[Diagnostic accuracy of endobronchial ultrasound-guided transbronchial needle aspiration in mediastinal lymph nodes].

INTRODUCTION

Endobronchial ultrasound-guided transbronchial needle aspiration in mediastinal lymph nodes is a safe method that can be performed for mediastinal and hilar lymph nodes sampling. Because it allows collecting only a small amount of materiel, an optimal processing of the samples is needed.

MATERIAL AND METHODS

Based on 150 consecutive procedures, we evaluate the overall diagnostic performances and of each technical methods used for the exploitation of the samples.

RESULTS

The global diagnostic yield is 64.0% for the 50 first exams (learning phase), 88.0% for the next 100 exams. The maximal sensitivity and negative predictive value (NPV) are 79.3%, 96.3% and 92.7%, respectively. The yield, sensitivity and NPV of smears are 68.0%, 75.0% and 66.0%, of monolayer preparation 77.8%, 62.1% and 50.0%, of sections from tissue cores of 65.8%, 94.4% and 86.7%. The combination of the different methods increases the yield comparing to tissue cores and smears when taken alone (P < 0.05), and the sensibility and the NPV comparing to smears (P < 0.0005 and P < 0.01, respectively) and monolayer preparation (P < 0.0001 and P < 0.0005, respectively). The sensitivity of tissue cores is greater than smears (P < 0.005) and monolayer preparations (P < 0.0001). This increase in sensitivity is significant for granuloma (sarcoidosis), but not for carcinoma.

CONCLUSION

Endobronchial ultrasound-guided transbronchial needle aspiration is an accurate and sensitive technique. Liquid based conditioning of samples and paraffin embedded tissue cores increases the diagnostic performances comparing to smears, notably for the diagnosis of sarcoidosis.