In 1978, Cooper and associates first described resecting

Median Sternotomy Versus Thoracotomy to Resect Primary Lung Cancer: Analysis of 815 Cases James W. Asaph, MD, John R. Handy, Jr, MD, Gary L. Grunkemeier, PhD, E. Charles Douville, MD, Andrew C. Tsen, MD, Richard C. Rogers, MD, and John F. Keppel, MD The Oregon Clinic, P.C., Earle A. Chiles Research Institute, and Medical Data Research Center, Providence Health System, Portland, Oregon Background. We sought to determine if median sternotomy (MS) is an equivalent incision to thoracotomy (TH) in the treatment of primary pulmonary carcinoma. Methods. We followed 801 patients undergoing 815 operations for primary lung carcinoma in a computer registry; 447 had MS, 368 had TH. Results. Both groups were similar in preoperative risk assessment. Complete staging lymph node dissections were performed in 42% of MS patients and 17% of TH patients. Operative mortality (3.8% for MS, 3.3% for TH) and postoperative complications were similar. MS patients had a shorter postoperative hospital stay (7.5 days vs 8.2 days). One hundred thirty-nine underwent pneumonectomy. Operative mortality was 12.5% for MS and 10.4% for TS (p NS). Five hundred eighty-one underwent lobectomy with an operative mortality of 2.1% for MS and 2.0% for TH. Mean length of stay for MS lobectomy was 7.5 days compared with 8.5 days for TH (p 0.06). Follow-up was 89% through 1998, comprising 1,339 MS and 1,463 TH patient-years. Survival for stage I at 5 and 10 years, respectively, was 51% and 34% for MS vs 54% and 32% for TH (p NS). Survival for other stages was also similar. Conclusions. Median sternotomy provides more complete staging, shorter postoperative hospitalization, and better patient acceptance with equivalent operative and long-term survival when compared with thoracotomy. Concerns regarding increased wound infections in MS patients appear unfounded. (Ann Thorac Surg 2000;70:373 9) 2000 by The Society of Thoracic Surgeons In 1978, Cooper and associates first described resecting primary pulmonary cancer through a median sternotomy (MS) [1]. In 1980, Urschel and Razzuk presented their experience with MS for resecting primary lung tumors at The Society of Thoracic Surgeons meeting [2]. Both recognized that cardiac surgery through MS appears to cause less incisional pain and fewer pulmonary complications compared with thoracotomy (TH). However, concerns about completeness of resection, hilar control, and the disastrous complication of deep sternal wound infections [3] have kept MS from becoming widespread as an approach for resection of lung cancer. We began using MS for resection of primary pulmonary carcinoma in 1980. Early in our experience, we compared the use of postoperative analgesics between MS and TH groups [4]. The MS patients used significantly fewer analgesics than patients undergoing TH. Subsequently, we have preferentially utilized MS for resection of lung neoplasms. Currently, MS is utilized in approximately 75% of our pulmonary resections (Fig 1). This report compares MS with TH as the primary incision for surgical treatment of primary lung carcinoma Accepted for publication Feb 18, 2000. Address reprint requests to Dr Handy, The Oregon Clinic, P.C., 507 NE 47th Ave, Portland, OR 97213. in terms of adequacy of resection, length of stay, operative morbidity and mortality, and long-term survival in 815 pulmonary operations. Material and Methods A computerized registry (Patient Analysis & Tracking System; Axis Softwear, Portland, OR) was used to track 2,324 consecutive patients undergoing thoracic surgery procedures by a single group of thoracic surgeons operating in two medical centers from January 1980 through December 1998. Patients undergoing surgery before 1984 were entered retrospectively by chart review. Since then, all patients were entered prospectively at the time of, or shortly after, their surgery. Only patients undergoing planned pulmonary resection for primary lung carcinoma were selected for analysis. Video-assisted thoracotomies for biopsy or local excisions were excluded. Longterm follow-up was tracked from hospital tumor registries and managed care systems data sources. Data collected included patient demographics, preoperative risk factors, diagnostic procedures, operative features, pathology, stage, postoperative complications, and mortality. Stage was determined according to the Manual for Staging of Cancer, 4th Edition [5]. Follow-up information included date of last follow-up, administration of radiation or chemotherapy, current tumor status, and 2000 by The Society of Thoracic Surgeons 0003-4975/00/$20.00 Published by Elsevier Science Inc PII S0003-4975(00)01364-3

