| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Thoracic and Cardiovascular Surgery, The University of Texas, M.D. Anderson Cancer Center, Houston, Texas
b Department of Pathology, The University of Texas, M.D. Anderson Cancer Center, Houston, Texas
c Department of Pulmonary Medicine, The University of Texas, M.D. Anderson Cancer Center, Houston, Texas
d Department of Gastroenterology, Hepatology, and Nutrition, The University of Texas, M.D. Anderson Cancer Center, Houston, Texas
e Department of Diagnostic Imaging, The University of Texas, M.D. Anderson Cancer Center, Houston, Texas
Accepted for publication May 8, 2009.
* Address correspondence to Dr Rice, Department of Thoracic and Cardiovascular Surgery, The University of Texas M. D. Anderson Cancer Center, Box 445, 1515 Holcombe Blvd, Houston, TX 77030 (Email: drice{at}mdanderson.org).
Presented at the Forty-fifth Annual Meeting of The Society of Thoracic Surgeons, San Francisco, CA, Jan 26–28, 2009.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
Methods: Eighty-five patients with mesothelioma who were potential candidates for radical surgery underwent preoperative staging with mediastinoscopy (n = 50) or EBUS (n = 38). Eleven patients also underwent EUS.
Results: Diagnostic yield (specimens containing lymphocytes or tumor cells) was 100% for mediastinoscopy and 84% for EBUS (p < 0.001). Mediastinoscopy identified 7 of 50 (14%) patients with nodal metastases. Thirty-eight (76%) mediastinoscopy-negative patients underwent surgery with nodal sampling and there were 18 false negatives. Endobronchial ultrasound identified 13 of 38 (34%) patients with nodal metastases. Twenty-two (58%) EBUS-negative patients underwent surgery with nodal sampling and there were 10 false negatives. Sensitivity and negative predictive value for mediastinoscopy were 28% and 49%, and 59% and 57% for EBUS. Eleven patients had EUS preoperatively, which revealed infradiaphragmatic nodal metastases in 5 patients.
Conclusions: Although this study is retrospective, EBUS had higher sensitivity than either mediastinoscopy or imaging studies for detection of nodal metastases. Nevertheless, the ability to accurately identify nodal involvement preoperatively in patients with mesothelioma remains suboptimal. Esophageal ultrasound may complement EBUS particularly in cases where infradiaphragmatic nodal metastases are suspected.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
Although histologic type can be determined by needle aspiration cytology or tissue biopsy, accurate assessment of tumor extent and nodal metastases is problematic. Because transdiaphragmatic invasion or peritoneal carcinomatosis may be missed with radiographic imaging, we routinely perform laparoscopy prior to radical resection [1–3]. Assessment of mediastinal nodal metastases is inaccurate using computed tomography (CT) and positron emission tomography (PET) due to difficulty differentiating lymph nodes from adjacent tumor nodules [4–6]. We have routinely performed cervical video-mediastinoscopy (CM) at the time of staging laparoscopy. As previously reported, extended surgical staging identified abdominal involvement or contralateral mediastinal nodal metastases in 13.6% of patients who were radiographically considered resectable [3].
While generally safe, CM requires a cervical incision and is associated with a small risk of recurrent nerve injury, pneumothorax, tracheal injury, hemorrhage and even death [7]. Endobronchial ultrasound (EBUS) and esophageal endoscopic ultrasound (EUS) guided fine needle aspiration (FNA) of mediastinal lymph nodes have been highly effective for staging non-small cell lung cancer (NSCLC) [8–11]. Diagnostic accuracy of these procedures is comparable with CM and they are less invasive with the added capability of sampling nodal stations beyond the reach of the mediastinoscope. Since 2006 we have performed EBUS for assessment of mediastinal nodes in patients being considered for radical resection of MPM. The utility of EBUS and EUS for the staging of MPM has not previously been described and is the purpose of this study.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
Cervical Mediastinoscopy
Mediastinoscopy was performed under general anesthesia using a video-mediastinoscope (Karl Storz, Tuttlingen, Germany). In all cases attempts were made to biopsy nodes from ipsilateral and contralateral paratracheal stations, as well as the subcarinal nodal packet. Biopsies were paraffin embedded, stained with hematoxylin and eosin, and examined by a thoracic pathologist.
