- Email: [email protected]
Extended cervical mediastinoscopy
J
THoRAc CARDIOVASC SURG
1987;94:673-8
Extended cervical mediastinoscopy A single staging procedure for bronchogenic carcinoma of the left upper lobe Despite a common misconception, bronchogenic carcinoma of the left upper lobe frequently metastasizes to lymph nodes not only in the anterior mediastinum (para-aortic and subaortic) but also in the superior mediastinum. Anterior (parasternal) mediastinotomy can be used to assess only the former compartment. This procedure alone, if not done in conjunction with standard cenical mediastinoscopy, will fail to disclose technically unresectable N2 or N3 disease of the left upper lobe involving the superior mediastinum. We have developeda technique to explore and sample nodes from both regions by extending a standard cenical mediastinoscopy, eliminating the need for a second incision when the anterior mediastinal compartment requires assessment. We have prospectively analyzed the first 100 procedures that we performed. This technique has been found to be accurate and exceptionally safe with one superficial wound infection as the only complication. We can recommend this single staging procedure for preoperative assessment of bronchogenic carcinomas of the left upper lobe.
Robert J. Ginsberg, MD, FRCS(C), Thomas W. Rice, MD, FRCS(C)* (by invitation), Melvyn Goldberg, MD, FRCS(C), Paul F. Waters, MD, FRCS(C), and Beverly J. Schmocker, RN (by invitation), Toronto, Ontario, Canada
Anterior mediastinotomy has been suggested as the invasive procedure of choice for preoperative staging of bronchogenic carcinoma of the left upper lobe." 2 However, a significant number of these tumors metastasize to the nodes of the superior mediastinum (Figs. 1 and 2), inaccessible by anterior (parasternal) mediastinotomy and accessible only by standard cervical mediastinoscopy.3.4 Left upper lobe tumors, with superior mediastinal node spread, are considered inoperable by most authorities. For bronchogenic carcinomas arising in the left upper lobe to be fully staged, both the superior and anterior mediastinal node-bearing areas must be sampled. Both standard cervical mediastinoscopy and anterior mediastinotomy are necessary to accomplish this. Anterior mediastinotomy does not always allow full assessment of the subaortic window, because frequently From the Division of Thoracic Surgery, Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada Read at the Sixty-seventh Annual Meeting of The American Association for Thoracic Surgery, Chicago, Ill., April 6-8, 1987. Address for reprints: Dr. Robert J. Ginsberg, Mount Sinai Hospital, Ste. 451, 600 University Ave., Toronto, Ontario, M5G 1X5 Canada. 'Present address: The Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH 44106.
one is unable to obtain adequate lymph node tissue. Significant complications include bleeding, damage to the internal mammary vessels, pneumothorax, and postoperative wound problems such as pain and infection. To avoid the problems of anterior mediastinotomy and to allow for assessment of both superior and anterior mediastinal lymph node areas, we have used a single surgical procedure, extended cervical mediastinoscopy, in staging left upper lobe tumors. A similar technique has been suggested as a possible staging approach by
others.-" Method In patients with left upper lobe tumors, after bronchoscopy has been performed, a standard cervical mediastinoscopy is done. Samples are taken from nodes in the subcarinal, left tracheobronchial angle, and left paratracheal groups, as well as from contralateral paratracheal nodes (Fig. 3). After the assessment of the superior mediastinum has been completed, an indexfinger is reintroduced into the pretracheal plane. The innominate artery is palpated and is traced to the aortic arch, where the "innominate triangle" is identified (Fig. 4, A), This space, enclosed in fascia, is bounded medially by the innominate artery, laterally by the left common carotid artery, anteriorly by the left innominate vein, and posteroinferiorly by the aortic arch. The fascia is opened with blunt digital dissection. A finger is slipped over the aortic arch, posterior to the innominate vein, and the anterolateralpara-aortic region is
673
674
Ginsberg et al.
The Journal of Thoracic and Cardiovascular Surgery
Fig. 2. Nodes in anterior mediastinum frequently involved in left upper lobe tumors.
