A simplified wall-based model for regional innervation/perfusion mismatch assessed by cardiac 123I-mIBG and rest 99mTc-tetrofosmin SPECT to predict arrhythmic events in ischaemic heart failure

Abstract

Abstract Aims Cardiac 123iodine-meta-iodobenzylguanidine (123I-mIBG) single-photon emission computed tomography (SPECT) imaging provides information on regional myocardial innervation. However, the value of the commonly used 17-segment summed defect score (SDS) as a prognostic marker is uncertain. The present study examined whether a simpler regional scoring approach for evaluation of 123I-mIBG SPECT combined with rest 99mTc-tetrofosmin SPECT myocardial perfusion imaging could improve prediction of arrhythmic events (AEs) in patients with ischaemic heart failure (HF). Methods and Results Five hundred and two ischaemic HF subjects of the ADMIRE-HF study with complete cardiac 123I-mIBG and rest 99mTc-tetrofosmin SPECT studies were included. Both SPECT image sets were read together by two experienced nuclear imagers and scored by consensus. In addition to standard 17-segment scoring, the readers classified walls (i.e. anterior, lateral, inferior, septum and apex) as normal, matched defect, mismatched (innervation defect > perfusion defect), or reverse mismatched (perfusion defect > innervation defect). Cox proportional hazards ratios (HRs) were used to determine if age, body mass index, functional class, left ventricular ejection fraction (LVEF), B-type natriuretic peptide (BNP), norepinephrine, 123I-mIBG SDS, 99mTc-tetrofosmin SDS, innervation/perfusion mismatch SDS, and our simplified visual innervation/perfusion wall classification were associated with occurrence of AEs (i.e. sudden cardiac death, sustained ventricular tachycardia, resuscitated cardiac arrest, appropriate implantable cardioverter-defibrillator therapy). At 2-year median follow-up, 52 subjects (10.4%) had AEs. Subjects with 1 or 2 mismatched walls were twice as likely to have AEs compared with subjects with either 0 or 3–5 mismatched walls (16.3% vs. 8.3%, P = 0.010). Cox regression analyses showed that patients with a visual mismatch in 1-2 walls had an almost two times higher risk of AEs [HR 2.084 (1.109–3.914), P = 0.001]. None of the other innervation, perfusion and mismatch scores using standard 17 segments were associated with AEs. BNP (ng/L) was the only non-imaging parameter associated with AEs. Conclusion A visual left ventricular wall-level based scoring method identified highest AE risk in ischaemic HF subjects with intermediate levels of innervation/perfusion mismatches. This simple technique for the evaluation of SPECT studies, which are often challenging in HF subjects, seems to be superior to the 17-segment scoring method.