Eur J Cardiothorac Surg 2008;33:1124-1128. doi:10.1016/j.ejcts.2008.03.014
Copyright © 2008, European Association for Cardio-thoracic Surgery. Published by Elsevier. All rights reserved.
Does endoesophageal ultrasound-guided fine-needle aspiration replace mediastinoscopy in mediastinal staging of thoracic malignancies?
Biruta Wittea,*,
Wolfgang Neumeisterb,
Martin Huertgena
a Department of Thoracic Surgery, Lungenzentrum, Katholisches Klinikum Koblenz, Koblenz, Germany
b Department of Pneumology, Lungenzentrum, Katholisches Klinikum Koblenz, Koblenz, Germany
Received 7 August 2007;
received in revised form 4 March 2008;
accepted 8 March 2008.
* Corresponding author. Address: Department of Thoracic Surgery, Katholisches Klinikum Koblenz, Kardinal-Krementz-Straße 1-5, 56073 Koblenz, Germany. Tel.: +49 261 496 9027; fax: +49 261 496 6469. (Email: b.witte{at}kk-koblenz.de).
 |
Abstract
|
|---|
Objective: To determine the impact of endoesophageal ultrasound-guided fine-needle aspiration (EUS-FNA) on management of thoracic malignancies. Methods: One hundred and twenty patients referred for invasive diagnostic and resection of thoracic malignancies were studied prospectively. Negative and inconclusive EUS-FNA findings were assessed by video-assisted mediastinoscopic lymphadenectomy (VAMLA) or open lymphadenectomy. Results: One hundred and twenty patients, aged 64.1 years (range 38–85) underwent 120 EUS-FNA, 53 video-assisted mediastinoscopic and 48 open lymphadenectomies for diagnosis and treatment of 99 lung carcinoma, six lung metastases, five mesothelioma, three lymphoma, and eight other conditions. EUS-FNA showed T4 in 15/120 and adrenal or hepatic metastases in 9/120 cases. Prevalence of mediastinal lymph node metastases was 51.7%. EUS-FNA false-negative rate was 25.3%. EUS-FNA sensitivity was 91.7%, 78.1% and 43.8% for bulky disease, enlarged mediastinal nodes or normal nodes on CT scan, 50% and 96.6% for right- and left-sided tumours, and 80.6%, 78.9%, 23.8% and 25.0% for the lymph node stations 7, 5/6, 4R, and 4L. A 38.3% respectively 100% cut-down of mediastinoscopies leads in 7.5% respectively 20.8% to incorrect treatment decisions. Conclusions: EUS-FNA sensitivity depends on the localisation of the primary tumour, and extent and location of mediastinal disease. For left-sided tumours, EUS-FNA improves mediastinal staging by assessing stations 5 and 6 inaccessible to conventional mediastinoscopy. For extended mediastinal disease, mediastinoscopy can be avoided or spared for restaging after neoadjuvant therapy. Exclusion of mediastinal involvement requires mediastinoscopy or open lymphadenectomy. Beyond mediastinal nodal staging, EUS-FNA may detect T4 and M1 situations. Thus, EUS-FNA is a useful supplement to and not the replacement of mediastinoscopy.
Key Words: Endoscopic ultrasound • Fine-needle aspiration • Lung carcinoma • Lymph node staging • Mediastinum • Mediastinoscopy • VAMLA
|
1. Introduction
|
|---|
Prognosis of lung carcinoma largely depends on lymphatic spread as long as the disease is confined to the thoracic cavity. Therapeutic decisions for non-metastatic disease should reflect the stage-dependent risk of local recurrence and distant metastases, and are therefore guided by nodal status. Despite rapid progress in genomic and proteomic research, the N denominator of the TNM system is still an extremely robust criterion for prognosis and therapy. Although non-surgical ways to pre-therapy lymph node staging such as computed tomography, PET scan, endoscopic ultrasound and fine-needle aspiration techniques emerged, mediastinoscopy has not been outperformed, but is still considered to be gold standard [1,2].