Endobronchial Ultrasound
Endobronchial ultrasound-guided FNA was usually performed at the time of staging laparoscopy by thoracic surgeons. A dedicated flexible ultrasonic bronchoscope was used with a linear scanning 7.5 MHz transducer (UC160F-OL8; Olympus America Inc, Melville, NY) as previously described in the NSCLC population [9]. It scans parallel to the insertion direction of the bronchoscope with a dedicated ultrasound scanner (EU-C60; Olympus America Inc). An examination of all mediastinal lymph node stations accessible by EBUS (stations 1, 2, 4, 7, 10, and 11) was performed. All lymph nodes visualized with ultrasound were measured and those with a short axis 5 mm or greater were sampled with a 22-gauge needle (NA-201SX-4022; Olympus America Inc) designed for the EBUS scope under real-time imaging EBUS guidance. Nodes were readily identifiable as discrete round or oval structures immediately adjacent to the trachea or bronchi. When contralateral (N3) nodes were visualized, they were sampled before N2 nodes. The N1 nodes were not routinely sampled, unless clinically suspicious (short axis >1 cm on CT or 18F-fluorodeoxyglucose [FDG] avid). A minimum of two, but most often three separate aspirations were performed at each nodal station. The aspirated material was applied to glass slides and fixed in Carnoy's solution (6 parts ethanol [absolute or 95%], 3 parts chloroform, 1 part glacial acetic acid). Additional material was then aspirated into Roswell Park Memorial Institute solution (Sigma-Aldrich, St Louis, MO) and analyzed using cytospin or cell blocks. Fixed specimens were stained with Papanicolaou staining and examined by an on-site cytopathologist to ensure adequate cellular (lymphoid or tumor) material in the specimens (Fig 1). If adequate tissue was not identified after five passes, the procedure was terminated for that node. Specimens were categorized as positive (tumor cells present), negative (lymphoid tissue but no tumor cells), or nondiagnostic (poor cellularity).
|
Therapy
Patients with negative mediastinal nodes generally underwent surgical resection (EPP or pleurectomy-decortication [PD], according to the patient's functional status). Patients with N2 disease were treated nonoperatively or with PD. Patients with N3 nodal or distant metastases were treated nonoperatively. Patients undergoing EPP underwent mediastinal node dissection. Nodal sampling was usually (but not uniformly) performed in patients undergoing PD or exploratory thoracotomy.
Statistical Analysis
Statistical comparisons between patients receiving CM or EBUS were performed using the Student t test, the 
2 test, or the Fisher exact test, as appropriate. Malignant cells identified on EBUS, or positive nodes from CM were deemed "true positives." If CM confirmed negative mediastinal nodes after EBUS, the results of EBUS were assumed to be a "true negative." In the case of negative CM or EBUS, pathologic nodal staging information obtained at thoracotomy was considered the "gold standard." Only cases where nodal tissue was resected were used for sensitivity and specificity calculations. The 3 patients who had both EBUS-FNA and CM were excluded from sensitivity calculations for EBUS as definitive pathology was not available for these patients and a negative CM was not considered to be sufficiently reliable.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
|
Station 7 (subcarinal) was biopsied in all patients and yielded a diagnostic specimen in 95%. Stations 4R and 4L (right and left lower paratracheal) were biopsied in approximately 50% of patients each, but the diagnostic yield was higher for 4R compared with 4L (91% vs 55%). Twenty-one (27%) mediastinal lymph nodes from 13 (34%) patients contained metastatic tumor. Contralateral (N3) involvement was diagnosed in 1 patient. There were 9 (26%) false negatives among 34 patients in whom confirmatory pathologic staging information was available.
