Ability of Rapid On-Site Cytologic Evaluation to Guide Lymph Node Sampling Reliably During Bronchoscopy for Granulomatous Disease.

BACKGROUND

Rapid on-site cytologic evaluation (ROSE) during endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) has the potential to influence specimen acquisition, but ROSE's ability to inform decision-making during bronchoscopy for granulomatous disease has not been assessed.

RESEARCH QUESTION: Does ROSE detect granulomas accurately enough to guide lymph node sampling reliably during EBUS-TBNA for granulomatous disease?

STUDY DESIGN AND

METHODS

We retrospectively studied individuals who underwent EBUS-TBNA with ROSE performed by cytotechnologists from 2013 through 2021 and either (1) a granuloma was identified by ROSE or final cytologic review or (2) they received a diagnosis of granulomatous disease during clinical follow-up. We assessed the accuracy of ROSE and EBUS-TBNA final cytologic results for a diagnosis of granulomatous disease during serial lymph node sampling.

RESULTS

Granulomatous disease was diagnosed in 176 individuals, including 150 patients with a granuloma identified by EBUS-TBNA final cytologic analysis and 26 patients who received a diagnosis by other means. Sarcoidosis was diagnosed in 143 patients and miscellaneous diseases were diagnosed in 33 patients. Results of ROSE and final cytologic analysis were concordant in 76% of procedures, similar to studies that have used cytopathologists. Using final cytologic analysis as the gold standard, ROSE showed a diagnostic yield of 0.74, false-positive rate of 2%, and positive and negative predictive values of 0.97 and 0.38, respectively. Final cytologic analysis showed a diagnostic sensitivity of 0.85 with a cumulative diagnostic yield of 98% when sampling just 3 lymph nodes. Although uncommon, false-positive diagnoses resulting from ROSE were less frequent when > 1 lymph node was sampled.

INTERPRETATION: ROSE showed limited practical value in guiding lymph node sampling in granulomatous disease. Diagnostic yield did not improve by sampling > 3-4 lymph nodes in patients with negative ROSE findings, and false-positive results are avoided by sampling at least 2 nodes. Negative ROSE results should not discourage submitting nodal aspirates for microbiologic studies when suspicion for granulomatous disease is high.