ABC | Volume 113, Nº6, December 2019

Original Article Ket et al. Myocardial ischemia by SPECT and CCTA Arq Bras Cardiol. 2019; 113(6):1092-1101 physician, except in case of identification of significant lesions in the trunk of the left coronary artery or in the LAD coronary artery detected by CCTA. The inclusion criteria were patients with medical request for stress/rest myocardial perfusion scintigraphy to assess CAD. Patients with creatinine above 1.5 mg/dl, obstructive pulmonary chronic disease, asthmatic patients, patients who were allergic to iodinated contrast material or for whom dipyridamole or metoprolol was contraindicated and any other aspect that the researcher deemed limiting to the method were excluded. The exams were performed with the following flow: first the patitent was selected at the Nuclear Medicine Service and, after signing the free and informed term of consent, the patient was referred to the service of radiology to undergo CCTA (perfusion at rest) followed by myocardial perfusion under pharmacological stress with dipyridamole. Before the infusion of iodinated contrast material, during stress-induced hyperemia, 2-methoxy- isobutyl-isonitrile-99mTc (sestamibi- 99m Tc) was infused at the computed tomography room. The CCTA protocol included two imaging acquisitions: one for coronary anatomy assessment by CTA, which is also used to assess myocardial perfusion at rest; and a second myocardial perfusion under pharmacological stress performed shortly after the first acquisition. The mean acquisition time was 30 ± 5 minutes. The first acquisition was volumetric and static, having been performed retrospectively using the following parameters: 120 KV, 240-400 mA and 512 × 512 matrix, 70 ml iodinated contrast media at a concentration of 350 mg/mL, infused at 5 ml/s. The second acquisition was performed following the same parameters and soon after 5 to 6 minutes from the beginning of dipyridamole infusion (Persantin ® , Boehringer Ingelheim España S.A., España) (0.56 mg/kg/4 minutes). We chose to infuse it by hand, after images of the ascending aorta were blurred using iodinated contrast media, because it facilitates the correct selection of the beginning of acquisition, especially in the stress phase, which must occur a little earlier than usual for other coronary studies. During dipyridamole infusion, the patients’ heart rate, blood pressure and symptoms were monitored every minute. Immediately after the conclusion of stress perfusion evaluation, 240 mg of aminophylline were administered (Minoton ® , Teuto Brasileiro S.A., Brazil) to reverse the vasodilatation effect of the stress agent. This CT protocol was idealized in a 64-detector tomographic angiography (Brilliance CT 64-slice, Philips, Netherlands) and the mean dose of radiation was 12.1 ± 5.2 mSv. Myocardial perfusion scintigraphy (SPECT) was performed with intravenous infusion of Tc- 99m sestamibi, using a single- day protocol (rest-stress). The patient was referred to the Radiology Sector, and the injection of the radiotracer was performed at the tomography room, in the Radiology Sector. Soon after CT was finished, the patient was referred to stress imaging acquisition (first-passage perfusion) with a maximum interval of 30 minutes. After this stage and an interval between 60 and 120 minutes, the rest phase was performed with a new injection of Tc-99m sestamibi. The mean dose administered in each stage was 925 MBq. The images were acquired 30 to 90 minutes after intravenous administration of the agent. A total of 64 projection images of the chest were acquired from an arc of 180 degrees, from the 45-degree right anterior oblique view to the 45-degree left posterior oblique view. In the rest phase, the acquisition time was 30 seconds per projection; in the stress phase, the acquisition time was 30 seconds per projection as well. In both the stress and rest phases, ECG-synchronized image acquisition was performed. To analyze the correlation between the myocardial perfusion techniques, the following criterion was used to characterize myocardial ischemia: there should be perfusion defects on stress images with no correspondent perfusion defect on rest images of both CCTA and SPECT. Myocardial perfusion and CCTA were assessed visually and semi-quantitatively by two blinded and independent observers, without any knowledge of clinical data or other exams. Disagreements were resolved by means of consensus. The degree of coronary stenosis was graded, according with visual and semi-quantitative assessment by CCTA, as non-significant (< 50% reduction in luminal diameter) and significant (> 50% reduction in luminal diameter). Statistical Analysis All continuous variables were expressed as mean ± standard deviation and the categorical variables as number and percentage. Fisher’s exact test was used to compare between proportions. Based on CCTA fidings, the patients were grouped according with the presence or not of significant CAD. The criterion used to define significant CAD was existence of obstruction > 50% of the lumen of coronary arteries. Sensitivity and specificity were estimated and displayed as number and percentage. The analysis of the area under the ROC curve was used to identify the efficacy of CCTA (CT perfusion) and scintigraphy (SPECT) in the diagnosis of perfusion data in this study. The research was conducted on two groups: one with stenosis > 50% on anatomical assessment by CCTA, as the "true positive" surrogate marker in this population, compared with the group with stenosis < 50% in the same method as the “true negative” (AUC ≥ 0.5 to < 0.7 = poor fit; AUC ≥ 0.7 to < 0.9 = good fit; AUC ≥ 0.9 to 1.0 = excellent fit). Intra‑ and interobserver agreement was obtained by using intraclass correlation coefficient reliability analysis (CCI < 0.40: poor agreement; CCI = 0.40 to 0.59: fair agreement; CCI = 0.60 to 0.74: good agreement; CCI = 0.75 to 1.00: excellent agreement). About 43% of perfusions performed using CCTA techniques (15/35) were reassessed by the same observer; the analysis was performed by a second independent observer to characterize the variability between the analyses. A total of 1,440 segments were assessed using the 16-segment model of the American College of Cardiology (ACC) and the American Heart Association (AHA), with 240 LV segments being analyzed by observer 1 at rest and, subsequently, under pharmacological stress, totaling 480 segments. Observer 1 repeated this analysis after a 3-month period, blinded to the previous analysis. Observer 2 performed the independent analysis, blind and with no previous agreement with the first observer. Both observers have more than 10 years experience in performing CCTA. 1094