Compared to mediastinoscopy, endoesophageal ultrasound-guided fine-needle aspiration (EUS-FNA) is less invasive and considered to be more cost-effective [3]. Previous investigations assessed EUS-FNA findings by clinical follow-up, mediastinoscopy, thoracoscopy or, if including resectable patients, by open lymphadenectomy, and the EUS-FNA false-positive and false-negative rates were reported to be about 2%, respectively 80% [4–7]. The presented study investigates the use of EUS-FNA in unselected surgical patients with thoracic malignancies in order to determine its accuracy, its impact on diagnostic and therapeutic decisions, and to what extent it could replace mediastinoscopy. As negative or inconclusive EUS-FNA findings are assessed by complete, either open or video-assisted, mediastinoscopic lymphadenectomy (VAMLA) [8], the focus is on false-negative EUS-FNA results.
|
2. Materials and methods
|
|---|
After its introduction into the Department of Pneumology of our institution in 2001, EUS-FNA became a regular part of the routine work-up of patients with thoracic malignancies. From January 2004 to December 2006 we prospectively collected data of all patients who were referred for mediastinoscopy or resection of thoracic tumours and underwent EUS-FNA prior to the intended surgical procedure. Informed written consent of each patient was obtained. As the protocol had impact only on data collection and did not change routine clinical proceedings, a decision was made not to apply for the votum of an external review board. Preliminary investigations were bronchoscopy and CT scans of the thorax and upper abdomen. Radiological mediastinal staging adhered to the accepted criteria for enlarged lymph nodes exceeding 10 mm short axis diameter. All patients underwent abdominal sonography, cranial CT or MRI, and bone scans to exclude distant metastases. PET scans and EBUS-TBNA were not performed. EUS-FNA was performed under local anaesthesia and intravenous sedation with commercially available 22 G aspiration needles and ultrasonic gastroscopes. All endosonographically visible hilar and mediastinal lymph nodes were assessed for size and structure. Enlarged or atypical nodes were punctured at least twice whenever accessible [4–6]. Fine-needle aspiration (FNA) specimens were sent for routine cytological and immunocytochemical examination. Patients with negative or inconclusive EUS-FNA findings underwent systematic mediastinal nodal dissection by VAMLA or open lymphadenectomy in the course of their further diagnostic work-up and treatment. VAMLA comprised of en bloc resection of the stations 7, 3 and 4R, systematic dissection of 4L, 2R+L, and, if feasible 5 and 6 via extended mediastinoscopy [9], as described previously [8,10]. Open lymphadenectomy was performed according to ESTS recommendations [11] as radical lymphadenectomy for right-sided and systematic nodal dissection for left-sided procedures. Lymphadenectomy specimens underwent routine histological and immunohistochemical examination. Prospective data collection covered patient age and sex, tumour location, histology, TNM stage, radiographic mediastinal staging based on CT scan, endoesophageal ultrasound (EUS) findings, FNA punctures, FNA cytology, further diagnostics and treatment, and extent and histological results of lymphadenectomy. The reliability of EUS-FNA was assessed by overall false-negative rate, and sensitivity stratified for site of primary (right/left), extent of mediastinal disease (normal nodes/enlarged nodes/bulky disease), and lymph node station (4 R/7/4 L/5 + 6). The false-negative rate was calculated as number of false-negative patients (negative mediastinum by EUS-FNA, but positive by lymphadenectomy), divided by the number of all (true and false) EUS-FNA negative patients. Sensitivity was calculated as number of true-positive EUS-FNA findings divided by the prevalence of positive nodes (true-positive + false-negative EUS-FNA findings). EUS-FNA impact on patient management was described, and compared to treatment decisions based on surgical staging.
