Tej Desai BA, Presenter: Nothing to Disclose

J. Brian Fowlkes PhD, Abstract Co-Author: Research support, General Electric Company Equipment support, General Electric Company Equipment support, Toshiba Corporation Research collaboration, Sonetics Inc Stockholder, HistoSonics, Inc Founder, HistoSonics, Inc

Man Zhang MD, PHD, Abstract Co-Author: Nothing to Disclose

Jonathan Matthew Rubin MD, PhD, Abstract Co-Author: Consultant, General Electric Company Consultant, ZONARE Medical Systems, Inc Equipment support, ZONARE Medical Systems, Inc Research grant, General Electric Company Equipment support, General Electric Company

Stephen Z Pinter PhD, Abstract Co-Author: Equipment support, General Electric Company, Milwaukee, WI

Oliver D. Kripfgans, Abstract Co-Author: Research support, General Electric Company Equipment support, General Electric Company

Anne L. Hall, Abstract Co-Author: Employee, General Electric Company

To test the accuracy of real-time 3D angle-independent measurements of volume flow at cardiac blood flow rates.

We utilized a 4D (real-time 3D) color Doppler ultrasound scanner (Vivid 7, GE Healthcare, Milwaukee, WI) equipped with a 2D phased array (3V, GE Healthcare) to acquire bench top cardiac-type output under pulsatile flow conditions. In 3D, the color flow constant-depth plane (C-plane) resolution was 3 and 2.25 pixels/cm in the lateral and elevational directions respectively, while recording 10 volumes/second. Experiments were performed with a 2.5-cm diameter, thin-walled latex tube mounted in a water tank. Blood-mimicking fluid (CIRS Inc., Norfolk, VA) was circulated at up to 108 mL/s using a calibrated cardiac bypass pump set at 60 and 80 beats/min. Data was acquired randomly throughout the pump cycle with sweeps of 10, 25, 50, and 100 volumes. For each acquisition, a mean volume flow estimate was computed from C-plane integration and Doppler power based partial volume correction as previously published for canine femoral and carotid flow. Linear regression analysis was performed between the ultrasound estimated flow and the pump setting. Root mean square (RMS) and maximum errors were computed.

For pulsatile flow at 60 beats/min, mean volume flow estimates showed R2-values of 0.96, 0.96, 0.99, 0.99 and RMS errors of 10.5, 9.62, 5.79, 2.03 mL/s for 10, 25, 50, and 100 sweeps respectively. The coefficient of variation (CV) of the RMS errors was 15.6%, 14.3%, 8.58%, and 3.00%, with respect to the average measured flow of 67.5 mL/s. At 80 beats/min, R2 values were 0.90, 0.93, 0.99, 0.99 and RMS errors were 13.9, 5.57, 2.62, 2.25 mL/s for 10, 25, 50, and 100 sweeps respectively. The CV of the RMS errors ranged from 20.6% for 10 sweeps to 3.33% for 100 sweeps.

Estimates of angle-independent volume flow can be achieved with less than 10% error when using a minimum of 50 volumes per acquisition. Imaging time is approximately 5 seconds for a 50-volume acquisition, which demonstrates the ability for essentially real-time non-invasive measurements of cardiac output and related flows. Future work will involve evaluation in preclinical models and human subjects.

Non-invasive monitoring of cardiac function in critically ill patients with 3D Doppler ultrasound is a promising alternative to current invasive methods that are associated with high mortality.

Desai, T, Fowlkes, J, Zhang, M, Rubin, J, Pinter, S, Kripfgans, O, Hall, A, Three-dimensional Doppler Ultrasound in Obtaining Cardiac Output.  Radiological Society of North America 2010 Scientific Assembly and Annual Meeting, November 28 - December 3, 2010 ,Chicago IL. http://archive.rsna.org/2010/9012731.html