Fig. 1. Nodes in superior mediastinum frequently involved in left upper lobe tumors.
inspected digitally (Fig. 4, B). The mediastinoscope is then introduced carefully through this window (Fig. 5). Para-aortic lymph nodes are usually easily identified and biopsy specimens can be obtained. The mediastinoscope is advanced further over the aortic arch, and the lymph nodes of the subaortic window are viewed. Biopsy specimens of this area are taken (Figs. 6 to 8). With the scope used as a probe, fixation of the subaortic window may be appreciated, and a primary tumor invading the mediastinum can be identified. We have not attempted this procedure in patients with aneurysmal dilatation or with severe calcification of the aortic arch. To gain familiarity with this approach, we used anterior mediastinotomy simultaneously in the initial few patients to verify the position of the dissection and the mediastinoscope at extended cervical mediastinoscopy. In four of our initial patients, we did not take samples from normal-appearing nodes for biopsy.
Results We have performed this staging procedure in over 125 patients with bronchogenic carcinoma of the left upper lobe and have analyzed the first 100 consecutive patients. There were 67 men and 33 women, their ages ranging from 30 to 83 years (median 61, mean 62.5). At extended cervical mediastinoscopy, 20 patients were found to have nodal spread to the superior or anterior mediastinum, or to both areas. A further nine
patients were found to have microscopic spread to mediastinal lymph nodes at subsequent thoracotomy (Table I), eight in the anterior mediastinum and one in the subcarinal space. Of 15 patients whose condition was deemed inoperable because of superior mediastinal involvement, eight had extension of the disease to both the anterior and superior mediastinum and seven to the superior mediastinum alone. Eight patients with "false negative" results of mediastinoscopy were found to have microscopic N2 disease in the anterior mediastinum only during the thoracotomy. In all eight the lesions were completely resectable. Two of these patients had no biopsy of normal-sized nodes, and in six the involved nodes would have been inaccessible by either extended cervical mediastinoscopy or anterior mediastinotomy, because they lay just under the vagus nerve at the level of the recurrent nerve takeoff. All "false negative" extended mediastinoscopies resulted in complete resections at thoracotomy. A total of 75 thoracotomies were performed, and in only one patient with invasion of the posterior portion of the aortic arch (T4) was complete resection not possible after extended cervical mediastinoscopy. We analyzed the relative value of chest x-ray assessment in determining ultimate N2 status. Before staging and thoracotomy, 29 patients were judged to have either an enlarged hilum or mediastinal involvement on the routine chest roentgenogram. Only 13 of these were
Volume 94 Number 5 November 1987
Extended cervical mediastinoscopy
675
Fig. 3. Standard cervical mediastinoscopy illustrating subcarinal node biopsy.
Table I. The distribution and method of diagnosis of N2 disease found in 100 consecutive patients with bronchogenic carcinoma of the left upper lobe Superior mediastinum diseased
Anterior mediastinum diseased
Both compartments diseased
Total diseased
Extended mediastinoscopy Thoracotomy
7
5
8
20
I
8
0
9
Total
8
13
8
29
subsequently found to have mediastinal N2 or N3 disease. Of the 71 patients with no abnormalities on chest roentgenograms, 16 were found to have N2 disease (Table 11). A single superficial wound infection was the only complication in 100 consecutive patients. This quickly resolved with local drainage and antibiotic therapy.
