Assessment of mediastinal lymph nodes and masses is important for both diagnostic purposes and lung cancer staging. The endobronchial application of ultrasound was first described in the English literature in 1992,1 and endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) has been commercially available since 2005.2 The initial literature suggested that diagnostic yields from EBUS-TBNA offered diagnostic accuracy of 89% to 97%.3–5
In conventional TBNA without ultrasonic guidance, the type of standard needle can influence results and larger gauge needles can obtain satisfactory sample volumes with higher diagnostic yield.6–8 To obtain cytology specimens, 20-gauge to 22-gauge needles are usually used, whereas 19-gauge needles are needed to obtain a “core” of tissue for histology.7,9
At the time of the initial completion of this study, only 22-gauge EBUS-TBNA needles are marketed because of the structural limitations of bronchoscopes. However, for this study, we had the opportunity to use 21-gauge EBUS-TBNA needles, and evaluated the safety and efficacy of obtaining histologic specimens and comparing diagnostic yields with conventional 22-gauge needles.
PATIENTS AND METHODS
Between July 2006 and September 2007, all patients with indications for EBUS-TBNA were included in this study. The purpose of EBUS-TBNA for mediastinal lesions was the evaluation of suspected mediastinal or hilar metastases in patients with lung cancer or the evaluation of unknown mediastinal or hilar lesions. Target lesions for puncture were selected from lymph nodes with a short axis of ≥10 mm on chest computed tomography or were identified as positive on positron emission tomography with 18F-fluorodeoxyglucose (FDG-PET). A board of thoracic surgeons, pulmonologists, and radiologists decided whether lymphadenopathy was suspected to represent a malignant or benign lesion based on clinical images and findings in regularly held clinical conferences. Of a total of 56 patients examined, 24 patients were enrolled into the 21-gauge needle group (21 G) and 32 patients were enrolled into the 22-gauge needle group (22 G). The ethics committees of the St Marianna University School of Medicine approved all study protocols, and written informed consent was obtained from all patients in this study.
Figures 1A, B illustrate the experimental protocol. Figure 1A illustrates the flow chart of EBUS-TBNA with onsite stereoscopic microscopy for aspirated specimens. Figures 1A b illustrates 2 types of the needle tips used in this study; the right is a 21-gauge needle and the left is a 22-gauge needle. All examinations were performed on patients under conscious sedation by 1 bronchoscopist. The bronchoscopist had previous experience in performing approximately 300 procedures of EBUS-TBNA. Local anesthesia was achieved with 4% lidocaine solution (5 mL) sprayed in the pharyngeal area combined with a bolus dose of 2 mL of 2% lidocaine applied to the trachea. EBUS-TBNA was performed using a linear curved-array ultrasonic bronchoscope (UC260F; Olympus, Tokyo, Japan) (Fig. 1B a). This instrument has a 30-degree oblique forward-viewing optical lens, an outer diameter of 6.9 mm, and an instrument channel of 2.0-mm width. Images were scanned at 7.5 MHz and then processed using an EU-C2000 ultrasound processor (Olympus). Needle aspiration was performed by a 22-gauge needle (NA-201SX-4022; Olympus) or a dedicated prototype of the 21-gauge needle.