|
3. Results
|
|---|
From January 2004 to December 2006, EUS was performed without complications in all 120 patients with a median age of 64.1 years (range, 38–85 years), among them 19 women. Ninety-nine had primary lung carcinoma (49 adenocarcinoma, 34 squamous cell carcinoma, nine SCLC, seven others), six lung metastases, five mesothelioma, three lymphoma, and seven other conditions. They presented with 64 right-sided, 46 left-sided and 10 bilateral respectively mediastinal lesions. CT scan showed enlarged mediastinal lymph nodes in 61 of 120 patients; among them 12 classified as bulky disease. With EUS, suspicious mediastinal lymph nodes were described in 99 patients. Beyond nodal staging, EUS detected T4 signs in 15 mediastinal structures (aorta 3, intrapericardial pulmonary artery 3, atrium 2, oesophagus 2, trachea 1, subclavian vein 1) of 10 patients, and M1 criteria in 21 structures (16 left and two right adrenal glands, two hepatic nodules, one retroperitoneal lymphoma) of 15 patients. In 89 patients, 115 lymph nodes and 21 abdominal structures were punctured by EUS-FNA (Fig. 1
). EUS-FNA samples cytologically confirmed 45 cases of mediastinal disease (42 N2, 3 N3), nine cases of distant metastases (eight adrenal and one hepatic) and seven benign adrenal tumours (six adenoma, one chemodectoma). Eighty-three patients underwent 53 VAMLA and 61 open lymphadenectomies to assess negative EUS-FNA results (n
= 66), inconclusive EUS-FNA cytology (n
= 1), and EUS-FNA proven N2 cases suspicious of contralateral nodal disease (n
= 16). In eight patients with negative EUS-FNA results, lymphadenectomy was inappropriate because of medical contraindications (n
= 4), EUS-FNA proven distant metastases (n
= 2), extended local tumour growth (n
= 2), and tumour-associated local complications (n
= 1). Restaging by lymphadenectomy detected 17 overlooked cases of mediastinal disease (16 N2, one N3), and staged up eight N2 cases to N3 (Tables 1 and 2
). From those data, the prevalence of mediastinal disease was calculated to be 51.7%, the false-negative rate for EUS-FNA prediction of the presence of mediastinal disease, regardless of extent, 25.3%, and the overall sensitivity 72.6%. Table 3
shows the stratification of EUS-FNA sensitivity data by mediastinal lymph node size, location of the primary tumour and lymph node station. Seventy-one EUS-FNA results had impact on patient management: detection of nine distant metastases, detection of mediastinal disease in 45 cases, among them eight overlooked N3 disease, and 17 false-negative cases. Without surgical reassessment, 20.8% of treatment decisions would have led to inadequate therapy.
View larger version (8K):
[in this window]
[in a new window]
|
Fig. 1. EUS-FNA punctures. For puncture and fine-needle aspiration, a structure had to be suspicious as well as accessible. In 89 patients, 115 lymph nodes and 21 abdominal structures were punctured, most often the subcarinal nodes (station 7), nodes of the aortopulmonary window (station 5), and the left adrenal gland (lADR).
|
|
|
4. Discussion
|
|---|
To determine the possible impact of EUS-FNA on the management of thoracic malignancies, 120 tumour patients referred for surgical interventions, mediastinoscopy as well as resection, were studied. Thus, a comparatively high number of advanced tumours and unresectable patients were included, reflected by a high prevalence of mediastinal disease (51.7%), and a low thoracotomy rate (50.8%). The radicality of VAMLA was considered to be equal to open lymphadenectomy, as VAMLA had been shown to harvest more mediastinal lymph nodes than open lymphadenectomy [9] and to remain highly accurate under routine clinical conditions [10].
EUS was found to be more sensitive than CT scan for detection of suspicious lymph nodes (82.5 % vs 50.8% of 120 patients), probably due to its higher resolution, resilience to motion artefacts, and ability to assess not only lymph node size but also nodal structure as a second criterion for malignancy. However, FNA cytology was positive in only 41.4% of patients with sonographically suspicious nodes. This is similar to the data of Fritscher-Ravens et al. [5], describing 66% suspicious nodes and 42% positive FNA findings in 33 patients, and supporting their conclusion that EUS imaging alone is not conclusive, and that additional fine-needle aspiration cytology is extremely helpful. This was also found to be true for distant metastases, especially for adrenal masses with a high rate of benign cytologies. The occasional EUS detection of T4 was of limited importance and reliability, and should be checked by multiplanar transoesophageal [12] or mediastinoscopic [13] ultrasonography in potentially resectable patients.