Discussion We and others have shown that mediastinal lymph node spread in left upper lobe tumors can involve the anterior or superior mediastinum.v' Adequate staging requires exploration of both of these areas. However, others have suggested that only anterior mediastinotomy is necessary for invasive staging. I. 2 In this present series of 100 consecutive patients, 29% had N2 disease. Fifteen percent of the patients had involvement of the superior mediastinal nodes, which certainly would not have been detected by anterior mediastinotomy without concomitant cervical mediasti-
Table II. Mediastinal nodal assessment by chest roentgenography and extended mediastinoscopy (N = 100 patients) Chest roentgenography
Sensitivity (%) Specificity (%) Positive predictive value (%) Negative predictive value (%) Accuracy (%)
45
78
Extended mediastinoscopy
69
45
100 100
78
89
68
91
noscopy. The condition of all 15 patients was deemed inoperable because of the nodal disease in the superior mediastinum. The technique of extended cervical mediastinoscopy allows complete staging of bronchogenic carcinomas of the left upper lobe. It avoids a second incision, which has significant morbidity. The test has acceptable specificity
676
The Journal of Thoracic and Cardiovascular Surgery
Ginsberg et aJ.
Anterior Vrew'Digital Perforation
A
B
Fig. 4. A, Digital palpation of "innominate triangle" (see text) through standard cervical incision.B, Index finger perforates fascia of "innominate triangle" and palpates para-aortic area.
Anterior Mediastinal Node .~
E ~,-.-l
f
~V Sub Aortic Nodes
Fig. 5. Mediastinoscope can inspect and sample para-aortic and subaortic lymph node area.
(100%), positive predictive value (100%), and negative predictive value (89%). The sensitivity of this test (the ability to predict lymph node spread to the mediastinum) is 69%, similar to that obtained by anterior mediastinotomy.' This somewhat low sensitivity would have been improved if biopsies had been done in all cases. All patients who were ultimately found to have N2 disease only at thoracotomy were able to have complete resection. Extended cervical mediastinoscopy can be a difficult procedure to learn. Concomitant use of anterior mediastinotomy until one is adept at the procedure is recommended. Although fixation of the subaortic window by tumor can be assessed, this may not be quite as accurate if one uses bimanual palpation through two incisions. The finding of a "normal" hilum or mediastinum on the routine chest roentgenogram does not preclude the necessity for invasive staging. Although computed tomographic scans and 55-degree oblique tomograms were not performed, there is evidence to suggest that these studies improve the accuracy only minimally over that of chest roentgenography.' We conclude that both the superior mediastinum and the anterior mediastinum must be examined for com-
Volume 94 Number 5 November 1987
Extended cervical mediastinoscopy
677
Fig. 6. Mediastinoscope is used to inspect (right) and sample (left) subaortic lymph nodes.
Fig. 7. Surgical clip (arrows) has been placed in area of subaortic node biopsy.
plete staging of bronchogenic carcinomas of the left upper lobe. Initial assessment of the superior mediastinum should be done by standard cervical mediastinoscope. If this examination shows abnormalities, then the anterior mediastinum need not be assessed, unless findings there will change the management decision. If the superior mediastinum is judged to be normal, then the anterior mediastinum can be staged simply by extending the mediastinoscopy as described, without a second incision. Microscopic N2 disease found in the anterior mediastinal nodes does not necessarily preclude a potentially curative resection." However, fixed extensive disease in this area usually contraindicates operation.
We find that extended cervical mediastinoscopy is a safe, reliable, highly specific, and relatively sensitive operative procedure. It avoids the necessity of anterior mediastinotomy and allows assessment of the superior mediastinum, anterior mediastinum, and the subaortic window. REFERENCES I. Jolly PC, Li W, Anderson RP. Anterior and cervical mediastinoscopy for determining operability and predicting resectability in lung cancer. J THORAC CARDIOVASC SURG 1980;79:366-71. 2. Page AP, Nakhle G, Mercier C, et al. Surgical treatment of bronchogenic carcinoma: the importance of staging in evaluating late survival. Can J Surg 1987;30:96-9.
6 7 8 Ginsberg et al.