374 ASAPH ET AL Ann Thorac Surg MEDIAN STERNOTOMY TO RESECT LUNG CANCER 2000;70:373 9 cause of death. Statistical calculations were performed with SPSS 9.0 (SPSS, Inc, Chicago, IL) and PATS (Axis Software, Portland, OR). Values greater than p 0.05 are designated NS. Postoperative length of stay (LOS) was compared by unpaired t test. Survival was calculated by the Kaplan-Meier technique. MS was performed utilizing standard techniques with the arms tucked by the sides. All patients had radial artery monitoring lines and double lumen endobronchial intubation. Mediastinoscopy was performed selectively; however, since 1991, complete mediastinal staging lymph node dissection has been performed routinely. No special sternal retractors were used but the retractor was tilted to facilitate exposure of the hemithorax. The surgeon was positioned opposite to the lung to be operated upon. The patient was turned away from the surgeon to allow the first assistant better visualization. Hilar exposure was accomplished by a combination of traction sutures placed into the perihilar pericardium anterior to the phrenic nerve and one or two laparotomy packs placed posterior to the lung. Hilar division began with pulmonary veins, proceeded to the pulmonary artery branches, and ended with the bronchus. Early completion of the interlobar fissure facilitated the dissection of the more peripheral vascular branches. Mediastinal lymphadenectomy was performed by exposing the trachea between the extrapericardial aorta and superior vena cava. Right and left paratracheal as well as subcarninal lymph nodes were easily removed. Aortopulmonary nodes were removed by dissecting between the extrapericardial aorta and extrapleural left pulmonary artery. With the MS approach, lateral or anterior chest wall en bloc dissections were performed after the hilar structures were divided. MS patients with central tumors were easily approached intrapericardially to control the pulmonary vessels. TH patients underwent standard posterolateral, thoracotomies, entering in the fifth or sixth interspace. Ribs were not divided nor removed, but muscle-sparing incisions have been utilized more frequently in the latter years of the series. Thoracic lymphadenectomy via TH was performed in the routine manner. Fig 1. Percentage of all patients undergoing planned resection of primary lung carcinoma through a median sternotomy between 1980 and 1998. Results Eight hundred and one patients underwent 815 operations for primary lung cancer. Two patients presented with synchronous tumors and 18 patients had metachronous carcinomas resulting in second or third operations. Four hundred forty-seven MS patients and 368 TH patients were analyzed. Three hundred forty-eight of the TH patients underwent posterolateral thoracotomy. A small number of the TH group had other incisions (16 anterior thoracotomies, three posterior thoracotomies, and one thoraco-abdominal incision). Patient characteristics are shown in Table 1. There was no difference in patient age, comorbidities, or smoking history between the groups. More MS patients had been treated for COPD (31% vs 22%; p 0.027). Table 2 outlines surgical procedures. Upper lobectomies were the most common operation in the MS group. En bloc chest wall resections were more frequent in the TH group. Complete staging lymph node dissections were performed 42% of MS patients but in only 17% of TH patients ( p 0.001). There were minor differences in tumor cell type and no difference in stage between the groups. Adenocarcinoma predominated in both groups (Table 3). Mortality and complications are shown in Table 4. Operative mortality and postoperative complications were similar in both groups. Cancer stage, advanced age, male gender, and poor pulmonary function were predictive of reduced operative survival. Throughout the study period, MS patients experienced a shorter mean postoperative LOS, and more were discharged in 5 days or less than were TH patients. Long-term follow-up was 89% complete through 1998, comprising 1,339 MS and 1,463 TH patient-years, and was similar regardless of incision. Survival for stage I and stage II was similar in both groups, as noted in Figures 2 and 3. One hundred thirty-nine patients underwent a pneumonectomy (72 MS, 67 TH). Preoperative risk factors were similar in both groups. Both pneumonectomy groups had similar mortality (MS 12.5%, TH 10.4%; p NS) and complication rates. Right and left pneumonectomies were analyzed separately. There were significantly more right pneumonectomies in the MS group (45 MS, 27 TH) and a predominance of left pneumonectomies in the TH group (45 MS, 27 TH). Operative mortality for right pneumonectomy was 15.6% for MS and 11.5% for TH. For left pneumonectomy, MS mortality was 7.4% and TH mortality was 9.8%. Respiratory failure occurred in 21% of right pneumonectomies but in only 3% of left pneumonectomies ( p 0.20). The postoperative pneumonectomy LOS was 8.28 days for the MS patients and 8.19 days for TH patients (p NS). There was no difference in long-term survival for pneumonectomy regardless of surgical approach (Table 5).

