Prognostic significance of surgical-pathologic N1 disease in non-small cell carcinoma of the lung

Prognostic Significance of Surgical-Pathologic N1 Disease in Non-Small Cell Carcinoma of the Lung Marc Riquet, MD, Dominique Manac’h, MD, Franc¸oise Le Pimpec-Barthes, MD, Antoine Dujon, MD, and Antoine Chehab, MD Service de Chirurgie Thoracique, Hoˆpital Laennec, Paris, and Centre Chirurgical du Ce`dre, Boisguillaume, France

Background. N1 disease represents a heterogeneous group of non-small cell lung carcinoma with varying 5-year survival rates. Specific types of N1 lymph node involvement need to be further investigated and their prognostic significance clarified. Methods. From 1984 to 1993, 1,174 patients with nonsmall cell lung cancer had complete mediastinal lymph node dissection: N0, 50.25% (n 5 590); N1, 21.8% (n 5 256); and N2, 27.95% (n 5 328). The N1 subgroup cases were reviewed. Four levels of N1 nodes were identified using the New Regional Lymph Node Classification for Lung Cancer Staging. Their prognostic significances were tested and 5-year survival rates were compared with those of N0 and N2 patients of the whole group. Results. The overall 5-year survival rate of N1 patients was 47.5%. Survival was not related to site of the primary lung cancer, pathologic T factor, histologic type, type of

resection, number of N1 station involved, nor type of N1 involvement (direct extension or metastases). Five-year survival was significantly better when N1 involvement was intralobar (levels 12 and 13, n 5 102), as compared with extralobar (hilar) involvement (levels 10 and 11, n 5 154): 53.6% versus 38.5% (p 5 0.02). Intralobar N1 5-year survival was similar to that of N0 (53.6% vs 56.5%, p 5 0.01), and extralobar 5-year survival with that of N2 (38.5 vs 28.3%, p 5 0.01) when N2 was present in only one station in the ipsilateral mediastinum. Conclusions. N1 disease is a compound of two subgroups: one located inside the lobes is related to N0, and the other (extralobar or hilar) behaves like an early stage of N2 disease. This offers further information for clinical, therapeutic, and research purposes. (Ann Thorac Surg 1999;67:1572– 6) © 1999 by The Society of Thoracic Surgeons

N

Patients and Methods

1 disease represents a heterogeneous group of nonsmall cell lung carcinoma with varying 5-year survival rates. According to the new Tumor, Node, Metastasis (TNM) staging [1], N1 status of lung cancer is encountered in stage IIA (T1 N1), IIB (T2 N1), IIIA (T3 N1), and IIIB (T4 N1). Five-year surgical-pathologic survival rates are, respectively, 55%, 39%, and 25% for the first three stages, and not published for the IIIB group; only clinical-pathologic stage IIIB is reported to offer a 7% 5-year survival rate. If only N status is considered, N1 5-year survival rate becomes 47.5%, an intermediate value between those of N0 (56.3%) and N2 (20%) [2]. N1 disease may involve one or more node levels in the lung [3]. The number of involved nodes [4] or their levels in the lung [5, 6] have an influence on the surgicalpathologic 5-year survival rates. Some authors even suggest that N1 disease is an underestimated N2 disease [5, 7]. The purpose of our study was to evaluate the significance of N1 disease in lung cancer and to better understand its place as compared with N0 and N2 involvement.

Accepted for publication Dec 22, 1998. Address reprint requests to Dr Riquet, Service de Chirurgie Thoracique, Hoˆpital Laennec, 42 Rue de Se`vres, 75007 Paris, France; e-mail: [email protected].