Fig. 8. Computed tomographic scan showing surgical clip (arrow) left in the area of biopsy of an inoperable subaortic mass. 3. Deneff EG, Lacquet LM, Gyselen A. Cervical mediastinoscopy and anterior mediastinotomoy in patients with lung cancer and radiologically normal mediastinum. Eur J Respir Dis 1983;64:613-9. 4. Luke WP, Pearson FG, Todd TRJ, et al. Prospective evaluation of mediastinoscopy for assessment of carcinoma of the lung. J THORAC CARDIOVASC SURG 1986;91:553-6. 5. Specht VG. Erweiteite mediastinoskopie. Thoraxchirurgie 1965;13:401-7. 6. Kirschner P. "Extended" mediastinoscopy. In: Jepsson 0, Ruhbek-Sorensen, H, eds. Mediastinoscopy. Odense, Denmark: Odense University Press, 1971. 7. McKenna RJ, Libshitz HI, Mountain CE, McMurtrey MJ. Roentgenographic evaluation of mediastinal nodes for preoperative assessment in lung cancer. Chest 1985;88:20610. 8. Patterson GA, Piazza 0, Pearson FG, et al. Significance of metastatic disease in subaortic lymph nodes. Ann Thorac Surg 1987;43: 155-9.
Discussion Dr. David B. Skinner (Chicago. Ill.). The papers by Drs. Ginsberg and Patterson [page 679] and their associates both come from the same section of thoracic surgery, albeit on opposite sides of the street. Dr. Ginsberg, I wonder if you might put the two in perspective for us. It seems to me that you have introduced a very helpful and ingenious way to sample the left anterior mediastinum, but when should it be used outside of study circumstances? For example, if you do find a 2 em node in that area on the computed tomographic or the magnetic resonance imaging scan, is it really necessary to perform a biopsy? Since you are resecting the small nodes in
The Journal of Thoracic and Cardiovascular Surgery
the anterior mediastinum anyway, it is not really helpful to know that they are abnormal before the resection. Dr. Willard A. Fry (Evanston. Ill.). I think Dr. Ginsberg has demonstrated how careful he was in initially developing the technique. I wonder how necessary this operation is if you take into consideration some of his group's" previously reported work: the situation in which the only nodes that are diseased are those in Groups V and VI and the paratracheal and subcarinal nodes are normal. The Toronto group has published that the results of surgical resection in that situation are really not so bad. Nobody with N2 disease has a wonderful prognosis, but on the basis of that previous study from Toronto, we have cut down quite a bit on staging of disease in the anterior mediastinum by mediastinotomy and prefer standard mediastinoscopy. If you have demonstrated, Dr. Ginsberg, that the paratracheal and subcarinal nodes are normal, then why bother to sample the nodes under the aortic arch (Groups V and VI) in otherwise straightforward cases? Isn't mediastinoscopy alone sufficient for proper mediastinal node staging before thoracotomy? Dr. Ginsberg (Closing). Both Dr. Fry and Dr. Skinner have similar questions: When do we do an extended mediastinoscopy and how do we act upon it? Our philosophy is based on the premise that lung cancer can be cured only if it can be completely resected. With left upper lobe tumors, 15% of patients can be eliminated by cervical mediastinoscopy alone if they have involved nodes in the superior mediastinum. If the computed tomographic scan suggests that there is an enlarged hilum that may be fixed, and if cervical mediastinoscopy fails to reveal metastases, then certainly one has to do something else to make sure that the tumor can be completely resected. Our results suggest that extended cervical mediastinoscopy is an appropriate examination to explore the subaortic space. On occasion, especially with adenocarcinoma, we have found multiple small involved nodes beyond the point of total resection in the anterior mediastinum despite negative computed tomographic results. We still perform extended mediastinoscopy on all patients to avoid thoracotomy when they have multiple metastatic node involvement or fixed 'nodes such that complete resection is important. The other advantage of mediastinoscopy is to allow us to stop and do a total staging for metastases if we unexpectedly identify stage III disease at this examination. It has not been our practise to do routine organ scans for metastatic disease if the patient is clinically in stage I or II. However, if we do identify resectable stage III disease during mediastinoscopy, we do not proceed with a thoracotomy at that time but stop and do complete organ scanning, including brain, bone, liver, and adrenals. I think that probably answers both questions. We perform mediastinoscopy on everyone. We tend to do it just before a thoracotomy if routine chest roentgenograms suggest that the results of mediastinoscopy will be normal.
