Comparative Diagnostic Accuracy of Blinded Lung Ultrasound and Clinicians in Pulmonary Consolidations

Bedside lung ultrasound (LUS) is an increasingly utilized technique as an adjunct to clinical history and physical examination. This study aims to assess the added value of LUS by comparing the diagnostic accuracy of a blinded sonographer’s ultrasound findings against the clinical suspicion of pulmonologists and emergency physicians at admission. A prospective single-center consecutive study of hospitalized patients (n=101) with imaging-detected consolidations was conducted. Ultrasound was performed by a blinded operator unaware of the physician’s diagnostic suspicions. Definitive diagnoses were established through clinical-radiologic follow-up at 12 months, biopsy, and/or microbiological studies. The pulmonologist and emergency physician correctly estimated the probability of benignity/malignancy in 88.1% and 80% of the cases, respectively, whereas the blind operator achieved a 90% accuracy rate. Combining LUS findings with pulmonologists’ clinical data improved benign/malignant classification accuracy to 96%. The agreement with the definitive dichotomous diagnosis (benign/malignant), assessed through Cohen kappa, was 0.664 for pulmonologists, 0.129 for emergency physicians, and 0.747 for the ultrasonographer. Ultrasound demonstrated a sensitivity (Se) of 85% and specificity (Spe) of 92% for diagnosing cancer. For pulmonologists, Se was 65% and Spe was 96%, whereas for emergency physicians, Se was 12% and Spe was 97%. The approach to hospitalized patients with imaging-detected consolidations that achieves the highest diagnostic accuracy (96%) combines clinical suspicion with ultrasound findings. Our findings confirm the importance of LUS, particularly in suspected malignancy.