The evolving role of interventional pulmonology with endobronchial ultrasonography in the diagnosis of lymphoma: a narrative review.

Introduction

Background
Undifferentiated lymphadenopathy is a common problem faced by interventional pulmonologists with a wide range of diagnosis including sarcoidosis, carcinoma and lymphomas. The most common primary mediastinal lymphomas are primary mediastinal B-cell lymphoma, Hodgkin lymphoma (HL), and T-lymphoblastic lymphoma, and these disease types often present as isolated mediastinal adenopathy or masses. Primary mediastinal B-cell lymphoma, a subtype of diffuse large B-cell lymphoma, arises from thymic medullary B cells and accounts for 2–4% of all non-HL (1,2). It typically presents as a rapidly enlarging mediastinal mass with local invasion or compression of airways and vessels, often followed by extra thoracic spread. HL involving the mediastinum as a primary site is less common and has distinct histological and clinical features. T-lymphoblastic lymphoma, the most common T-cell lymphoma presenting in the mediastinum, usually affects adolescents and young adults (2). Additionally, although not malignant, hyaline-vascular Castleman disease frequently involves the mediastinum and can present with lymphadenopathy (3). Given the possibility of lymphoma, mediastinal adenopathy is typically evaluated by an interventional pulmonologist for consideration for biopsy to render a diagnosis.
Interventional pulmonology has grown significantly as a field in the last two decades, offering diagnostic procedures using minimally invasive techniques. In this field, several new tools to assist with the diagnosis and management of thoracic diseases have been developed recently (4). Among them, endobronchial ultrasonography (EBUS) has become a foundational technique, used for accessing and sampling mediastinal and hilar lymph nodes. By integrating ultrasonography with traditional bronchoscopy, we can visualize and sample peri-bronchial structures and lymph nodes around the mediastinal and hilar regions using techniques such as transbronchial needle aspiration (TBNA) (5). With real-time ultrasound, a needle can be inserted into the desired location and samples can be obtained in a safe and efficient manner (6). Since the first clinical experience of using EBUS-TBNA for lymph node sampling, described by Krasnik et al. in 2003, the technique has been well established in lung cancer diagnosis and staging, largely replacing mediastinoscopy (7,8).
Herth et al. [2006], ASTER trial [2010] and Yasufuku et al. [2011] established EBUS as a key tool in mediastinal staging for lung cancer, showing that EBUS offers similar or superior sensitivity compared with mediastinoscopy, with fewer complications (9-11). These findings led to American College of Chest Physicians (ACCP) guidelines recommending minimally invasive needle-based techniques over surgical approaches in patients with radiographically normal mediastinum but central tumors or N1 disease (12).

Rationale and knowledge gap
Although EBUS has an established role in lung cancer staging and diagnosis, its use is less clearly defined in the diagnosis of lymphoma. Excisional lymph node biopsy remains the gold standard to diagnose lymphoma since it preserves the architecture of the lymph node (2). With needle-based techniques such as EBUS-TBNA, the underlying architecture of the lymph node is inadequate in the sample, limiting the utility of EBUS-TBNA to diagnose and subtype lymphomas. No current guideline has clearly defined when EBUSTBNA alone is sufficient for lymphoma diagnosis versus when escalation to core or excisional biopsy is required. British Thoracic Society (BTS) guidelines state that there is insufficient evidence to recommend the routine use of EBUS-TBNA for the diagnosis of lymphoma (13). However, a consensus-based statement from the ACCP recommends that EBUS-TBNA may be considered as an initial, minimally invasive diagnostic test for patients with suspicion for lymphoma (14).
In recent years, newer techniques, such as EBUS-forceps biopsy and EBUS-cryobiopsy, have been suggested to obtain larger tissue specimens and possibly increase the sensitivity of EBUS for diagnosing lymphomas. There are several prospective studies on the use of these newer techniques for the diagnosis of lymphadenopathy; however, they include cohorts that encompass a mixed population, including sarcoidosis, benign diseases, lymphoma, carcinoma, etc. The data specifically on patients diagnosed with lymphoma by these techniques is largely limited to small, single-center retrospective studies, which makes it difficult to determine if these techniques are more useful than EBUS-TBNA alone in the diagnosis of lymphomas.

