SPECKLE TRACKING QUANTIFICATION PARASTERNAL INTERCOSTAL MUSCLE LONGITUDINAL STRAIN TO PREDICT WEANING OUTCOMES: A MULTICENTRIC OBSERVATIONAL STUDY.

Background: The purpose of this study was to determine the feasibility, reliability, and reproducibility of parasternal intercostal muscle longitudinal strain (LSim) quantification by speckle tracking and the value of maximal LSim to predict weaning outcomes. Methods: This study was divided into three phases. Phases 1 and 2 comprehended prospective observational programs to evaluate the feasibility, reliability, and repeatability of speckle tracking to assess LSim in healthy subjects and mechanically ventilated patients. Phase 3 was a multicenter retrospective study to evaluate the value of maximal LSim, intercostal muscle thickening fraction (TFim), diaphragmatic thickening fraction, diaphragmatic excursion, and rapid shallow breathing index to predict weaning outcomes. Results: A total of 25 healthy subjects and 20 mechanically ventilated patients were enrolled in phases 1 and 2, respectively. Maximal LSim was easily accessible, and the intraoperator reliability and interoperator reliability were excellent in eupnea, deep breathing, and mechanical ventilation. The intraclass correlation coefficient ranged from 0.85 to 0.96. Moreover, 83 patients were included in phase 3. The areas under the receiver operating characteristic curve of maximal LSim, TFim, diaphragmatic thickening fraction, diaphragmatic excursion, and rapid shallow breathing index were 0.91, 0.79, 0.71, 0.70, and 0.78 for the prediction of successful weaning, respectively. The best cutoff values of LSim and TFim were >-6% (sensitivity, 100%; specificity, 64.71%) and <7.6% (sensitivity, 100%; specificity, 50.98%), respectively. Conclusions: The quantification of LSim by speckle tracking was easily achievable in healthy subjects and mechanically ventilated patients and presented a higher predictive value for weaning success compared with conventional weaning parameters. Trial registration no. ChiCTR2100049817.