THoRAc CARDIOVASC SURG
1987;94:673-8
Extended cervical mediastinoscopy A single staging procedure for bronchogenic carcinoma of the left upper lobe Despite a common misconception, bronchogenic carcinoma of the left upper lobe frequently metastasizes to lymph nodes not only in the anterior mediastinum (para-aortic and subaortic) but also in the superior mediastinum. Anterior (parasternal) mediastinotomy can be used to assess only the former compartment. This procedure alone, if not done in conjunction with standard cenical mediastinoscopy, will fail to disclose technically unresectable N2 or N3 disease of the left upper lobe involving the superior mediastinum. We have developeda technique to explore and sample nodes from both regions by extending a standard cenical mediastinoscopy, eliminating the need for a second incision when the anterior mediastinal compartment requires assessment. We have prospectively analyzed the first 100 procedures that we performed. This technique has been found to be accurate and exceptionally safe with one superficial wound infection as the only complication. We can recommend this single staging procedure for preoperative assessment of bronchogenic carcinomas of the left upper lobe.
Robert J. Ginsberg, MD, FRCS(C), Thomas W. Rice, MD, FRCS(C)* (by invitation), Melvyn Goldberg, MD, FRCS(C), Paul F. Waters, MD, FRCS(C), and Beverly J. Schmocker, RN (by invitation), Toronto, Ontario, Canada
Anterior mediastinotomy has been suggested as the invasive procedure of choice for preoperative staging of bronchogenic carcinoma of the left upper lobe." 2 However, a significant number of these tumors metastasize to the nodes of the superior mediastinum (Figs. 1 and 2), inaccessible by anterior (parasternal) mediastinotomy and accessible only by standard cervical mediastinoscopy.3.4 Left upper lobe tumors, with superior mediastinal node spread, are considered inoperable by most authorities. For bronchogenic carcinomas arising in the left upper lobe to be fully staged, both the superior and anterior mediastinal node-bearing areas must be sampled. Both standard cervical mediastinoscopy and anterior mediastinotomy are necessary to accomplish this. Anterior mediastinotomy does not always allow full assessment of the subaortic window, because frequently From the Division of Thoracic Surgery, Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada Read at the Sixty-seventh Annual Meeting of The American Association for Thoracic Surgery, Chicago, Ill., April 6-8, 1987. Address for reprints: Dr. Robert J. Ginsberg, Mount Sinai Hospital, Ste. 451, 600 University Ave., Toronto, Ontario, M5G 1X5 Canada. 'Present address: The Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH 44106.
one is unable to obtain adequate lymph node tissue. Significant complications include bleeding, damage to the internal mammary vessels, pneumothorax, and postoperative wound problems such as pain and infection. To avoid the problems of anterior mediastinotomy and to allow for assessment of both superior and anterior mediastinal lymph node areas, we have used a single surgical procedure, extended cervical mediastinoscopy, in staging left upper lobe tumors. A similar technique has been suggested as a possible staging approach by
others.-" Method In patients with left upper lobe tumors, after bronchoscopy has been performed, a standard cervical mediastinoscopy is done. Samples are taken from nodes in the subcarinal, left tracheobronchial angle, and left paratracheal groups, as well as from contralateral paratracheal nodes (Fig. 3). After the assessment of the superior mediastinum has been completed, an indexfinger is reintroduced into the pretracheal plane. The innominate artery is palpated and is traced to the aortic arch, where the "innominate triangle" is identified (Fig. 4, A), This space, enclosed in fascia, is bounded medially by the innominate artery, laterally by the left common carotid artery, anteriorly by the left innominate vein, and posteroinferiorly by the aortic arch. The fascia is opened with blunt digital dissection. A finger is slipped over the aortic arch, posterior to the innominate vein, and the anterolateralpara-aortic region is
673
674
Ginsberg et al.
The Journal of Thoracic and Cardiovascular Surgery
Fig. 2. Nodes in anterior mediastinum frequently involved in left upper lobe tumors.
