1. Using 3D-Wall Motion Tracking
ABSTRACT
Background: Circumferential and longitudinal strain (CS,LS) are both important indicators of myocardial function. Prior 2D echo research has verified speckle tracking to assess cardiac mechanics despite lost information from “out of plane motion.” This study evaluated strain determination by a novel 4D echo image-based 3D Wall Motion Tracking (3D-WMT) program.
Methods: A pulsatile pump apparatus was used to drive each phantom heart model with constant rotation (5°) and compaction (10 mm) to simulate physiological cardiac motion. Stroke volumes of ml in 10 ml increments were used. A latex balloon was sutured into the LV of 5 fresh pig hearts. The hearts were affixed to a rotating plate at the base and lightly fixed at the apex. The model was submerged in a water tank and imaged with the Toshiba matrix array PST-25SX 2.5-MHz transducer. Three sono crystals were sutured into the myocardium adjacent to the LAD. The full volume data was analyzed by 3D-WMT focusing on the mid anterior segment.
Results: Linear regression analysis suggest a strong correlation between sono data and 3D-WMT derived strain values, with high correlation coefficients for both CS (R² = .76) and LS (R² =.79), P < Bland-Altman analyses revealed 96% of the data points were within the 95% CI as well as a consistent overestimation.
Conclusion: The results suggest that Toshiba 3D-WMT can accurately measure CS and LS despite overestimation.
Figure 3. Linear Regression analyses of Toshiba 3D-WMT compared to sono derived strain values and Bland-Altman Analyses
CONCLUSIONS
This was a validation study to determine the accuracy of the Toshiba 3D-WMT strain-determination in quantifying CS and LS in a beating-heart model. The results suggest that the Toshiba 3D-WMT can accurately measure circumferential and longitudinal strain, though values were consistently higher than the sonomicrometry baseline. The beating-heart model introduced compaction and rotation, which are components of the cardiac mechanics of contraction. The quality of the Toshiba ultrasound imaging system and UltraExtend analysis software suggests that it can be utilized for the determination of CS and LS as an accurate indicator of cardiac health in a clinical setting. Further investigation is necessary to validate the accuracy of Toshiba 3D-WMT in detecting changes in strain after a simulated myocardial infarction. This would have important clinical applications to assess damage to cardiac tissue after myocardial infarctions.
DISCLOSURES
No relationships to disclose:
David J. Sahn
Rosie Bleck
Kacie Amacher
Meihua Zhu
Cole Streiff
Sumito Kimura
Sreyas Ravi
Austin Park
Muhammad Ashraf
BACKGROUND
Circumferential and Longitudinal Strain (CS, LS) are reliable indicators of cardiac function and are measured clinically using ultrasound imaging systems. Prior 2D echo research has verified speckle tracking to assess cardiac mechanics despite lost information from “out of plane motion.” The accuracy of ultrasound imaging systems has been verified through comparison to sonomicrometry, the “gold standard” of heart strain measurements. The sonomicrometry system for CS and LS determination utilizes piezoelectric crystals that are implanted within the outer myocardium adjacent to the LAD. Each sonomicrometry (sono) crystal is capable of transmitting an ultrasonic signal, detected by nearby crystals, thereby measuring distance. By measuring the changes in distance between the crystals during cardiac contraction, CS and LS can be quantified.
4D ultrasound imaging is not subject to the geometric assumptions and out of plane motion errors present in 2D echo. Prior research on cardiac strain has verified that speckle tracking and three-dimensional imaging systems can accurately detect strain. In addition, the relationship between cardiac twist and strain has been accurately measured using two-dimensional ultrasound imaging. The pumping heart models used in prior studies have not simulated the natural compaction and rotation of a beating heart, thus limiting the physiological accuracy of CS and LS determination.
This study evaluated the accuracy of Toshiba's 3D-WMT Ultrasound system and software in circumferential and longitudinal strain determination.
METHODS
A latex balloon was connected to a custom calibrated pulsatile pump apparatus. The major vessels and atria were removed from five adult porcine hearts and the balloon was sutured into the Left Ventricle (LV). The pump apparatus was used to drive the heart with constant rotation (5°) and compaction (10mm) to simulate physiologically realistic mechanics of cardiac motion. Stroke Volumes of ml in 10 ml increments were used to account for physiological stroke volume variance. The hearts were fixed to a rotating plate at the base and to a ring at the apex. The model was submerged in a water tank and imaged with the Toshiba matrix array PST-25SX 2.5-MHz transducer from the apex (Figure 1).
Figure 1. Experimental Setup
Three sonomicrometry crystals were sutured into the outer myocardium at the Mid Anterior segment. Strain measurements from the sono crystals were collected and analyzed with SonoMetrics Sonosoft software.
Full 4D images were acquired by a Toshiba Artida ultrasound system. The full volume data was analyzed by Toshiba UltraExtend 3D-Wall Motion Tracking (3D-WMT), focusing on the Mid Anterior segment of the heart (Figure 2). The strain determination of the Toshiba ultrasound system and analysis software was compared to sono-derived strain measurements.
Evaluating Cardiac Mechanics from 4D Echocardiographic Imaging for Longitudinal and Circumferential Strain
Using 3D-Wall Motion Tracking
David J. Sahn, MD, MACC; Rosie Bleck; Kacie Amacher; Meihua Zhu, MD, PhD; Cole Streiff, BA; Sumito Kimura, MD;
Sreyas Ravi; Austin Park; Muhammad Ashraf, MD
Oregon Health & Science University, Portland, Oregon, USA
Figure 2. Toshiba 3D-WMT UltraExtend software analysis of ultrasound images of the LV for strain determination
RESULTS
Linear regression analyses resulted in a strong linear correlation between the sonomicrometry strain measurements and the Toshiba 3D-Wall Motion Tracking derived strain values. High correlation coefficients were found for both CS, with an R² value of 0.76, and LS, with an R² value of The P-value verified these results, with P < .001.(Figure 3) Bland-Altman analyses revealed 96% of the data points to be within a 95% Confidence Interval for both CS and LS (Figure 3). A consistent overestimation by the Toshiba 3D-WMT system was observed, but the strong linear correlation suggests that the Toshiba Artida ultrasound imaging system and the Toshiba UltraExtend 3D-Wall Motion Tracking software accurately quantified both Circumferential and Longitudinal strain, despite overestimation.