In our study the EUS-FNA sensitivity for detection of mediastinal disease regardless of extent, was 72.6%, and lower than most published data. A recent meta-analysis [7] reported a pooled sensitivity of 83%, suggesting sensitivity in the range of conventional mediastinoscopy. Stratification for lymph node size (Table 3) correspondingly showed lower sensitivities for normal nodes on CT scan (39–75% [7]), and higher sensitivities for enlarged nodes (84–94% [7]). For geometrical reasons, a relation between diagnostic accuracy and volume of diseased as well as sampled tissue is not surprising (Fig. 2
). This means the more extensive a disease, the smaller the samples could be. This is of practical importance as risk, invasiveness and expenses all rise with sample size. However, even for enlarged nodes EUS-FNA sensitivity did not catch up with conventional mediastinoscopy, and even for bulky disease it did not reach 100%. The comparatively low overall sensitivity observed in our cohort can be explained with the high prevalence of normal size mediastinal nodes, and the thorough surgical mediastinal reassessment. EUS-FNA sensitivity for left-sided primaries appeared to be excellent. However, in seven of 23 left-sided true-positive, EUS-FNA missed N3 disease. These are effects of the endoesophageal location of the ultrasound probe, and its consequences for the range of EUS-FNA. High sensitivity for left-sided tumours is due to the good accessibility and the frequent involvement of station 5. On the contrary, the stations 2, 3, and 4 and parts of 7 are located in front of the central airways and therefore out of range unless heavily enlarged. Thus, EUS-FNA is prone to overlook N3 with left-sided and N2 with right-sided primaries. This is reflected by single station sensitivity calculations (Table 3) showing high sensitivities for the better visible and accessible stations 5/6, and 7, and low sensitivities for 4R and 4L. A detailed look at the cases with false-negative EUS-FNA (Table 1) suggests two main causes of fine-needle failure: no puncture of sonographically normal, but nevertheless involved, lymph nodes and metastases in partially inaccessible stations. Remarkably, failure was not restricted to microscopic or single node disease, and happened even with puncture of involved stations.
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 2. Relation between diagnostic accuracy and volume of diseased as well as sampled tissue. Intuitively, accuracy of mediastinal staging is dependent on tumour mass (first line, range: isolated cells to extranodal tumour growth/bulky disease) and specimen size (second line, range: fine-needle aspiration to en bloc resection/radical lymphadenectomy). For a given accuracy, specimen size inversely correlates with tumour mass.
|
|
In conclusion, there are several options to improve the sensitivity of EUS-FNA. We suggest routine fine-needle aspirations of station 7, and 5/6 for left-sided primaries, a complementary conventional mediastinoscopy [14] or EBUS-TBNA [15,16] to assess the stations 3, 4 R+L, and parts of 7, and a preceding PET scan to guide the sampling of confirmatory cytologies [17].
EUS-FNA had considerable, not always positive, impact on patient management. Detection and cytological confirmation of distant metastases changed therapy and prognosis in 7.5%, and staged mediastinal disease correctly in another 30.8%, associated with a 38.3% cut-down of mediastinoscopies. A negligible rate of false-positive results given; this is associated with overlooked N3 disease and an unfavourable change of treatment in 7.5%. Complete abandonment of mediastinoscopy would have had the consequence of another 13.3% of overlooked mediastinal disease and a negative impact on treatment. To summarise, any cut-down of mediastinoscopies is associated with some amount of incorrect staging and treatment, in this example 38.3% and 100% with 7.5% and 20.8% respectively.
To conclude, EUS-FNA sensitivity depends on the localisation of the primary tumour, and localisation and extent of mediastinal disease. Exclusion of mediastinal involvement requires mediastinoscopy or open lymphadenectomy. For left-sided tumours, EUS-FNA improves mediastinal staging by assessing stations 5 and 6 inaccessible to conventional mediastinoscopy. To prove extended mediastinal disease, mediastinoscopy can be avoided or spared for restaging after neoadjuvant therapy. Beyond mediastinal nodal staging, EUS-FNA may detect T4 and M1 situations. Complete replacement of mediastinoscopy by EUS-FNA is associated with a high false-negative rate for mediastinal staging and incorrect treatment decisions, and therefore inappropriate. Thus, EUS-FNA is a useful supplement to and not the replacement of mediastinoscopy.
|
Appendix A
|
|---|
Conference discussion
Dr J. Kuzdzal (Zakopane, Poland): I would like to congratulate you on the excellent study and would like to comment on one point. This study shows how important is the quality of the confirmation test we use to assess the diagnostic yield of different kinds of needle biopsies or endoscopic techniques. Of course if we use a mediastinoscopy as a confirmatory test, we know that mediastinoscopy itself has a relative low diagnostic yield – it causes better results of the needle aspiration biopsy. As we have seen in your presentation a more radical lymphadenectomy using the VAMLA technique, which is for sure much more accurate than the mediastinoscopy, shows that the real diagnostic yield of the needle puncture is relatively low. So congratulations on the excellent study.