Fig. 1. Nodes in superior mediastinum frequently involved in left upper lobe tumors.
inspected digitally (Fig. 4, B). The mediastinoscope is then introduced carefully through this window (Fig. 5). Para-aortic lymph nodes are usually easily identified and biopsy specimens can be obtained. The mediastinoscope is advanced further over the aortic arch, and the lymph nodes of the subaortic window are viewed. Biopsy specimens of this area are taken (Figs. 6 to 8). With the scope used as a probe, fixation of the subaortic window may be appreciated, and a primary tumor invading the mediastinum can be identified. We have not attempted this procedure in patients with aneurysmal dilatation or with severe calcification of the aortic arch. To gain familiarity with this approach, we used anterior mediastinotomy simultaneously in the initial few patients to verify the position of the dissection and the mediastinoscope at extended cervical mediastinoscopy. In four of our initial patients, we did not take samples from normal-appearing nodes for biopsy.
Results We have performed this staging procedure in over 125 patients with bronchogenic carcinoma of the left upper lobe and have analyzed the first 100 consecutive patients. There were 67 men and 33 women, their ages ranging from 30 to 83 years (median 61, mean 62.5). At extended cervical mediastinoscopy, 20 patients were found to have nodal spread to the superior or anterior mediastinum, or to both areas. A further nine
patients were found to have microscopic spread to mediastinal lymph nodes at subsequent thoracotomy (Table I), eight in the anterior mediastinum and one in the subcarinal space. Of 15 patients whose condition was deemed inoperable because of superior mediastinal involvement, eight had extension of the disease to both the anterior and superior mediastinum and seven to the superior mediastinum alone. Eight patients with "false negative" results of mediastinoscopy were found to have microscopic N2 disease in the anterior mediastinum only during the thoracotomy. In all eight the lesions were completely resectable. Two of these patients had no biopsy of normal-sized nodes, and in six the involved nodes would have been inaccessible by either extended cervical mediastinoscopy or anterior mediastinotomy, because they lay just under the vagus nerve at the level of the recurrent nerve takeoff. All "false negative" extended mediastinoscopies resulted in complete resections at thoracotomy. A total of 75 thoracotomies were performed, and in only one patient with invasion of the posterior portion of the aortic arch (T4) was complete resection not possible after extended cervical mediastinoscopy. We analyzed the relative value of chest x-ray assessment in determining ultimate N2 status. Before staging and thoracotomy, 29 patients were judged to have either an enlarged hilum or mediastinal involvement on the routine chest roentgenogram. Only 13 of these were
Volume 94 Number 5 November 1987
Extended cervical mediastinoscopy
675
Fig. 3. Standard cervical mediastinoscopy illustrating subcarinal node biopsy.
Table I. The distribution and method of diagnosis of N2 disease found in 100 consecutive patients with bronchogenic carcinoma of the left upper lobe Superior mediastinum diseased
Anterior mediastinum diseased
Both compartments diseased
Total diseased
Extended mediastinoscopy Thoracotomy
7
5
8
20
I
8
0
9
Total
8
13
8
29
subsequently found to have mediastinal N2 or N3 disease. Of the 71 patients with no abnormalities on chest roentgenograms, 16 were found to have N2 disease (Table 11). A single superficial wound infection was the only complication in 100 consecutive patients. This quickly resolved with local drainage and antibiotic therapy.
Discussion We and others have shown that mediastinal lymph node spread in left upper lobe tumors can involve the anterior or superior mediastinum.v' Adequate staging requires exploration of both of these areas. However, others have suggested that only anterior mediastinotomy is necessary for invasive staging. I. 2 In this present series of 100 consecutive patients, 29% had N2 disease. Fifteen percent of the patients had involvement of the superior mediastinal nodes, which certainly would not have been detected by anterior mediastinotomy without concomitant cervical mediasti-
Table II. Mediastinal nodal assessment by chest roentgenography and extended mediastinoscopy (N = 100 patients) Chest roentgenography
Sensitivity (%) Specificity (%) Positive predictive value (%) Negative predictive value (%) Accuracy (%)
45
78
Extended mediastinoscopy
69
45
100 100
78
89
68
91
noscopy. The condition of all 15 patients was deemed inoperable because of the nodal disease in the superior mediastinum. The technique of extended cervical mediastinoscopy allows complete staging of bronchogenic carcinomas of the left upper lobe. It avoids a second incision, which has significant morbidity. The test has acceptable specificity
676
The Journal of Thoracic and Cardiovascular Surgery
Ginsberg et aJ.
