Assessment of five different probes for lung ultrasound in critically ill patients: A pilot study

The present study was aimed at comparing the diagnosis concordance of five echo probes of lung ultrasound (LUS) with CT scans in intensive care and emergency patients with acute respiratory failure.

This prospective, observational, pilot study involved 10 acute patients in whom a thoracic CT scan was performed. An expert performed an LUS reference exam using five different probes: three probes with a high-quality conventional echo machine (cardiac phased-array probe, abdominal convex probe, linear probe) and two probes (cardiac and linear) with a pocket ultrasound device (PUD). Then, a trained physician and a resident performed 'blinded' analyses by viewing the video results on a computer. The primary objective was to test concordance between the blinded echo diagnosis and the CT scan.

In the 100 LUS performed, the phased-array probe of the conventional machine and linear array probe of the PUD have the best concordance with the CT scan (Kappa coefficient=0.75 [CI 95%=0.54-0.96] and 0.62 [CI 95%=0.37-0.86], respectively) only for experts and trained physicians. The agreement was always poor for residents. Convex (abdominal) and linear transducers of conventional machines and the phased-array transducers (cardiac) of PUD have poor or very poor agreement, regardless of the physician's experience.

Among the probes tested for LUS in acute patients, the cardiac probe of conventional machines and the linear probes of PUDs provide good diagnosis concordance with CT scans when performed by an expert and trained physician, but not by residents.