Survival related to nodal status after sleeve resection for lung cancer

General Thoracic Surgery

Survival related to nodal status after sleeve resection for lung cancer Sleeve lobectomy is a lung-saving procedure indicated for central tumors for which the alternative is a pneumonectomy. The relation between survival and nodal status is controversial because, in most series, the presence of Nl disease adversely affects the prognosis with few or no long-term survivors. During the period 1972 to 1992, 142 patients underwent sleeve resection for lung cancer at our institution. Mean age (±standard deviation) was 60.7 ± 9.1 years (range 11 to 78 years), and indications for operation were a central tumor in 112 patients (79 %~ a peripheral tumor in 18 patients (13% ~ and compromised pulmonary function in 12 patients (8%). Histologic type was predominantly squamous (72.5 %) foUowed by nonsquamous (24.6 %) and carcinoid tumors (2.8%). Resection was complete in 124 patients (87%) and incomplete in 18 (13% ~ and the operative mortality was 2.1 % (n = 3). FoUow-up was complete for the 139 remaining patients. Including operative deaths, survivals at 5 and 10 years for aU patients were 46 % (95 % confidence intervals 38 % to 55 %) and 33 % (95% confidence intervals 24 % to 42 % ~ respectively. For patients with NO status (n = 73~ 5- and 100year survivals were 57 % (95 % confidence intervals 45 % to 69 %) and 46 % (95 % confidence intervals 32 % to 60 %); for patients with Nl status (n = 5~ these rates were 46 % (95 % confidence intervals 32 % to 60%) and 27% (95% confidence intervals 14% to 40%) (p = 0.13). No patient with N2 status (n = 14) survived 5 years. Local recurrences occurred in 23% of cases, but the prevalence was not statisticaUy different between patients with NO disease (16.6%) and Nl disease (23.1 %) (p = 0.43). These data suggest that sleeve resection is an adequate operation for patients with resectable lung cancer and NO Nl status. The presence of N2 disease significantly worsens the prognosis and may contraindicate the use of the procedure. (J 'fHORAC CARDIOVASC SURG 1994;107:576-83)

Reza John Mehran, MDCM, MSc, FRcs(C)a (by invitation), Jean Deslauriers, MD, FRCS(C),a, b Michel Piraux, MD, FRcs(C)a (by invitation), Maurice Beaulieu, MD, FRcs(C)a (by invitation), Chantal Guimont, MSc* (by invitation), and Jacques Brisson, MD, DSc** (by invitation), Ste-Foy, Quebec, Canada

From the Division of Thoracic Surgery; Centre de Pneumologie de Laval; and Department of Surgery," Laval University, Ste-Foy, Quebec, Canada. Read at the Seventy-third Annual Meeting of The American Association for Thoracic Surgery, Chicago, Ill., April 25-28, 1993. Address for reprints: Jean Deslauriers, MD, FRCS(C), 2725 Chemin Ste-Foy, Ste-Foy, Quebec, Canada, G I V 4G5. "Statistician. Laval University Epidemiology Research Group. **Epidemiologist, Laval University Epidemiology Research Group. Copyright © 1994 by Mosby-Year Book, Inc. 0022-5223/94 $1.00 + .10

576

12/6/51770

Sleeve lobectomy is a lung-saving procedure indicated for central tumors for which the alternative is a pneumonectomy. It preserves otherwise normal lung tissue and enables surgical resection to be done in patients with inadequate pulmonary reserve. The first successful bronchoplasty is credited to PriceThomas,' but the first significant and comprehensive report on the use ofthese techniques is that of Paulson and Shaw- in 1955. They reviewed the case histories of 18 patients in whom sleeve resection had been done for a variety of reasons, including benign conditions, and they