Anterior Vrew'Digital Perforation
A
B
Fig. 4. A, Digital palpation of "innominate triangle" (see text) through standard cervical incision.B, Index finger perforates fascia of "innominate triangle" and palpates para-aortic area.
Anterior Mediastinal Node .~
E ~,-.-l
f
~V Sub Aortic Nodes
Fig. 5. Mediastinoscope can inspect and sample para-aortic and subaortic lymph node area.
(100%), positive predictive value (100%), and negative predictive value (89%). The sensitivity of this test (the ability to predict lymph node spread to the mediastinum) is 69%, similar to that obtained by anterior mediastinotomy.' This somewhat low sensitivity would have been improved if biopsies had been done in all cases. All patients who were ultimately found to have N2 disease only at thoracotomy were able to have complete resection. Extended cervical mediastinoscopy can be a difficult procedure to learn. Concomitant use of anterior mediastinotomy until one is adept at the procedure is recommended. Although fixation of the subaortic window by tumor can be assessed, this may not be quite as accurate if one uses bimanual palpation through two incisions. The finding of a "normal" hilum or mediastinum on the routine chest roentgenogram does not preclude the necessity for invasive staging. Although computed tomographic scans and 55-degree oblique tomograms were not performed, there is evidence to suggest that these studies improve the accuracy only minimally over that of chest roentgenography.' We conclude that both the superior mediastinum and the anterior mediastinum must be examined for com-
Volume 94 Number 5 November 1987
Extended cervical mediastinoscopy
677
Fig. 6. Mediastinoscope is used to inspect (right) and sample (left) subaortic lymph nodes.
Fig. 7. Surgical clip (arrows) has been placed in area of subaortic node biopsy.
plete staging of bronchogenic carcinomas of the left upper lobe. Initial assessment of the superior mediastinum should be done by standard cervical mediastinoscope. If this examination shows abnormalities, then the anterior mediastinum need not be assessed, unless findings there will change the management decision. If the superior mediastinum is judged to be normal, then the anterior mediastinum can be staged simply by extending the mediastinoscopy as described, without a second incision. Microscopic N2 disease found in the anterior mediastinal nodes does not necessarily preclude a potentially curative resection." However, fixed extensive disease in this area usually contraindicates operation.
We find that extended cervical mediastinoscopy is a safe, reliable, highly specific, and relatively sensitive operative procedure. It avoids the necessity of anterior mediastinotomy and allows assessment of the superior mediastinum, anterior mediastinum, and the subaortic window. REFERENCES I. Jolly PC, Li W, Anderson RP. Anterior and cervical mediastinoscopy for determining operability and predicting resectability in lung cancer. J THORAC CARDIOVASC SURG 1980;79:366-71. 2. Page AP, Nakhle G, Mercier C, et al. Surgical treatment of bronchogenic carcinoma: the importance of staging in evaluating late survival. Can J Surg 1987;30:96-9.
6 7 8 Ginsberg et al.