Objective
In this review, we discuss the current evidence for these various techniques in the diagnosis of lymphoma. We present this article in accordance with the Narrative Review reporting checklist (available at https://med.amegroups.com/article/view/10.21037/med-2025-1-51/rc).

Methods
We conducted a narrative review by searching PubMed/MEDLINE, Embase and Google Scholar for articles published during 1980–2025 using search terms including “EBUS”, “EBUS-TBNA”, “endobronchial ultrasound”, “bronchoscopy”, “lymphoma diagnosis”, “mediastinal lymphadenopathy” and “interventional pulmonology” (Table 1). The review was not designed as a systematic review and therefore did not follow a predefined protocol or PRISMA methodology. The study selection was intended to be comprehensive but not exhaustive. Articles written in English that focused on the use of EBUS for the diagnosis, classification or evaluation of lymphoma and guideline documents by various societies were considered for inclusion. Eligible studies included articles on adult patients and reported outcomes such as diagnostic yield, sensitivity, adequacy of specimens, comparison with surgical techniques etc. Studies were excluded if they focused on a pediatric population or malignancies other than lymphomas. Types of studies included were observational, comparative, meta-analyses and systematic reviews and select review articles. Case reports, editorial pieces without primary data and abstract only publications were generally excluded. Relevant articles were manually reviewed by the authors, and the reference list was also checked to add additional studies not captured in the initial screening process. Disagreements regarding study relevance were resolved via discussion and consensus among the authors. No formal arbitration was needed. Data extraction was performed qualitatively by the authors. Since this was a narrative review, no formal risk of bias assessment was done.

Findings
In isolated mediastinal/hilar lymphadenopathy, biopsies have been traditionally obtained via surgical procedures such as mediastinoscopy, thoracotomy, or video-assisted thoracoscopic surgery, which allow for excisional biopsies of these lymph nodes. Although excisional biopsy of these group of lymph nodes is considered the gold standard for the diagnosis of lymphoma, the technique is much more invasive than EBUS-TBNA and has a higher risk of complications and higher cost. In addition, surgical procedures often cannot sample the mediastinum and bilateral hila at the same time (15). In comparison, EBUS is much less invasive, is less painful, and may be performed under either moderate sedation or general anesthesia (13). However, biopsying station 5 or 6 lymph nodes with the standard linear EBUS probe is technically challenging and often only accessible via surgical techniques (16).
While excisional biopsies are highly reliable for diagnosing and subtyping lymphomas, EBUSTBNA shows fair to good detection sensitivity for lymphoma but is limited in reliably subtyping it, which is important for guiding therapy. In addition, some lymphoma types are more likely than others to be reliably diagnosed or subtyped via EBUSTBNA. The lower sensitivity of EBUS-TBNA compared with excisional biopsy may be explained by the fact that EBUS-TBNA aspirates are usually small volume and disperse cells into the collection medium, which often fails to show the cell morphology, background architecture, and immunophenotype, which is of particular importance in lymphomas (14).

