Effective prediction of response to cardiac resynchronization therapy using a novel program of gated myocardial perfusion single photon emission computed tomography

    loading  Checking for direct PDF access through Ovid



The limitation of QRS duration as a surrogate measure for left ventricular (LV) mechanical dyssynchrony (LVMD) in cardiac resynchronization therapy (CRT) patient selection encourages seeking alternatives to QRS duration. Exploring the potential of an analysis program of electrocardiographically gated myocardial perfusion single photon emission computed tomography (SPECT) (GMPS) for the estimation of LVMD to predict CRT response.

Methods and results

Twenty-four patients undergoing CRT for advanced heart failure caused by non-ischaemic cardiomyopathy were studied. Gated myocardial perfusion single photon emission computed tomographies were performed in the setting of temporary CRT suspension after 1 week of CRT adoption. The GMPS data were computed with a novel program capable of segmental LV time–volume analysis. When a brain natriuretic peptide (BNP) value decreased >50% at 6-month follow-up, the patient was defined as a CRT responder. Receiver operating characteristic (ROC) curves for identification of responders were analysed for standard deviation of time to end systole (TES-SD) among 17 LV segments. Linear regression analyses demonstrated that an increase in percentage reduction in BNP level at 6-month follow-up was predicted by an increase in TES-SD (R2 = 0.21, P = 0.023). The TES-SD in responders (n= 15, 62.5%) was higher than that in non-responders (100 ± 51 vs. 41 ± 17 ms, P= 0.0008). A cutoff value of TES-SD >49 ms predicted responders with 100% sensitivity and 78.8% specificity and the area under the ROC curve was 0.881 for TES-SD (P= 0.002).


The estimation of LVMD using this novel GMPS program could be an alternative or a complementary approach to QRS duration in CRT patient selection. This finding warrants further assessment of our approach in larger studies.

Related Topics

    loading  Loading Related Articles