Objective: To determine the accuracy of ultrasound in the detection of pneumothorax.
Methods: Prospective blinded study comparing ultrasound, CT and radiographic findings in 29 patients following 41 CT-guided lung biopsies. Ultrasound examination of the chest was limited to the biopsy needle entry site.
Results: Thirteen patients developed a post-biopsy pneumothorax demonstrated by CT. Seven of these were detected by ultrasound and six were visible on erect chest radiographs. Six of the 13 pneumothoraces were not detected by ultrasound, but five of these were loculated away from the biopsy needle entry site and were therefore in areas not examined during the limited ultrasound examination. There were no false-positive diagnoses of pneumothorax using ultrasound. The positive predictive value for ultrasound was 100% and the negative predictive value was 82%.
Conclusion: In this patient group, ultrasound was more sensitive than erect chest radiography in the detection of pneumothorax. Both have a specificity of 100%. This study suggests that ultrasound may prove valuable in pneumothorax detection when rapid conventional radiography is not possible or practical, and in circumstances where ultrasound is readily available, such as during ultrasound-guided interventional procedures.