Diagnostic sensitivity of EBUS-TBNA for lymphoma
Several studies have investigated the utility of EBUS-TBNA for patients with a concern for lymphoma, but most of the studies were small, single-center, and retrospective. In a retrospective study by Grosu et al. involving 181 patients with suspected lymphoma, EBUS-TBNA established the diagnosis in 63 of the 75 (84%) patients ultimately diagnosed with lymphoma. In 7 patients, EBUS-TBNA was used to diagnose lymphoma, but the disease could not be adequately subtyped and required confirmatory biopsy. The diagnosis and subtype rate for lymphoma was lower in the de novo group (67%) compared with the relapsed group (81%), and EBUS-TBNA had the best sensitivity for low-grade non-HL (92%) (17).
A study by Moonim et al. collected data from 2,256 patients undergoing EBUS at a tertiary care center in England. EBUS-TBNA helped to establish a definite diagnosis of lymphoma in 59 (89%) of 66 cases ultimately diagnosed with lymphoma, of which 51 were de novo and 15 were relapsed lymphoma. The sensitivity for relapsed lymphoma was 100%. Interestingly, EBUS-TBNA was used to diagnose high-grade non-HL in 10 of 12 cases and HL in 19 of 24 cases. Although these results are encouraging, they involved the use of rapid onsite evaluation (ROSE), immunohistochemistry, and other advanced diagnostic techniques along with a multi-disciplinary team to establish a diagnosis, which may not be possible in resource-limited settings (18).
In a metanalysis involving 14 studies and 425 patients, EBUS-TBNA had an overall sensitivity of 66.2% and a specificity of 98.2% for the diagnosis of lymphoma. The sensitivity was higher for recurrent compared with de novo disease (77.8% vs. 67.1%). However, only 63% of positive samples were of sufficient quality to undergo subtype testing. The authors also found that rapid onsite examination, volume of the sample, and flow cytometry increased the sensitivity of EBUS-TBNA in detecting recurrent disease (19). Flow cytometry adds substantial value to the diagnosis of lymphoma. The American Society for Clinical Pathology and the College of American Pathologists strongly recommend that immunophenotyping by flow and/or immunohistochemistry with morphology be used for diagnosis and subclassification of lymphomas (20).
Altogether, while EBUS-TBNA was often reasonably sensitive in the diagnosis of lymphoma, these results varied from 32.5% to as high as 89.2% (19). The discordance in some of these results may be explained by factors such as operator and pathologist experience, size of needle, lymph node site, number of aspirates, type of lymphoma, and de novo versus recurrent disease, all of which make it challenging to study the true utility of EBUS-TBNA in the diagnosis of lymphoma (21).

Limits of EBUS-TBNA in subtyping lymphoma
Lymphoma subtype is known to affect the ease of diagnosis when sample size is limited, as when EBUS-TBNA is used. In a study by Farmer et al., chronic lymphocytic leukemia, mantle cell lymphoma, and T-lymphoblastic lymphoma were diagnosed with high sensitivity using small-volume samples obtained via needle-based techniques, whereas follicular lymphoma and marginal zone B-cell lymphoma were difficult to diagnose using small-volume specimens. The authors argued that although the former group has small cells that look morphologically similar, they have a characteristic immunophenotypic profile, which is particularly important in the diagnosis of lymphoma. Conversely, the latter group lacks a distinct immunophenotypic profile, explaining the difficulty in diagnosing these subtypes using small-volume samples (22).
Similarly, HL is difficult to diagnose using EBUS because the characteristic Reed-Sternberg cells may be scarce, the lymph node may be sclerotic, and mimickers of Reed-Sternberg cells and HL may be present. In addition, flow cytometry is generally less helpful in HL due to its lack of a clonal population (23-26).

Needle characteristics and diagnostic yield of EBUS-TBNA
It is unclear whether the needle size used to perform EBUS-TBNA affects diagnostic yield. Although the volume of sample obtained using a 19-gauge needle is greater than that obtained using a 21-gauge needle, whether this difference translates into better sensitivity or diagnostic yield is uncertain, and further studies are needed to evaluate this question (27,28).
In the metanalysis by Labarca et al., no statistically significant difference was found between needle size subgroups in patients with new and de novo lymphoma. However, for recurrent lymphoma, the use of a 21-gauge needle was associated with lower sensitivity compared with a 22-gauge needle (63.8% vs. 82.3%) (19). This difference may be due to better penetration of the 22-gauge needle into the lymph node and less contamination with blood (29). The results were similar to those of Chaddha et al., who found that specimens obtained via 19-guage needles were significantly less adequate and bloodier; however, this study was not specific to patients with lymphomas, and the results may not be generalizable to this specific patient population (30).
More recently, there has also been interest in whether differently designed needles may help obtain core biopsies more consistently and achieve better diagnostic yield in diagnosing lymphomas. Some of these needles have been used commonly by gastroenterologists and radiologists but have only recently been used via EBUS. In a single-center study of 100 patients, Franseen needles, which have three beveled edges, were used to biopsy 70 locations. Franseen needles can be rotated to cut in a cylindrical manner which may provide core tissue biopsies. The authors reported that “core biopsy” specimens were obtained in 87% of the patients. Since only 2 of these patients were diagnosed with lymphoproliferative disease, it is difficult to determine if this needle will help in improving diagnostic yield for lymphomas (31). Similarly, in another retrospective study, Franseen needles showed higher diagnostic yield in benign lymphadenopathy, due to greater identification of granulomatous inflammation, and more sample adequacy and tumor volume in patients with non-small cell lung cancer (32). Head-to-head comparison is needed with more commonly used EBUS-TBNA needles to determine the effects of needle design.