Ann Thorac Surg ASAPH ET AL 2000;70:373 9 MEDIAN STERNOTOMY TO RESECT LUNG CANCER 375 Table 1. Preoperative Risk Factors Total cases 447 54.8% 368 45.2% Male 240 53.7% 205 55.7% NS Female 207 46.3% 163 44.3% NS Mean age 64.1 63.6 NS Comorbidities Hypertension 101 22.6% 96 26.1% NS Coronary artery disease 57 12.8% 51 13.9% NS Congestive heart failure 12 2.7% 7 1.9% NS History of cardiac arrhythmia 44 9.8% 48 13.0% NS Peripheral vascular disease 35 7.8% 34 9.2% NS Cerebral vascular disease 23 5.1% 15 4.1% NS Chronic renal insufficiency 3 0.7% 7 1.9% NS Diabetes mellitus 35 7.8% 27 7.3% NS Chronic corticosteroid use 15 3.4% 10 2.7% NS Smoker: Never 30 6.7% 29 7.9% NS Former 204 45.6% 170 46.2% NS Current 208 46.5% 169 45.9% NS Treated for COPD 139 31.1% 80 21.7% 0.027 Impaired FEV-1: Severe 5 1.1% 6 1.6% NS Moderate 85 17.9% 56 15.2% NS Mild 43 33.1% 108 29.3% NS History of other malignancies 79 17.7% 85 23.1% 0.105 Previous chest surgery: Pulmonary 17 3.8% 34 9.2% 0.004 Cardiac 4 0.9% 35 9.2% 0.001 COPD chronic obstructive pulmonary disease; FEV-1 forced expiratory volume in 1 second. Five hundred eighty-one patients underwent lobectomy (331 MS, 250 TH). Preoperative risks factors and pulmonary function were similar. Postoperative complications were similar in both groups, except that reoperation for postoperative bleeding was more frequent in the MS patients (2.7% vs 0.4%). While four serious wound problems occurred in the TH groups (1.6%), there were no wound complications in the MS lobectomy patients. Operative mortality was 2.1% for MS and 2.0% for TH. Mean postoperative LOS for MS was 7.45 days compared Table 2. Surgical Procedures Surgical site Right 292 65.3% 171 45.5% 0.005 Left 134 30.0% 196 53.3% 0.001 Bilateral 21 4.7% 1 0.3% 0.001 Procedure Explore/biopsy only 19 4.3% 22 6.0% NS Local excision 30 6.7% 20 5.4% NS Segmental resection 9 2.0% 9 2.4% NS Upper lobectomy 232 51.9% 107 29.1% 0.001 Middle lobectomy 13 2.9% 10 2.7% NS Lower lobectomy 54 12.1% 110 29.9% 0.001 Bilobectomy 30 6.7% 23 6.3% NS Pneumonectomy 72 16.1% 67 18.2% NS Bronchial sleeve resection 6 1.3% 3 0.8% NS Mediastinoscopy 226 50.6% 109 29.6% 0.001 Staging mediastinal lymph node dissection 187 41.8% 61 16.6% 0.001 En bloc chest wall resection 8 1.8% 21 5.7% 0.007

376 ASAPH ET AL Ann Thorac Surg MEDIAN STERNOTOMY TO RESECT LUNG CANCER 2000;70:373 9 Table 3. Pathology Histology Squamous cell 145 32.4% 126 34.2% NS Adenocarcinoma 204 45.6% 154 41.8% NS Bronchoalveolar carcinoma 5 1.1% 13 3.5% 0.041 Large cell carcinoma 75 16.8% 46 12.5% NS Small cell carcinoma 5 1.1% 9 2.4% NS Carcinoid tumor 12 2.7% 13 3.4% NS Pulmonary sarcoma 1 0.2% 7 1.9% 0.042 Staging Stage I 281 62.9% 206 56.0% NS Stage II 67 15.0% 66 17.9% NS Stage IIIA 73 16.3% 75 20.4% NS Stage IIIB 11 2.5% 11 3.0% NS Stage IV 14 3.1% 10 2.7% NS with 8.54 days for TH ( p 0.028). Lobectomy long-term survival, adjusted for stage, was the same in both groups (Table 5). Comment We analyzed 815 cases of MS and TH incisions for resection of primary lung cancer. Posterolateral thoracotomy is the gold standard incision for pulmonary resection for primary cancer. Other incisions, such as anterior thoracotomy, posterior