The Journal of Thoracic and Cardiovascular Surgery Volume 107, Number 2

stressed the importance of preserving functional parenchyma in lung cancer operations. Sleeve lobectomy was then considered a compromise operation for patients whose pulmonary reserve was believed to be inadequate to permit pneumonectomy. Since then, many reports have suggested that sleeve resection accomplishes tumor and nodal clearance similar to that of pneumonectomy with the possible advantages oflower operative mortality rates, equal if not better survival rates, and improved quality of life. 3- 18 In most series, the overall 5-year survival after sleeve resection carried out for lung cancer is 40% to 45% but the presence of N 1 disease adversely affects the prognosis with few or no long-term survivors. This report presents a comparative survival analysis based on nodal status for 142 patients who underwent sleeve resection for lung cancer during a 20-year interval. The results indicate that there is no statistically significant difference in survival between patients with NO and Nl status.

Patients and methods General characteristics of the study population. Between January 1972 and December 1991, 142 patients underwent sleeve resection as primary treatment for bronchogenic carcinoma. There were 123 men and 19 women whose ages at the time of operation ranged from 11 to 78 years with a mean of 60.7 ± 9.1 years (mean ± standard deviation) (median age of 61 years). The first bronchoplasty was done in May 1972 and, from that date until January I, 1980, 22 patients had a sleeve resection. The remaining 119 patients had operation during the eighties and early nineties. We now average 8 to 10 cases per year, which is approximately 5% of our total volume of resectional surgery for lung cancer. Sleeve resection was usually done by choice in patients who would have been able to tolerate pneumonectomy but in whom the lesion could be completely resected by a lesser bronchoplastic procedure. In this group, indications for operation were a central tumor in 112 patients (79%) or a peripheral tumor with nodal involvement at the bronchopulmonary level (N I) in 18 patients (13%). In all of those patients, the tumor could not have been removed by conventional lobectomy because the line of resection would have either transected the tumor or provided an inadequate margin. In 12 additional patients, all of whom had severe respiratory impairment, the bronchoplasty was done as a compromise to pneumonectomy. Mediastinoscopy, with random node sampling at three different levels, was done in nearly every case (120/142) for pretreatment staging of the superior mediastinum. Only 6 patients had a bronchoplasty when nodal metastasis was identified at mediastinoscopy. Operative procedures. Although the initial finding that identified a possible candidate for sleeve resection was the bronchoscopic appearance of the tumor that extended to the lobar orifice and toward the mucosa of the adjacent main bronchus, the final decision was made at the time of thoracotomy. Operative findings such as evidence of disease along the resection line of the lobectomy, extraluminal extension ofthe carcinoma to the main bronchus, or presence of metastatic nodes at the hilum of

Mehran et al.

577

Table I. Types of sleeve resections Side procedure Right lung Upper lobe Upper and middle lobes Middle and lower lobes Left lung Upper lobe Lower lobe Main bronchus

No. ofpatients 93 9 2 104 (73.2%)

31 5

2 38 (26.8%)

the lobe to be removed were considered possible indications for sleeve resection. Procedures were considered complete if all gross carcinoma had been removed, the resection margin was free of disease, and the highest mediastinal node was free of tumor. The majority of sleeve resections were done for central tumors involving the bronchial origin of either upper lobe (n = 124, 87%) (Table 1). In 9 patients, the middle lobe had to be taken in continuity with the right upper lobe because of tumor growth across the fissure or because of nodal disease around the bronchus intermedius. Seven patients had a lower lobe sleeve resection, and in 2 patients a tumor located in the left main bronchus could be resected without sacrificing any of the distal lung. No patient in this series required a concomitant full sleeve resection of the pulmonary artery (double sleeve resection) and no patient had preoperative radiotherapy. Resection was complete in 124 patients and incomplete in 18 patients usually because of diseased bronchial resection margins. Histopathology and stage of disease. Pathologic diagnosis was obtained by the study of surgical specimens in all patients (Table II). The majority of resected neoplasms (72.5%) were squamous carcinomas and their prevalence was similar in the NO group and in the N 1 group. Thirty-five patients had nonsquamous carcinomas and 4 patients had a typical carcinoid tumor, all with NO status. For the purpose of survival analysis, the squamous group was considered to include patients having pure squamous disease, whereas the nonsquamous group included all other patients, including some with a mixture of cell types. All tumors were pathologically staged in accordance with the TNM terminology as modified in 1986 by Mountain 19 and the American Joint Committee on Cancer Staging. From 1972 to 1978 staging was done retrospectively, but for the remaining years (1978 to 1992) staging was based on complete clinical and pathologic analysis of the primary tumor (T status) and intraoperative sampling of at least the bronchopulmonary, hilar, and mediastinal (superior, posterior, and anterior mediastinum) nodal areas (N status). Seventy-three patients (51.4%) had NO disease and 55 (38.7%) had N 1disease, and the T status distribution shows that it is nearly identical for the two groups (Table II). Stage distribution of the patients included in the series also shows that the majority are either in stage I (43%, 61/142) or in stage II (38.8%,48/142) categories. Among the 31 patients with stage IIIa disease, 14 had N2 disease. Two patients had T4 tumors (stage IIIb) because of a nonmalignant exudative pleural effusion.

