Accuracy of transthoracic ultrasound for the detection of pleural adhesions

Abstract

In the era of minimally invasive surgery, preoperative detection of pleural adhesions can be very useful for the assessment of surgical approach, because pleural adhesions are the main contraindication to video-assisted thoracoscopy. The aim of this study was to assess the sensitivity and specificity of transthoracic ultrasound in the detection of pleural adhesions prior to thoracic surgery. From February 2010 to January 2011, 142 consecutive patients (male, 98; female, 44; age range, 36-83 years, mean age, 63.4 years) undergoing surgical thoracic intervention (except for pneumothorax) were preoperatively scanned by two different surgeons. According to thoracic wall projections of lung segments, we created a nine-region topographic map, in which every pulmonary area was scanned to assess the presence or the absence of 'gliding sign' (lesion-by-lesion analysis). During operations the surgeon, blinded to the prediction, confirmed or excluded each suspected adhesion or documented other adhesions not previously identified. A total of 1192 predictions were made. Ultrasound predictions were confirmed 1124 times and refuted 68 times. Sensitivity was 80.6% (95% confidence interval, 0.740-0.872) and specificity 96.1% (95% confidence interval, 0.949-0.973). The positive predictive value was 73.2% and the negative predictive value was 97.4%. Transthoracic ultrasound is an effective method for predicting pleural adhesions before thoracic surgery in experienced hands. Its safety, feasibility and low cost make it a useful method for the planning of minimally invasive surgical interventions.