- Email: [email protected]
Is there a role for routine mediastinoscopy in patients with peripheral T1 lung cancers?
Is There a Role for Routine Mediastinoscopy in Patients with Peripheral T1 Lung Cancers? Robert W. Tahara, MD, Omaha, Nebraska, Rudy P. Lackner, MD, L. Michael Graver, MD, New Hyde Park, New York
The role of surgical staging of patients with non-small cell lung cancer (NSCLC) continues to evolve. This report describes our findings utilizing routine cervical mediastinoscopy in the evaluation of peripheral T1 (<3 cm) lung tumors. METHODS: Retrospectively 30 patients with peripheral T1 lesions and CT scans negative for pathologic adenopathy were identified over a 3-year period. Cervical mediastinoscopy was performed prior to VATS/thoracotomy during the same operative session. RESULTS: Mediastinoscopy was performed in 29 of 30 patients. For patients with malignancy (27 of 30), 3 of 27 (11%) had mediastinoscopy positive for malignancy and no further resection performed. Overall the subgroup of patients with bronchogenic carcinomas had positive mediastinal involvement identified in 5 of 24 (21%) after mediastinoscopy or complete resection. CONCLUSIONS: A significant number of patients with small peripheral lung cancers harbor radiographically occult lymph node involvement. Mediastinoscopy facilitates identification of patients with regionally advanced disease prior to resection, allowing neoadjuvant therapy and avoiding unnecessary resections. Am J Surg. 2000;180: 488 – 492. © 2001 by Excerpta Medica, Inc. BACKGROUND:
T
he approach to staging lung cancer continues to evolve. Although there is consensus regarding the definition of each pathologic stage and the pivotal role mediastinal lymph node involvement plays in both prognosis and therapeutic decision making,1–5 considerable variance remains in how preoperative assessment of the mediastinum is accomplished. The last 2 decades have brought an increasing emphasis on the radiographic evaluation of the mediastinum in the form of computed tomography (CT) and, more recently, positron emission (PET)
From the Department of Surgery (RWT), University of Nebraska Medical Center, Omaha, Nebraska; and the Department of Surgery (RPL, LMG), Division of Cardiothoracic Surgery, Long Island Jewish Medical Center, New Hyde Park, New York. Requests for reprints should be addressed to Rudy P. Lackner, MD, Long Island Jewish Medical Center, Cardiothoracic Surgery, Suite 2123, 270-05 76th Avenue, New Hyde Park, New York 11040. Presented at the 52nd Annual Meeting of the Southwestern Surgical Congress, Colorado Springs, Colorado, April 9 –12, 2000.
488
© 2001 by Excerpta Medica, Inc. All rights reserved.
scanning,6 with emphasis on replacing or minimizing invasive sampling of the mediastinal nodes. This has occurred despite the proven accuracy, safety, and relative minimally invasive nature of cervical mediastinoscopy. In fact, when reviewing the literature mediastinoscopy is usually the standard to which these other modalities and strategies are compared. Commonly used preoperative staging strategies range from CT or PET alone, to selective mediastinoscopy or radiographically guided biopsy, to routine mediastinoscopy, mediastinotomy, and thoracoscopy.3,7,8 Bronchoscopy and, more recently, endobronchial ultrasound have also been used to assist in staging. These are employed in an effort to identify those patients already having regionally advanced disease and in whom surgical resection would not currently be the treatment of choice. Finding these patients preoperatively allows the use of neoadjuvant chemotherapy and/or radiation therapy. Conversely, patients who can be shown to have only local involvement benefit from a markedly better rate of cure with surgical resection.9 This is especially true when evaluating patients with small peripheral lung tumors. Patients with T1 (⬍3 cm) tumors and no superior mediastinal involvement or distant metastases are by definition stage I and should have the best chance for cure with surgical resection (67% at 5 years). The survival for patients in this category who are found to have superior mediastinal involvement, and hence N2 disease (stage IIIA), would drop markedly and is only 20% to 25% at 5 years.1,2 The optimum approach to best exclude mediastinal node metastases and how to best use these various strategies remains quite controversial, although several large series have recently been reported in the European literature that would seem to demonstrate the superiority of routine mediastinoscopy.10 –12 This report details our experience using routine cervical mediastinoscopy in assessing the subset of patients with peripheral T1 lung cancers.
PATIENTS AND METHODS Patients A retrospective review from November 1994 through September 1998 identified 30 patients with peripheral T1 lung tumors. All patients had CT scans performed, which showed mediastinal lymph nodes ⬍1 cm in size. Percutaneous needle biopsies or flexible bronchoscopy with transbronchial biopsies were not standardly performed in the preoperative assessment of solitary pulmonary nodules. Cervical mediastinoscopy was routinely performed in all patients to accurately stage the mediastinum prior to resection. Patients with negative nodes proceeded on to video-assisted thoracic surgery (VATS) or thoracotomy, or 0002-9610/00/$–see front matter PII S0002-9610(00)00509-2
ROUTINE MEDIASTINOSCOPY IN PERIPHERAL T1 LUNG CANCER/TAHARA ET AL
TABLE I Summary of Final Pathologic Staging Final Pathologic Stage IA IIA IIIA Limited stage SCLC
TNM Classification
Number of Patients
T1N0M0 T2N0M0* T1N1M0 T1N2M0 T2N2M0*
13 5 1 3 1 1
* Patient had a history of Hodgkin’s disease, prior chest radiotherapy, and highly calcified nodes by radiography.
