Jacob Bjerring Olesen

Volume Flow in Arteriovenous Fistulas Using Vector Velocity Ultrasound

Volume flow in arteriovenous fistulas for hemodialysis was measured using the angle-independent ultrasound technique Vector Flow Imaging and compared with flow measurements using the ultrasound dilution technique during dialysis. Using an UltraView 800 ultrasound scanner (BK Medical, Herlev, Denmark) with a linear transducer, 20 arteriovenous fistulas were scanned directly on the most superficial part of the fistula just before dialysis. Vector Flow Imaging volume flow was estimated with two different approaches, using the maximum and the average flow velocities detected in the fistula. Flow was estimated to be 242 mL/min and 404 mL/min lower than the ultrasound dilution technique estimate, depending on the approach. The standard deviations of the two Vector Flow Imaging approaches were 175.9 mL/min and 164.8 mL/min compared with a standard deviation of 136.9 mL/min using the ultrasound dilution technique. The study supports that Vector Flow Imaging is applicable for volume flow measurements.

Vector Flow Imaging Compared with Conventional Doppler Ultrasound and Thermodilution for Estimation of Blood Flow in the Ascending Aorta

Transverse oscillation (TO) is a real-time ultrasound vector flow method implemented on a commercial scanner. The TO setup was examined on a flowrig with constant and pulsatile flow. Subsequently, 25 patients undergoing cardiac bypass surgery were scanned intraoperatively with TO on the ascending aorta and compared to transesophageal echocardiography (TEE) and pulmonary artery catheter thermodilution (PACTD). On the flowrig, TO had a precision of 5.5%, 9.4% and 14.7%, a percentage error of 18.2%, 14.6% and 40.7%, and a mean bias of 0.4 cm/s, 36.8 ml/min and 32.4 ml/min for velocity and flow rate (constant and pulsatile) estimation. The correlation coefficients for all flowrig evaluations were 0.99 indicating systematic bias. After bias correction, the percentage error was reduced to 11.5%, 12.6% and 15.9% for velocity and flow rate (constant and pulsatile) estimation. In the in vivo setup, TO, TEE, and PACTD had a precision of 21.9%, 13.7%, and 12.0%. TO compared with TEE and PACTD had a mean bias of 12.6 cm/s and −0.08 l/min, and a percentage error of 23.4%, and 36.7%, respectively. The percentage error was reduced to 22.9% for the TEE comparison, but increased to 43.8% for the PACTD comparison, after correction for the systematic bias found in the flowrig. TO is a reliable and precise method for velocity and flow rate estimation on a flowrig. However, TO with the present setup, is not interchangeable with PACTD for cardiac volume flow estimation, but is a reliable and precise angle-independent ultrasound alternative for velocity estimation of cardiac flow.

Noninvasive estimation of 2-D pressure gradients in steady flow using ultrasound.

A noninvasive method for estimating 2-D pressure gradients from ultrasound vector velocity data is presented. It relies on vector velocity fields acquired using the transverse oscillation method during steady flow conditions. The pressure gradients are calculated from the velocity fields using the Navier-Stokes equations. Scans of a carotid bifurcation phantom with a 70% constriction are performed using a linear transducer connected to a scanner. The performance of the estimator is evaluated by comparing its results to those of a computational fluid dynamics model of the carotid bifurcation phantom. The geometry of the model is determined from magnetic resonance imaging. The presented study is conducted assuming steady flow using velocity data acquired at 18 frames per second. The proposed method shows pressure gradients at the constricted region from -8 kPa/m to 9 kPa/m, with a maximum bias of -7% for the axial component and -8% for the lateral component. The relative standard deviation of the estimator is 5% (axial component) and 30% (lateral component) when studying the pressure gradient across the constriction using 3 velocity frames per pressure estimate. The study shows that 2-D pressure gradients can be achieved noninvasively using ultrasound data in a constant flow environment.