Dr M. Krasnik (Copenhagen, Denmark): It was a very nice presentation and it is also nice to compare it with the (VEMLA) because we also use the normal mediastinoscopy. A lot of studies have already shown that EUS-FNA is not enough for staging of mediastinum diseases, it is a good tool but the anterior part of the mediastinum cannot be reached. If you combine we can see that mediastinum anterior and hilar, reaching at both sides, if you combine them you could have good results and maybe a little bit better because you get some entry which you get of course with the VEMLA but it is easier to do FNA. There is no doubt that even we proved the combination. And with FNA even if the departure of the same lymph node is from two different directions, there will be positive results and vice versa. We can never use EUS-FNA alone, only if it is positive. A negative EUS-FNA is not enough even the gastroenterologist thinks it is.
Dr B. Witte (Koblenz, Germany): With regard to the range of the methods you are absolutely right. However what can you do with EBUS-TBNA. It may have the optimal range, but it will always be a fine-needle biopsy as pathologic test and never a complete lymphadenectomy. We are going to adopt EBUS-FNA this summer, and we will continue to do VAMLA.
Dr S. Cassivi (Rochester, MN): It was a great presentation but permit me to be the ‘Doubting Thomas’. I don’t know if you can get from the oesophageal lumen to a lymph node in the true station 5. Can you explain how, from the oesophagus, you get in to the anterior mediastinum?
Dr B. Witte (Koblenz, Germany): Probably your comment is on the range of the two methods. Let me explain to you how we assess station 5. For any left sided tumour with visible nodes at station 5 or 6 on CT scan, we combine our preoperative mediastinoscopic assessment with an extended mediastinoscopy.
Dr S. Cassivi (Rochester, MN): But you are not getting it from a EUS approach? From the oesophagus, you are not sampling a true station 5 node.
Dr B. Witte (Koblenz, Germany): EUW-FNA is a good method to get station 5 and sometimes even station 6.
Dr R. Berrisford (Exeter, U.K.): I can believe you can get at station 5. When we do minimal invasive resection in the chest, you always see the arch and we actually biopsy station 5 occasionally when we see the nodes in the oesophageal bed. So I can believe you can sometimes reach it.
Dr B. Witte (Koblenz, Germany): This node we biopsied quite often, I cannot believe it is invisible.
Unidentified discussant I am doing this for 15 years. I have seen lymph nodes number 5 several times. It is difficult because the oesophagus is sometimes moving, so you press it and then you see the whole node but you can always pass through it. And if you are very brave, I have done it once in the aorta and once in the pulmonary artery, you can put a needle just through it. We puncture both the aorta and the pulmonary artery, nothing happens.
|
Footnotes
|
|---|
Presented at the 15th European Conference on General Thoracic Surgery, Leuven, Belgium, June 3–6, 2007.
|
References
|
|---|
- Rusch VW. Mediastinoscopy: an endangered species?. J Clin Oncol 2005;23:8283-8285.[Free Full Text]
- De Leyn P, Lardinois D, Van Schil P, Rami-Porta R, Passlick B, Zielinski M, Waller D, Lerut T, Weder W. ESTS guidelines for preoperative lymph node staging for non-small cell lung cancer. Eur J Cardiothorac Surg 2006;30:787-792.[Abstract/Free Full Text]
- Kramer H, van Putten JWG, Post WJ, van Dullemen HM, Bongaerts AHH, Pruim J, Suurmeijer AJH, Klinikenberg TJ, Groen H, Groen HJM. Oesophageal endoscopic ultrasound improves and simplifies the staging of lung cancer. Thorax 2004;59:596-601.[Abstract/Free Full Text]
- Silvestri GA, Hoffman BJ, Bhutani MS, Hawes RH, Coppage L, Sanders-Cliette A, Reed CE. Endoscopic ultrasound with fine needle aspiration in the diagnosis and staging of lung cancer. Ann Thorac Surg 1996;61:1441-1445.[Abstract/Free Full Text]
- Fritscher-Ravens A, Bohuslavizki KH, Brandt L, Bobrowski C, Lund C, Knöfel WT, Pforte A. Mediastinal Lymph node involvement in potentially resectable lung cancer. Chest 2003;123:442-451.[CrossRef][Medline]
- Annema J, Verstegh MI, Veselic M, Voigt P, Rabe KF. Endoscopic ultrasound-guided fine-needle aspiration in the diagnosis and staging of lung cancer and its impact on surgical staging. J Clin Oncol 2005;33:8357-8361.