Using EBUS, a lymph node survey was performed in all patients at the beginning of the procedure to identify the largest and most accessible lymph node station (Fig. 1B b). Power Doppler mode was used to exclude intervening vessels immediately before needle puncture. Under real-time ultrasonic guidance, the needle was placed within the lesion. A stylet was used to prevent plugging of the bronchial wall and was moved back and forth with the needle to remove any bronchial mucosal plug. After withdrawal of the stylet, suction was applied using a syringe, and the needle was moved back and forth inside the lesion obtaining the longest needle strokes possible. After withdrawal of the TBNA needle, the stylet was inserted into the TBNA needle to push out the aspirated specimen on the filter paper for histology (Fig. 1B c). The stylet was then removed and residual aspirates were blown onto glass slides using air through a syringe, and then fixed in 95% ethanol for cytology. An observer who was not aware of images of the chest computed tomography and EBUS, immediately evaluated the specimen on paper under stereoscopic microscopy (SZX12; Olympus) (Fig. 1B d), using a magnifying power ranging from 7 to 90×, to confirm whether the specimen was adequate for diagnosis. This information was then transmitted to the bronchoscopist. The criteria to judge a specimen as adequate for diagnosis were decided after discussions or correlations between pathologies and stereoscopic microscopic features before the beginning of this study as follows. A positive result was obtained when we could detect whitish components on the surface of the sample, sometimes containing several black spots of anthoracosis; the sample seemed adequate for diagnosis (Figs. 1B e, f, i, j, m, n). If we could not find whitish components anywhere on the paper, which tended to include a higher ratio of blood to tissue, the sample was categorized as negative. If adequate tissue was identified under stereoscopic microscopy, additional punctures were terminated. Specimens blown onto the glass slide were used for cytological examination after Papanicolaou staining. Specimens on the filter paper were used for pathohistologic examination after hematoxylin and eosin (H&E) staining (Figs. 1B g, h, k, l, o, p). We measured the width of the histologic “core,” which must be described as the inner lumen diameter of the needle used herein, from microscopic images. Figure 2 shows representative histologic views of the section using the H&E stain, which contains measurable tissue in the width. Top panels (Figs. 2A, B) are findings of specimens obtained by the 22-gauge needle and lower panels (Figs. 2C, D) by the 21-gauge needle.
In patients with malignant lymph nodes, determination was based on malignant cytologic or histologic results of EBUS-TBNA. In patients with benign lymphadenopathy, determination was based on surgical pathologic confirmation, or lack of disease progression according to clinical and radiologic findings after ≥6 months as an end point. If clinical and radiologic features consistent with sarcoidosis were observed, then the diagnosis was made by EBUS-TBNA on the basis of pathologic evidence of non-necrotizing epithelioid cell granulomas, with negative stain for acid-fast bacilli and fungal organisms.
Data are expressed as mean±standard deviation. Statistical differences in lymph node diameter, maximal standard uptake value of target lesion from FDG-PET [maximum standardized uptake value (SUVmax)], and width of core sample were analyzed using Student t test. Statistical differences in diagnostic yields and inadequate material rates were analyzed using Fisher exact probability test. Other statistical differences were determined using the 2-sided Mann-Whitney U test. Values of P<0.05 were considered statistically significant. Statistical analyses were performed using Statcel version 2 software (OMS Publishing, Tokorozawa, Saitama, Japan). Sensitivity, specificity, and diagnostic accuracy rates were calculated using the standard definitions. Patients with inconclusive EBUS-TBNA cytologic and/or histologic results due to inadequate tissue volume were included in this analysis, as we regarded the efficacy of sampling as reliability itself for the diagnostic algorithm through EBUS-TBNA.
Patient characteristics are noted in Table 1. Twenty-four patients in 21 G consisted of 17 men and 7 women with a mean age of 65.0 years (range, 38 to 84 y). Thirty-two patients in 22 G consisted of 20 men and 12 women with a mean age of 65.6 years (range, 25 to 82 y).
The final diagnosis was available for all patients enrolled in this study. Malignancy was detected in 18 patients (75%) in 21 G and 23 patients (71.9%) in 22 G. No significant difference in proportion of malignancy versus benign disease was identified. A specific diagnosis was made in all 21 G patients except 1, who revealed no malignancy after a follow-up of 10 months after EBUS-TBNA. However, 5 patients (15.6%) in 22 G remained undiagnosed and 2 of these underwent radical surgery with mediastinal lymph node dissection that revealed malignancy in mediastinal lymph nodes. The remaining 3 patients were diagnosed on follow-up by tissue sampling at other sites or by long-term stable disease on radiology, despite absence of therapy. One of these cases was regarded as malignant lymphadenopathy from the emergence of malignant pleural effusion, 1 as sarcoidosis based on other clinical criteria, and 1 was regarded as reactive lymphadenopathy of unknown cause.