thoracotomy, and clam shell incisions, are advocated only in selected cases [6, 7]. While these incisions give excellent exposure for surgical resection, compromise of pulmonary function, disability, and chronic pain associated with thoracotomy has been considerable [1, 7]. Additionally, adequate lymph node staging has been difficult, particularly with left-sided lesions. MS has been used for years in cardiac surgery and has not been associated with the pain and disability of TH [7]. Although cardiac surgery and lung cancer surgery patients are typically similar in terms of age, smoking histories, and comorbidities, the incidence of postoperative pulmonary complications is significantly less in those Table 4. Mortality and Complications Mortality 17 3.8% 12 3.3% NS Complications Atelectasis requiring bronchoscopy 61 13.6% 34 9.2% 0.013 Pneumonia 10 2.2% 3 0.8% 0.191 Respiratory failure 29 6.5% 16 4.3% 0.268 Ventilator more than 24 hours 22 4.9% 12 3.3% NS Air leak more than 7 days 33 7.4% 30 8.2% NS Space/reexpansion 28 5.8% 22 6.0% NS Other pulmonary complications 27 6.0% 21 5.7% NS No pulmonary complications 312 69.8% 276 75.0% NS Major wound complication 4 0.9% 7 1.9% NS Pneumonectomy space infection 9 12.5% 7 10.4% NS Cardiac arrhythmia 68 15.2% 44 12.0% NS Myocardial infarction 6 1.3% 4 1.1% NS Other cardiac complications 8 1.8% 8 2.2% NS Stroke 5 1.1% 2 0.5% NS Renal failure 7 1.6% 5 1.4% NS Psychiatric 15 3.4% 9 2.4% NS Other complications 34 7.6% 32 8.7% NS Required reoperation 31 6.9% 14 3.8% NS Length of stay Mean 7.49 8.24 0.601 5 days or less 175 39.1% 100 27.2% 0.001 More than 10 days 63 14.1% 67 18.2% 0.134

Ann Thorac Surg ASAPH ET AL 2000;70:373 9 MEDIAN STERNOTOMY TO RESECT LUNG CANCER 377 patients undergoing heart surgery as compared with pulmonary operations. In the 1970s and early 1980s, surgeons began resecting bullous lesions and multiple metastatic pulmonary lesions through a sternotomy [8, 9]. As Cooper and others have shown, MS is a feasible incision to resect pulmonary tumors [1, 2, 4]. Using this approach, we previously demonstrated a shorter length of stay, less postoperative pain, and significantly less analgesic use [4]. However, the number of MS patients in each of these series was small, many common complications of pulmonary and cardiac surgery were not evaluated, there was no information regarding late pain syndromes, and follow-up was limited to the immediate postoperative period. Most significantly, there have been no data available to determine if resection through MS is as effective as TH in controlling cancer. This study was embarked upon to answer these questions and determine the place of MS in the surgical therapy of lung cancer. The cases were not randomized. Surgeon preference and anatomic concerns influenced the selection process. We consider superior sulcus tumor, left lower lobectomy, and posterior chest wall resection to be contraindications to pulmonary resection via MS. Additionally, we avoided left lung resection in the setting of previous coronary artery bypass using the left internal mammary artery because of the risk of injury to that conduit. We readily perform left and right