Envelope Statistics Analysis of M-mode Signals in Lung Ultrasound for Distinguishing Stratosphere from Seashore Signs: A Preliminary Study

M-mode ultrasound images of a healthy lung exhibit a seashore pattern containing a sandy sign resulted from the lung sliding while those of a pneumothorax have a stratosphere pattern (barcode sign) indicating the lung sliding absence. In this study, the envelope statistics analysis, which is commonly used for B-mode images, was applied to lung ultrasound M-mode images to quantify the difference between the seashore and stratosphere patterns. A probability density function of the envelope data of the M-mode images was modeled by a Rice distribution to quantify the difference of coherent component in the chest wall region (mimicking pseudo barcode sign) and incoherent component in the healthy lung region (sandy sign). An in-vivo experiment with a healthy volunteer showed that the Rice scale parameters of the pseudo barcode sign in the chest wall region were greater than those of the sandy sign in the lung region. The result indicates the potential of the envelope statistics analysis of M-mode signals using the Rice distribution for distinguishing the barcode sign from the sandy sign quantitatively, to automatically detect the pneumothorax.