Noninvasive right ventricular pressure estimation in vivo using the subharmonic emissions from ultrasound contrast agents

Abstract

In this work, the ability of subharmonic aided pressure estimation (SHAPE) to noninvasively estimate cardiac right ventricular pressures was investigated. Five canines received Sonazoid (GE Healthcare, Oslo, Norway) infusion (0.015 μl/kg/min) and were scanned using a Sonix RP ultrasound scanner (Ultrasonix Medical Corp, Richmond, BC, Canada) with a PA4-2 phased array (transmit/receive: 2.5/1.25 MHz). Unprocessed radiofrequency (RF) data post pulse inversion (but before envelope modulation) were acquired synchronously with Millar pressure catheter (reference standard) from the aorta, the right ventricle (RV) and the right atrium (RA) at five incident acoustic power (IAP) levels (5 s per acquisition; n = 3). Subharmonic signal amplitudes were extracted from the RF data as the mean amplitude within a bandwidth of 1 to 1.5 MHz; the resulting temporally varying subharmonic signals were median filtered. The IAP level eliciting subharmonic signals most sensitive to ambient pressure changes was selected for each scanned location in each canine. Based on data obtained from the aorta, a calibration factor (in mmHg/dB) was calculated for each canine, and combined with the respective RA pressures and RV subharmonic data to obtain RV pressures. The resulting RV pressures and RV relaxation rate (peak -dP/dt) were compared to data obtained with the Millar pressure catheter. Paired comparisons revealed absolute errors ranging from 0.0 to 3.4 mmHg (mean difference: 2.3 ± 1.3 mmHg; p = 0.02) for RV systolic peak pressures, from 0.1 to 1.8 mmHg (mean difference: 0.8 ± 0.7 mmHg; p = 0.06) for RV diastolic minimum pressures and from 1.2 to 5.9 mmHg/s (mean difference: 2.9 ± 3.1 mmHg/s; p = 0.10) for RV relaxation rate. These results show that RV pressures obtained with SHAPE were in agreement with the Millar pressure catheter. Thus, SHAPE is a promising technique for noninvasive RV pressure estimation.