|
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
Radiographic imaging is relatively poor at identifying nodal metastases in patients with MPM. Computed tomographic scanning has an accuracy of approximately 60%, and magnetic resonance imaging has not shown superiority [20, 21]. Although useful for identifying distant disease, the sensitivity of PET for N2 metastases has been reported to be as low as 11% and integrated PET-CT has been shown to be only 30% accurate in determining nodal stage [5, 6]. The high incidence of nodal involvement with MPM, coupled with the low sensitivity of imaging, has led many to recommend mediastinoscopy prior to radical surgery [14, 18, 22]. Although mediastinal staging with CM is 100% specific, we have found the procedure to have a false negative rate of 22% and a sensitivity of only 36% for staging MPM [22]22. The current series revealed an even higher false negative rate (51%) for CM and a sensitivity of 28%. Others have reported similar results; for instance, Schouwink and colleagues [20] noted 25 patients with negative CM of which 9 (36%) had positive nodes at extrapleural pneumonectomy. Six (24%) patients had metastases at stations not accessible by CM but 3 (12%) were positive at CM accessible stations. Similarly, de Perrot and colleagues [23] found positive nodal disease in 10 of 19 patients (53%) who had been staged as node-negative with CM. Six (32%) patients had N2 disease, and in 4 patients metastases were noted at stations that had been previously biopsied by CM. Edwards and colleagues [24] also reported a high rate of false negatives with CM in 10 of 30 (33%) who underwent EPP. Of 35 patients with N2 metastases, positive nodes were noted at sites inaccessible to CM in 17 (49%). In the current series there were 18 (36%) patients with falsely negative CM and 9 had metastases to N2 stations that had been previously biopsied, despite the fact that diagnostic specimens were obtained in all patients, and most patients had bilateral nodal stations sampled.
Endobronchial ultrasound-fine needle aspiration is rapidly gaining acceptance for mediastinal staging of NSCLC [9–11]. It has been shown to be more sensitive than PET and CT for diagnosing N2 disease, is highly specific, and appears to be as accurate as mediastinoscopy, yet is less invasive. The use of EBUS for nodal staging in MPM has not been reported but has several potential attractions. First, the procedure is associated with few, if any serious complications, whereas CM has a small but definite risk of major hemorrhage, recurrent laryngeal nerve paresis, tracheobronchial injury, and even death, none of which have been described with EBUS [7]. Second, CM violates peritracheal tissue planes, potentially making dissection at the time of EPP or PD more difficult. The EBUS-FNA causes minimal trauma to peritracheal tissue. Third, hilar nodes usually inaccessible to CM are easily reached with EBUS. Whether N1 metastases portend the same poor prognosis as positive N2 nodes has been questioned; however, most series have shown no difference in survival between the two [25]. Therefore, preoperative identification of N1 metastases could potentially influence management. For these reasons we investigated the use of EBUS-FNA for preoperative mediastinal staging in MPM.
A major limitation of this study is that EBUS and CM were generally performed on different patients and during different time periods. Nevertheless, radiographic staging and treatment indications remained more or less constant during the study period. We identified some noteworthy differences between CM and EBUS. Although CM had a diagnostic yield of 100% and more nodal stations were biopsied per patient compared with EBUS (median 3 vs 2.5), a greater proportion of patients had positive nodes on EBUS (34% vs 10%). Although there were no significant differences in clinical stage between patients having EBUS and those having CM, there was a slightly higher prevalence of nodal metastases in the EBUS group (65% vs 58%, p = not significant). Another difference between the two groups that may have influenced the results was the time interval between staging and definitive surgery was slightly longer for the CM group than patients staged with EBUS (mean 57 vs 31 days, p = 0.029). Interestingly, the lower diagnostic yield of EBUS (84%) did not lessen the ability to diagnose nodal metastases. In fact, nodes that were nondiagnostic were usually small (median diameter 5 mm), often contralateral, and no false negatives occurred in nodes that were nondiagnostic on EBUS. Although subtle imbalances in stage and nodal prevalence might have contributed to the observed differences in sensitivity of the two procedures, it is also possible that the greater yield of nodal metastases with EBUS might be due to inherent differences in nodal sampling between the two techniques. With EBUS, multiple passes of the needle (about 30 to 50) are taken through each node, perhaps resulting in a more representative sample. Although more nodal tissue is obtained with CM, often only a portion of a node is sampled. This is particularly true with respect to the subcarinal station where usually only the cephalad and anterior portions of the nodal packet are biopsied at CM. In addition, pathologic examination of biopsies from CM usually entails evaluation of one or two representative cuts through the tissue submitted, whereas cytologic specimens obtained with EBUS are smeared and completely visualized.
