1. Validation Against Sonomicrometry
ABSTRACT
Background: We tested the accuracy and feasibility of a new 3D echo based strain analysis method.
Methods: We studied 10 pig hearts in a custom designed water tank to facilitate ultrasound scanning with a pulsatile pump at 5 different calibrated stroke volumes (30-70ml) through a balloon in the left ventricular (LV) cavity. Three sonomicrometry (sono) crystals were secured in the myocardium at the anterior of the LV at mid cavity level, with crystals 1 & 2 in long axis and 2 & 3 along the LV short axis. Cardiac motion was scanned from the apex to acquire gated full volume 3D image loops on a Philips iE33 ultrasound system with an X7.2 matrix probe at frame rates >25/sec. Sono data was also obtained at sampling rates >250/sec. The study was repeated in each heart after injecting 2ml of glutaraldehyde into the myocardium between the sono crystals to simulate infarct. Cartesian raw volumes were obtained by processing VolDicom image data in TomTec 4D LV-Analysis prototype software and analyzed for 3D strain.
Results: Increase in SV resulted in increased myocardial stretching that was detected by both sono and the 3D echo based method. 3D echo slightly overestimated the 3D strain values (6.5 ± 3.5%) when compared to sono in corresponding segments, but both methods showed a strong positive linear correlation. (r = 0.85). Simulated infarct showed significantly lower strain values by both methods.
Conclusions: The new 3D strain program from TomTec is an accurate imaging based method for computing strain and defining infarct.
CONCLUSIONS
The new 3D strain program from TomTec showed a strong linear relationship with sono-derived strain with Bias consistent with persistent overestimation in this controlled study. It may provide another clinically feasible imaging based method for computing strain and defining infarct.
DISCLOSURE
No relationships to disclose:
Muhammad Ashraf
Jill Panosian
Keith DesRochers
Employed by TomTec:
Thomas Balbach
Occasional Consultant to TomTec:
David J. Sahn
BACKGROUND
Strain imaging has been shown an important tool in non-invasive evaluation of segmental left ventricular (LV) contractile dysfunction and mechanical dyssynchrony. Accurate and reproducible assessment of LV deformation is important for evaluation of dynamic heart function.
Conventional 2D echo provides only subjective assessment of dysfunction being limited to scanning window. Out of plane myocardial motion during 2D acquisition by static scanner may result in computational errors due to in plane decorrelation of speckles.
Recently 3D echocardiography (3DE) has been rapidly integrated into clinical practice due to improved three dimensional dynamic display of cardiac chambers and valves allowing accurate volume and area measurements without geometrical assumptions. Recent improvements in spatio-temporal resolution has made it possible to track speckle motion in pyramidal volumes of 3D image loop and compute strain parameters in all segments of heart simultaneously without concerns about out of plane myocardial motion.
Goal of our study is to test the accuracy and feasibility of a new 3D echo based strain method developed by TomTec.
METHODS
We designed a digitally controlled phantom and studied freshly harvested pig hearts in a water tank to facilitate ultrasound scanning . Each pig heart was attached to a calibrated pulsatile pump through a latex balloon in left ventricular cavity (LV) secured at mitral inlet, and driven at 5 different stroke volumes (30-80ml). Three sonomicrometry (sono) crystals were secured in the LV myocardium anteriorly at mid cavity level, with crystals 1 & 2 in long axis and 2 & 3 along the LV short axis. Cardiac motion was scanned from the apex to acquire gated full volume 3D image loops on the latest high resolution 3D ultrasound systems ( GE Vivid E9 and Philips iE33) with matrix probes at frame rates >25/sec. Sono data was also obtained at sampling rates >250/sec. The study was repeated in each heart after injecting 2ml of glutaraldehyde into the myocardium between the sono crystals to simulate infarct. Cartesian raw volumes were obtained by processing VolDicom image data in TomTec 4D LV-Analysis prototype software and analyzed for 3D strain. Sono data was exported as MS Excel compatible files to a custom software (Sono Compare) to compute strain.
A New 3D Strain Method for Processing of 4D Echo Images Can Delineate Regional Myocardial Dysfunction:
Validation Against Sonomicrometry
Muhammad Ashraf, MD; Jill Panosian; Thomas Balbach, MD; Keith DesRochers; David J. Sahn MD, MACC
Oregon Health & Science University, Portland, OR; TomTec Imaging Systems, Unterschleissheim, Germany
RESULTS
Increase in injected stroke volume into LV cavity by calibrated pulsatile pump resulted in increased myocardial stretching that was detected by both sono and the 3D echo based strain method. All myocardial segments showed higher strain values at higher stroke volumes.
3D echo-derived strain values were persistently higher at each stroke volume (6.5 ± 3.5%) when compared to sono-derived strain in corresponding myocardial segments. However, both methods showed a strong positive linear correlation. (r = 0.85).
At each stroke volume simulated infarct showed significantly lower strain values by both methods when compared to strain in intact myocardial segments.
This Figure shows our digitally controlled phantom (left) and 3D images from GE Vivid E9 (top right ) and Philips IE33 (Bottom right)
This figure shows 4D LV-analysis software (left) processing a 3D image to obtain Cartesian raw volumes used to compute strain displayed as color coded parametric maps (right) and as graphs (top middle). Sono-derived strain is also shown (bottom middle)
Echo-derived strain in simulated infarct
Bland-Altman Plot between 3D echo and sono derived longitudinal segmental strain