578

The Journal of Thoracic and Cardiovascular Surgery February 1994

Mehran et al.

Table II. Pathology and stage of 142 patients included in the comparison of survival by nodal status Nodal status NO

NI

N2

(n = 73)

(n =55)

Histologic type Squamous cell carcinoma Nonsquamous cell carcinoma Carcinoid tumor T status

55 (75.3%) 14 (19.2%) 4 (5.5%)

43 (78.2%) 12 (21.8%) 0

5 (35.7%) 9 (64.3%) 0

103 (72.5%) 35 (24.6%) 4 (2.8%)

6 (8.2%) 55 (75.3%) 12 (16.5%)

5 (9.1%) 43 (78.2%) 7 (12.6%)

I (7.1%) 9 (64.3%) 4 (28.6%)

12 (8.5%) 107 (75.3%) 23 (16.2%)

Tl T2 T3-T4

Table III. Major operative complications after sleeve

resection Complication

Early (n = 16) Atelectasis/pneumonia Atelectasis/ empyema

Atelectasis/bronchopleural fistula

Respiratory failure Others Late (n=4) Granulation at sutureline Bronchial stricture

(n =

Total

Characteristic

14)

(n = /42)

by the log-rank test. 22 , 23 Survival data are given with the 95% confidence interval (Cl).

No. ofpatients

Results 4 2 I

2 7

2 2

Follow-up. Patients were observedfrom the date of operation to the time of death or end of observation (July 1992), whichever came first. No patient was lost to follow-up. The potential duration of follow-up was at least 12 months for all patients under review and the mean follow-up time was 2149 days. A local recurrence was defined as tumor growth within the ipsilateral hemithorax or mediastinum, or both. For patients who died, the cause of death and the site of first recurrence were determined from clinical records, death certificates, or from information provided by family physiciansor by relatives ofthe patient. Each death was classifiedas being due to (I) lung cancer, including second primary lesions and operative deaths, (2) causes unrelated to lung cancer, or (3) unknown causes. Sites of firstrecurrences wereclassifiedas being (I) local,(2) distant, (3) local and distant, (4) unknown site, or (5) second primary defined as a tumor of different histologic type or a tumor of similar histologictype if it occurred more than 2 years after the first operation or if its origin could be traced to a carcinoma in situ.2o, 2\ Operative mortality was defined as a death occurring within 30 days of the operation or a death directly related to the procedure even if it occurred more than 30 days after the operation. Statistical analysis. Results are presented as means (± standard deviation) for continuous variables and as percentages for categoric data. The x 2 test was used to compare these proportions, The analysis of survival was based on deaths from all causes. Survival was analyzed according to the Kaplan-Meier method and differences between survivalcurves were compared

