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

Transbronchial needle aspiration (TBNA) is a minimally invasive diagnostic modality for patients with mediastinal and hilar diseases [1]. Endobronchial ultrasound (EBUS) allows for real-time localization and aspiration of lymph nodes (LNs) during bronchoscopy, which improves the safety and accuracy of TBNA. Both conventional TBNA (cTBNA) and EBUS-guided TBNA (EBUS-TBNA) are associated with a high diagnostic yield for malignant mediastinal LN enlargement [2‐5]. However, the diagnostic value of the two modalities for benign mediastinal lymphadenopathy has been less thoroughly investigated. The diagnostic efficiency of the various biopsy methods remains controversial. The reported diagnostic rate of cTBNA in patients with benign mediastinal LN enlargement varies between 21.4 and 76% [6‐11], while that of EBUS-TBNA varies between 74.5 and 96% [12‐19]. However, these studies used single-arm designs and had relatively small sample sizes. The main determinants of the diagnostic rate for cTBNA are the operative technique and the type of specimen (tissue or cytology) obtained by needle aspiration. Although EBUS improves the diagnostic rate of cTBNA, puncture specimens are usually a mixture of cell masses and blood clots that are not always morphologically representative. A few studies have introduced the technique of transbronchial forceps biopsy (TBFB) as a potential option to obtain a large volume of biopsied tissue, which could aid in the diagnosis of benign mediastinal lymphadenopathy [20‐23]. However, the sample size of these studies was somewhat limited and the reported diagnostic rates varied greatly. Most studies were conducted under EBUS guidance and the diagnostic efficiency of TBFB without EBUS remains unknown. More importantly, no study has directly compared TBFB and TBNA. Thus, the role of TBNA, including cTBNA and EBUS-TBNA, in the diagnosis of benign thoracic LN enlargement is not known. Whether TBFB is superior to TBNA for evaluating mediastinal and hila lymphadenopathy remains to be determined. We retrospectively analyzed 3540 patients with benign mediastinal lymphadenopathy. Systematic comparison of cTBNA and EBUS-TBNA, as well as of cTBNA and TBFB, was carried out. We aimed to determine the diagnostic role of TBNA and TBFB in patients with benign mediastinal lymphadenopathy.

The current study provides strong evidence of the effectiveness of cTBNA, EBUS-TBNA, and TBFB for diagnosing benign mediastinal lymphadenopathy. The major findings of this retrospective comparative study were as follows: TBNA was effective for revealing enlargement of benign LNs; EBUS-TBNA had a higher diagnostic yield than cTBNA, but cTBNA was not inferior for LNs > 20 mm; and TBFB without EBUS guidance is a safe and feasible option for diagnosing benign LN adenopathy. However, combined cTBNA and TBFB achieved a higher diagnostic yield. Our study provides solid evidence that EBUS-TBNA is the optimal approach for diagnosing benign LN enlargement, although combined cTBNA and TBFB is a viable alternative when EBUS guidance is not available.

Both cTBNA and EBUS-TBNA are important tools for diagnosing mediastinal and hilar lymphadenopathy, including benign and malignant diseases [25]. A large number of prospective trials have demonstrated that EBUS-TBNA, as the gold standard, is similar to surgical mediastinoscopy in sensitivity and pathological staging [26‐28]. EBUS-TBNA had a sensitivity of 89% for lung cancer and was recommended by the guidelines to stage the mediastinum [29]. Likewise, for peripheral pulmonary nodules, ultrasonography (US) also plays a critical role. US-guided transthoracic fine-needle aspiration demonstrated elevated clinical applicability and diagnostic utility compared to CT-guided transthoracic fine-needle aspiration [30]. Moreover, the materials obtained by needle aspiration can be further evaluated for molecular pathological diagnosis through various approaches, such as next generation sequencing, multiplex fluorescence in situ hybridization and so on [31‐33]. However, there have been few studies on TBNA for patients with benign lymphadenopathy. TBNA has been shown to be useful for diagnosing benign lymphadenopathy, with diagnostic rates of 51.6–96% [12‐19, 23, 34‐36], but these studies had obvious limitations. First, most used a single-arm design, reporting only the diagnostic rate of EBUS-TBNA for benign disease, without any comparison to cTBNA [13, 14, 17‐19, 23, 34‐36]. Second, the sample size of some of the studies was small; further studies with larger samples are needed size to confirm the diagnostic accuracy of EBUS-TBNA [13, 15, 16, 18, 19, 23, 34, 35]. Third, most studies enrolled both malignant and benign lymphadenectasis patients, rather than specifically focusing on benign lymphadenopathy [13, 17, 19, 23, 36]. In addition, studies performed in western countries were mainly concerned with sarcoidosis, whereas those done in Asian countries focused on tuberculosis [12, 14‐16, 18, 34, 35]. In our study, we tried to address the shortcomings of the previous investigations. We enrolled patients with various kinds of benign disease, and also performed a subgroup analysis to determine the indications for cTBNA in individuals with suspected benign LN adenopathy. In our study, the diagnostic yield was 73.37% for TBNA; more specifically, it was 61.31% for cTBNA and 76.84% for EBUS-TBNA. Although EBUS-TBNA was superior, cTBNA also showed good diagnostic power, and was not for LNs > 20 mm. In the univariate and multivariate logistic regression analyses, only granulomatous inflammation was significantly associated with the diagnostic efficiency of TBNA. Based on both previous studies and our findings, EBUS-TBNA should be considered as the first-line modality for diagnosing enlargement of benign mediastinal LNs. cTBNA remains a valuable technique with acceptable (> 60%) sensitivity, particularly for larger LN lesions.

