Comparison of transbronchial needle aspiration with and without ultrasound guidance for diagnosing benign lymph node adenopathy

Abstract
Background Transbronchial needle aspiration (TBNA) is a minimally invasive procedure performed to diagnose lymph node (LN) adenopathy. TBNA with and without endobronchial ultrasound (EBUS) guidance has a high diagnostic yield for malignant LN enlargement, but the value for diagnosing benign LN enlargement has been less thoroughly investigated. Methods We retrospectively evaluated 3540 patients with mediastinal LN enlargement who received TBNA. One hundred sixty-six patients with benign mediastinal lymphadenopathy were included and 293 LNs were biopsied. A positive result was defined as a specific histological abnormality. Conventional TBNA (cTBNA) and EBUS-TBNA, as well as cTBNA and transbronchial forceps biopsy (TBFB), were compared. The subgroup analysis was stratified by disease type and LN size. Results A diagnosis was made in 76.84% of the EBUS-TBNA and 61.31% of the cTBNA (P < 0.05). EBUS-TBNA was superior to cTBNA for both granulomatous (65.18% vs. 45.45%, P < 0.05) and non-granulomatous disease (96.92% vs. 84.06%, P < 0.05). In contrast, the diagnostic yield of EBUS-TBNA was higher than that of cTBNA for LNs < 20 mm (79.44% vs. 64.29%, P < 0.05), but for LNs > 20 mm the difference was marginal. These findings were confirmed in a group of independent patients who received cTBNA plus EBUS-TBNA. The diagnostic yield did not differ between cTBNA and TBFB, but significantly increased to 76.67% when both modalities were employed. Conclusions EBUS-TBNA is the preferred minimally invasive diagnostic method for benign mediastinal LN disease. Combined cTBNA and TBFB is a safe and feasible alternative when EBUS is unavailable.
Funding Information
  • Natural Science Foundation of Zhejiang Province (LY20H010004)
  • National Natural Science Foundation of China (81870022)
  • the Medical and Health Technology Program of Zhejiang Province (2017204226)

This publication has 39 references indexed in Scilit: