On the necessary paradigm-shift from qualitative to quantitative lung ultrasound

Abstract

Lung ultrasound (LUS) is spread in the medical world since the nineties. In fact, despite ultrasound imaging is not designed for the lungs, there exists a clear difference between images of healthy and diseased lung tissue. This difference relies on specific imaging artifacts, which have been studied and classified in the literature. However, this differentiation is performed by visual inspection of clinical experts. Moreover, standardization of imaging protocols is lacking. As a result, three strong limitations affect LUS: (1) only qualitative analyses are performed, which are influenced by intra and inter operator variability; (2) intra and inter ultrasound-scanner variability is not taken into account; (3) although the sensitivity is reported to be high, specificity remains low. To tackle these problems, a paradigm shift from qualitative to quantitative LUS is needed. In this talk, current efforts in this direction will be reviewed. A standardized image acquisition protocol and scoring system developed for COVID-19 patients will be presented, together with dedicated algorithms able to automatically score and segment LUS images. Data from a multicenter study involving 200 patients will be discussed. Moreover, the results from a recent study on quantitative lung ultrasound spectroscopy applied to the diagnosis of pulmonary fibrosis will be introduced.