pneumonectomies, and right-sided resection, left upper lobectomy, left superior segmentectomy, and anterior or lateral en bloc chest wall resections via MS. In patients with a previous thoracotomy, MS is an easier approach because it allows full entry into the chest away from the densest adhesions and early, safe hilar and vascular control [1]. In the present report, we did not specifically investigate late incisional pain syndromes. However, in 1987, we Fig 2. Kaplan-Meier survival curve of all patients with pathologically confirmed stage I (A and B) non-small cell carcinoma. Solid line represents sternotomy survival curve; dashed line represents thoracotomy survival curve. Represented are 281 stage I sternotomy patients and 206 stage I thoracotomy patients. Fig 3. Kaplan-Meier survival curve of all patients with pathologically confirmed stage II (A and B) non-small cell carcinoma. Solid line represents sternotomy survival curve; dashed line represents thoracotomy survival curve. Represented are 67 stage II sternotomy patients and 66 stage II thoracotomy patients. reviewed the charts of all pulmonary TH and MS patients operated upon by our group between 1980 and 1987. We found that chronic postsurgery pain syndrome was present in 12% (25 of 203) of TH patients as compared with 1% (1 of 114) of MS patients ( p 0.001). MS and TH were not different in terms of operative mortality and are similar to other large reports [10 13]. Complication rates were not statistically different between the groups. Of interest, the incidence of postoperative new atrial fibrillation was 15% in MS and 12% in TH patients. This compares with an 18% incidence in our coronary revascularization patients. As others and we have observed, MS patients experienced a shorter postoperative LOS compared with TH patients [2, 4]. More than 39% of MS patients were discharged in 5 days or less as compared with 27% of TH patients. In the later years of the study, as our incision bias tended towards MS, managed care mandated earlier discharge. However, the difference in LOS between the two groups was present in every time period during the study. The mortality and complications of pneumonectomy are similar to that reported in several large studies [10, 13] but are higher than that observed in others [14, 15]. A difference in mortality between right and left pneumonectomies has previously been noted. El-Oakley and Cameron reported an operative mortality of 37% in elderly patients undergoing right pneumonectomy compared with 6% for left pneumonectomy [16]. Nagasaki and associates observed 13.3% right pneumonectomy mortality with none for left pneumonectomy [17]. Morbidity and mortality rates for lobectomy through either incision were similar or lower than in other large reports [10, 11, 13]. Reoperations for control of bleeding were more frequent in the MS patients (2.4%), but most