Figure 1. Lymph node stations. Modified from Mountain and Dresler.2
both, to perform resection of the lesion. Those found to have positive mediastinal nodes were offered standard chemoradiation or the opportunity to undergo neoadjuvant therapy followed by resection. In node-negative patients resection was performed during the same operative session. There were 23 males and 7 females with a mean age of 63 ⫾ 8 years (range 45 to 75) in our patient population. Procedure Following routine preoperative workup, counseling, and informed consent, patients were brought to the operating room and had a single-lumen endotracheal tube placed. Cervical mediastinoscopy was then performed in the following fashion: a suprasternal incision was made and dissection carried down to the platysma. The platysma was opened transversely and the cervical fascia opened vertically. The pretracheal space was then entered. Blunt finger dissection and the mediastinoscope were used to further develop this plane. The scope was then inserted and advanced anterior to the trachea and lymph node stations 2R, 4L, 4R, and 7 (see Figure 1) were identified and sampled bilaterally. Frozen section analysis was performed and if positive for malignancy the procedure was terminated without further resection. Otherwise, the patient was reintubated with a double lumen endotracheal tube and VATS/thoracotomy was performed. Patients undergoing mediastinoscopy only were discharged home on the same day.
RESULTS Cervical mediastinoscopy was carried out prior to VATS/ thoracotomy as planned in 29 of 30 patients. Mediastinoscopy was not performed in 1 patient because of a history of
Hodgkin’s disease, previous chest radiotherapy, and radiographic evidence of heavily calcified mediastinal nodes. Of the remaining patients, 3 of 29 (10%) had nodes positive for malignancy found on frozen section at mediastinoscopy. There was one small cell lung cancer (SCLC) and two non-small cell lung cancers (NSCLC). Twenty-seven patients then underwent VATS with or without thoracotomy. The final pathologic findings are shown in Table I. Histology was distributed as follows: adenocarcinoma (10); squamous cell carcinoma (10); large cell carcinoma (1); synchronous adenocarcinoma (1); synchronous squamous cell carcinoma (1); synchronous adenocarcinoma and nonsmall cell carcinoma (1). Six patients were found to have T2 disease on final pathologic examination. Five of these were on the basis of microscopic visceral pleural involvement. The remaining patient’s tumor was measured at 3.1 cm in its greatest dimension on final sections. Three patients were found to have a presumed synchronous primary lesion on final pathologic examination. All of these lesions were found to be T1N0M0, and remained stage IA. Two patients with left upper lobe lesions were found to have N2 disease (pathologic stage IIIA) on the basis of aortopulmonary window (station 5) node involvement discovered during lymph node dissection at the time of complete resection (see Figure 2 for summary of patient procedures). There were no false negative mediastinoscopies (when comparing frozen with permanent sections), and there were no complications observed from mediastinoscopy. For patients with a malignancy (27 of 30), the false negative rate for CT evaluation of mediastinal nodes found to be positive for malignancy at mediastinoscopy was 11% (3 of 27). When considering only SCLC and NSCLC, the rate is 13% (3 of 24). Overall in this group of patients with peripheral T1 SCLC and NSCLC, 21% (5 of 24) had mediastinal nodes positive for malignancy identified after complete resection despite negative CT scans. Follow-up and survival data are available for these patients. One patient with a stage IA adenocarcinoma was lost to follow-up. A survey of the remaining 29 patients showed that 11 have died. Six of these 11 (55%) had stage IA disease with an average survival of 13 months (range 3 to 24). Overall survival for all patients with stage IA disease in this series was 67% with an average follow-up of
THE AMERICAN JOURNAL OF SURGERY® VOLUME 180 DECEMBER 2000
489
ROUTINE MEDIASTINOSCOPY IN PERIPHERAL T1 LUNG CANCER/TAHARA ET AL
Figure 2. Summary of patient procedures.
approximately 36 months. One patient with stage IIA disease survived 5 months. Patients who died with stage IIIA disease (3 of 11) survived an average of 8 months (range 5 to 10). There is currently 1 patient with stage IIIA (T2N2M0) with a survival of 29 months to date. The sole patient in our series with SCLC died after 10 months. All failures were secondary to distant metastases.