EBUS with forceps biopsy
EBUS with forceps biopsy is performed using a standard EBUS scope along with a forceps. Commonly, the 1-mm mini-forceps are used. Typically, a small puncture is created in the airway using the TBNA needle. Using ultrasound guidance, the forceps is advanced into the lymph node and a biopsy is taken (33).
In a retrospective study involving 213 patients with lymphadenopathy, of which 10 were diagnosed with lymphoma, Ray et al. showed that EBUS specimens obtained via forceps were superior in quality compared with those obtained by EBUS-TBNA. A total of 44 lymph nodes were sampled for 16 patients ultimately diagnosed with lymphoma. The specimens were graded using a validated scoring system involving amount of cellular material, background blood, retention of normal architecture, and degree of cellular degradation and trauma. Significantly higher quality specimens were obtained with forceps biopsy compared with TBNA (mean score of 7 vs. 5.4). This increased quality resulted in an improved diagnostic yield with forceps biopsy compared with TBNA (93.8% vs. 62.5%, P=0.042) (34).
Additionally, this technique is safe and efficient. In one study, 74 lymph nodes were biopsied with both EBUS-TBNA and EBUS-forceps biopsy. The authors reported only minor bleeding as a complication, which occurred in 2.9% of patients (35). Furthermore, EBUS-forceps adds a mean of less than 4 minutes to the total procedure time compared with TBNA alone (36).
One study suggested that the benefit of EBUS-forceps biopsy is greatest when used in combination with EBUS-TBNA; the increase in diagnostic yield associated with the combination was most impressive in lymphoma cases (81% vs. 35%, P=0.038) (37). The authors argued that in patients with suspected lymphoma, the addition of forceps biopsy to TBNA resulted in an improved diagnostic yield and the procedure was safe (37).
A meta-analysis compared the use of EBUS-TBNA combined with EBUS-forceps biopsy compared with EBUS alone for patients with intrathoracic adenopathy. EBUS-TBNA had a yield of 30% for lymphoma, whereas the combined approach had a yield of 86% (38). The reported complications were higher for the combined approach than for EBUS-TBNA alone and included pneumothorax (1%), bleeding (0.8%), and pneumomediastinum (1%). However, the authors reported several potential biases in the studies they analyzed, including selection bias, variable number of passes, variable use of ROSE, variable needle and forceps type and size, which may have affected the results (38).

EBUS with cryobiopsy
EBUS with cryobiopsy involves inserting a small cryoprobe via EBUS into the lymph node and using compressed gases to rapidly freeze, crystallize, and recover larger pieces of tissue than typically obtained via EBUS-TBNA (39).
Cheng et al. showed much larger pieces of tissue obtained via EBUS-cryobiopsy compared with EBUS-forceps biopsy (8.1 vs. 2.1 mm3) (33). The authors reported successful lymphoma diagnosis in 11 of 15 patients using EBUS-forceps biopsy and in 13 of 15 patients using EBUS cryobiopsy (33). Typically, a 1.1-mm cryoprobe is used, but owing to the paucity of literature, no standard duration of freezing has been reported. This may result in heterogeneity among studies of EBUS-cryobiopsy. It has been reported that longer freezing duration results in a larger volume of cells and volume of sample obtained (40). Whether this translates into a better diagnostic yield remains to be seen.
In a retrospective, multicenter study of 40 patients with lymphoma who underwent both EBUS-TBNA and EBUS-cryobiopsy in the same procedure, EBUS-cryobiopsy demonstrated a high sensitivity (>95%) for the diagnosis of lymphoma, whereas EBUS-TBNA had a sensitivity of only 15% (41). However, the study had several confounding factors such as variable use of flow cytometry and ROSE and a variable needle size for TBNA. This study reported no complications from the use of EBUS-cryobiopsy (41). Given the confounders however, the results from this study should be interpreted cautiously.
In a systematic review comprising 555 patients among 7 studies, EBUS-cryobiopsy was especially useful in diagnosing lymphoma compared with EBUS-TBNA. The authors found that EBUS-cryobiopsy was diagnostic in 87% of cases, in contrast to 12% for EBUS-TBNA. Furthermore, EBUS-cryobiopsy allowed characterization of lymphoma subtype, genetic studies, and immunohistochemical determination. The only complication reported was light bleeding (39). The authors concluded that in benign etiologies and lymphomas, EBUS-cryobiopsy was much more effective to determine diagnosis and was safe. Overall, the data is promising but limited and cryobiopsy is not yet standard of care.

