A Novel Normalized Cross-Correlation Speckle-Tracking Ultrasound Algorithm for the Evaluation of Diaphragm Deformation

Abstract

Objectives: To develop a two-dimensional normalized cross-correlation (NCC)-based ultrasonic speckle-tracking algorithm for right diaphragm deformation analysis.Methods: Six healthy and eight mechanical ventilation patients were enrolled in this study. Images were acquired by a portable ultrasound system in three sections. DICOM data were processed with NCC to obtain the interframe/cumulative vertical and horizontal displacements, as well as the global strain of the right diaphragm, with continuous tracking and drift correction.Results: The NCC algorithm can track the contraction and relaxation of the right diaphragm by following the respiratory movement continuously. For all three sections, the interframe and accumulated horizontal displacements were both significantly larger than the corresponding vertical displacements (interframe p values: 0.031, 0.004, and 0.000; cumulative p values: 0.039, 0.001, and <0.0001). For the global strain of the right diaphragm, there was no significant difference between each pair of sections (all p > 0.05), regardless of whether the horizontal interval of the initial diaphragm point was 1, 3, 5, or 10 times in the sampling interval.Conclusions: This study developed a novel diaphragm deformation ultrasound imaging method. This method can be used to estimate the diaphragm interframe/accumulated displacement in the horizontal and vertical directions and the global strain on three different imaging planes, and it was found that the strain was not sensitive to the imaging plane.