As this study is retrospective, caution should be exercised in interpreting differences between the sensitivities of EBUS and CT and PET. Although radiographic studies were compared only against formal pathology obtained with EBUS, CM, or surgery, there was no uniformity with respect to the radiographic protocols used, scanners, or radiologists interpreting the images. Nonetheless, both CT and PET had relatively low accuracy for determining nodal involvement, which is consistent with our earlier experience as well as that of others [5, 5, 21]. The EBUS allowed accurate measurement of nodes that were sampled and it is important to note that the median short axis dimension of tumor-positive nodes was 8 mm. Only 7 of 21 (33%) nodes were greater than 10 mm in size on short axis which would meet axial CT criteria for nodal enlargement. The smallest positive node biopsied measured 4 mm. This corroborates results from the Leicester group who found no difference in size between nodes that harbored tumor and those that did not, and may explain the relatively low sensitivity of CT and PET for detecting nodal metastases in MPM [22].
Like EBUS, esophageal endoscopic ultrasound has been effective for nodal staging in patients with NSCLC [8, 11, 26] but its use for staging MPM has only recently been reported [27]. Tournoy and colleagues [27] performed EUS on 32 patients with early stage MPM. Pathologic confirmation was possible for 21 patients and EUS diagnosed nodal metastases in 4 (19%). The authors reported a sensitivity of 80%, and similar to our experience with EBUS, mediastinoscopy did not increase the yield of positive nodes. Importantly, one patient had esophageal perforation secondary to EUS-FNA. In our series EUS was used on a selective basis. For patients with suspicious nodes in the periesophageal or infradiaphragmatic region we have found EUS-FNA to be very useful and metastatic disease was confirmed in 5 of 11 patients preoperatively. Esophageal ultrasound, however, is limited in its ability to biopsy right-sided paratracheal nodes, and as the recent report by Tournoy and colleagues shows, is not without potential morbidity. As combined use of EBUS and EUS-FNA has been reported to have higher accuracy than either procedure alone for the staging of lung cancer it seems reasonable to consider a combined approach for the staging of mesothelioma [28].
Despite the fact that EBUS was more sensitive than CM in preoperatively identifying patients with nodal metastases, a significant proportion of patients had occult N2 metastases at the time of surgery. For this reason we have modified our operative approach so that clinical node-negative patients who are candidates for EPP undergo intraoperative nodal assessment after initial extrapleural mobilization, but prior to pneumonectomy. If frozen section histology reveals nodal metastases then PD is preferred instead of EPP. The question of course arises whether preoperative invasive staging should be performed at all given its low yield. However, in light of the fact that patients with N2 metastases do not benefit from radical resection, it is our belief that every effort should be made to identify these patients prior to thoracotomy. Endobronchial ultrasound and esophageal ultrasound, though not perfect, improve our ability to accurately stage patients preoperatively.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
I don't see quite as many mesothelioma patients as you do, but I would argue that you should probably do both EUS [esophageal ultrasound] and EBUS [endobronchial ultrasound] routinely rather than selective use of EUS. After all, mesothelioma is a pleural disease, and those paraesophageal, inferior pulmonary ligament, level 8 and 9 nodes are equally likely to be involved.
So rather than use it selectively, EUS should be used routinely with EBUS as a combined procedure. I would like to hear your comments.
DR STELIGA: You know, when looking at this data, we've started to discuss that. Not only are they located in EUS accessible areas, but combined with the fact that many positive nodes are small and non-PET [positron emission tomography] avid would suggest that there may be a lot of positive nodes that we're missing.
We used EUS only selectively when it was lighting up hot on the PET scan or enlarged, and if we started using it universally in all these patients, we may detect more people with disease in those regions.
DR PHILIPPE R. NAFTEUX (Leuven, Belgium): Thank you for this very nice presentation. I have a question. If I caught the paper produced by your group a few years ago using extensive surgical staging, the sensitivity for mediastinoscopy was about 40%. The question is, why is it that low in this study?
The second part of the question is, have you ever used laparoscopy to stage the abdomen also for lymph node biopsies?
DR STELIGA: The first question, why is it so low in mediastinoscopy in our group. We found only 28%. We reported 36% sensitivity in our earlier series and 28% with mediastinoscopy now. The two studies were based on different time periods and different groups of patients, but I don't think anything had significantly changed between the two ways that we used the video-mediastinoscope to stage patients.
And your second question, I'm sorry, I didn't catch it.
DR NAFTEUX: Did you use laparoscopy to do lymph node biopsy in the belly?
