Tracking pleural sliding motion to assess lung overdistention using an open source algorithm: a proof-of-concept study on lung ultrasound scans

TL;DR

This study shows that tracking pleural line movement with open-source software can help detect lung overdistention and pneumothorax using ultrasound.

Contribution

A novel open-source algorithm is proposed to quantify pleural line movement for assessing lung overdistention.

Findings

PL movement increased significantly during quiet breathing compared to expiratory hold in healthy volunteers.

PL movement correlated with lung sliding severity in ICU patients and with overdistention in COVID-19 patients.

Motion tracking correctly identified all cases of absent lung sliding with no false positives.

Abstract

Pleural line (PL) movement, assessed by lung ultrasound, is crucial for the detection of pneumothorax and might also indicate overdistention, but research is limited by the lack of a quantitative tool. We set out to answer two research questions: can PL movement be quantified using open-source motion tracking software, and can PL movement be used to identify overdistention? We hypothesize that motion tracking of the PL is feasible and represents an accurate estimation of lung sliding.

Lung ultrasound video clips from three patient groups were used: (1) healthy volunteers during expiratory hold maneuvers (functional residual capacity) and quiet breathing, (2) ICU patients, blindly assessed for lung sliding (absent, doubtful, evident but limited or evident and extensive) and (3) Severe COVID-19 viral pneumonia patients undergoing PEEP titration and electrical-impedance tomography. Open-source software that implements the “Channel and Spatial Reliability Tracking” tracker algorithm was used for motion tracking, identifying the PL at three points and a soft tissue reference. Each motion-time curve was subsequently smoothed and normalized to account for soft tissue displacement. The maximum lateral movement on the transversal plane among the three normalized PL landmarks defined PL movement.

In 143 video clips from 7 healthy individuals, PL movement increased from functional residual capacity (1.2 ± 0.6 mm) to quiet breathing (5.4 ± 2.5 mm; p < 0.01). In 336 video clips from 40 ICU patients, PL movement increased from absent (2.7 ± 1.2 mm) to extensive lung sliding (14.7 ± 5.8 mm; p < 0.01). Ordered logistic regression predicted Absent sliding with 71% balanced accuracy, with motion tracking correctly identifying all cases and no patients without lung sliding misclassified as extensive, based on visual inspection of the pleural line. In 358 video clips from 30 patients undergoing PEEP titration, there was an association between overdistention quantified by electrical-impedance tomography and PL movement (Spearman-rho=−0.6). PL movement decreased from low to high PEEP levels (p < 0.01).

Pleural line motion tracking is feasible and provides quantitative insight into pleural movement based on data from healthy volunteers and visual inspection of images from ICU patients. Moreover, pleural line movement allows accurate assessment of overdistention during mechanical ventilation when compared with electrical-impedance tomography.

The online version contains supplementary material available at 10.1186/s13054-025-05742-8.

Figures