Zhang et al pooled 20 studies that compared ultrasound, chest X-ray, or both against a reference standard (usually CT scan) for the diagnosis of pneumothorax.
- Chest X-ray had a pooled sensitivity of 52% and specificity 99% for diagnosis of pneumothorax.
- Ultrasound's pooled sensitivity was 88% and specificity, 100%.
Unsurprisingly, the accuracy of ultrasonography to diagnose PTX was dependent on the operator's experience, in the studies that allowed multiple non-radiologist examiners. How do you rule-in or rule-out a pneumothorax on ultrasound?
- The lung sliding sign and the comet tail sign are what to look for; if both are present, there's no pneumothorax. (Here's another example of normal lung sliding.)
- The lung-point sign, or absence of lung sliding, suggests pneumothorax is present (I can't tell you the specificity of this sign in isolation).
Patients with COPD may have more false-positive PTX's on ultrasound, some say.
I'm impressed, and it's attractive to think we can carry these devices around and make rapid, accurate diagnoses without waiting for test results or subjecting our unstable patients to road-trips. But because this is such a new, hot area with plenty of people justifiably eager to prove ultrasound's potential, I wonder about publication bias (toward positive studies by experienced examiners). I'd like to know, how accurate would my bedside ultrasound exams be after a weekend course and a few nights on call, for example?
For a great resource on ultrasound in the ICU, check out the Critical Care Ultrasonography Forum (critcaresono.com) by Pierre Kory and friends at Beth Israel in NYC. Lots of great ultrasound video clips for reference and self-education -- all for free of course.
Ding W et al. Diagnosis of Pneumothorax by Radiography and Ultrasonography: A meta-analysis. CHEST 2011;140:859-866.