Vector velocity volume flow estimation: Sources of error and corrections applied for arteriovenous fistulas

A method for vector velocity volume flow estimation is presented, along with an investigation of its sources of error and correction of actual volume flow measurements. Volume flow errors are quantified theoretically by numerical modeling, through flow phantom measurements, and studied in vivo. This paper investigates errors from estimating volumetric flow using a commercial ultrasound scanner and the common assumptions made in the literature. The theoretical model shows, e.g. that volume flow is underestimated by 15%, when the scan plane is off-axis with the vessel center by 28% of the vessel radius. The error sources were also studied in vivo under realistic clinical conditions, and the theoretical results were applied for correcting the volume flow errors. Twenty dialysis patients with arteriovenous fistulas were scanned to obtain vector flow maps of fistulas. When fitting an ellipsis to cross-sectional scans of the fistulas, the major axis was on average 10.2mm, which is 8.6% larger than the minor axis. The ultrasound beam was on average 1.5mm from the vessel center, corresponding to 28% of the semi-major axis in an average fistula. Estimating volume flow with an elliptical, rather than circular, vessel area and correcting the ultrasound beam for being off-axis, gave a significant (p=0.008) reduction in error from 31.2% to 24.3%. The error is relative to the Ultrasound Dilution Technique, which is considered the gold standard for volume flow estimation for dialysis patients. The study shows the importance of correcting for volume flow errors, which are often made in clinical practice.

Accuracy and sources of error for an angle independent volume flow estimator

This paper investigates sources of error for a vector velocity volume flow estimator. Quantification of the estimators accuracy is performed theoretically and investigated in vivo. Womersleys model for pulsatile flow is used to simulate velocity profiles and calculate volume flow errors in cases of elliptical vessels and not placing the transducer at the vessel center. Simulations show, i.e., that volume flow is underestimated with 5 %, when the transducer is placed 15 % from the vessel center. Twenty patients with arteriovenous fistulas for hemodialysis are scanned in a clinical study. A BK Medical UltraView 800 ultrasound scanner with a 9 MHz linear array transducer is used to obtain Vector Flow Imaging sequences of a superficial part of the fistulas. Cross-sectional diameters of each fistula are measured on B-mode images by rotating the scan plane 90 degrees. The major axis of the fistulas was on average 8.6 % larger than the minor axis, so elliptic dimensions should be taken into account in volume flow estimation. The ultrasound beam was on average 1.5 ± 0.8 mm off-axis, corresponding to 28.5 ± 11.3 % of the major semi-axis of a fistula, and this could result in 15 % underestimated volume flow according to the simulation. Volume flow estimates were corrected for the beam being off-axis, but was not able to significantly decrease the error relative to measurements with the reference method.

New developments in vector velocity imaging using the transverse oscillation approach

Vector velocity imaging using the Transverse Oscillation (TO) approach has recently been FDA approved for linear array transducers on a commercial platform. It can now be used clinically for studying the complex ow at e.g. bifurcations, valves, and the heart in real time. Several clinical examples from venous ow to rotational ow in the heart will be shown. The technique is also being further developed and adapted for convex and phased array probes, for spectral velocity estimation, pressure estimation, and for three dimensional velocity tensor imaging. It is shown how the methods are optimized using Field II simulations along with several examples of their performance.

Non-invasive measurement of pressure gradients using ultrasound

A non-invasive method for estimating 2-D pressure gradients from ultrasound vector velocity data is presented. The method relies on in-plane vector velocity elds acquired using the Transverse Oscillation method. The pressure gradients are estimated by applying the Navier-Stokes equations for isotropic uids to the estimated velocity elds. The velocity elds were measured for a steady ow on a carotid bifurcation phantom (Shelley Medical, Canada) with a 70% constriction on the internal branch. Scanning was performed with a BK8670 linear transducer (BK Medical, Denmark) connected to a BK Medical 2202 UltraView Pro Focus scanner. The results are validated through nite element simulations of the carotid ow model where the geometry is determined from MR images. This proof of concept study was conducted at nine ultrasound frames per second. Estimated pressure gradients along the longitudinal direction of the constriction varied from 0 kPa/m to 10 kPa/m with a normalized bias of -9.1% for the axial component and -7.9% for the lateral component. The relative standard deviation of the estimator, given in reference to the peak gradient, was 28.4% in the axial direction and 64.5% in the lateral direction. A study made across the constriction was also conducted. This yielded magnitudes from 0 kPa/m to 7 kPa/m with a normalized bias of -5.7% and 13.9% for the axial and lateral component, respectively. The relative standard deviations of this study were 45.2% and 83.2% in the axial and lateral direction, respectively.