© 1999 by The Society of Thoracic Surgeons Published by Elsevier Science Inc

From 1984 to 1993, 1,174 patients underwent pulmonary resection at Laennec Hospital and Boisguillaume Surgical Center for bronchogenic carcinoma. All patients had non-small cell carcinoma; the surgical procedure was a complete potentially curative resection with an extensive mediastinal lymph node dissection similar to that described by Martini and Flehinger [8]. Overall 5-year survival rate was 45%; “zero time” was the date of surgery and March 31, 1996 was the closing date. Surgical-pathologic distribution of N disease was: N0, 50.25% (n 5 590); N1, 21.8% (n 5 256); and N2, 27.95% (n 5 328). The group of patients with N1 disease formed the basis of this study. There were 237 men and 19 women. Mean age was 60.3 years (range 37 to 81 years). Tumors were histologically classified as: squamous cell carcinoma (n 5 181), adenocarcinoma (n 5 53), undifferentiated large cell carcinoma (n 5 10), adenosquamous (n 5 8), and miscellaneous (n 5 4). Tumor location was the right lung in 132 cases and the left one in 124; 147 were located inside one or two segments, and 109 involved a lobar, an intermediate, or a main bronchus. Complete resection consisted of: pneumonectomy (n 5 179), lobectomy (n 5 64), and bilobectomy (n 5 13). Tumors presented as: T1 in 64 patients (25%), T2 in 125 (48.8%), T3 in 60 (23.4%), and T4 in 7 (2.3%). Postoperatively, 160 patients received adjuvant external radiotherapy; this was performed according 0003-4975/99/$20.00 PII S0003-4975(99)00276-3

Ann Thorac Surg 1999;67:1572– 6

RIQUET ET AL SIGNIFICANCE OF N1 INVOLVEMENT IN LUNG CANCER

Table 1. Number of N1 Node Levels and Their Location Inside the Lung

Level Involved (location) 10 11 12 13 Total Number of patients

One Level Involved: No. (%)

Two Levels or More Involved: No. (%)

Total

9 (6.25) 41 (28.47) 46 (31.94) 48 (33.33) 144 144

17 (15.2) 96 (85.8) 100 (89.3) 40 (35.7) 253 112

26 (10.15) 137 (53.51) 146 (57.03) 88 (34.37) 397 256

10 5 hilar; 11 5 interlobar; 12 5 lobar; 13 5 segmental [3]. Percentage of patients suffering one level of N1 node involvement is 56.25%; total number of involved node levels is 397.

to specific management regimens adopted by the different referring physicians, but was not randomized. Four levels of N1 nodes were identified as revised in the New Regional Lymph Node Classification for Lung Cancer Staging [3], initially based on the mapping by Naruke and associates [9]: hilar (level 10), interlobar or peribronchial (level 11), lobar (level 12), and segmental (level 13). Of the 256 resected tumors, 144 (56.25%) had only single-level metastases (Table 1). Lymph nodes were characterized as being invaded by either direct extension (defined as infiltration of the neoplastic process into an adjacent lymph node) or by metastases. Difference between the different lymph node levels, even in very large tumors with direct extension, was always possible to establish, either macroscopically or by histology. However, the number of involved nodes was not counted. N1 node levels were further divided between intralobar levels (12 and/or 13) and extralobar or hilar levels (10 and/or 11). Disease limited to the intralobar levels was observed in 101 patients. Actuarial survival curves were calculated by the Kaplan Meier method [10]. The statistical comparisons were made using the log-rank test [11]. Survival was analyzed according to the following: 1) primary tumor characteristics including site, histology, T classification, type of resection performed, and performing postoperative adjuvant external radiotherapy; and 2) N1 characteristics including location and number of involved levels, invasion by either direct extension or metastases, modality of recurrence, and cause of death. The results obtained were compared with N0 and N2 survival of the whole population of patients. The N2 population was divided in two subgroups: “single station” N2, which is N2 involving superior mediastinal nodes (2R 1 4R, 3, 4L) or aortic nodes (5, 6) or inferior mediastinal nodes (7, 8, 9) [3]; and “dual station” N2, which is involvement of any combination of the abovementioned stations. N0 5-year survival rate was 56.5%; “single station” N2 was 28.3%, and “dual station” N2 was 11.3%.