Fig. 8. Computed tomographic scan showing surgical clip (arrow) left in the area of biopsy of an inoperable subaortic mass. 3. Deneff EG, Lacquet LM, Gyselen A. Cervical mediastinoscopy and anterior mediastinotomoy in patients with lung cancer and radiologically normal mediastinum. Eur J Respir Dis 1983;64:613-9. 4. Luke WP, Pearson FG, Todd TRJ, et al. Prospective evaluation of mediastinoscopy for assessment of carcinoma of the lung. J THORAC CARDIOVASC SURG 1986;91:553-6. 5. Specht VG. Erweiteite mediastinoskopie. Thoraxchirurgie 1965;13:401-7. 6. Kirschner P. "Extended" mediastinoscopy. In: Jepsson 0, Ruhbek-Sorensen, H, eds. Mediastinoscopy. Odense, Denmark: Odense University Press, 1971. 7. McKenna RJ, Libshitz HI, Mountain CE, McMurtrey MJ. Roentgenographic evaluation of mediastinal nodes for preoperative assessment in lung cancer. Chest 1985;88:20610. 8. Patterson GA, Piazza 0, Pearson FG, et al. Significance of metastatic disease in subaortic lymph nodes. Ann Thorac Surg 1987;43: 155-9.
Discussion Dr. David B. Skinner (Chicago. Ill.). The papers by Drs. Ginsberg and Patterson [page 679] and their associates both come from the same section of thoracic surgery, albeit on opposite sides of the street. Dr. Ginsberg, I wonder if you might put the two in perspective for us. It seems to me that you have introduced a very helpful and ingenious way to sample the left anterior mediastinum, but when should it be used outside of study circumstances? For example, if you do find a 2 em node in that area on the computed tomographic or the magnetic resonance imaging scan, is it really necessary to perform a biopsy? Since you are resecting the small nodes in
The Journal of Thoracic and Cardiovascular Surgery
the anterior mediastinum anyway, it is not really helpful to know that they are abnormal before the resection. Dr. Willard A. Fry (Evanston. Ill.). I think Dr. Ginsberg has demonstrated how careful he was in initially developing the technique. I wonder how necessary this operation is if you take into consideration some of his group's" previously reported work: the situation in which the only nodes that are diseased are those in Groups V and VI and the paratracheal and subcarinal nodes are normal. The Toronto group has published that the results of surgical resection in that situation are really not so bad. Nobody with N2 disease has a wonderful prognosis, but on the basis of that previous study from Toronto, we have cut down quite a bit on staging of disease in the anterior mediastinum by mediastinotomy and prefer standard mediastinoscopy. If you have demonstrated, Dr. Ginsberg, that the paratracheal and subcarinal nodes are normal, then why bother to sample the nodes under the aortic arch (Groups V and VI) in otherwise straightforward cases? Isn't mediastinoscopy alone sufficient for proper mediastinal node staging before thoracotomy? Dr. Ginsberg (Closing). Both Dr. Fry and Dr. Skinner have similar questions: When do we do an extended mediastinoscopy and how do we act upon it? Our philosophy is based on the premise that lung cancer can be cured only if it can be completely resected. With left upper lobe tumors, 15% of patients can be eliminated by cervical mediastinoscopy alone if they have involved nodes in the superior mediastinum. If the computed tomographic scan suggests that there is an enlarged hilum that may be fixed, and if cervical mediastinoscopy fails to reveal metastases, then certainly one has to do something else to make sure that the tumor can be completely resected. Our results suggest that extended cervical mediastinoscopy is an appropriate examination to explore the subaortic space. On occasion, especially with adenocarcinoma, we have found multiple small involved nodes beyond the point of total resection in the anterior mediastinum despite negative computed tomographic results. We still perform extended mediastinoscopy on all patients to avoid thoracotomy when they have multiple metastatic node involvement or fixed 'nodes such that complete resection is important. The other advantage of mediastinoscopy is to allow us to stop and do a total staging for metastases if we unexpectedly identify stage III disease at this examination. It has not been our practise to do routine organ scans for metastatic disease if the patient is clinically in stage I or II. However, if we do identify resectable stage III disease during mediastinoscopy, we do not proceed with a thoracotomy at that time but stop and do complete organ scanning, including brain, bone, liver, and adrenals. I think that probably answers both questions. We perform mediastinoscopy on everyone. We tend to do it just before a thoracotomy if routine chest roentgenograms suggest that the results of mediastinoscopy will be normal.