Operative morbidity and mortality. Three patients died (2.1% operative mortality rate) during the immediate postoperative period. Two of the three deaths were the result of infectious complications in the reimplanted lung and the third operative death was due to a pulmonary embolus. Early nonfatal complications were seen in 16 additional patients and these are listed in Table III. Four patients had a late complication at the site of the bronchial anastomosis. Among those, 2 had granulation tissue that was successfully removed with a bronchoscope, and 2 had a fibrous stricture that was dilated in I but necessitated completion pneumonectomy in the other. Survival data. The observed survival rates for the entire group and each subset of the N descriptor are shown in Figs. 1 and 2. The 5- and 10-year survivals for all patients included in the analysis are 46% (95% CI 38% t055%)and 33% (95% CI 24% t042%), respectively, with a median survival time of 3.90 years. For patients with NO status, 5- and 10-year survivals were 57% (95% CI 45% to 69%) and 46% (95% CI 32% to 60%) (median survival of 8.02 years); for patients with N 1 status, these survivals were 46% (95% CI 32% to 60%) and 27% (95% CI 14% to 40%), respectively (median survival of 3.90 years). Among the 14 patients with N2 status, none survived more than 3 years after the resection (median survival of 1.53 years). The difference in survival among these three groups is highly significant (p < 0.001), but the difference in survival between patients with NO disease and patients with Nl disease is statistically not significant (p = 0.13). Survival data obtained by comparison of stage (Fig. 3) shows that for those patients with stage I disease, the 5and 1O-year survivals are 63% (95% CI 50% to 76%) and 52% (95% CI 37% to 68%) and for patients with stage II

The Journal of Thoracic a-nd Cardiovascular Surgery Volume 107, Number 2

Mehran et al.

579

% 100

80

60

46% 40

33%

20

O<.---------------------~

o

2

3

4

5

6

7

8

9

10

45

37

28

24

20

YEAR

Patients at risk 142

113

89

74

59

54

Fig. 1. Life-table analysis showing the percentage of all patients remaining alive after sleeve resection of bronchogenic carcinoma. % 100

80

60

~% 40

No 27%

N,

20 P

0

< 0.001

N2 2

0

3

4

5

6

7

8

9

10

YEAR

Patients at risk

No

73

60

48

43

36

32

28

24

17

15

12

N,

55

44

35

30

23

22

17

13

11

9

8

N2

9

6

0

0

0

0

0

0

0

0

Fig. 2. Life-table analysis by nodal status showing the percentage of all patients remaining alive after sleeve resection of bronchogenic carcinoma.

disease, these rates are 48% (95% CI 33% to 63%) and 29% (95% CI 14% to 43%) respectively. Only 14% of patients with stage III disease survived 5 years and all of them were in the T3 NO or T3 N I categories.

Site of first recurrence and cause of death. From the beginning of the study, 86 patients died, and the cause of death was related to cancer (n = 66) in the majority of cases. The sites of first recurrences according to the N

The Journal of Thoracic and

580

Cardiovascular Surgery February 1994

Mehran et al.

% 100

80 63% 60

STAGE 1

40

20 7%

p < 0.001

0

a

STAGE 3

2

3

4

5

6

7

8

9

10

YEAR

Patients at risk STAGE 1

61

49

42

39

33

30

26

23

16

14

11

STAGE 2

48

40

32

28

21

20

16

12

10

8

7

STAGE 3

33

24

5

4

3

2

2

2

2

15

7

Fig. 3. Life-table analysis by stage showing the percentage of all patients remaining alive after sleeve resection of bronchogenic carcinoma.

