Ferraro AM, Adar A, Ghelani SJ, Sleeper LA, Levy PT, Rathod RH, Marx GR, Harrild DM.Cardiovasc Ultrasound. 2020 May 21;18(1):15. doi: 10.1186/s12947-020-00199-x.PMID: 32438907 Free PMC article.
Abstract
Background: Strain and synchrony can be calculated from a variety of software packages, but there is a paucity of data with inter-vendor comparisons in children. To test the hypothesis that different packages may affect results, independent of acquisition, we compared values obtained using two commercially available analysis tool (QLAB and TomTec), with several different settings.
Methods: The study population included 108 children; patients were divided into three groups: (1) normal cardiac structure and conduction; (2) ventricular paced rhythm; and (3) flattened ventricular septum (reflecting right ventricular pressure or volume load lesions). We analyzed the same image acquired from the apical 4-chamber (AP4) and short-axis at the mid-papillary level (SAXM) views in both QLAB (versions 10.5 and 10.8) and TomTec (version 1.2). In QLAB version 10.8, low, medium, and high quantification smoothness settings were employed. In TomTec, images were analyzed with both low and high frame rates. Tracking quality for each package was graded. AP4 and SAXM strain and synchrony values were recorded. A mixed-effects linear regression model was used, with main effect considered significant if the p-value was < 0.05.
Results: Tracking scores were high for all packages except QLAB 10.5 in the SAXM view. AP4 and SAXM strain values varied significantly between QLAB 10.5 and the other packages. Synchrony values varied widely for all strain values (p < 0.001 for both) in all packages. Quantification smoothness changes in QLAB 10.8 did not impact strain significantly in any patient group; temporal resolution changes in TomTec resulted in strain differences in children with flat ventricular septums, but not those with normal or ventricular paced hearts.
Conclusion: Synchrony values varied substantially among all packages in children. Strain values varied widely between QLAB 10.5 and all other software packages, recommending avoidance of QLAB 10.5 for future studies. Quantification smoothness settings in QLAB 10.8 resulted in minimal strain differences. In TomTec, low and high frame rate strain values differed only in a subset of patients (flattened septum). These data suggest that reliable comparisons between strain values derived from QLAB and TomTec is possible in certain cases, but that caution should be used especially in different hemodynamics conditions.
Fig. 1 Study methods. One hundred and eight patients were included, with equal representation from 3 groups of children. Analyses were conducted using 2 commercially available vendors and three software packages with a variety of settings (6 variations in total). The same apical 4 chamber (AP4) and parasternal short axis views at the mid-ventricular papillary myscler level (SAXM) views were analyzed for each of the 6 variations. AP4 (Apical four chamber view), FR (frame rate), QLAB 10.8-Low (Low quantification smoothness setting), QLAB 10.8-Med (Medium quantification smoothness settings), QLAB 10.8-High (High quantification smoothness setting), SAXM (Short axis mid-ventricular papillary muscle)
Fig. 2 Speckle Tracking Packages Used for Assessment. Screenshots from the various speckle tracking packages. a) QLAB 10.5 (AP4 view), b) QLAB 10.5 (SAXM view), c) QLAB 10.8 (AP4 view), d) QLAB 10.8 (SAXM view), e) TomTec (AP4 view), f) TomTec (SAXM view). AP4 (Apical four chamber view), SAXM (Short axis mid-ventricular papillary muscle)
Fig. 3 Tracking Assessment Scores. A tracking score was generated for six variations of the software packages included in this study for the AP4 and SAXM views. A tracking score > 2.5 (dotted line) was consider to be excellent. AP4 (Apical four chamber view), FR (frame rate), SAXM (Short axis mid-ventricular papillary muscle)
Fig. 4 Mean strain and TTPSD values in each software configuration. Bars indicate mean values (±2 standard deviation) in each software configuration based on 108 patients. a, AP4 strain; b, SAXM strain; c, AP4 TTPSD; d, SAX-M TTPSD. An asterisk indicates a p-value < 0.05. There were many differences in mean AP4 (4C) and SAXM (4D) TTPSD for QLAB 10.5, QLAB 10.8, and TomTec at each frame rate and smoothness setting for the entire cohort. AP4 (Apical four chamber view), HFR (high frame rate), LFR (low frame rate), SAXM (Short Axis Mid papillary muscle); TTPSD (Time To Peak Standard Deviation)
Fig. 5 Mean strain values in each software configuration, by group. Mean values are presented (±2 standard deviation) by group (36 patients in each). a) AP4 strain; b) SAXM strain. An asterisk indicates a p-value < 0.05. AP4 Apical four chamber view, FR (frame rate), LS (Low quantification smoothness setting), HS (High quantification smoothness setting), MS (Medium quantification smoothness setting), SAXM (Short Axis Mid papillary muscle). Dark gray bars (group 1) are patients with normal hearts; white bars (group 2) are patients with ventricular paced rhythms; light gray bars (group 3) patients have flattened septal wall
Fig. 6 Mean TTPSD values in each software configuration, by group. Mean values are presented (±2 standard deviation) by group (36 patients in each). a) AP4 TTPSD; b) SAX-M TTPSD. An asterisk indicates a p-value < 0.05. There were many differences in mean AP4 and SAXM TTPSD for QLAB 10.5, QLAB 10.8, and TomTec at each frame rate and smoothness setting for all of the subgroups AP4 (Apical four chamber view), LFR (Low frame rate), LS (Low quantification smoothness setting), HS (High quantification smoothness setting), MS (Medium quantification smoothness setting), SAXM (Short Axis Mid papillary muscle). Dark gray bars (group 1) patients have normal hearts; white bars (group 2) patients have ventricular paced rhythms; light gray bars (group 3) patients have flattened ventricular septum.