COMMENTS Our experience demonstrates that a significant number of patients with small peripheral lung cancers may harbor radiographically occult lymph node metastases. Several reported series have found the accuracy of CT scan to be lacking when applied to evaluation of mediastinal nodes. CT scan criteria for positive nodes are usually based on nodal size ⬎1 cm to 1.5 cm; however, CT cannot reliably distinguish between benign and malignant adenopathy based only on size. As many as 20% to 30% of nodes ⬎2 cm in size are benign.13 This may be especially true in the subgroup of patients with small peripheral tumors; Seely et al14 have shown that for patients with T1 lesions, the sensitivity and specificity of CT scan was only 77% and 91%, respectively. More recently, Gdeedo et al12 have shown in a prospective study of 100 patients with NSCLC that even the latest generation CT scanners lack the ability to reliably stage mediastinal lymph nodes demonstrating a sensitivity and specificity of only 63% and 57%, respectively. Their experience shows that routine mediastinoscopy found 12 of 100 (12%) patients with lymph nodes positive for malignancy following negative mediastinal evaluation by CT.12 These observations correspond closely with our observed CT scan false negative rate of 13% in patients who have SC/NSCLC. Other authors have demonstrated the advantages of combining mediastinoscopy and frozen section analysis immediately prior to, but during the same operative session as a potentially curative resection. This approach was found to be safe, highly accurate, and cost effective when compared 490
with doing sequential but temporally separate mediastinoscopy with permanent section followed by later resection.15,16 Our experience supports this approach; we had no additional morbidity or complications from mediastinoscopy and no false negative frozen sections when compared to permanent sections. We did not specifically address cost in our review The 2 patients with left upper lobe tumors and aortopulmonary window involvement found following node dissection and complete resection raise the question of whether performing another staging procedure would be useful. Extended cervical mediastinoscopy, a Chamberlain procedure, or VATS biopsy of this region may have been helpful.3,17,18 However, completely resected tumors of this category are very similar in 5-year survival when compared with N1 disease,3,15 thus an additional staging procedure to preoperatively identify them is not necessarily warranted and adds little to their management.3 However, since the 5-year survival of patients with positive cervical mediastinoscopies is less than 10%, and adjuvant therapy adds little to survival, resection of mediastinoscopy positive patients is not indicated without some attempt to downstage the disease. PET scanning deserves a comment regarding its place in the staging of lung cancer. Although there are multiple promising studies detailing the use of PET to find not only lymph node involvement but also occult distant metastases, its use is not routine for several reasons.19 –21 PET is not commonly available in all areas, it remains expensive, and its accuracy is not precisely defined.3,20 Although Medicare and some insurance carriers have begun to pay for PET scans to evaluate patients with lung cancer, it will probably be some time before this modality becomes a standard of care. In conclusion, routine cervical mediastinoscopy should be employed when staging small peripheral lung cancers. In our experience it is highly accurate and finds a significant number of patients with radiographically occult mediasti-
THE AMERICAN JOURNAL OF SURGERY® VOLUME 180 DECEMBER 2000
ROUTINE MEDIASTINOSCOPY IN PERIPHERAL T1 LUNG CANCER/TAHARA ET AL
nal metastases. There is no increased operative risk. Routine use of mediastinoscopy helps to properly identify those patients who will benefit most from curative resection while allowing the use of neoadjuvant therapies and avoiding unnecessary and potentially morbid resections in patients with advanced regional disease.
REFERENCES 1. Mountain CF. Revisions in the international system for staging lung cancer. Chest. 1997;111:1710 –1717. 2. Mountain CF, Dresler CM. Regional lymph node classification for lung cancer staging. Chest. 1997;111:1718 –1723. 3. Ginsberg RJ, Kaiser LR, LoCicero J, et al. Symposium: current concepts in the treatment of lung cancer—Part 1. Contemp Surg. 1999;55:154 –166. 4. Takizawa T, Terashima M, Koike T, et al. Lymph node metastasis in small peripheral adenocarcinoma of the lung. J Thorac Cardiovasc Surg. 1998;116:276 –280. 5. Jolly PC, Hutchinson CH, Detterbeck F, et al. Routine computed tomographic scans, selective mediastinoscopy, and other factors in evaluation of lung cancer. J Thorac Cardiovasc Surg. 1991;106:266 –271. 6. American Thoracic Society/European Respiratory Society. Pretreatment evaluation of non-small-cell lung cancer. Am J Respir Crit Care Med. 1997;156:320 –332. 7. Akamatsu H, Terashima M, Koike T, et al. Staging of primary lung cancer by computed tomography-guided percutaneous needle cytology of mediastinal lymph nodes. Ann Thorac Surg. 1996;62: 352–355. 8. Roberts JR, Blum MG, Arildsen R, et al. Prospective comparison of radiologic, thoracoscopic, and pathologic staging in patients with early non-small cell lung cancer. Ann Thorac Surg. 1999;68:1154 – 1158. 9. Miller JD, Gorenstein LA, Patterson A. Staging: the key to rational management of lung cancer. Ann Thorac Surg. 1992;53: 170 –178.