Table IV. Site offirst recurrence in 142 patients included in the comparison of survival by nodal status Nodal status

NO

Site offirst recurrence

Local recurrences (local alone plus local and distant) Distant recurrences Unknown site Second primary carcinoma

In = 73) 12 6 I 10

status are given in Table IV. The overall rate of local recurrence was 23%, but the distribution was not significantly different between patients with NO and patients with Nl disease (p = 0.43). When the resection was complete, 21 patients (17%) had a local recurrence. By comparison, when the resection was incomplete, II patients (61%) had a local recurrence (p < 0.001). Local recurrence was also more common in patients with N2 disease (p < 0.001). Discussion

Sleeve resection of the bronchus was introduced to avoid pneumonectomy in proximal lung tumors and, in most series of bronchoplasties done for lung cancer, lesions in the hilum of the right upper lobe constitute the commonest indication for operation. The frequency of this location relates to the anatomic structure of the right main bronchus and the relatively long bronchus interme-

(16.6%) (8.2%) (1.4%) (13.9%)

NI 55)

In = 12 II 2 7

(23.1 %) (20%) (3.6%) (13.5%)

N2

In = 14) 8 (57.1%) 3 (21.4%) I (7.1%)

o

Total

In = 32 20 4 17

142)

(23%) (14.1%) (2.8%) (12.2%)

dius. The anatomic justification for sleeve resection of the upper lobes was provided by Nohl,24 who described, in 1956, the lymphatic and vascular spread of bronchogenic carcinoma and emphasized the predictability of upper lobe lymphatic drainage. He showed that right upper lobe lesions metastasize to nodes located below the right upper lobe bronchus and seldom involve nodes located below a line drawn from the middle lobe bronchus to the apical bronchus of the lower lobe. Anatomic justification was further provided by Maeda and colleagues.P who reported on 47 patients who were pathologically studied to identify local tumor spread. They showed that lung cancer not only infiltrates around the primary tumor for a distance averaging 7.2 mm (range of 3 to 20 mm) but also that metastatic clusters of cancer cells are often found in the lymphatics, nodes, bronchial arterials, and nerve sheaths in the peribronchial and intrabronchial wall. In 1981, Immerman and colleagues/" reviewed surviv-

The Journal of Thoracic and Cardiovascular Surgery Volume 107, Number 2

Mehran et al.

58 1

Table V. Survival after sleeve resection for lung cancer First author (yr)

Bennett" (1978) Faber!" (1984) Frist 14 (1987)

Watanabe" (1990) Van Schill I 7 (1991)

Years of study

No. ofpatients

Five-year survival (%)

Ten-year survival (%)

1958-1973 1962-1982 1962-1986 1975-1990 1960-1989

80 101 63 79 145

34 30 58 45 49

20 22

al and sites of recurrences in 99 patients with non-smallcell stages I and II carcinoma of the lung resected for cure from 1967 through 1975. The overall5-year disease-free survival was 45%. In patients with stage I disease (Tl NO, T2 NO, Tl N 1) the 5-year disease-free survival was 55%, and in those patients with stage II disease (T2 N 1) it was 14%. Treatment was unsuccessful in 44 (44%) of the 99 patients, and the site of the first relapse was a local recurrence in 18 patients (18%) and a distant metastasis in 26 (26%). In the Ludwig Lung Cancer Study Group analysis of patterns of failure in patients with resected stage I and II non-small-cell carcinoma of the lung,27 1012 patients were analyzed and the site of initial failure for the 426 patients with objective evidence of failure recorded. This site was intrathoracic in 231 patients (23% of patients having resection) and extrathoracic including supraclavicular nodes in 195 patients (19%). Other studies have shown similar patterns of recurrences after complete resections of early-stage lung cancer. 28-30 In a review of the literature over a 12-year period (1980 to 1992), Tedder and colleagues'f reported that the 5-year survival of patients undergoing sleeve lobectomy for bronchogenic carcinoma was 40% (Table V). Sleeve lobectomy for stage I, II, and III disease yielded 63%, 37%, and 21% 5-year survivals, and absence of nodal involvement was associated with a 60% 5-year survival. These survival figures are comparable with those seen after conventional lobectomy but much better than the 5-year survivals reported after pneumonectomy.I' The use of sleeve resection when the carcinoma has spread to bronchopulmonary or hilar nodes (N 1 disease) is an area of controversy even when both tumor and nodes can be completely resected. In a group of 76 patients with squamous carcinoma, Firmin and associates? reported 5and 1O-year survivals of 71% and 48.5%, respectively, for patients with uninvolved hilar nodes. When the tumor had metastasized into the hilar nodes, the survival was significantly worse, with a 5-year survival of 17% and a 10-year survival of only 10% (p < 0.001). In 1991, Van Schill and coworkers'? also studied the impact on survival of nodal status in 112 patients with T2 and T3 squamous cell carcinoma who had had a sleeve lobectomy between 1960 and 1989. The 5- and 10-year survivals for patients with

