Purpose: To evaluate the relationship between vFAI and myocardial perfusion and test the value of the former for predicting inducible perfusion abnormalities on SPECT-MPI.
Methods: vFAI and SPECT-MPI were assessed in 201 patients (mean age 60.2 ± 9.43 years) with intermediate pre-test likelihood of CAD. Coronary artery stenoses =50% were considered angiographically significant. SPECT-MPI was considered positive for significant CAD, when the summed difference score (SDS) between stress and rest images was =2.
Results: Hybrid imaging analysis was performed in CTCA and SPECT datasets and from 603 coronary arteries in total, 566 vessels could be analysed for vFAI assessment. An inducible perfusion abnormality was present in 55 coronary territories. vFAI was lower in vessels with downstream inducible perfusion defects (0.71±0.34 vs. 0.84±0.18, p=0.008). Overall, there was a weak negative correlation between vFAI and myocardial perfusion defects (r= -0.181 p=0.001). SPECT and vFAI were both normal in 61.13%, abnormal in 5.83% and discordant in 33.04% of coronary territories, respectively, ?= 0.128. By ROC curve analysis, the optimal cut off value of vFAI for predicting a positive SPECT result was =0.85 with a sensitivity of 60% and specificity of 67.7% (AUC=0.653, 95% CI: 0.612- 0.692). The corresponding values for CTCA were 41.8% and 88.3% respectively. Addition of vFAI to anatomical CTCA data increased the ability for predicting a positive SPECT result (AUCctca+vfa=0.715, 95% CI 0.676- 0.752 vs. AUCctca= 0.650, 95% CI, 0.610-0.690, p=0.009). The combined model showed a sensitivity of 72.7% and a specificity of 62%.
Conclusion: In this population with a low prevalence of myocardial ischaemia, there is a weak relationship between vFAI and myocardial perfusion. CTCA demonstrates a low sensitivity for predicting inducible myocardial perfusion defects on SPECT-MPI, but excellent specificity. In combination with vFAI predictive performance improves significantly driven by an increase in sensitivity.