FDG-PET was performed for 25 patients (21 G, n=12; 22 G, n=13), and SUVmax of the target lymph node was measured. No significant difference was seen between 21 G and 22 G in the size of lymph nodes, level of SUVmax, or number of needle passes (Table 2). Despite using different sizes of needles, no difference in the width of tissue stained with H&E was seen, with 463±71 μm in 21 G and 460±50 μm in 22 G. Histologic specimen using a 22-gauge needle shows noncaseating granulomas diagnosed as sarcoidosis, which indicates that the granuloma may be scraped by the needle (Figs. 2A, B). Histologic specimen using a 21-gauge needle demonstrates similar features of a granuloma indicating the granuloma may be aspirated entirely into the needle (Figs. 2C, D).
We did not encounter any major complications such as significant bleeding, massive hematoma, pneumothorax, or pneumomediastinum with EBUS-TBNA using either the 21-gauge or 22-gauge needle. Self-limited bleeding from the puncture site was observed in some patients, but no differences were noted between groups.
Comparison of EBUS-TBNA Cytologic Results With Final Diagnosis
Inadequate materials were only obtained from 1 of 32 patients in 22 G (Table 3). False-negative results were found in 2 patients in 21 G and in 10 patients in 22 G.
Comparison of EBUS-TBNA Histologic Results With Final Diagnosis
Inadequate materials were found in 1 of 24 patients in 21 G and in 6 of 32 patients in 22 G. The major reason for insufficiency was based on clot and cartilage (Table 4). No false-negative or false-positive results were identified in 21 G.
Comparison of Diagnostic Yields for Suspected Malignancy
Cytologic sensitivities of 21 and 22 G were 88.9% and 52.2%, respectively (P=0.01). Histologic sensitivities of 21 and 22 G were 100% and 82.6%, respectively (P=0.09). The sensitivity in 21 G of combining both cytology and histology samples (100%) was higher than that for 22 G (87.0%), although the difference was not significant (Table 5).
Comparison of Diagnostic Yields
The cytologic diagnostic value of EBUS-TBNA to obtain a diagnostic lymph node sample has an accuracy of 91.7% in 21 G, which was significantly higher than an accuracy of 65.6% in 22 G (P=0.02). The histologic diagnostic value of EBUS-TBNA had an accuracy of 95.8% in 21 G and 81.3% in 22 G (P=0.11). The accuracy in 21 G combining both cytology and histology samples was 100%, which was significantly higher than the 84.4% in 22 G (P=0.04). The ratio of inadequate material with 22 G was higher than that with 21 G for both cytology and histology, but no significant differences were apparent between groups (Table 6).
This is the first comparison study on a per patient basis, to the best of our knowledge, of 21-gauge and 22-gauge needles for EBUS-TBNA in terms of safety and in terms of efficacy. Nakajima et al10 recently reported a comparison study on a per lymph node basis, revealing that the diagnostic yield was the same between both size needles. The supply of 21-gauge needles were not on the market at the time this study began, making it difficult to randomize the total patient population in this study.