DISCUSSION Dr. William F. Sasser (St. Louis, MO): In the pursuit of improving survival in lung cancer, combinations of therapy have been established. Exact and accurate staging is mandatory for the selection of treatment and for comparison of survival data. Diagnostic procedures such as needle aspiration biopsy, bronchoscopy, mediastinoscopy, and thoracoscopy are part of the surgeon’s armamentarium. The gold standard to evaluate the mediastinal lymph nodes is mediastinoscopy. With evidence of N2 or N3 disease, surgical resection as the primary treatment is not appropriate. Treatment with chemotherapy and/or radiation is justified. If there is a response to neoadjuvant treatment, surgery may be done in selected cases. Many surgeons do not do routine mediastinoscopy on peripheral T1 lesions, because they feel that the procedure is overly invasive with significant morbidity and occasional mortality. They prefer to rely on non-invasive imaging, which is not accurate for either small or large lymph nodes. This paper, as well as others, provides evidence that it should be done, because of the 10 –15% chance of a small unsuspected mediastinal node being positive. The authors did the mediastinoscopy at the time of the planned resection surgery. The thoracotomy was cancelled if the node was positive. Others would do the procedure as a preoperative work-up because there is a small percent of false
10. De Leyn P, Vansteenkiste J, Cuypers P, et al. Role of cervical mediastinoscopy in staging non-small cell lung cancer without enlarged mediastinal lymph nodes on CT scan. Eur J Cardiothorac Surg. 1997;12:706 –712. 11. Dillemans B, Deneffe G, Verschakelen J, Decramer M. Value of computed tomography and mediastinoscopy in preoperative evaluation of mediastinal nodes in non-small cell lung cancer. A study of 569 patients. Eur J Cardiothorac Surg. 1994;8:37– 42. 12. Gdeedo A, Van Schil P, Corthouts B, et al. Prospective evaluation of computed tomography and mediastinoscopy in mediastinal lymph node staging. Eur Respir J. 1997;10:1547–1551. 13. Arita T, Matsumoto T, Kuramitsu T, et al. Is it possible to differentiate malignant mediastinal nodes from benign nodes by size? Chest. 1996;110:1004 –1008. 14. Seely JM, Mayo JR, Miller RB, Muller NL. T1 lung cancer: prevalence of mediastinal nodal metatases and diagnostic accuracy of CT. Radiology. 1993;186:129 –132. 15. Ginsberg RJ. The role of preoperative surgical staging in left upper lobe tumors. Ann Thorac Surg. 1994;57:526 –527. 16. Gephardt GN, Rice TW. Utility of frozen-section evaluation of lymph nodes in the staging of bronchogenic carcinoma at mediastinoscopy and thoracotomy. J Thorac Cardiovasc Surg. 1990;100: 853– 859. 17. Gossot D, Toledo L, Fritsch S, Celerier M. Mediastinoscopy vs thoracoscopy for mediastinal biopsy. Chest. 1996;110:1328 –1331. 18. Ginsberg RJ. Extended cervical mediastinoscopy. Chest Surg Clin North Am. 1996;6:21–30. 19. Lowe VJ, Naunheim KS. Positron emission tomography in lung cancer. Ann Thorac Surg. 1998;65:1821–1829. 20. Chin R Jr, Ward R, Keyes JW, et al. Mediastinal staging of non-small-cell lung cancer with positron emission tomography. Am J Respir Crit Care Med. 1995;152:2090 –2096. 21. Stokkel MP, Bakker PF, Heine R, et al. Staging of lymph nodes with FDG dual headed PET in patients with non-small-cell lung cancer. Nucl Med Comm. 1999;20:1001–1007.
negative results. Each approach is acceptable. I agree that CAT scans and/or PET scans do not replace a biopsy. You must have tissue to diagnose the cancer in the node. A study by the Clinical Oncology group of the American College of Surgeons is underway to evaluate PET scans in this role. I would like to ask several questions regarding this paper: 1) Did you attempt to establish a pre-operative diagnosis by performing a needle biopsy of the nodule or, because of the presence of the nodule, did you go straight to mediastinoscopy, and if it was negative, do the thoracotomy? If you knew in advance that the primary was adenocarcinoma, would you be more suspicious of the smaller nodes? Or if the peripheral lesion was a squamous cell cancer, would you skip the mediastinoscopy? 2) When do you use an anterior thoracotomy (or Chamberlain procedure)? I prefer this for evaluation of left upper lobe lesions rather than the extended cervical mediastinoscopy. 3) How many lymph nodes did you usually sample? Did you make an attempt to include states 2RL, 4RL, and 7? 4) You did not mention cell type other than small cell vs. non-small cell cancer. In most instances peripheral adenocarcinomas produce smaller mediastinal nodes than squamous cell cancer. Did you find this to be true?
THE AMERICAN JOURNAL OF SURGERY® VOLUME 180 DECEMBER 2000
491
ROUTINE MEDIASTINOSCOPY IN PERIPHERAL T1 LUNG CANCER/TAHARA ET AL
5) In your own practice do you always do a mediastinoscopy on every peripheral nodule? I agree with your conclusions. Mediastinoscopy is the preferred examination for the mediastinum. With rare exception it should be done on all patients before a definitive resection is contemplated.
CLOSING Dr. R. W. Tahara: With regard to the performance of preoperative biopsy, the answer is no. The technique that we employed was to go straight to the operating room. We would perform bronchoscopy, and mediastinoscopy, and then follow our pathway from there. We did not perform needle biopsies in any type, whether they were CT scan or ultrasound guided, unless there was a bronchial lesion. And in these cases there were none. With respect to the cell type, most of our tumors were
492
peripheral adenocarcinomas. I’m sorry, I can’t answer your exact survival question because I have not broken the data down, based on that. With respect to the question on which lymph nodes were biopsied, we made every attempt to do a very thorough, a very meticulous mediastinoscopy. In general, we were able to identify and biopsy the 2R2L, 4R4L and Station 7 nodes. We attempted to do this in every patient. Finally, Dr. Sasser asked whether our take home message is, “Should mediastinoscopy be performed in most, if not all, patients”, and I think that the answer, is “Yes”. People have been relying more and more on radiographic studies, and I think that our data and the previously published data of other authors, certainly implies that excessive reliance on radiographic tests may, indeed, cause us to miss 10 to 15 percent of these patients.