378 ASAPH ET AL Ann Thorac Surg MEDIAN STERNOTOMY TO RESECT LUNG CANCER 2000;70:373 9 Table 5. Long-Term Survival Sternotomy Thoracotomy No. % No. % p Value 5 Year: Pneumonectomy 16 30.7% 19 29.7% NS Stage I 4 21.5% 8 38.1% NS Stage II 9 43.6% 6 25.0% NS Stage IIIA 3 21.4% 4 16.9% NS Stage IIIB 0 0.0% 1 33.3% NS Stage IV 2 43.2% 0 0.0% NS 10 Year: Pneumonectomy 6 10.1% 13 19.8% NS Stage I 2 7.2% 5 27.5% NS Stage II 1 8.7% 4 15.0% NS Stage IIIA 3 21.4% 3 11.25% NS Stage IIIB 0 0.0% 1 33.3% NS Stage IV 0 0.0% 0 0.0% NS 5 Year: Lobectomy 84 46.5% 57 46.3% NS Stage I 72 56.5% 35 56.9% NS Stage II 6 26.4% 14 32.7% NS Stage IIIA 6 25.4% 6 16.2% NS Stage IIIB 0 0.0% 2 40.0% NS Stage IV 0 0.0% 0 0.0% NS 10 Year: Lobectomy 36 6.0% 44 5.7% NS Stage I 31 39.1% 35 34.9% NS Stage II 5 26.4% 6 13.7% NS Stage IIIA 0 0.0% 1 7.6% NS Stage IIIB 0 0.0% 2 40.0% NS Stage IV 0 0.0% 0 0.0% NS occurred early in our series, and we attribute that to our learning curve. A major concern with MS in the treatment of lung cancer has been the fear of sternal wound infections and dehiscence [1, 3]. In our series, deep sternal wound infections and dehiscence were extraordinarily rare, particularly when compared with cardiac surgery. Lung cancer resections through a median sternotomy were associated with four of 447 (0.9%) deep sternal wound infections and no dehiscences, while there were eight (2.2%) deep wound infections and two (0.5%) dehiscences in the 368 thoracotomy patients. There were no wound complications in the 330 patients undergoing MS lobectomy, but in 249 TH lobectomies, there were three deep infections (1.2%) and one dehiscence (0.4%). By comparison, there were 44 (1.1%) deep wound infections and 16 (0.4%) dehiscences in 4,057 patients undergoing cardiac surgery through a median sternotomy at our institution between 1987 and 1997. While there is no statistical difference in wound complication rates between the two pulmonary resection groups or between cardiac and pulmonary MS operations, clearly, the wound complication risk is no higher for patients undergoing transternal pulmonary resection than it is with traditional thoracotomy or cardiac surgery. Our experience, showing an apparent difference in staging lymph node dissection between MS and TH, is a function of practice change and not real. We adopted routine thoracic lymphadenectomy in 1991 simultaneously with increase use of MS for pulmonary resection (Fig 1). However, even small or peripheral cancers can be associated with unsuspected N2/N3 lymph nodes metastases [18, 19], and skip nodal metastases are not rare [20]. Due to the technical ease, we routinely perform complete bilateral staging lymph node dissections on all patients approached through MS. Although we have not accrued sufficient 5-year survival data to confirm our impression, we believe complete lymphadenectomy will lead to up-staging and an improved survival for patients who are truly stage I. There is no compromise in cancer control using MS instead of TH. Our long-term survival results are similar to those reported in the literature [6, 21, 22]. Five-year survival for pneumonectomy and lobectomy is similar regardless of incision. Conclusions Median sternotomy provides shorter postoperative hospitalization, less postoperative pain, equivalent operative morbidity, operative mortality, and long-term survival with the potential for more complete staging when compared with thoracotomy in the surgical treatment of primary lung carcinoma.

