Quantitative Gated SPECT–Derived Phase Analysis on Gated Myocardial Perfusion SPECT Detects Left Ventricular Dyssynchrony and Predicts Response to Cardiac Resynchronization Therapy

Abstract

The significance of left ventricular (LV) dyssynchrony for the prediction of response to cardiac resynchronization therapy (CRT) has been demonstrated. Parameters reflecting LV dyssynchrony (phase SD, histogram bandwidth) can be derived from gated myocardial perfusion SPECT (GMPS) using phase analysis. The feasibility of LV dyssynchrony assessment with phase analysis on GMPS using Quantitative Gated SPECT (QGS) software has not been demonstrated in patients undergoing CRT. The aim of the present study was to validate the QGS algorithm for phase analysis on GMPS in a direct comparison with echocardiography using tissue Doppler imaging (TDI) for LV dyssynchrony assessment. Also, prediction of response to CRT using GMPS and phase analysis was evaluated. Methods: Patients (n 5 40) with severe heart failure (New York Heart Association class III-IV), an LV ejection fraction of no more than 35%, and a QRS complex greater than or equal to 120 ms were evaluated for LV dyssynchrony using GMPS and echocardiography with TDI. At baseline and after 6 mo of CRT, clinical status, LV volumes, and LV ejection fraction were evaluated. Patients with functional improvement were classified as CRT responders. Results: Both histogram bandwidth (r 5 0.69, r 2 5 0.48, SEE 5 25.4, P , 0.01) and phase SD (r 5 0.65, r 2 5 0.42, SEE 5 26.8, P , 0.01) derived from GMPS correlated significantly with TDI for assessment of LV dyssynchrony. At baseline, CRT responders showed a significantly larger histogram bandwidth (94°6 23°vs. 68°6 21°, P , 0.01) and a larger phase SD (26°6 6°vs. 18°6 5°, P , 0.01) than did nonresponders. Receiver-operating-characteristic curve analysis identified an optimal cutoff value of 72.5°for histogram bandwidth to predict CRT response, yielding a sensitivity of 83% and a specificity of 81%. For phase SD, sensitivity and specificity similar to those for histogram bandwidth were obtained at a cutoff value of 19.6°. Conclusion: QGS phase analysis on GMPS correlated significantly with TDI for the assessment of LV dyssynchrony. Moreover, a high accuracy for prediction of response to CRT was obtained using either histogram bandwidth or phase SD.

Myocardial Viability Assessment is Still Alive and an Important Element in Predicting Prognosis and Providing Optimal Management for Ischemic Heart Failure

“…Phase analysis of gated SPECT data determined the onset of mechanical contraction by measuring regional count variation over cardiac cycle, which is consistent with wall thickening, with Fourier transformation [52]. Phase standard deviation and the phase histogram bandwidth predict response to CRT (Figure 7) [48]. Responders demonstrated larger histogram bandwidth (94 ± 23° vs 68± 21°, p < 0.01) and stander deviation (26±6° vs 18±5°, p < 0.01).…”

“…Responders demonstrated larger histogram bandwidth (94 ± 23° vs 68± 21°, p < 0.01) and stander deviation (26±6° vs 18±5°, p < 0.01). Cutoff value of 102.5° for histogram bandwidth and 19.6°for phase SD demonstrated that sensitivity, specificity, and area under the curve(AUC) for predicting response to CRT were 83%, 81%, and 0.83; 83%, 81%, and 0.85, respectively [48]. These dyssynchrony parameters derived from gated SPECT were superior in predicting response to CRT than those obtained by echocardiography demonstrated in PROSPECT study (AUC ≤ 0.62) [47].…”

“…[49,53,54,55] Summed rest score (SRS) is quantified as an index of scar extent on myocardial perfusion SPECT. SRS ≥ 27 predicts lack of response to CRT (Figure 8) [48]. Pacing site and degree of localized scarring also affect the response to CRT.…”

“…Nuclear imaging also provides mechanical dyssynchrony parameters [48]. In contrast echocardiography, nuclear imaging technique is less variable and determined the responders to CRT more accurately [48].…”

“…In contrast echocardiography, nuclear imaging technique is less variable and determined the responders to CRT more accurately [48]. In addition, the extent and location of scar tissue can be determined [49][50][51].…”