Three-dimensional myocardial strain estimation from volumetric ultrasound: Experimental validation in an animal model

Abstract

Abstract-Although real-time three-dimensional echocardiography has the potential to allow for more accurate assessment of global and regional ventricular dynamics compared to the more traditional two-dimensional ultrasound examinations, it still requires rigorous testing and validation against other accepted techniques should it breakthrough as a standard examination in routine clinical practice. Very few studies have looked at a validation of regional functional indices in an in-vivo context. The aim of the present study therefore was to validate a previously proposed 3D strain estimation-method based on elastic registration of subsequent volumes on a segmental level in an animal model. Volumetric images were acquired with a GE Vivid7 ultrasound system in five open-chest sheep instrumented with ultrasonic microcrystals. Radial ( RR ), longitudinal ( LL) and circumferential strain ( CC ) were estimated during four stages: at rest, during esmolol and dobutamine infusion, and during acute ischemia. Moderate correlations for LL (r=0.63; p<0.01) and CC (r=0.60; p=0.01) were obtained, whereas no significant radial correlation was found. These findings are comparable to the performance of the current state-of-the-art commercial 3D speckle tracking methods.

Current State of Three-Dimensional Myocardial Strain Estimation Using Echocardiography

“…More than a decade ago, Papademetris et al 20,49 were the first to report on the in vivo validation of a model-based 3D STE approach against sonomicrometry 20 and opaque MRI markers. 49 In more recent years, different implementations of block matching-based approaches, [50][51][52][53][54] a new implementation of a modelbased approach, 40 and an elastic registration 3D STE approach 56 have been validated in an in vivo setting. All these studies reported moderate to good correlations (r = 0.49-0.91) between regional strain values obtained by 3D STE methods and sonomicrometry 40,51,56 (Table 2).…”

“…49 In more recent years, different implementations of block matching-based approaches, [50][51][52][53][54] a new implementation of a modelbased approach, 40 and an elastic registration 3D STE approach 56 have been validated in an in vivo setting. All these studies reported moderate to good correlations (r = 0.49-0.91) between regional strain values obtained by 3D STE methods and sonomicrometry 40,51,56 (Table 2). However, care should be taken in interpreting and comparing those results, as not all of the studies reported strain values corresponding to a local cardiac coordinate system (i.e., RS, LS, and CS).…”

“…true, as reflected by their good correlations with sonomicrometry in an in vivo experimental setting. 20,40,[49][50][51]56,57…”

Three-dimensional myocardial strain estimation from volumetric ultrasound data using a novel transformation model adapted to the heart

“…Similar experiments were performed previously using a CFFD model. In phantom data, correlations were 0.98, 0.96 and 0.92 for RR , LL and CC respectively [4], while correlations in sheep data were 0.21, 0.63 and 0.60 respectively [9]. Seo et al reported similar findings in an in-vivo setup using a state-of-the-art commercially available 3D speckle tracking method that typically relies on block-matching based algorithms to track tissue motion (r=0.59-0.70 for RR , r=0.65-0.68 for LL and r=0.71-0.78 for CC [10]).…”

“…For a full description of the experimental protocol in these setups, the reader is referred to [4] and [9] respectively. Briefly, volumetric data was recorded in gel phantoms and in 5 open-chest sheep, from an apical view using a GE VividE9 equipped with a 2.5 MHz transducer (3V) at a frame rate of 13-17 Hz and 25-32 Hz respectively.…”