THE AMERICAN JOURNAL OF SURGERY® VOLUME 180 DECEMBER 2000
The role of surgical staging of patients with non-small cell lung cancer (NSCLC) continues to evolve. This report describes our findings utilizing routine cervical mediastinoscopy in the evaluation of peripheral T1 (<3 cm) lung tumors. METHODS: Retrospectively 30 patients with peripheral T1 lesions and CT scans negative for pathologic adenopathy were identified over a 3-year period. Cervical mediastinoscopy was performed prior to VATS/thoracotomy during the same operative session. RESULTS: Mediastinoscopy was performed in 29 of 30 patients. For patients with malignancy (27 of 30), 3 of 27 (11%) had mediastinoscopy positive for malignancy and no further resection performed. Overall the subgroup of patients with bronchogenic carcinomas had positive mediastinal involvement identified in 5 of 24 (21%) after mediastinoscopy or complete resection. CONCLUSIONS: A significant number of patients with small peripheral lung cancers harbor radiographically occult lymph node involvement. Mediastinoscopy facilitates identification of patients with regionally advanced disease prior to resection, allowing neoadjuvant therapy and avoiding unnecessary resections. Am J Surg. 2000;180: 488 – 492. © 2001 by Excerpta Medica, Inc. BACKGROUND:
T
he approach to staging lung cancer continues to evolve. Although there is consensus regarding the definition of each pathologic stage and the pivotal role mediastinal lymph node involvement plays in both prognosis and therapeutic decision making,1–5 considerable variance remains in how preoperative assessment of the mediastinum is accomplished. The last 2 decades have brought an increasing emphasis on the radiographic evaluation of the mediastinum in the form of computed tomography (CT) and, more recently, positron emission (PET)
From the Department of Surgery (RWT), University of Nebraska Medical Center, Omaha, Nebraska; and the Department of Surgery (RPL, LMG), Division of Cardiothoracic Surgery, Long Island Jewish Medical Center, New Hyde Park, New York. Requests for reprints should be addressed to Rudy P. Lackner, MD, Long Island Jewish Medical Center, Cardiothoracic Surgery, Suite 2123, 270-05 76th Avenue, New Hyde Park, New York 11040. Presented at the 52nd Annual Meeting of the Southwestern Surgical Congress, Colorado Springs, Colorado, April 9 –12, 2000.
488
© 2001 by Excerpta Medica, Inc. All rights reserved.
scanning,6 with emphasis on replacing or minimizing invasive sampling of the mediastinal nodes. This has occurred despite the proven accuracy, safety, and relative minimally invasive nature of cervical mediastinoscopy. In fact, when reviewing the literature mediastinoscopy is usually the standard to which these other modalities and strategies are compared. Commonly used preoperative staging strategies range from CT or PET alone, to selective mediastinoscopy or radiographically guided biopsy, to routine mediastinoscopy, mediastinotomy, and thoracoscopy.3,7,8 Bronchoscopy and, more recently, endobronchial ultrasound have also been used to assist in staging. These are employed in an effort to identify those patients already having regionally advanced disease and in whom surgical resection would not currently be the treatment of choice. Finding these patients preoperatively allows the use of neoadjuvant chemotherapy and/or radiation therapy. Conversely, patients who can be shown to have only local involvement benefit from a markedly better rate of cure with surgical resection.9 This is especially true when evaluating patients with small peripheral lung tumors. Patients with T1 (⬍3 cm) tumors and no superior mediastinal involvement or distant metastases are by definition stage I and should have the best chance for cure with surgical resection (67% at 5 years). The survival for patients in this category who are found to have superior mediastinal involvement, and hence N2 disease (stage IIIA), would drop markedly and is only 20% to 25% at 5 years.1,2 The optimum approach to best exclude mediastinal node metastases and how to best use these various strategies remains quite controversial, although several large series have recently been reported in the European literature that would seem to demonstrate the superiority of routine mediastinoscopy.10 –12 This report details our experience using routine cervical mediastinoscopy in assessing the subset of patients with peripheral T1 lung cancers.
PATIENTS AND METHODS Patients A retrospective review from November 1994 through September 1998 identified 30 patients with peripheral T1 lung tumors. All patients had CT scans performed, which showed mediastinal lymph nodes ⬍1 cm in size. Percutaneous needle biopsies or flexible bronchoscopy with transbronchial biopsies were not standardly performed in the preoperative assessment of solitary pulmonary nodules. Cervical mediastinoscopy was routinely performed in all patients to accurately stage the mediastinum prior to resection. Patients with negative nodes proceeded on to video-assisted thoracic surgery (VATS) or thoracotomy, or 0002-9610/00/$–see front matter PII S0002-9610(00)00509-2
ROUTINE MEDIASTINOSCOPY IN PERIPHERAL T1 LUNG CANCER/TAHARA ET AL
TABLE I Summary of Final Pathologic Staging Final Pathologic Stage IA IIA IIIA Limited stage SCLC
TNM Classification
Number of Patients
T1N0M0 T2N0M0* T1N1M0 T1N2M0 T2N2M0*
13 5 1 3 1 1
* Patient had a history of Hodgkin’s disease, prior chest radiotherapy, and highly calcified nodes by radiography.