Ann Thorac Surg ASAPH ET AL 2000;70:373 9 MEDIAN STERNOTOMY TO RESECT LUNG CANCER 379 References 1. Cooper JD, Nelems JM, Pearson FG. Extended indications for median sternotomy in patients requiring pulmonary resection. Ann Thoracic Surg 1978;26:413 20. 2. Urschel HC, Razzuk MA. Median sternotomy as the standard approach for pulmonary resection. Ann Thoracic Surg 1986;41:130 4. 3. Serry C, Bleck PC, Javid H, et al. Sternal wound complications. J Thoracic Cardiovasc Surg 1981;80:861 7. 4. Asaph JW, Keppel JF. Midline sternotomy for the treatment of primary pulmonary neoplasms. Am J Surg 1984;147: 589 92. 5. Beahrs OH, Henson DE, Hutter RVP, Kennedy BJ. Thorax. In: Beahrs OH, Henson DE, Hutter RVP, Kennedy BJ, eds. Manual for staging cancer, 4th ed. Philadelphia: JB Lippincott Co, 1992; 115 21. 6. Shield TW. Surgical treatment of non-small cell carcinoma of the lung. In: Shield TW, ed. General thoracic surgery, 4th ed. Philadelphia: Williams & Wilkins, 1994;1159 87. 7. Heitmiller RF. Thoracic incisions. In: Baue AE, ed. Glenn s thoracic and cardiovascular surgery, 6th ed. Stamford, CT: Appleton & Lange; 1996;73 89. 8. Lima O, Ramos L, DiBiasi P, Jadice L, Cooper JD. Median sternotomy for bilateral resection of emphysematous bullae. J Thoracic Cardiovasc Surg 1981;82:892 7. 9. Takita H, Merrin C, Didolkar MS, Douglass HO, Edgerton F. The surgical management of multiple lung metastases. Ann Thoracic Surg 1977;24:359 64. 10. Duque JL, Ramos G, Castrodeza J, et al. Early complications in surgical treatment of lung cancer: A prospective, multicenter study. Ann Thoracic Surg 1997;63:944 50. 11. Deslauries J, Ginsberg RJ, Piantadosi S, Fournier B. Prospective assessment of 30-day morabidity for surgical resections in lung cancer. Chest 1994 106 (Suppl):329 30. 12. Wang J, Olak J, Ferguson MK. Diffusing capacity predicts operative mortality but not long-term survival after resection for lung cancer. J Thoracic Cardiovasc Surg 1999;117:581 7. 13. Silvestri GA, Handy JR, Lackland D, Corley E, Reed CE. Specialists achieve better outcomes than generalists for lung cancer surgery. Chest 1998;114:675 80. 14. Harpole DH, Liptay MJ, DeCamp MM, Mentzer SJ, Swanson SJ, Sugarbaker DJ. Prospective analysis of pneumonectomy: risk factors for major morbidity and cardiac dysrhythmias. Ann Thoracic Surg 1996;61:977 82. 15. Deneffe G, Lacquet LM, Verbeken E, Vermaut G. Surgical treatment of bronchogenic carcinoma: a retrospective study of 710 thoracotomies. Ann Thoracic Surg 1988;45:380 3. 16. El-Oakley AJ, Cameron EWJ. Pneumonectomy for bronchogenic carcinoma in the elderly. Eur J Cardiothorac Surg 1994; 8:247 50. 17. Nagasaki F, Flehinger BJ, Martini N. Complications of surgery in the treatment of carcinoma of the lung. Chest 1982; 82:25 9. 18. Asamura H, Nakayama H, Kondo H, Tsuchiya K, Shimosato Y, Naruke T. Lymph node involvement recurrence, and prognosis in resected small, peripheral, non-small-cell lung carcinomas: are these carcinomas candidates for videoassisted lobectomy? J Thoracic Cardiovasc Surg 1996;111: 1125 34. 19. Graham ANJ, Chan KJM, Pastorino U, Goldstraw P. Systematic nodal dissection in the intrathoracic staging of patients with non-small cell lung cancer. J Thoracic Cardiovasc Surg 1999;117:246 251. 20. Watanabe W, Hayashi Y, Takabatake I, et al. Clinical significance of extended mediastinal lymph node dissections on the basis of clinico-pathological analysis of nodal involvement in bronchogenic carcinoma. Jpn J Thorac Surg 1994;47: 4 9. 21. Shah R, Sabanathan S, Richardson J, Mearns AJ, Goulden C. Results of surgical treatment of stage I and II lung cancer. J Cardiovasc Surg 1996;37:169 72. 22. Mountain CF. Revisions in the International System for Staging Lung Cancer. Chest 1997;111:1710 7.