Non-contact laser Doppler vibrometry for measurement of heart sounds in patients undergoing cardiac resynchronization therapy

Background

Conventionally, auscultation by stethoscope has been used to assess cardiac function. The subjectivity of this method has limited its applicability for diagnostic and monitoring purposes. With the development of new acquisition techniques and signal processing tools, the reliability and accuracy of heart sound signals increased, increasing their potential for clinical use. As part of the EU-funded CARDIS project a contact-less Laser Doppler Vibrometer (LDV) was developed to register skin vibrations, induced by mechanical activity, much alike heart sounds.

Purpose

To evaluate LDV-based heart sound characteristics in patients during different cardiac contraction patterns induced by a biventricular pacemaker.

Methods

Ten patients implanted with a pacemaker for cardiac resynchronization therapy (CRT) were included. The LDV registers skin vibrations through optical interference using a laser beam (1550nm). LDV recordings were obtained approximately at a distance of 5 cm from the thorax’s skin surface at the level of the 4th or 5th intercostal space using a retroreflective patch. Measurements were performed during atrial (AAI), right ventricular (RV), left ventricular (LV), and biventricular (BiV) pacing, whilst heart rate was kept the same.

LDV signals were analysed using a semi-automated algorithm for segmenting heart sounds and quantifying amplitude and timing characteristics. We obtained amplitude and width of the first and second heart sound (S1 and S2), the ratio of S1-S2 amplitude, and S1-S2 time interval (Figure 1).

Values are presented as median +- IQR. Statistical significance was evaluated using the Friedman test followed by Dunn’s test of multiple comparisons. P<0.05 was considered statistically significant.

Results

Compared to AAI, BiV increased S1 amplitude (p<0.05) while LV pacing showed an increasing trend in S1 amplitude (p = 0.15) and RV pacing showed no trend. S1-S2 ratio showed an increasing trend for BiV (p=0.15), compared to AAI, but not for RV and LV. S2 amplitude, S1-S2 time and widths of S1 and S2 did not show a significant change during the various pacing modes. (Figure 2)

Conclusions

Contactless LDV captures heart sounds from thoracic skin vibrations. The larger S1 amplitude during BiV pacing corresponds to previous findings connecting improved activation synchrony to increased rate of LV pressure rise during contraction, which is associated with amplification of S1. These findings warrant for further studies to investigate whether LDV can be used for non-contact assessment of the effect of CRT pacing strategies.LDV heart sound recordingBox plots of relative changes