COMPARISON OF LEFT AND RIGHT VENTRICULAR STRAIN MEASURED BY PHILIPS ACMQ AND TOMTEC USING TRANSESOPHAGEAL ECHOCARDIOGRAPHY.

To investigate how left and right ventricular strain measured by two different vendors compare to each other and how they compare in terms of interrater and intrarater reliability. In this secondary analysis, we examined transesophageal images acquired after induction of patients undergoing cardiac surgery during a period of hemodynamic stability. The same images were analyzed using Philps aCMQ and Tomtec analysis programs by two assessors. Specifically, global and segmental strain of the left ventricular (LV) was measured by two assessors in the midesophageal 4 chamber view (ME4C), the midesophageal 3 chamber view (ME4C), and midesophageal 2 chamber view (ME4C). Additionally, strain of the free wall (FW), septum (S), and segments were measured in the right ventricle (RV) in an RV focused ME4C. Correlation and agreement were examined between vendors in addition to interrater and intrarater reliability for each software. In a convenience sample of 50 patients, strain in a total of 181 views was examined. Median [interquartile range] strain values for aCMQ and TomTec for the ME4C, ME2C, ME3C, RVFW, and RVS were ME4C: -15.9% [-8.9% to -18.4%] vs -15.1% [-11.3% to -17.8%] p=0.83, ME2C: -14.0% [-11.9% to -17.0%] vs -16.7% [-12.2% to -18.6%] p=0.13, ME3C: -15.1% [-11.2% to -15.7%] vs -14.9% [-12.2% to -18.1%] p=0.39, RVFW: -18.3% [-14.1% to -21.2%] vs -24.9% [-21.5% to -29.3%] p<0.01, -17.1% [-11.7% to -19.8%] vs -19.8% [-16.2% to -22.8%] p=0.03, respectively. Correlation (Pearsons r = 0.79, 0.72, 0.91) and agreement (mean differences of -1%, -1%, -2%) were similar for LV views, while significant differences were found for the RV with correlation (Pearsons r = 0.62 and 0.73) and agreement (mean differences of -6% and -3%). On a segmental level, correlation was poorer for the basal, mid, and apical segments of the LV (Pearsons r = 0.37, 0.67, and 0.72). Interrater reliability (Pearsons r) for the ME4C, ME2C, ME3C, RVFW, and RVS were 0.91, 0.86, 0.81, 0.77, and 0.83 as well as 0.85, 0.89, 0.77, 0.88, and 0.89, for TomTec and Philips aCMQ respectively. Relevant differences in strain were found between the Philips aCMQ and TomTec Software tools for the RV, but not the LV. Segmental correlation between the two softwares was poorest in the basal segments. Similarly interrater reliability was observed. Clinicians should exercise caution in comparing strain measurements for the RV and on a segmental level; vendor-based systematic variation may exist. [ABSTRACT FROM AUTHOR]