT is limited, and in a considerable number of patients with suspected perfusion deficits conventional coronary angiography shows no significant coronary stenoses. Multislice computed tomography (MSCT) coronary angiography allows the noninvasive assessment of coronary artery stenoses, and has a high negative predictive value. Thus, MSCT might be applied as a secondary gate-keeper in patients with positive SPECT, prior to deciding whether to perform conventional coronary angiography in these patients. We aimed to investigate the role of MSCT (Aquilion, Toshiba), using 16 simultaneous detector rows (according to a previously described protocol) to triage 38 patients (63 8 years) with positive SPECT myocardial perfusion imaging (Tc, n1⁄4 24; Tl, n1⁄4 14; both rest and stress) who were scheduled for conventional coronary angiography. Patients with an equivocal SPECT examination (e.g. examinations in which neither the target heart rate nor a perfusion deficit was achieved) were not included in the study, and SPECT was considered positive when at least either the rest or the stress examination showed a significant perfusion deficit in any myocardial area. Readers of the three tests were blinded against the results of the other tests. The 17 myocardial segments/areas on SPECT were assigned to one of the three coronary arteries seen on MSCT, with post-hoc adjustment for the coronary artery distribution type as defined by conventional coronary angiography according to the detailed recommendations in the American Heart Association segmentation scheme publication. The study was approved by the institutional review board, and all patients gave written informed consent. A contingency analysis with a 2 or Fisher’s exact test was used to compare the per-patient positive predictive value and per-artery/territory diagnostic performance of SPECT and MSCT with conventional coronary angiography as the reference standard. Twenty-one of the 38 patients (55%, Figure 1) with a positive SPECT result showed significant coronary artery disease (at least 50% diameter stenosis on the worst view) on quantitative analysis of conventional coronary angiography. MSCT correctly identified 90% (19/21) of the patients with coronary disease, while the positive predictive value was significantly superior to that of SPECT: 95% (19/20) vs. 55% (21/38), p<0.002. The triage of patients with positive SPECT was significantly influenced by using MSCT, which would have avoided 15 of 17 unnecessary conventional coronary angiographies (specificity 88%, 15/17) (Figure 1). One of the false-negative patients with MSCT did not undergo revascularization because this patient had single-vessel disease with a 50% stenosis of the right coronary artery, which was considered not to require revascularization. One patient who had no significant disease on conventional coronary angiography was nondiagnostic on MSCT because of a singular motion artifact in the right coronary artery. In the per-coronary artery analysis, MSCT also had significantly higher diagnostic accuracy (85% [97/114] vs. 66% [75/114], p<0.001), specificity (90% [62/69] vs. 62% [43/69], p<0.001), and positive and negative predictive values (92% [35/ 38] vs. 55% [32/58], p<0.001 and 91% [62/68] vs. 77% [43/56], p<0.05, respectively) than SPECT. Of the 45 significant stenoses on conventional coronary angiography, 13 were not correctly detected using SPECT. The mean percentage diameter stenosis of these 13 false-negative lesions was 87 7% (range: 72–100%), and 11 (85%) of these required subsequent revascularization with coronary artery bypass surgery (n1⁄4 7) or stent placement (n1⁄4 4), indicating that these lesions missed by SPECT were considered clinically relevant. None of the 26 falsepositive myocardial territories on SPECT showed borderline or significant stenoses of the respective