Authors' reply.

Sir,

We greatly appreciate the comments[1] on our recent article, ‘Utility of on-site evaluation of endobronchial ultrasound-guided transbronchial needle aspiration specimens (EBUS-TBNA)’.[2] The perspective put forth by the authors of the letter is extremely valuable, as it illustrates the variation in the practice of EBUS-TNBA, with and without rapid on-site evaluation (ROSE) across institutions.

While we acknowledge that performing EBUS-TBNAs in an Operating Room under general anesthesia allows for immediate transition to mediastinoscopy, this is not the common practice at our institution, for several reasons. By performing EBUS-TBNA procedures under conscious sedation our patients are able to avoid the expense as well as the morbidity associated with the use of an Operating Room and general anesthesia. The success of our experienced EBUS operators, as shown by our overall diagnostic rate of greater than 90%, further supports our institutional practice, as few patients subsequently undergo the more invasive mediastinoscopy.

We have reviewed the Gibson, et al. article cited by the authors.[3] Although we acknowledge that the article by Gibson, et al., addresses the use of ROSE as a means to triage a patient for mediastinoscopy, it fails to critically analyze the role ROSE plays in triaging the specimens for ancillary studies, as compared to those performed without ROSE. Furthermore, while the authors of the letter suggest that ROSE has greatly assisted them in triaging their specimens, they do not offer evidence to support this statement.

We believe that our data delivers an important message; that the work-up of a patient's specimen is not compromised by a clinician's decision not to pursue ROSE. As shown in our study, a cell block was produced and a diagnosis achieved in nearly all specimens obtained with and without ROSE at similar rates (92 vs. 88% and 94 vs. 90%, respectively). As explained in our article the specimens from patients who did not undergo ROSE were routinely processed as liquid-prep and cell-block preparations, thus avoiding ‘unnecessary slides,’ and conserving material for ancillary testing. Furthermore, the finding that more of our malignant specimens with ROSE had immunostains performed (63 vs. 37%), likely reflects the differences in the patient's clinical presentation, the underlying diagnosis, and the need for immunostaining, to render a definitive diagnosis rather than the material available. We would also recommend against drawing conclusions from our data on flow cytometry, given its use in only 17 of 294 (6%) specimens.

We do acknowledge that our article did not address the use of EBUSspecimens for molecular testing. However, a recent article from our institution evaluated our experience with using cytology specimens for molecular testing and showed these to be a suitable alternative to surgical biopsies.[4]

The authors’ letter stresses the importance of each institution's experience in the formulation of an EBUS-TBNA protocol, with respect to ROSE. We firmly agree, and feel that our study has assisted us by refining our practice and providing better patient care.