In our cohort, we demonstrated that the diagnostic value of EBUS-TBNA combining both cytologic and histologic results had significantly higher accuracy with a 21-gauge needle than with a 22-gauge needle. Although diagnostic yields of accuracy in EBUS-TBNA using conventional 22-gauge needles for mediastinal lymph node have reportedly reached 89% to 97%,3,5,11 when every selected patient was examined for lung cancer or suspected lung cancer, our results for accuracy using a 22-gauge needle remained unexpectedly low, at 84.4%. We consider this discrepancy in diagnostic yield as attributable to population bias and to the absence of rapid onsite cytopathologic evaluation (ROSE) in our study, in addition to the small number of needle passes (1.7 to 1.8) per patient compared with other studies.10,12 Holty et al13 reported that sensitivity depends heavily on the population under investigation, and the higher the prevalence of mediastinal metastasis in a population, the higher the sensitivity. ROSE of the specimen for sample adequacy has been shown to increase diagnostic yield.14 To date, our institution has unfortunately been unable to prepare ROSE for samples obtained. We did perform onsite stereoscopic microscopy in an attempt to confirm the presence of tissue in samples during bronchoscopy. Onsite stereoscopic microscopy increases diagnostic yield and decreases needle passes (data not shown).
Traditional TBNA using transbronchial histology (19 G) and cytology (21 G) needles has demonstrated that diagnostic yield is lower for benign conditions (37%) than for malignant lesions (69%).15 However, Wong et al16 recently achieved high diagnostic results for sarcoidosis using EBUS-TBNA by a 22-gauge needle (91.3%). Our results for the 22-gauge needle showed a 77.8% diagnostic rate, even in benign disease (7 of 9 patients), whereas the 21-gauge needle reached 100% (6 of 6 patients). As it is unlikely that these were due to ultrasound guidance alone, they may demonstrate developments in needle manufacturing. The inner diameter of a 22-gauge needle is nearly equal to that of a conventional 21-gauge needle and allows the sampling of histologic cores in some cases.16 In addition, a newly developed 21-gauge needle will obtain more sample for precise diagnosis.
Inadequate material shows either few or no representative cells, or cellular material that is damaged, necrotic, blood-stained, or poorly fixed, thus making interpretation difficult.17 Kurimoto et al18 reported the following intranodal features for puncture: (1) metastatic mediastinal and hilar lymph nodes having a short axis >5 mm, harvested from patients with lung cancer, revealed that metastasis was located on the marginal area in 22% to 24% of lymph nodes and (2) histologic examination of all the lymph nodes with necrosis showed that necrotic cells were located far from the blood vessels, whereas viable tumor cells were located adjacent to the vessels. Therefore, we estimated the internal component of the lymph node before puncture using both B mode and power Doppler mode scanning, which might reveal viable areas and intranodal vessel sites, for the purposes of sampling representative cells and avoiding puncture of vessels. Clearing intraluminal residue on the needle by stylet just before suction was a matter of course. Despite these efforts to obtain appropriate samples, the majority of inadequate materials in this study were cartilage or massive blood contamination. We could not completely resolve until now the fact that needle penetration through the anterior tracheal wall is difficult due to the hard cartilage ring, and complete clearance of cartilage from needles is impossible. However, we have made an effort to overcome these problems, routinely widening the intercartilage space by pushing with the outer sheath before puncture to avoid puncturing the cartilage.
We measured the width of the sample, which seemed to represent the internal lumen of the TBNA needle, from a slide instead of directly measuring the weight of the sample. We measured width because stereoscopic microscopy revealed that a majority of samples contain a liquid component of blood. Kreula et al19 showed that sample weights correlated significantly (P<0.001) with cell counts from liver taken for fine-needle aspiration cytology. However, their target tissue was the resected organ after autopsy, which thus has no blood flow in the target organ. We speculated that this correlation between sample weight and cell counts might not apply in clinical settings dealing with viable organs such as mediastinal lymph nodes. We, thus evaluated the width of histologic tissue. Despite using different size needles, no difference was seen in the width of tissue in either group. Width of histologic tissue may depend on the manner of extraction from the needle and the degree of viscosity.
EBUS-TBNA displayed higher diagnostic yield for a 21-gauge needle than for a 22-gauge needle. This study revealed greater efficacy of using 21-gauge needles in obtaining diagnostic nodal specimens. We believe that the 21-gauge needle has great potential for widespread acceptance for mediastinal lesions containing benign conditions such as sarcoidosis and tuberculosis.
