Diagnostic performance of LUS
LUS demonstrated remarkable sensitivity and negative predictive value
(NPV) among reviewed studies, with sensitivity ranging from 68% to
93.3% and a NPV ranging from 52%to 94.1% [78,80,82,87],
suggesting the utility of LUS as a screening test to rule out COVID-19
lung infection. Data regarding specificity, positive predictive value
(PPV) and diagnostic accuracy has been conflicting, with some studies
demonstrating values as low as 21.3%, 19.2% and 33.3% for
specificity, PPV and diagnostic accuracy, respectively[87], while
others demonstrating higher values up to 92.9%, 84.6% and 93.3% for
specificity, PPV and diagnostic accuracy, respectively[78]. Lu W et
al. demonstrated higher sensitivity and NPV of LUS in patients with
moderate disease (77.8% and 88.9%, respectively) with both values
reaching 100% in patients with severe disease. Overall, LUS showed a
higher efficacy in assessing patients with no and severe lung lesions,
with diagnostic accuracy for patients with no, mild, moderate, and
severe lung lesions of 93.3 %, 76.7 %, 76.7 %, and 93.3 %,
respectively[78]. Bar S et al. suggested four ultrasound signs that
were independently associated with a positive COVID-19 RT-PCR; Upper
sites B lines ≥ 3 [OR 1.52 (1.31‐1.79]), lower sites thickened
pleura [OR 1.73 (1.49‐1.98)], lower sites consolidation [OR 2.39
(2.07‐2.69)], and posterolateral sites thickened pleura [OR 1.97
(1.72‐2.22)][88], suggesting a role for LUS in triaging patients
with suspected COVID-19 lung infection.