Non-invasive estimation of pressure changes along a streamline using vector velocity ultrasound

A non-invasive method for estimating pressure changes along a streamline using ultrasound is presented. The suggested method estimates pressure gradients from 2-D vector velocity fields. Changes in pressure are derived using a model based on the Navier-Stokes equations. Scans of a carotid bifurcation phantom with a 70% constriction are performed using a linear array transducer connected to the experimental scanner, SARUS. 2-D fields of angle-independent vector velocities are acquired to a depth of 3 cm using directional synthetic aperture vector flow imaging. The performance of the suggested estimator is evaluated by comparing its results to a 3-D numerical simulation model. The study showed pressure drops across the constricted phantom varying from -5 Pa to 7 Pa with a standard deviation of 4%. The proposed method had a normalised root-mean-square error of 10% in reference to the simulation model. Further, an in-vivo scan of the carotid bifurcation is made to show the feasibility of the technique in a less experimental environment.

Surveillance of hemodialysis vascular access with ultrasound vector flow imaging

The aim of this study was prospectively to monitor the volume flow in patients with arteriovenous fistula (AVF) with the angle independent ultrasound technique Vector Flow Imaging (VFI). Volume flow values were compared with Ultrasound dilution technique (UDT). Hemodialysis patients need a well-functioning vascular access with as few complications as possible and preferred vascular access is an AVF. Dysfunction due to stenosis is a common complication, and regular monitoring of volume flow is recommended to preserve AVF patency. UDT is considered the gold standard for volume flow surveillance, but VFI has proven to be more precise, when performing single repeated instantaneous measurements. Three patients with AVF were monitored with UDT and VFI monthly for five months. A commercial ultrasound scanner with a 9 MHz linear array transducer with integrated VFI was used to obtain data. UDT values were obtained with Transonic HD03 Flow-QC Hemodialysis Monitor. Three independent measurements at each scan session were obtained with UDT and VFI each month. Average deviation of volume flow between UDT and VFI was 25.7 % (Cl: 16.7% to 34.7%) (p= 0.73). The standard deviation for all patients, calculated from the mean variance of each individual scan sessions, was 199.8 ml/min for UDT and 47.6 ml/min for VFI (p = 0.002). VFI volume flow values were not significantly different from the corresponding estimates obtained using UDT, and VFI measurements were more precise than UDT. The study indicates that VFI can be used for surveillance of volume flow.

Non-invasive estimation of pressure gradients in pulsatile flow using ultrasound

This paper investigates how pressure gradients in a pulsatile flow environment can be measured non-invasively using ultrasound. The presented set-up is based on vector velocity fields measured on a blood mimicking fluid moving at a peak flow rate of 1 ml/s through a constricted vessel. Fields of pressure gradients are calculated using the Navier-Stokes equations. Flow data are acquired to a depth of 3 cm using directional synthetic aperture flow imaging on a linear array transducer producing 1500 image frames of velocity estimates per second. Scans of a carotid bifurcation phantom with a 70% constriction are performed using an experimental scanner. The performance of the presented estimator is evaluated by comparing its results to a numerical simulation model, which geometry is reconstructed from MRI data. The study showed pressure gradients varying from 0 kPa/m to 4.5 kPa/m with a maximum bias and standard deviation of 10% and 13%, respectively, relative to peak estimated gradient. The paper concludes that maps of pressure gradients can be measured non-invasively using ultrasound with a precision of more than 85%.