37

NO disease (n = 52) were 59% and 47%, for those with Nl disease (n = 51) 21% and 0%, and for those with N2 disease (n = 9) 44% and 0%, respectively. They raised the question as to whether patients with N 1 disease should be regarded as having systemic disease and be treated accordingly with neoadjuvant therapy, but they did not specifically conclude that sleeve resection should be avoided in patients with Nl disease. The results in the present series are concordant with those of Naruke, 15 who reported 5-year survivals of 50% for patients with NO disease and 45.9% for patients with Nl disease, and they indicate that sleeve lobectomy should be considered in any case of lung cancer with NO and N 1 status that can be completely resected by this technique. Because the definition of complete resection is essential to this decision, intraoperative frozen section evaluation of lobar, hilar, and mediastinal nodes, and ultimately frozen section biopsy specimens of both resection margins, are a critical feature of the operation. The presence of diseased interlobar nodes between the right upper and middle lobes is not a contraindication to sleeve resection, but the middle lobe might have to be included in the bronchial sleeve. On the basis of our results, lymph node involvement has an adverse effect on survival, but this effect is significant only in the N2 category, in which no patient survived 5 years after operation. For the 55 patients included in the N 1 subset, the 5-year survival figure of 46% is not significantly different from that observed for the NO subset (57%) and indicates that patients with NO or Nl status can expect a survival after sleeve resection similar to that expected after conventional procedures for similar stage tumors.P: 27, 30 The difference in survival at 10 years (NO: 46%; N 1:27%) and the difference in median survival (NO: 8.02 years; Nl: 3.90 years) are greater and in keeping with the pathologic stage of disease (stage I versus stage

11).

The prevalence of local recurrence of 23% is also not significantly different between these two groups and is similar to what is reported after all types of resection. 26-30 In this series, the prevalence of second primary tumors is notable at 12.2% and it is higher than that reported by Van Schill and associatesl- (7.6%). This observation

582

The Journal of Thoracic and Cardiovascular Surgery February 1994

Mehran et af.