DR STELIGA: No, we did not use laparoscopy to do lymph node biopsies in the abdomen. The way we do the laparoscopy is with two ports in the abdomen, one camera port above the umbilicus. We look around, look for any disease in the liver, in the peritoneum or transdiaphragmatic. And if we find no evidence of disease, we lavage the abdomen with saline, approximately a liter of saline and suction it out and send that for cytology.
It doesn't happen very often, but when that cytology does come back positive, those patients we've deemed have advanced disease, and we don't offer them resection.
So we do use laparoscopy and peritoneal lavage, but we don't do it for needle biopsy, or for nodal biopsy. Those nodes we can access with EUS, though. In a transesophageal, transgastric way, we've biopsied celiac nodes and left gastric nodes and found it by that method.
DR NAFTEUX: Thank you.
DR DANIEL L. MILLER (Atlanta, GA): I have one concern in regards to the EUS biopsies of stations 8 and 9, where you may get a high false positive rate. Because, as you know, the majority of the tumor burden for mesothelioma is usually located very low within the pleural cavity along the recess of the esophagus. I was wondering if that was the only location where you had N2 disease, in stations 8 and 9, which would be falsely positive and therefore up-stage possibly resectable patients?
DR STELIGA: I don't know of any data on that. But we have talked about the possibility of biopsying through the esophagus and either going beyond the node or biopsying tumor down there. And we watch very carefully where our needle goes when we pass it with a scope, and we do it under real-time image guidance. And we make sure that the needle lands just in the node (which would be seen as discrete round nodules) and not beyond it in the tumor, which looks like amorphous poorly marginated densities next to the esophagus.
DR MILLER: We have tried to do it a couple of times after a talc pleurodesis, and with the tumor burden and talc effect there was no plane to freely biopsy the node without being concerned for a false positive. So I would take caution in your station 8 and 9 data.
DR RAFAEL S. ANDRADE (Minneapolis, MN): I'm just struggling a little bit with the sensitivity issue: why is it so low for mediastinoscopy and so much better for EBUS?
Did you sample the same number of stations on average? Did you check how many stations you sampled with each technique? Are you being more meticulous with EBUS? Something just doesn't add up perfectly, and I don't know exactly what it is.
DR STELIGA: When we were looking at our data on this, we looked back at how we were doing our mediastinoscopy, and we had 153 nodes for 50 patients, a little over 3 nodes per patient. Those nodes were all adequately biopsied with mediastinoscopy and had adequate tissue on pathology.
Whether we are being more thorough with EBUS, the way we typically did the EBUS was to look at the tracheobronchial nodes, biopsy anything that we could more than 5 mm.
Now, we did have 15 nondiagnostic nodes that we tried to biopsy and missed. And we had fewer nodes actually biopsied with EBUS, but we ended up with more positives even though we were biopsying fewer nodes.
We were discussing different ways that we can biopsy the node. By biting the node with a little biopsy on mediastinoscopy, you're sampling just a very proximal part of that node.
With an EBUS, a needle transgresses the full dimension of the node, and even though you're only taking a smaller amount of tissue, maybe the needle is getting a more representative sample of that node. It's difficult to tell. I'm not sure why we were much higher with the EBUS sensitivity, but that may explain it.
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Nakas, D. Waller, K. Lau, C. Richards, and S. Muller The new case for cervical mediastinoscopy in selection for radical surgery for malignant pleural mesothelioma Eur J Cardiothorac Surg, July 1, 2012; 42(1): 72 - 76. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. S. Ettinger, W. Akerley, H. Borghaei, A. Chang, R. T. Cheney, L. R. Chirieac, T. A. D'Amico, T. L. Demmy, A. K. P. Ganti, R. Govindan, et al. Malignant Pleural Mesothelioma J Natl Compr Canc Netw, January 1, 2012; 10(1): 26 - 41. [Full Text] [PDF]
|
|
|
|
|
|
A. Nakas and D. A. Waller Reply to West and Walker Eur J Cardiothorac Surg, May 1, 2011; 39(5): 800 - 801. [Full Text] [PDF]
|
|
|
|
 |
|
 A. R. Haas, A. Vachani, and D. H. Sterman Advances in Diagnostic Bronchoscopy , September 1, 2010; 182(5): 589 - 597. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
