Diaphragm assessment by two dimensional speckle tracking imaging in normal subjects

Abstract

Background: Conventionally, ultrasonographic assessment of diaphragm contractility has involved measuring respiratory changes in diaphragm thickness (thickening fraction) using B-mode or caudal displacement with M-mode. Two-dimensional speckle-tracking has been increasingly used to assess muscle deformation ('strain') in echocardiography. We sought to determine in a pilot study if this technology could be utilized to analyze diaphragmatic contraction.

Methods: Fifty healthy adult volunteers with normal exercise capacity underwent ultrasound imaging. A linear array transducer was used for the assessment of diaphragm thickness, thickening fraction (TF), and strain in the right anterior axillary line at approximately the ninth intercostal space. A phased array transducer was applied subcostally for the assessment of diaphragm displacement on the right mid-clavicular line. Diaphragmatic images were recorded from the end of expiration through the end of inspiration at 60 % maximal inspiratory capacity. Diaphragm strain was analyzed off-line by speckle tracking imaging. Blinded inter- and intra-rater variability was tested in 10 cases.

Results: Mean right diaphragm thickness at end-expiration (±SD: standard deviation) was 0.24 cm (±0.1), with TF of 45.1 % (±12) at 60 % peak inspiratory effort. Mean right diaphragm caudal displacement was 4.9 cm (±1). Mean right diaphragm strain was -40.3 % (±9). A moderate correlation was seen between longitudinal strain and TF (R(2) 0.44, p < 0.0001). A weak correlation was seen between strain and caudal displacement (R(2) 0.14, p < 0.01), and an even weaker correlation was seen between caudal displacement and TF (R(2) 0.1, p = 0.04). Age, gender, and body mass index were not significantly associated with right diaphragm strain or TF. Although inter- and intra-rater variability was overall good for TF, caudal displacement, and strain (inter-rater R(2); 0.8, 0.9, and 0.7, respectively [p < 0.01], intra-rater R(2); 0.9, 0.7, and 0.9, respectively [p < 0.01]), strain values did have a slightly lower inter-rater repeatability.

Conclusions: Diaphragmatic strain estimated by speckle tracking imaging was associated with conventional ultrasound measures of diaphragmatic function (TF and caudal displacement). Further clinical studies are warranted to investigate its clinical utility.

Keywords: Diaphragm; Speckle tracking; Ultrasound.

Fig. 1

Ultrasound image of normal diaphragm in zone of apposition. A linear array transducer was applied with the use of M-mode ultrasound at the right anterior axillary line at approximately the ninth intercostal space

Fig. 2

Examples of conventional assessment of diaphragm function. a Diaphragm thickening Fraction = diaphragm diameter at end-inspiration minus diaphragm diameter at end-expiration divided by diaphragm diameter at end-expiration, expressed as a percentage: [(DDinsp – DDexp)/DDexp] x 100. b Diaphragm caudal displacement with inspiration measured with a phased array transducer with the use of M-mode ultrasound in mid-axillary line in subcostal position

Fig. 3

Representation of diaphragm strain assessment. D1 = Distance between diaphragm ‘kernels’ (unique groups of grey-scale pixels) at end-expiration. D2 = Distance between diaphragm ‘kernels’ at end-inspiration

Fig. 4

Longitudinal strain of diaphragm during inspiration. The x-axis represents time (millisecond), and the y-axis represents longitudinal strain (%). Strain is a measure of relative deformation and is a negative value. In this example, the central portion of the region of interest (depicted in blue in the ultrasound image at the upper-left corner) was traced and measured as – 46.4 % (displayed in the column below the x-axis). The more negative value means the higher degree of deformation (contraction)

Fig. 5

Graphical representation of correlation among various diaphragm assessment methods. a Log(Longitudinal strain) vs. thickening fraction (Log Strain = 3.1 – 0.01*Thickening fraction; R² 0.44, p < 0.0001), b Log(Longitudinal strain) vs. caudal displacement (Log Strain = 3.4 + 0.6*Displacement; R² 0.14, p < 0.01), and c Thickening fraction vs. caudal displacement (Thickening fraction = 33.5 + 2.4*Displacement; R² 0.1, p = 0.04)