illustrates one of the most important features of sleeve resection, which is to preserve otherwise normal lung tissue enabling further surgical resection to be done. REFERENCES 1. Price-Thomas e. Conservative resection of the bronchial tree. JR Coll Surg Edin 1955;1:169-86. 2. Paulson DL, Shaw RR. Preservation of lung tissue by means of bronchoplastic procedures. Am J Surg 1955; 89:347-55. 3. Paulson OL, Urschel HC, McNamara JJ, Shaw RR. Bronchoplastic procedures for bronchogenic carcinoma. J THORAC CARDlOVASC SURG 1970;59:38-48. 4. Bennett WF, Smith RA. A twenty-year analysis of the results of sleeve resection for primary bronchogenic carcinoma. J THORAC CARDlOVASC SURG 1978;76:840-5. 5. Weisel RD, Cooper JD, Delarue NC, Theman TE, Todd TRJ, Pearson FG. Sleeve lobectomy for carcinoma of the lung. J THORAC CARDlOVASC SURG 1979;78:839-49. 6. Shaw KM, Luke DA. Lobectomy with sleeve resection of the bronchus for malignant disease of the lung and the influence of the suture material used for the bronchial repair. Thorac Cardiovasc Surg 1979;27:325-9. 7. Ungar I, Gyeney I, Scherer E, Szarvas I. Sleeve lobectomy: an alternative to pneumonectomy in the treatment of bronchial carcinoma. Thorac Cardiovasc Surg 1981;29: 41-6. 8. Lowe JE, Bridgman AH, Sabiston DC. The role of bronchoplastic procedures in the surgical management of benign and malignant pulmonary lesions. J THORAC CARDlOVASC SURG 1982;83:227-34. 9. Firmin RK, Azariades M, Lennox SC, Lincoln JCR, Paneth M. Sleeve lobectomy (lobectomy and bronchoplasty) for bronchial carcinoma. Ann Thorac Surg 1983; 35:442-9. 10. Faber LP, Jensik RJ, Kittle CF. Results of sleeve lobectomy for bronchogenic carcinoma in 101 patients. Ann Thorae Surg 1984;37:279-85. II. Deslauriers J, Gaulin P, Beaulieu M, Piraux M, Bernier R, Cormier Y. Long-term clinical and functional results of sleeve lobectomy for primary lung cancer. J THORAC CARDlOVASC SURG 1986;92:871-9. 12. Keszler P. Sleeve resection and other bronchoplasties in the surgery of bronchogenic tumors. Int Surg 1986;71:229-32. 13. Vogt-Moykopf I, Fritz TH, Meyer G, Biilzerbruck H, Daskos G. Bronchoplastic and angioplastic operation in bronchial carcinoma: long-term results of a retrospective analysis from 1973 to 1983. Int Surg 1986;71:211-20. 14. Frist WH, Mathisen OJ, Hilgenberg AD, Grillo He. Bronchial sleeve resection with and without pulmonary resection. J THORAC CARDlOVASC SURG 1987;93:350-7. 15. N aruke T. Bronchoplastic and bronchovascular procedures of the tracheobronchial tree in the management of primary lung cancer. Chest I989;96(suppl):53s-6s. 16. Watanabe Y, Shimizu J, ada M, et al. Results in 104 patients undergoing bronchoplastic procedures for bronchiallesions. Ann Thorac Surg 1990;50:607-14.

17. Van Schill PE, de la Riviere AB, Knaepen PJ, van Swieten HA, Defaww JJ, van den Bosch JM. TNM staging and long-term follow-up after sleeve resection for bronchogenic tumors. Ann Thorac Surg 1991;52:1096-101. 18. Tedder M, Anstadt MP, Tedder SD, Lowe JE. Current morbidity, mortality, and survival after bronchoplastic procedures for malignancy. Ann Thorac Surg 1992;54:38791. 19. Mountain CF. A new international staging system for lung cancer. Chest 1986;89(Suppl):225-33. 20. Martini N, Melamed MR. Multiple primary lung cancers. J THORAC CARDIOVASC SURG 1975;70:606-12. 21. Martini N, Ghosn P, Melamed MR. Nodal recurrence and new primary carcinoma after resection. In: Delarue NC, Eschapasse H, eds. International trends in general thoracic surgery. vol I. Philadelphia: WB Saunders, 1985:164-9. 22. Fleiss JL. Statistical methods for rates and proportions. 2nd ed. New York: John Wiley, 1981. 23. Lawless JF. Statistical models and methods for lifetime data. New York: John Wiley, 1982. 24. Nohl He. An investigation into the lymphatic and vascular spread of carcinoma of the bronchus. Thorax 1956; 11:172-85. 25. Maeda M, Nanjo S, Nakamura K, and Nakamoto K. Tracheobronchoplasty for lung cancer. Int Surg 1986;71: 221-8. 26. Immerman SC, Vanecko RM, Fry WA, Head LR, Shields TW. Site ofrecurrence in patients with stages I and II carcinoma of the lung resected for cure. Ann Thorac Surg 1981;32:23-7. 27. The Ludwig Lung Cancer Study Group: Patterns of failure in patients with resected stage I and II non-small-cell carcinoma of the lung. Ann Surg 1987;205:67-71. 28. Pairolero PC, Williams DE, Bergstralh EJ, Piehler JM, Bernatz PE, Payne WS. Postsurgical stage I bronchogenic carcinoma: morbid implications of recurrent disease. Ann Thorac Surg 1984;38:331-8. 29. Field R, Rubinstein LV, Weisenberger TH, Lung Cancer Study Group. Sites of recurrences in resected stage I nonsmall cell lung cancer: a guide for future studies. J Clin Oncol 1984;2:1352-8. 30. Iascone C, DeMeester TR, Albertacci M, Little AG, Golomb HM. Local recurrence of resectable non-oat cell carcinoma of the lung: a warning against conservative treatment for NO and N I disease. Cancer 1986;57:471-6. 31. Shields TW. Carcinoma of the lung. In: Shields TW, ed. General thoracic surgery. Philadelphia: Lea & Febiger, 1972:837. 32. Van Schill PE, dela Riviere Als, Knaepen PJ, van Swieten HA, Defauw JJ, van den Bosch JMM. Second primary lung cancer after bronchial sleeve resection: treatment and results in eleven patients. J THORAC CARDIOVASC SURG 1992;104:1451-5.

