Abstract 14627: Do Different Frame Rates Affect Timing of Peak Strain in a Rapidly Beating Rat Heart?

Abstract

Background: In echocardiography, time to peak strain measurements by 2D speckle tracking are used both clinically and experimentally. Theoretically, lower frame rates undersample data, which may affect accuracy in the timing of myocardial motion events, which could be exacerbated by spatial and temporal smoothing, particularly for the high heart rates in laboratory rats (∼300 bpm). We investigated the effect of different frame rates on time to peak strain measurements in a rat ischemia model.

Methods: Ischemia was induced for 45 to 60 minutes by left anterior descending coronary artery occlusion in 18 rats. Short-axis images of the mid left ventricle were obtained at 220 and 86 frames per second (fps). Adequate tracking was verified in 6 standard segments including ischemic and control areas. Time to peak radial strain measurements in the ischemic and control areas were compared between the two frame rate settings, using the same smoothing, by linear regression and paired t test.

Results: Time to peak strain intervals in the control areas between the different frame rates correlated better, r = 0.83, than in the ischemic areas, r = 0.68 (Figure). There were no significant differences of time to peak strain between different frame rates (86 fps: 107 ± 32 ms vs 220 fps: 95 ± 27 ms, p = 0.06, ischemic area; 86 fps: 51 ± 14 ms vs 220 fps: 48 ± 18 ms, p = 0.37, control area). However, in the ischemic area, measurements at the lower frame rate tended to overestimate time to peak strain intervals with respect to the high frame rate measurements.

Conclusions: Both higher and lower frame rates detect changes in regional deformation intervals induced by coronary occlusion in high heart rates in our rodent model. However, our comparison of different frame rates in the ischemic areas led us to find that it is not just temporal resolution but also variability in the representation of ischemic areas by standard segments and the effect of default smoothing that could potentially lead to inaccurate results.