1573

Results The cumulative postoperative survival rate at 5 years was 47.5% (Fig 1). Nine patients died within the first postoperative month, deaths resulting from surgery-related causes (3.5% of mortality rate); 10 patients were lost at follow-up (4%), 96 patients are still alive (37.5%), and 141 patients died during follow-up (55%). Cause of death was unknown in 16 patients (11.3%); not related to lung cancer in 34 (24.1%); due to another cancer in 7 (5%); and recurrence in 84 patients (59.6%). No significant differences in the survival rates were found with regard to site of primary bronchial cancer, pathologic T factor, histologic type, type of resection, and adjuvant radiotherapy (Table 2). In T2 cases, there was no difference in survival whether the tumor invaded the visceral pleura or not (Table 2). The 5-year survival rate was not influenced by the level of N1 disease, or by the number of levels involved (Table 2). Comparing N1 involvement by direct extension (n 5 84) with involvement by metastases (n 5 172), 5-year survival rates were not statistically different ( p 5 0.458). When the patients were further subdivided on the basis of Extralobar-Hilar and intralobar metastases and compared according to the type of nodal involvement either by direct or metastatic extension, differences between 5-year survival rates were also found not to be statistically significant (Table 3). In cases of intralobar N1 involvement alone, 5-year survival was better than in cases with extralobar hilar N1 involvement alone or in association (Table 4). Modality of recurrence (n 5 84) was not affected by the type of N1 involvement (Table 5). Comparing survival rates in both intralobar and hilar N1 groups with those of N0 and N2 disease yielded the following (Fig 2): no difference was observed between intralobar N1 involvement and N0; or between extralobar hilar N1 and N2 disease, especially when N2 involvement was limited to one station in the ipsilateral mediastinum (Table 6); these results, assimilating intralobar N1 in-

Fig 1. Survival curve after resection for 256 patients with N1 nonsmall cell lung cancer.

1574

RIQUET ET AL SIGNIFICANCE OF N1 INVOLVEMENT IN LUNG CANCER

Table 2. Univariate Analysis of Factors Associated With Postoperative Survival of N1 Disease Variable

No. of Patients

5-Year Survival (%)

147 109

39.6 52.6

Table 4. Five-Year Survival According to the Level of Involvement (Intralobar Versus Extralobar “Hilar”)

p Value

a

Location of cancer Segment Bronchus Histologic class Squamous Adenocarcinoma Operation Pneumonectomy Lobectomy Radiotherapyb Not done Done T Statusc T1 T2 T3 Visceral pleurald Not invaded Invaded Number of N1 level One level Two levels and more

0.120

181 52

50 37

0.379

189 77

53.3 61.3

0.416

77 160

56.5 42.8

0.397

64 125 60

49.5 47.3 36.2

0.129

98 26

50.3 35

0.65

144 112

49.4 38.6

0.195

a

Segment location means that the tumor was intralobar or peripheral (distal), and bronchus location means that most of the tumor in that b Only patients having location was extralobar or central (proximal). c T4 (n 5 7) was survived more than 3 months were included. d 5-year survival in excluded, the number of patients being too small. case of T2 tumor when visceral pleura invasion by the tumor was considered; there was no significant difference between both groups although 5-year survivals seemed lower in case of invasion.

volvement to N0 and extralobar N1 involvement to N2 in a single station, were significant ( p 5 0.01).

Comment The results of this study confirm that N1 disease is actually a compound of two types of lymph node involvement. This suggests that tumors with intralobar N1 extension remain a local disease and behave as N0 tumors, and that the hilar (or extralobar) N1 involvement Table 3. Five-Year Survival According to the Level of N1 Involvement and the Type of Involvement (Direct Extension Versus Metastases)

Level of Involvement Intralobar involvement Direct extension Metastases Extralobar involvement Direct extension Metastases

No. of Patients

5-Year Survival (%)

Median (months)

p

53 49

53.7 54.4

78 91

0.370

32 122

27.3 41.2

18 41

0.137

Observed differences are not statistically significant.