Discussion Dr. Willard A. Fry (Evanston, Ill.). I raise one question. You have highlighted the dismal prognosis of N2 disease, but I sub-

The Journal of Thoracic and Cardiovascular Surgery Volume 107, Number 2

mit that if the disease has been eradicated with a sleeve lobectomy, palliation for those patients is probably better than otherwise. I am not sure that I would agree with the conclusion that that treatment should be avoided. I would certainly hesitate to perform a right pneumonectomy on a patient who is going to have a short survival time. Dr. Deslauriers. The answer to that question is very personal because I seldom do palliative surgery of any kind. In my opinion, resection should always be potentially curative even if it is done as a compromise procedure. The second reason why I think bronchoplastic procedures may not be adequate for N2 disease is that, at least in theory, not as many nodes are being removed as would otherwise be removed with a pneumonectomy. Nodes are being left caudally along the bronchus intermedius, and this possiblyis the reason that the failure rate is so high. Dr. Thomas R. J. Todd (Ottawa, Ontario, Canada). It seems to me that what we classify as an N I lymph node is a very large nodal sump area that can extend anywhere from a lymph node that is intra parenchymal near the tumor out to a lymph node right in the hilum near the pleural reflection. In your series of patients, did you make any attempt to differentiate the level of those nodes, specifically those that might have been at the level of the main-stem bronchus or the lobar orifice versus those that were identified by the pathologist when the specimen was examined? Dr. Deslauriers: This was not done because it is extremely

Mehran et al.

583

difficult, even prospectively, to identify the specific sites of N I subsets. This identification is also difficult for the pathologist who dissects the resected lung. Dr. Van Schill from the Netherlands did look at survival related to location ofNI disease and he found no differences between patients with metastatic disease in various N I sites. Dr. Thomas Shields (Chicago, Ill.). What is your opinion concerning a concomitant angioplastic procedure? Vogt-Moykopf is a very strong advocate of this procedure, particularly with the left upper lobe. The Japanese, like you, advocate doing a pneumonectomy. Is that your philosophy or do you not see this involvement very often? Dr. Deslauriers. We sometimes see this degree of involvement, but if the patient can tolerate a pneumonectomy, we prefer to do a pneumonectomy. The double sleeve procedure, which is a bronchial sleeve with a main artery sleeve, is a good operation but carries a significantly higher mortality rate. VogtMoykopf reports a 16% to 18% mortality rate with a double sleeve operation. On the left side, a pneumonectomy is well tolerated by most individuals, so we have no hesitation to carry it out whenever necessary. In a patient with compromised function, a partial resection of the pulmonary artery can be accomplished with proximal and distal clamping. The arterial wall can then be replaced with a pericardial patch. This procedure is fairly easy and is well tolerated. If there are any doubts about tumor clearance, a pneumonectomy should always be done.