The authors thank Takashi Nishisaka from Hiroshima Prefectural Hospital for the excellent pathologic advice, Shinobu Tatsunami from the Institute of Radioisotope Research at St Marianna University for statistical advice, and Hiroaki Osada from Shonan Central Hospital for critical reading of the article.
1. Hürter T, Hanrath P. Endobronchial sonography: feasibility and preliminary results. Thorax. 1992;47:565–567
2. Herth FJ, Eberhardt R. Actual role of endobronchial ultrasound (EBUS). Eur Radiol. 2007;17:1806–1812
3. Rintoul RC, Skwarksi KM, Murchison JT, et al. Endobronchial and endoscopic ultrasound-guided real real-time fine-needle
aspiration for mediastinal staging. Eur Respir J. 2005;25:416–421
4. Herth F, Eberhardt R, Vilmann P, et al. Real-time endobronchial ultrasound guided transbronchial needle
aspiration for sampling mediastinal lymph nodes. Thorax. 2006;61:795–798
5. Yasufuku K, Chiyo M, Sekine Y, et al. Real-time endobronchial and hilar lymph nodes. Chest. 2004;126:122–128
6. Wang KP, Marsh BR, Summer WR, et al. Transbronchial needle
aspiration for diagnosis of lung cancer
. Chest. 1981;80:48–50
7. Shenk DA, Chambers SL, Derdak S, et al. Comparison of the Wang 19-gauge and 22-gauge needles in the mediastinal staging of lung cancer
. Am Rev Respir Dis. 1993;147:1251–1258
8. Harrow EM, Abi-Saleh W, Blum J, et al. The utility of transbronchial needle
aspiration in the staging of bronchogenic carcinoma. Am J Respir Crit Care Med. 2000;161:601–607
9. Wang KP, Brower R, Haponik EF, et al. Flexible transbronchial needle
aspiration for the staging of bronchogenic carcinoma. Chest. 1983;84:571–576
10. Nakajima T, Yasufuku K, Takahashi R, et al. Comparison of 21-gauge and 22-gauge aspiration needle
during endobronchial ultrasound-guided transbronchial needle
aspiration. Respirology. 2011;16:90–94
11. Herth FJ, Eberhardt R, Vilmann P, et al. Real-time endobronchial ultrasound guided transbronchial needle
aspiration for sampling mediastinal lymph nodes. Thorax. 2006;61:795–798
12. Tremblay A, David RS, MacEachern P, et al. A randomized controlled trial of standard versus endobronchial ultrasonography-guided transbronchial needle
aspiration in patients with suspected sarcoidosis. Chest. 2009;136:340–346
13. Holty JE, Kuschner WG, Gould MK. Accuracy of transbronchial needle
aspiration for mediastinal staging of non-small cell lung cancer
: a meta-analysis. Thorax. 2005;60:949–955
14. Davenport RD. Rapid on-site evaluation of transbronchial aspirates. Chest. 1990;98:59–61
15. Sharafkhaneh A, Baaklini W, Gorin AB, et al. Yield of transbronchial needle
aspiration in diagnosis of mediastinal lesions. Chest. 2003;124:2131–2135
16. Wong M, Yasufuku K, Nakajima T, et al. Endobronchial ultrasound: new insight for the diagnosis of sarcoidosis. Eur Respir J. 2007;29:1182–1186
17. Hartley MN, Tuffnell DJ, Hutton JL, et al. Fine needle
aspiration cytology: an in vitro study of cell yield. Br J Surg. 1988;75:380–381
18. Kurimoto N, Osada H, Miyazawa T, et al. Targeting area in metastatic lymph nodes in lung cancer
for endobronchial ultrasonography-guided transbronchial needle
aspiration. J Bronchol. 2008;15:134–138
19. Kreula J, Bondestam S, Virkkunen P. Sample size in fine needle
aspiration biopsy. Br J Surg. 1989;76:1270–1272