Manabu Nakamura, Masahisa Onoguchi and Takayuki Shibutani
In cardiac resynchronization therapy (CRT), even if patient selection is made according to Japanese adaptive criteria, there are non-responders. Its main factor is considered to be the lack of adequate preoperative assessment against mechanical left ventricular dyssynchrony. Recently, phase analysis was enabled on gated myocardial perfusion SPECT (GMPS). The purpose of this study was to examine the relationship between the index of phase analysis using the two software (cardioREPO® and QGS) and the left ventricular reverse remodeling index (ΔLVESV) for the evaluation of left ventricular dyssynchrony in CRT patients is there. It also evaluated whether it could be an index of adaptation decision and effect determination.
Methods: For 15 patients with severe heart failure who underwent CRT, GMPS was performed before (baseline) and after CRT. In cardioREPO®, standard deviation of the time to end systolic phase of 17 segments of the left ventricle (SDTES) and Bandwidth and Phase SD, Entropy of phase histogram were used as left ventricular dyssynchrony index. In QGS, standard deviation of the time to maximum displacement of each segment (SDTTMD) was used as an index. An example in which ΔLVESV (%Reduction) after 6 months of CRT decreased by 15% or more was defined as a CRT responder.
Results: 10 of 15 patients were responders. Bandwidth at baseline of the responder group was significantly higher. SDTES, Phase SD, Entropy and SDTTMD of the responder group tended to be higher. All indexes decreased significantly in the responder group after 6 months of CRT but not in the non-responder group. Excluding SDTES, positive correlation was shown between baseline and ΔLVESV, and the optimal cutoff value of responder prediction was SDTES 7.637%, Bandwidth 218°, Phase SD 50.0°, Entropy 0.785, SDTTMD 19.85 ms.
Conclusion: Phase analysis by GMPS showed that quantitative assessment of left ventricular dyssynchrony of CRT was possible and that the index was related to response prediction to CRT. In particular, SDTTMD showed good correlation between baseline and ΔLVESV, suggesting that it may be a more sensitive index of reaction prediction.