It is worth noting that EBUS-TBNA requires specialized, relatively expensive equipment, and thus is not available in many hospitals, particularly those in developing countries. Hence, it is important to explore ways to achieve higher diagnostic efficiency without EBUS. We aimed to devise a safe, efficient, feasible and easy-to-perform method to meet this need. In this study, we confirmed the diagnostic value of cTBNA; the poorer performance of cTBNA in the granulomatous cases was mainly due to the small volume of specimens available to support the histological findings. Forceps biopsy is a classic clinical approach used to obtain larger-volume specimens than needle biopsies. Several pilot studies reported a diagnostic yield of EBUS-guided mediastinal LN forceps biopsy above 80% for nonmalignant disease [20‐23]. However, in our study the diagnostic yield of TBFB without EBUS for benign disease was only 53.33% (16/30). No significant difference in yield was noted between cTBNA and TBFB. The low diagnostic rate of TBFB could have occurred for several reasons. First, the tip of the biopsy forceps is too blunt to effectively penetrate the targeted LNs; thus, a positive TBNA result but negative TBFB (fibrofatty tissue) may be seen. Second, whether the forceps cups opened successfully following forceps insertion into the LNs could not be confirmed. TBFB requires a new “forceps-needle” integrated tool to overcome its shortcomings. Of note, we found that cTBNA and TBFB combined had a high diagnostic rate, even without EBUS. The two modalities are complimentary and their combined use does not result in additional damage or cost based on currently available tools. Further validation of the efficiency and safety of this protocol is needed.

In addition to sampling techniques, preparation technique is also a pronounced determinant for diagnostic performance. Accumulated evidence shows that conventional cytological evaluation is capable to provide diagnostic clues in non-granulomatous diseases, such as amyloidosis, necrotizing lymphadenitis [37, 38]. On the other hand, cytology is a more potent diagnostic tool for granulomatous disease. However, the efficacy varied, ranging from ~ 10 to 80% [6, 39]. Notably, our cytological efficacy is close to the lower range of reported results. It could be partially explained by our protocol that we pick out all tissue fragment for histology evaluation which deprive off much indicating information for cytologists. Additionally, it is noteworthy to highlight that cell block is an essential complement to conventional smear in providing a reliable diagnosis especially in resource limited settings, presenting dramatic advantages in histomorphology analysis.

Our study had several limitations. First, used a retrospective design and various confounders may have affected the results. Second, we performed EBUS-TBNA with a 21-gauge TBNA needle only, rather than a 19-gauge histology needle, so could not address the potential effect of needle size on sensitivity. Third, the number of patients with benign lymphadenopathy enrolled to determine the efficacy of TBFB was relatively small. The type of forceps used for TBFB is also worthy of more in-depth investigation. Several factors, such as stiffness, size of the forceps cup, design of the forceps tip, and even the putative forceps-needle integrated design, should be considered in future studies.

Both cTBNA and EBUS-TBNA are feasible and safe, and can provide histological specimens of enlarged mediastinal or hilar LNs. Due to its high sensitivity, EBUS-TBNA is the preferred option for diagnosing enlargement of benign LNs. Combined cTBNA and TBFB is a viable alternative when EBUS is not available, due to its relatively high diagnostic rate. We strongly advocate the use of minimally invasive techniques for diagnosing of enlargement of benign mediastinal LNs.