Between January 2007 and December 2018, 144 consecutive patients post-PCI with chronic symptoms were referred for hybrid imaging with CCTA and 2-day rest/stress SPECT MPI and were retrospectively analyzed Figure 1. The exclusion criteria included: Patients who lost their follow up, patients who had only CCTA with no SPECT MPI imaging and patients who had PCI after coronary artery bypass surgery (CABG). However, patients who had PCI as a part of hybrid revascularization during CABG were included.

All patients signed an informed consent to be enrolled in a hybrid imaging registry (CCTA and SPECT MPI) with long-term follow-up. This study was approved by the local ethics committee.

A 64-slice CT scan was used for all patients (GE Healthcare, LightSpeed VCT, Chicago, IL, USA). All patients with heart rate >60 beats/minute received intravenous or oral beta blocker before the scan to lower heart rate. Sublingual 400 ug nitroglycerin was given before the scans if no contraindications. The scan protocol included a non-enhanced CT scan for calcium score measurement followed by the CCTA acquisition using a timing bolus technique to determine the contrast delay time.

CCTA images were reconstructed and analyzed on GE Healthcare Advantage Workstations (version 4.4). For stent image processing, a sharp kernel was employed for edge enhancement. Agatston score was measured in all the coronary arteries excluding the stent locations. For the purpose of the study, two cut-off values were used to define obstructive lesions (≥50% and ≥70% stenosis) in coronary segments greater than 1.5 mm in diameter.

A 2-day rest/stress protocol with Tc-99 m Sestamibi was used in all patients. The rest scans were performed on the same day as the CCTA. Stress scans were done the following day. Treadmill exercise was used for physically active patients while pharmacological stress with dipyridamole was used for those who were unlikely to perform adequately on the treadmill. Vasodilator stress was supplemented with supine cycle ergometer exercise if patient capable. A CZT gamma camera (GE Discovery NM 530c) was used for image acquisition in patients acquired from 2014 to 2018. A two headed sodium iodide gamma camera (GE Hawkeye) was used for patients acquired from 2007 to 2013. SPECT images were analyzed using cardiac software (Cedars QGS/QPS) and/or Emory Cardiac Toolbox (ECT) on GE Xeleris Nuclear Medicine workstations. Scans were defined as normal/near normal if perfusion defects were <5% of LV myocardial mass. Abnormal perfusion defects were defined as mild if involving 5%–10%, moderate if 10%–20% and large if >20% of the LV mass.

Patients were followed for any major adverse cardiovascular event (MACE) including: (1) All-cause mortality, (2) Non-fatal myocardial infarction (MI), (3) Unplanned revascularization with PCI or CABG (unplanned if the procedure was performed more than three months after the indexed CTA), (4) Cerebrovascular accident, (5) Hospitalization for arrhythmia or heart failure. We defined hard cardiac events (HCE) as: (1) cardiac death, (2) non-fatal MI or (3) unplanned revascularization.

Continuous variables are presented as means and standard deviations while categorical variables are presented as frequencies and percentages. We did descriptive statistics analysis using Wilcoxon rank-sum test to compare continuous variables and Pearson's chi-square test to compare categorical variables between patients with and without MACE. We also used univariable regression models to analyze the association of each independent clinical variables, CCTA measurements and SPECT MPI finding with outcome of MACE and HCE. Finally, we ran Kaplan Meier analyses using hazard-models to analyze the prognostic value of SPECT MPI and CCTA.

In total, 144 were included in the study. 25 patients were lost to follow-up and were excluded. 119/144 (84%) were included in the final analysis, Figure 1. The mean age at enrollment was 62 ± 9 years. Men were 87 (73.1%), other baseline characteristics are shown in Table 1.

Majority of patients 75/119 (63%) had one stent. 36/119 (30.3%) had two stents. 8/119 (6.7%) had three stents or more. 46/175 (26.3%) of the stents had a diameter less than 3 mm. 76 stent were deployed in left anterior descending artery (LAD), 52 stents in the right coronary artery (RCA), 42 stents in left circumflex artery (LCx) and 5 stents in the left main coronary artery. 81/119 (68%) of the patients had either patent stents or non-critical stenosis, while 14/119 (11.8%) had significant in-stent restenosis >50% (8 patients with >70% / total occlusion). In 24/199 (20.2%) the stent was non-interpretable.

In total, 67 perfusion defects were reported (38 reversible and 29 fixed). 17/38 (44.7%) of the reversible defects were in the stented coronary territories and 21/38 (55.3%) in non-stented coronary territories. 7 (24%) of the fixed perfusion defects were in the stented coronary artery territories while 22/38 (76%) were in the non-stented territories.

Endpoints (events) were during a follow-up period of 7.2 ± 3.4 years. In total, 45/119 (37.8%) patients experienced 57 MACE: Ten deaths (2 cardiac deaths and 8 non-cardiac deaths), 29 acute coronary syndrome including non-fatal MI (19 required PCI and 6 required CABG), 7 hospitalizations for heart failure, 6 cerebrovascular accidents, 5 new atrial fibrillation.

HCE were 31 (2 cardiac deaths and 29 of nonfatal MI and/or acute contrary syndrome requiring revascularization). Coronary artery disease risk factors and stent number did not differ between patients with and without MAEC or HCEs; other baseline characteristics of patients with and without MACE/HCE are illustrated in (Tables 2, 3).

Univariate Cox regression analysis showed that coronary artery stenosis of ≥50% and ≥70% and abnormal MPI were significant predictors of MACE (p = 0.037, 0.018 and 0.026), respectively, Table 2 In contrast, HCEs were significantly predicted with coronary stenosis of ≥50% and ≥70% (p = 0.004, 0.007), respectively; while abnormal SPECT did not predict HCEs (p = 0.062,) Table 3. Multivariable regression models including all variables that were significant with the univariable regression analysis and found that none of these variables have a significant prediction of MACE or HACE.

Kaplan-Meier survival analysis showed that coronary stenosis of greater than 50% or 70% and presence of revisable ischemia with SPECT are significant predictors of total MACE (p = 0.016, 0.033 and 0.023), respectively, Figure 2. However, CT stenosis were significant predictors of HCE (p = 0.008 and 0.009), while abnormal MPI was not a predictor of HCE (p = 0.068), Figure 3.