Figure 1. Lymph node stations. Modified from Mountain and Dresler.2
both, to perform resection of the lesion. Those found to have positive mediastinal nodes were offered standard chemoradiation or the opportunity to undergo neoadjuvant therapy followed by resection. In node-negative patients resection was performed during the same operative session. There were 23 males and 7 females with a mean age of 63 ⫾ 8 years (range 45 to 75) in our patient population. Procedure Following routine preoperative workup, counseling, and informed consent, patients were brought to the operating room and had a single-lumen endotracheal tube placed. Cervical mediastinoscopy was then performed in the following fashion: a suprasternal incision was made and dissection carried down to the platysma. The platysma was opened transversely and the cervical fascia opened vertically. The pretracheal space was then entered. Blunt finger dissection and the mediastinoscope were used to further develop this plane. The scope was then inserted and advanced anterior to the trachea and lymph node stations 2R, 4L, 4R, and 7 (see Figure 1) were identified and sampled bilaterally. Frozen section analysis was performed and if positive for malignancy the procedure was terminated without further resection. Otherwise, the patient was reintubated with a double lumen endotracheal tube and VATS/thoracotomy was performed. Patients undergoing mediastinoscopy only were discharged home on the same day.
RESULTS Cervical mediastinoscopy was carried out prior to VATS/ thoracotomy as planned in 29 of 30 patients. Mediastinoscopy was not performed in 1 patient because of a history of
Hodgkin’s disease, previous chest radiotherapy, and radiographic evidence of heavily calcified mediastinal nodes. Of the remaining patients, 3 of 29 (10%) had nodes positive for malignancy found on frozen section at mediastinoscopy. There was one small cell lung cancer (SCLC) and two non-small cell lung cancers (NSCLC). Twenty-seven patients then underwent VATS with or without thoracotomy. The final pathologic findings are shown in Table I. Histology was distributed as follows: adenocarcinoma (10); squamous cell carcinoma (10); large cell carcinoma (1); synchronous adenocarcinoma (1); synchronous squamous cell carcinoma (1); synchronous adenocarcinoma and nonsmall cell carcinoma (1). Six patients were found to have T2 disease on final pathologic examination. Five of these were on the basis of microscopic visceral pleural involvement. The remaining patient’s tumor was measured at 3.1 cm in its greatest dimension on final sections. Three patients were found to have a presumed synchronous primary lesion on final pathologic examination. All of these lesions were found to be T1N0M0, and remained stage IA. Two patients with left upper lobe lesions were found to have N2 disease (pathologic stage IIIA) on the basis of aortopulmonary window (station 5) node involvement discovered during lymph node dissection at the time of complete resection (see Figure 2 for summary of patient procedures). There were no false negative mediastinoscopies (when comparing frozen with permanent sections), and there were no complications observed from mediastinoscopy. For patients with a malignancy (27 of 30), the false negative rate for CT evaluation of mediastinal nodes found to be positive for malignancy at mediastinoscopy was 11% (3 of 27). When considering only SCLC and NSCLC, the rate is 13% (3 of 24). Overall in this group of patients with peripheral T1 SCLC and NSCLC, 21% (5 of 24) had mediastinal nodes positive for malignancy identified after complete resection despite negative CT scans. Follow-up and survival data are available for these patients. One patient with a stage IA adenocarcinoma was lost to follow-up. A survey of the remaining 29 patients showed that 11 have died. Six of these 11 (55%) had stage IA disease with an average survival of 13 months (range 3 to 24). Overall survival for all patients with stage IA disease in this series was 67% with an average follow-up of
THE AMERICAN JOURNAL OF SURGERY® VOLUME 180 DECEMBER 2000
489
ROUTINE MEDIASTINOSCOPY IN PERIPHERAL T1 LUNG CANCER/TAHARA ET AL
Figure 2. Summary of patient procedures.
approximately 36 months. One patient with stage IIA disease survived 5 months. Patients who died with stage IIIA disease (3 of 11) survived an average of 8 months (range 5 to 10). There is currently 1 patient with stage IIIA (T2N2M0) with a survival of 29 months to date. The sole patient in our series with SCLC died after 10 months. All failures were secondary to distant metastases.