Cost effectiveness of EBUS-TBNA in diagnosis of lymphoma
EBUS-TBNA is generally more cost effective in diagnosing mediastinal/hilar lymphadenopathy as compared to upfront surgical biopsy since it is often done in an outpatient setting, may avoid operating room costs and can be done under moderate sedation (42,43). In patients with lymphoma however the cost savings may be attenuated due to the lower sensitivity of EBUS-TBNA in diagnoses of lymphoma which may require repeat procedures and/or surgical biopsy. This is an important consideration in low resource settings where ROSE and flow cytometry may not be available which may affect the diagnostic yield of EBUS-TBNA.

Limitations and implications
As stated, the published data on the use of EBUS and other techniques mentioned in this review to diagnose lymphoma are largely limited to small, single-center, retrospective studies. These studies also include several confounding factors such as the use of ROSE, experience of the proceduralist, needle type and size, ancillary techniques and testing, lymphoma subtype, number of passes, and number of lymph nodes sampled, which make it difficult to determine the true sensitivity of EBUS in diagnosing and subtyping lymphomas. Furthermore, sensitivity is a characteristic of cytology/pathology and influenced by the test characteristics and the selected criteria for positivity. These limitations have important implications for interpreting studies and determining the usefulness of EBUS in diagnosing lymphomas.
When reviewing literature, the definition of diagnostic yield is variable. While some authors define diagnostic yield as a positive diagnostic test for lymphoma, others consider only patients in whom EBUS techniques were adequate to establish both the diagnosis and subtype of lymphoma. The context and setting while defining diagnostic yield are particularly important, as diagnostic yield is dependent on the pretest probability of disease. The importance here is that diagnostic yield is highly dependent on the prevalence of every single disease in that population that could show up in the numerator and not just lymphoma. Settings are diverse in the numerous studies in literature and that’s why the diagnostic yield for lymphoma will differ from one setting to another.
As an example, if a certain cohort has a higher percentage of a specific type of lymphoma, e.g., HL, which is more difficult to diagnose via EBUS than other lymphomas, then the diagnostic yield of EBUS to diagnose lymphoma in that population will be lower than compared to other cohorts where the prevalence of HL may be lower. If only diagnostic yield is reported and not sensitivity or specificity for lymphoma, we cannot generalize the results, as we did not know if that same test result would apply to a different setting, in which the population may be different.
In some publications, likelihood ratios are reported. When evaluating diagnostic tests like EBUS-TBNA, it is important to go beyond sensitivity and specificity, especially when more than two outcomes exist, as in cases of isolated mediastinal lymphadenopathy. In such scenarios, likelihood ratios provide a more nuanced interpretation by accounting for multiple result categories (e.g., lymphoma, granuloma, nondiagnostic). Applying likelihood ratios in the context of pretest probability allows clinicians to better estimate post-test disease likelihood and make more informed decisions. Pretest probability in mediastinal lymphadenopathy is also influenced by radiologic parameters. Recent clinical data show that Hounsfield Unit values alone have limited ability to discriminate malignant from benign nodes, although maximum standardized uptake value (SUVmax) retains moderate predictive value (44).

Conclusions
EBUS is a vital tool in staging lung cancer, but larger and prospective studies are needed to clarify its utility in the diagnosis of lymphoma. In the interim, EBUS may be a reasonable first step to diagnose lymphoma in patients with mediastinal or hilar lymphadenopathy, as the technique is less invasive with a lower complication rate compared with surgical excision. Further advances in EBUS techniques and histological examination may further improve the sensitivity of EBUS for lymphoma and establish best practices in using EBUS to diagnose and adequately subtype lymphomas.

Supplementary
The article’s supplementary files as