1. ABSTRACT
Background: The right ventricle (RV) is thin walled and has presented problems for both tissue Doppler and speckle tracking methods of determining mechanics. Methods: We studied 6 fresh pig hearts in a model that allowed known volumes to be pumped into the RV and left ventricle simultaneously while ECG gated full volume RV images were obtained using a 3.5 MHz transducer on a Toshiba Applio Artida scanner. Calibrated stroke volumes of 20 to 50 cc were delivered. Two triangular arrays of sonomicrometry crystals were placed around the RV free wall 1/3 and 2/3 of the distance from the tricuspid annulus to the RV apex. 4D strain data were computed on the system using 4D wall motion tracking to yield the full field of RV and longitudinal and circumferential strains. Results were compared to sonomicrometry in the free wall zones corresponding to sonomicrometry crystal locations, as analyzed by SonoView. Results: In the mid wall zones both the ultrasound and the sono longitudinal (r = 0.87, p< 0.05) and circumferential strains (r = 0.76, p< 0.05) and interobserver variability for strains was less that 5% since the algorithm tracks data within identified wall boundaries. Conclusions: This new method is easy to use and robust. It provides data on RV mechanics accurately and reproducibly.
CONCLUSIONS
The results of this model study were derived from analysis of very high resolution images, which suggested that the new 3D wall motion tracking is an accurate, reproducible and easy to use method for non invasive quantification of RV mechanics.
DISCLOSURE
No relationships to disclose:
S. Krishnakumar A. Rajendran
S. Ashraf S. Yang
N. Chang G. Kovch
W. Shentu Z. Zhou
M. Ashraf
Employees of Toshiba America Medical Systems:
B. Kenny B. Carpenter
Occasional consultant to TAMS:
D.J. Sahn
BACKGROUND
There is much interest in the study of tissue mechanics and volume measurements of the RV chamber to evaluate heart function. The study of RV mechanics could provide novel insight to dynamic right heart function, especially in patients with congenital heart disease and pulmonary hypertension. Recent developments in cardiac imaging have made it possible to compute cardiac mechanics non-invasively. Angle independent speckle tracking based motion detecting methods have recently attracted the attention of research scientists and cardiologists. These methods use high resolution echo images for analysis of cardiac mechanics and have been validated against more invasive sonomicrometry and more expensive MRI. Despite some potential for error due to in-plane decorrelation caused by through plane myocardial motion during acquisition, 2D speckle tracking based methods seem to provide reasonable accuracy for LV mechanics. Symmetrical shape and thick LV walls provide a sufficient number of speckle targets for reproducible tracking. Unique shape and geometry of the RV, along with relatively thin walls, present a challenge to conventional 2D based motion detecting methods. A lack of identifiable anatomical landmarks also presents another difficulty in reproducible selection of 2D imaging planes. We tested the feasibility and accuracy of a new 3D method based on wall motion tracking within a full volume dynamic 3D heart image loop to compute RV mechanics.
METHODS
We studied 6 freshly harvested pig hearts.
Both atriums of each heart were excised
for easier access to the ventricles. Each
heart was then mounted on a rotary
actuator in a water bath.
Both the LV and RV were connected to
separate pulsatile pumps through clear
plastic tubing and a latex balloon was
inserted into each ventricular cavity. Both
pumps were capable of delivering a
calibrated stroke volume at a controlled
rate. Three sonomicrometry crystals were
placed around the free wall of the RV, 1/3
and 2/3 of the distance from the tricuspid annulus to the RV apex in a triangular arrangement. The ventricles were pumped in synchrony at five different stroke volumes from 20 to 50cc. ECG gated full volume RV images were obtained using a 3.5 MHz transducer on a Toshiba Applio Artida scanner. After acquiring baseline data at each stroke volume, we injected 2-4 mL of gluteraldehyde to simulate myocardial infarct between sono crystals and the study was repeated. The image data was processed on a PC, using 3D wall motion tracking to yield a full view of the RV and longitudinal and circumferential strains. Sonomicrometry data was processed for longitudinal and circumferential strains in Sonosoft. The strain results from the images at each stroke volume were compared to the sonomicrometry derived strain values.
Simultaneously Derived Total Cavity Circumferential and Longitudinal Right Ventricular Strains
Using a New Method for 4D Cardiac Mechanics: Validation by Sonomicrometry
Sheetal Krishnakumar; Anugraha Rajendran; Shahryar Ashraf; Sarah Yang; Nicole Chang; Galyna Kovch;
Weihui Shentu; Zhiwen Zhou; Bill Kenny; Berkeley Carpenter; Muhammad Ashraf; David J. Sahn, FASE
Oregon Health & Science University, Portland, OR; Toshiba America Medical Systems, Tustin, CA
RESULTS
At all stroke volumes 3D Echo derived longitudinal and circumferential RV strain showed a linear correlation with sonomicrometry derived longitudinal strain (r = 0.87, p< 0.05) and circumferential strains (r = 0.76, p< 0.05).
At each stroke volume modeled infarct showed significantly lower strain as compared to normal myocardial tissue.
This 3D Echo based method showed high reproducibility (r=0.96) and inter-observer variability for strain computation was less than 5 percent.
This picture shows sonomicrometry crystals anchored to myocardium of RV and out model setup for this study