COMMENTS Our experience demonstrates that a significant number of patients with small peripheral lung cancers may harbor radiographically occult lymph node metastases. Several reported series have found the accuracy of CT scan to be lacking when applied to evaluation of mediastinal nodes. CT scan criteria for positive nodes are usually based on nodal size ⬎1 cm to 1.5 cm; however, CT cannot reliably distinguish between benign and malignant adenopathy based only on size. As many as 20% to 30% of nodes ⬎2 cm in size are benign.13 This may be especially true in the subgroup of patients with small peripheral tumors; Seely et al14 have shown that for patients with T1 lesions, the sensitivity and specificity of CT scan was only 77% and 91%, respectively. More recently, Gdeedo et al12 have shown in a prospective study of 100 patients with NSCLC that even the latest generation CT scanners lack the ability to reliably stage mediastinal lymph nodes demonstrating a sensitivity and specificity of only 63% and 57%, respectively. Their experience shows that routine mediastinoscopy found 12 of 100 (12%) patients with lymph nodes positive for malignancy following negative mediastinal evaluation by CT.12 These observations correspond closely with our observed CT scan false negative rate of 13% in patients who have SC/NSCLC. Other authors have demonstrated the advantages of combining mediastinoscopy and frozen section analysis immediately prior to, but during the same operative session as a potentially curative resection. This approach was found to be safe, highly accurate, and cost effective when compared 490
with doing sequential but temporally separate mediastinoscopy with permanent section followed by later resection.15,16 Our experience supports this approach; we had no additional morbidity or complications from mediastinoscopy and no false negative frozen sections when compared to permanent sections. We did not specifically address cost in our review The 2 patients with left upper lobe tumors and aortopulmonary window involvement found following node dissection and complete resection raise the question of whether performing another staging procedure would be useful. Extended cervical mediastinoscopy, a Chamberlain procedure, or VATS biopsy of this region may have been helpful.3,17,18 However, completely resected tumors of this category are very similar in 5-year survival when compared with N1 disease,3,15 thus an additional staging procedure to preoperatively identify them is not necessarily warranted and adds little to their management.3 However, since the 5-year survival of patients with positive cervical mediastinoscopies is less than 10%, and adjuvant therapy adds little to survival, resection of mediastinoscopy positive patients is not indicated without some attempt to downstage the disease. PET scanning deserves a comment regarding its place in the staging of lung cancer. Although there are multiple promising studies detailing the use of PET to find not only lymph node involvement but also occult distant metastases, its use is not routine for several reasons.19 –21 PET is not commonly available in all areas, it remains expensive, and its accuracy is not precisely defined.3,20 Although Medicare and some insurance carriers have begun to pay for PET scans to evaluate patients with lung cancer, it will probably be some time before this modality becomes a standard of care. In conclusion, routine cervical mediastinoscopy should be employed when staging small peripheral lung cancers. In our experience it is highly accurate and finds a significant number of patients with radiographically occult mediasti-
THE AMERICAN JOURNAL OF SURGERY® VOLUME 180 DECEMBER 2000
ROUTINE MEDIASTINOSCOPY IN PERIPHERAL T1 LUNG CANCER/TAHARA ET AL
nal metastases. There is no increased operative risk. Routine use of mediastinoscopy helps to properly identify those patients who will benefit most from curative resection while allowing the use of neoadjuvant therapies and avoiding unnecessary and potentially morbid resections in patients with advanced regional disease.
REFERENCES 1. Mountain CF. Revisions in the international system for staging lung cancer. Chest. 1997;111:1710 –1717. 2. Mountain CF, Dresler CM. Regional lymph node classification for lung cancer staging. Chest. 1997;111:1718 –1723. 3. Ginsberg RJ, Kaiser LR, LoCicero J, et al. Symposium: current concepts in the treatment of lung cancer—Part 1. Contemp Surg. 1999;55:154 –166. 4. Takizawa T, Terashima M, Koike T, et al. Lymph node metastasis in small peripheral adenocarcinoma of the lung. J Thorac Cardiovasc Surg. 1998;116:276 –280. 5. Jolly PC, Hutchinson CH, Detterbeck F, et al. Routine computed tomographic scans, selective mediastinoscopy, and other factors in evaluation of lung cancer. J Thorac Cardiovasc Surg. 1991;106:266 –271. 6. American Thoracic Society/European Respiratory Society. Pretreatment evaluation of non-small-cell lung cancer. Am J Respir Crit Care Med. 1997;156:320 –332. 7. Akamatsu H, Terashima M, Koike T, et al. Staging of primary lung cancer by computed tomography-guided percutaneous needle cytology of mediastinal lymph nodes. Ann Thorac Surg. 1996;62: 352–355. 8. Roberts JR, Blum MG, Arildsen R, et al. Prospective comparison of radiologic, thoracoscopic, and pathologic staging in patients with early non-small cell lung cancer. Ann Thorac Surg. 1999;68:1154 – 1158. 9. Miller JD, Gorenstein LA, Patterson A. Staging: the key to rational management of lung cancer. Ann Thorac Surg. 1992;53: 170 –178.
