False negative results and endobronchial ultrasound-guided transbronchial needle aspiration.

Abstract

Tian et al. have reported on the possible reasons for false negative results of endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA).1 A number of other points are relevant to their study. Firstly, the prevalence of disease is important to consider in order to interpret diagnostic performance. Specifically, their reported sensitivity of EBUS-TBNA for malignancy of 96% is impressive compared to other published studies and systematic reviews,2 and their lower negative predictive value (NPV) of 33%2 can be explained by the significantly higher disease prevalence of 92% in their study. It is also helpful to know the expertise of the centre as the learning curve for EBUS-TBNA is known to be highly variable, even amongst experienced bronchoscopists, and may influence diagnostic performance.3, 4 In this particular study, given the impressive sensitivity achieved, such factors are unlikely to have been an influence. Secondly, it is important to consider the local pathology services in interpreting results with EBUS-TBNA. They have suggested that the use of the higher gauge (21G) EBUS-TBNA needle may be a useful strategy to minimise false negative EBUS-TBNA results by providing a tissue core and more tissue for immunostaining. We too, have reported our initial findings with the 21G needle and also found a greater ability of the 21G needle to characterise benign and malignant EBUS-TBNA samples.5 However, we must emphasise this was in an institution with a histopathology service used for the analysis of EBUS-TBNA samples which analysed tissue cores in formalin, rather than in liquid cytology bottles or smeared slides analysed by cytopathologists. Although the endobronchial ultrasound-guided transbronchial needle forceps (EBUS-TBNF) sound like a promising alternative to the 21G needle to further increase the sample size, they are unlikely to become mainstream because of concerns over the risk of bleeding. Therefore, it is likely that the 21G needle will be the main strategy to increase sample size. It should be appreciated that the sample volume is also likely to be proportionate to the amount of blood admixed to the sample. The use of the smaller, lighter 22G needle might be preferable still in nodal locations where greater flexibility of the EBUS-TBNA bronchoscope is required, such as station 4L, and also where the nodes are very small. Finally, it would have been informative to have included a breakdown of the size and station of the lymph nodes sampled, especially those where false negative EBUS-TBNA results were obtained, to inform the readership. It is anticipated that nodes of the smallest size in the more technically difficult locations, for example, station 2R or 2L as compared with station 7, might be associated with negative results, but it would be informative to know this. No authors report any conflict of interest.