On-site Cytopathologic Analysis of Bronchoscopic Needle Aspiration

Bronchoscopic needle aspiration (BNA) is a valuable diagnostic and staging tool in intrathoracic malignancies and other mediastinal diseases but is underused. Impediments to its popular use include lack of formal training, technical problems with the procedure, unavailability of expert cytopathologists, concerns of false-positive results, failure to reproduce published results, risk of damage to the bronchoscope, concerns regarding the safety of the procedure, and reservations regarding the clinical utility of BNA.1 Some of these problems may be able to be overcome with the use of on-site cytopathologic review. It has been documented that training can improve diagnostic yield,2 but to maintain proficiency continuous feedback is invaluable. Immediate feedback from on-site review can help improve technique and needle positioning. False positives can be reduced by enhanced communication with the cytopathologists as well as independent review of the slides. The presence of lymphocytes mixed with malignant cells or large clumps of malignant cells helps determine proper acquisition of metastatic deposits and excludes the specimen with rare cells, which may indicate endobronchial contamination. Inadequate specimens with the presence of numerous bronchial epithelial cells without lymphocytes or malignant cells can help decision making during the procedure: either to redirect the needle or to proceed with other diagnostic techniques if further needle aspiration is felt to be futile. The use of rapid on-site evaluation (ROSE) in training can be amply demonstrated in the education of pulmonary fellows in the technique. The visual feedback that they can see in review of slides soon after the performance of the procedure can boost their confidence in the proper technique and can reinforce it as well as allow them to see the “disappointing” results of improper sampling. However, increased yield is the standard to which ROSE is commonly held. Both Davenport3 and Diette et al.4 have shown that ROSE increases yield. Davenport3 found an increase in the percentage of specimens containing malignant cells to increase from 31 to 56% with immediate interpretation compared with routinely processed specimens, and a decrease in percentage of inadequate specimens from 56% to 18%. It is not clear in this study whether ROSE affected patient management. Diette et al.4 also showed, in an observational study, an increase in diagnostic yield from 50 to 81% in all bronchoscopic procedures with the presence of on-site cytopathologic assessment (70% of patients had a “new” diagnosis established by needle aspiration or biopsy). Another small study found an increase in diagnostic yield in bronchoscopic needle biopsy from 19 to 81% after adding on-site evaluation.5 Unfortunately, to date there has not been a robust, prospective, randomized study to examine this. This is similar to the literature regarding fineneedle aspiration (FNA) in general, but in most centers ROSE is used widely for nonbronchoscopic FNA. In endoscopic ultrasound-guided FNA, it has been reported that the presence of a cytopathologist increases the yield by 10%.6,7 A small meta-analysis examining transthoracic FNA demonstrated a notable increase in yield.8 A comparison of two studies of FNA of breast lesions showed a better yield from the study with methods that included on-site review.9 Although a prospective randomized study favoring on-site cytopathologic review with needle aspiration is lacking, the current literature does at least suggest that the yield is better with on-site evaluation than without. Other benefits of ROSE in BNA can be seen in small cell lung carcinoma, true-negative aspirations, multistation lymphadenopathy, and other malignant and nonmalignant diagnoses. We have shown in our institution that the bronchoscopic diagnosis of small cell lung carcinoma is commonly made by BNA.10 Once small cell lung carcinoma is diagnosed by ROSE, further sampling procedures (forceps biopsy, bronchial brushings and washings) become superfluous and the procedure can be shortened. The presence of abundant lymphocytes without malignant cells decreases the likelihood of a falsenegative aspiration.11–12 In multistation lymphadenopathy associated with bronchogenic carcinoma, the presence of ROSE can help sampling selection. If a diagnosis is established in the more distant site (contralateral), then further nodal sampling may not be necessary; however, if adequate sampling cannot be established or does not show the presence of malignancy, sampling of the other sites can be done serially. The presence of granulomas in a field of lymphocytes suggests sarcoid or other granulomatous processes and can help support a diagnosis or prompt the collection of appropriate culture material. If lymphoma is suggested, then a sample for flow cytometry can be obtained for further confirmation. Drawbacks to ROSE are costs, the time and effort of cytopathologists (which may or not be reimbursed), and the availability of an experienced cytopathologist.1,13 It has been shown that BNA is cost-effective.14 Therefore, if ROSE can help popularize the use of BNA for diagnosing and staging peritracheal/bronchial disease, the increase in cost of ROSE can be minimized by the avoidance of the cost associated with the other more invasive tissue-accessing procedures. ROSE does require cytotechnologists and cytopathologists, whose availability may differ in various settings and whose familiarity with the smear technique may differ as well. The use of ROSE could also increase bronchoscopic time while awaiting drying, staining, and careful reviewing of the slides. The current literature favors the use of on-site cytopathologic evaluation in BNA, primary for its increase in yield.3–5 In our own study, we found a difference in obtaining a malignant diagnosis from 25 to 71%15 without and with ROSE respectively, but the use of ROSE was not controlled and most malignant diagnoses were established by the fourth aspiration with a plateau after seven aspirations regardless of ROSE.15 Even if new randomized studies do not show a notable improvement in yield, the other advantages outlined earlier, especially adequacy of the sample, are equally compelling and favor the use of ROSE if this service is available. I would still favor the use of ROSE selectively in BNA cases, but I may have a difficult time in convincing my colleagues in pathology to continue to provide the support.