DISCUSSION Dr. William F. Sasser (St. Louis, MO): In the pursuit of improving survival in lung cancer, combinations of therapy have been established. Exact and accurate staging is mandatory for the selection of treatment and for comparison of survival data. Diagnostic procedures such as needle aspiration biopsy, bronchoscopy, mediastinoscopy, and thoracoscopy are part of the surgeon’s armamentarium. The gold standard to evaluate the mediastinal lymph nodes is mediastinoscopy. With evidence of N2 or N3 disease, surgical resection as the primary treatment is not appropriate. Treatment with chemotherapy and/or radiation is justified. If there is a response to neoadjuvant treatment, surgery may be done in selected cases. Many surgeons do not do routine mediastinoscopy on peripheral T1 lesions, because they feel that the procedure is overly invasive with significant morbidity and occasional mortality. They prefer to rely on non-invasive imaging, which is not accurate for either small or large lymph nodes. This paper, as well as others, provides evidence that it should be done, because of the 10 –15% chance of a small unsuspected mediastinal node being positive. The authors did the mediastinoscopy at the time of the planned resection surgery. The thoracotomy was cancelled if the node was positive. Others would do the procedure as a preoperative work-up because there is a small percent of false
10. De Leyn P, Vansteenkiste J, Cuypers P, et al. Role of cervical mediastinoscopy in staging non-small cell lung cancer without enlarged mediastinal lymph nodes on CT scan. Eur J Cardiothorac Surg. 1997;12:706 –712. 11. Dillemans B, Deneffe G, Verschakelen J, Decramer M. Value of computed tomography and mediastinoscopy in preoperative evaluation of mediastinal nodes in non-small cell lung cancer. A study of 569 patients. Eur J Cardiothorac Surg. 1994;8:37– 42. 12. Gdeedo A, Van Schil P, Corthouts B, et al. Prospective evaluation of computed tomography and mediastinoscopy in mediastinal lymph node staging. Eur Respir J. 1997;10:1547–1551. 13. Arita T, Matsumoto T, Kuramitsu T, et al. Is it possible to differentiate malignant mediastinal nodes from benign nodes by size? Chest. 1996;110:1004 –1008. 14. Seely JM, Mayo JR, Miller RB, Muller NL. T1 lung cancer: prevalence of mediastinal nodal metatases and diagnostic accuracy of CT. Radiology. 1993;186:129 –132. 15. Ginsberg RJ. The role of preoperative surgical staging in left upper lobe tumors. Ann Thorac Surg. 1994;57:526 –527. 16. Gephardt GN, Rice TW. Utility of frozen-section evaluation of lymph nodes in the staging of bronchogenic carcinoma at mediastinoscopy and thoracotomy. J Thorac Cardiovasc Surg. 1990;100: 853– 859. 17. Gossot D, Toledo L, Fritsch S, Celerier M. Mediastinoscopy vs thoracoscopy for mediastinal biopsy. Chest. 1996;110:1328 –1331. 18. Ginsberg RJ. Extended cervical mediastinoscopy. Chest Surg Clin North Am. 1996;6:21–30. 19. Lowe VJ, Naunheim KS. Positron emission tomography in lung cancer. Ann Thorac Surg. 1998;65:1821–1829. 20. Chin R Jr, Ward R, Keyes JW, et al. Mediastinal staging of non-small-cell lung cancer with positron emission tomography. Am J Respir Crit Care Med. 1995;152:2090 –2096. 21. Stokkel MP, Bakker PF, Heine R, et al. Staging of lymph nodes with FDG dual headed PET in patients with non-small-cell lung cancer. Nucl Med Comm. 1999;20:1001–1007.
negative results. Each approach is acceptable. I agree that CAT scans and/or PET scans do not replace a biopsy. You must have tissue to diagnose the cancer in the node. A study by the Clinical Oncology group of the American College of Surgeons is underway to evaluate PET scans in this role. I would like to ask several questions regarding this paper: 1) Did you attempt to establish a pre-operative diagnosis by performing a needle biopsy of the nodule or, because of the presence of the nodule, did you go straight to mediastinoscopy, and if it was negative, do the thoracotomy? If you knew in advance that the primary was adenocarcinoma, would you be more suspicious of the smaller nodes? Or if the peripheral lesion was a squamous cell cancer, would you skip the mediastinoscopy? 2) When do you use an anterior thoracotomy (or Chamberlain procedure)? I prefer this for evaluation of left upper lobe lesions rather than the extended cervical mediastinoscopy. 3) How many lymph nodes did you usually sample? Did you make an attempt to include states 2RL, 4RL, and 7? 4) You did not mention cell type other than small cell vs. non-small cell cancer. In most instances peripheral adenocarcinomas produce smaller mediastinal nodes than squamous cell cancer. Did you find this to be true?
THE AMERICAN JOURNAL OF SURGERY® VOLUME 180 DECEMBER 2000
491
ROUTINE MEDIASTINOSCOPY IN PERIPHERAL T1 LUNG CANCER/TAHARA ET AL
5) In your own practice do you always do a mediastinoscopy on every peripheral nodule? I agree with your conclusions. Mediastinoscopy is the preferred examination for the mediastinum. With rare exception it should be done on all patients before a definitive resection is contemplated.
CLOSING Dr. R. W. Tahara: With regard to the performance of preoperative biopsy, the answer is no. The technique that we employed was to go straight to the operating room. We would perform bronchoscopy, and mediastinoscopy, and then follow our pathway from there. We did not perform needle biopsies in any type, whether they were CT scan or ultrasound guided, unless there was a bronchial lesion. And in these cases there were none. With respect to the cell type, most of our tumors were
492
peripheral adenocarcinomas. I’m sorry, I can’t answer your exact survival question because I have not broken the data down, based on that. With respect to the question on which lymph nodes were biopsied, we made every attempt to do a very thorough, a very meticulous mediastinoscopy. In general, we were able to identify and biopsy the 2R2L, 4R4L and Station 7 nodes. We attempted to do this in every patient. Finally, Dr. Sasser asked whether our take home message is, “Should mediastinoscopy be performed in most, if not all, patients”, and I think that the answer, is “Yes”. People have been relying more and more on radiographic studies, and I think that our data and the previously published data of other authors, certainly implies that excessive reliance on radiographic tests may, indeed, cause us to miss 10 to 15 percent of these patients.
THE AMERICAN JOURNAL OF SURGERY® VOLUME 180 DECEMBER 2000