- Micames CG, McCrory DC, Pavey DA, Jowell PS, Gress FG. Endoscopic ultrasound-guided fine needle aspiration for non-small lung cancer staging. A systematic review and metaanalysis. Chest 2007;131:539-548.[CrossRef][Medline]
- Hürtgen M, Friedel G, Toomes H, Fritz P. Radical video-assisted mediastinoscopic lymphadenectomy (VAMLA) – technique and first results. Eur J Cardiothorac Surg 2002;21:348-351.[Abstract/Free Full Text]
- Ginsberg RJ, Rice TW, Goldberg M, Waters PF, Schmocker BJ. Extended cervical mediastinoscopy. A single staging procedure for bronchogenic carcinoma of the left upper lobe. J Thorac Cardiovasc Surg 1987;94:673-678.[Abstract]
- Witte B, Wolf M, Huertgen M, Toomes H. Video-assisted mediastinoscopic surgery: clinical feasibility and accuracy of mediastinal lymph node staging. Ann Thorac Surg 2006;82:1821-1827.[Abstract/Free Full Text]
- Lardinois D, De Leyn P, Van Schil P, Rami Porta R, Waller D, Passlick B, Zielinski M, Junker K, Rendina EA, Ris HB, Hasse J, Detterbeck F, Lerut T, Weder W. ESTS guidelines for intraoperative lymph node staging in non-small cell lung cancer. Eur J Cardiothorac Surg 2006;30:787-792.[Abstract/Free Full Text]
- Schlotterbeck K, Schmid J, Klein F, Alber G. Transesophageal sonography in the staging of lung cancer. Ultraschall Med 1997;18:153-158.[Medline]
- Huertgen M, Metzler B, Friedel G, Toomes H. Mediastinoscopic ultrasonography (MUS). Eur J Cardiothorac Surg 2004;26:842-844.[Abstract/Free Full Text]
- Annema J, Versteegh MI, Veselic M, Welker L, Mauad T, Sont JK, Willems LNA, Rabe KF. Endoscopic ultrasound added to mediastinoscopy for preoperative staging of patients with lung cancer. JAMA 2005;294:931-936.[Abstract/Free Full Text]
- Rintoul RC, Skwarski KM, Murchinson JT, Hill A, Walker WS, Penman ID. Endoscopic and endobronchial real-time fine-needle aspiration for staging of the mediastinum in lung cancer. Chest 2004;126:2020-2022.[CrossRef][Medline]
- Herth FJF, Lunn W, Eberhardt R, Becker HD, Ernst A. Transbronchial versus transesophageal ultrasound-guided aspiration of enlarged mediastinal lymph nodes. Ann J Respir Crit Care Med 2005;171:1164-1167.[CrossRef]
- Eloubeidi MA, Cerfolio RJ, Chen VK, Desmond R, Syed S, Buddhiwardhan O. Endoscopic ultrasound-guided fine needle aspiration of mediastinal lymph nodes in patients with suspected lung cancer after positron emission tomography and computed tomography scans. Ann Thorac Surg 2005;79:263-268.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
I. Bar, M. Papiashvilli, G. Fink, J. Sandbank, and D. Stav
Cervical Mediastinoscopic Lymphadenectomy for Accurate Staging in Lung Cancer
Asian Cardiovasc Thorac Ann,
August 1, 2009;
17(4):
357 - 361.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
1. Introduction
