Randomized Cross-over Trial of Endobronchial Ultrasound Transbronchial Needle Aspiration With or Without Suction in Suspected Malignant Lymphadenopathy.

Abstract

The benefit from applying suction during endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is unclear. The purpose of this study was to evaluate the impact of applying suction during EBUS-TBNA of lymph nodes (LN) on the diagnostic yield and specimen quality of EBUS-TBNA smears and cellblocks. We conducted a randomized blinded cross-over trial comparing the diagnostic yield and quality of EBUS-TBNA samples obtained with suction (S+) and without suction (S-) from suspected malignant mediastinal LN. Each LN was aspirated 4 times in a randomly assigned alternating order: S+/S-/S+/S- or S-/S+/S-/S+, resulting in 2 matched S+/S- smears pairs and 1 matched S+/S- cellblocks pairs. Specimen were assessed by a blinded cytopathologist. The primary outcome was the diagnostic yield of EBUS-TBNA smears and cellblocks. Secondary analyses accounted for LN size (≤2 vs. >2 cm) and station (mediastinal vs. hilar). Of 234 EBUS-TBNA smears and 114 cellblocks were obtained from 59 LN. The concordance of S+ and S- cellblocks diagnostic yield was low (43.8%; 95% confidence interval: 30.7-57.6) with a significantly higher diagnostic yield among S+ compared with S- cellblocks (56.1% vs. 10.5%, respectively, P<0.001), regardless of LN station and size. However, S+ and S- smears were moderately concordant (66.7%; 95% confidence interval: 57.4-75.1) with no overall difference in their respective diagnostic yield. Numerically higher diagnostic yield among S+ smears in LN≤2 cm or hilar LN was not statistically significant. Results were similar for specimen adequacy, sensitivity and diagnosis of malignancy. Applying suction during EBUS-TBNA of suspected malignant LN enhanced the diagnostic yield of cellblocks regardless of mediastinal LN station or size, but not of smears.