Ann Thorac Surg 1999;67:1572– 6

All patients Intralobar N1 only Extralobar N1 (6 intra) Lung cancer confined to segment only Intralobar Extralobar

No. of Patients

Survival (%)

102

53.6

154

38.5

69 66

49.7 30.9

p

0.02

0.02

Differences in survival are statistically significant. When location of cancer was distal (segmental) only, 5-year survival was 39.6% (Table 2), but as may be seen here, this 5-year survival rate was also related to levels of N1 involvement.

corresponds to a lymphatic spread and has the same prognostic significance as N2 disease limited to ipsilateral single station of either upper or lower mediastinum. This N1 classification into two groups was proposed by Naruke and associates [12]: group a, intrapulmonary lymph or peribronchial node metastases in which spread to subsegmental or segmental bronchus is identified; and group b, hilar lymph node metastases in which spread to interlobar, lobar, or main bronchus is identified. The first group was thus termed Lobar N1 and the second Hilar N1 [5, 6]. Rational criticism has been made by Shields [13] about this terminology. The term “hilar” N1 is particularly inadequate from the anatomical point of view, the hilum being the depression at the site where ducts, vessels, and nerves enter an organ. The term “hilar” lymph nodes is usually used to describe those found along the main bronchus (No. 10) [13]. The extension of this concept to include the interlobar, lobar, and main bronchus nodes was introduced by the Japan Lung Cancer Society. As it appears now that both groups may be different entities, it would be preferable to use more appropriate terms based on, and referring to, the anatomic location of the nodes in concern; thus, lobar should be replaced by intralobar, which encompasses segmental nodes that are not only lobar; and hilar should be replaced by extralobar, which includes interlobar, fissure, and hilar nodes. The use of Table 5. Site of Recurrence According to the Type (Intralobar Versus Extralobar Hilar) of Metastasis

Local recurrence Brain metastasis Liver metastasis Lung metastasis Miscellanous Total

Intralobar N1

Extralobar Hilar N1

Total

6 8 7 7 2 30

10 19 12 11 2 54

16 27 19 18 4 84

Differences are not significant but the number of cases in each group was small.

Ann Thorac Surg 1999;67:1572– 6

RIQUET ET AL SIGNIFICANCE OF N1 INVOLVEMENT IN LUNG CANCER

Fig 2. Survival curves after resection for 1,174 patients with N0 (n 5 588), intralobar N1 (n 5 102), extralobar hilar N1 (n 5 154), N2 single station (n 5 200), and N2 dual stations (n 5 128). N0 and intralobar N1 curves as well as extralobar hilar N1 and singlestation N2 are superimposed.

inappropriate terms can be confusing and misleading to readers not dealing frequently with such topics, whereas the use of terms with anatomical significance is clear and comprehensible to most readers. We encountered intralobar N1 involvement in up to 39.5% of the entire population of N1 lung cancers we have studied. Frequencies reported in stage II disease were, respectively, 38.5% for Yano and associates [5] and 25% for Van Velzen and associates [6] (in stage IIb disease). An anatomic study based on injection of the pulmonary segments in 360 cadavers demonstrated that the dye injected right segmental nodes in 24.5% of cases and left ones in 31% of cases [14]. Thus, the frequency of N1 involvement pattern seems at first glance not to be very different from the distribution observed in basic anatomy. Yano and associates [5] reported that the survival rate associated with intralobar N1 disease was significantly better than that associated with extralobar hilar N1 Table 6. Comparison Between N0, Intralobar N1, Extralobar Hilar N1, and N2 Involvement

N0 Lobar N1 Hilar N1 N2 (one level) N2 (two levels)

No. of Patients

5-Year Survival (%)

588 102 154 200 128

56.5 53.6 38.5 28.3 11.3

] ]

p NS NS

0.001 0.001

NS 5 not statistically significant. Worst prognosis is observed in case of N2 involvement in both upper and lower ipsilateral mediastinum (dual station). When only one level of ipsilateral mediastinum is concerned (lower or upper single station), the 5-year survival is similar to that observed in cases with extralobar “hilar” N1 involvement.

1575

involvement; this is in agreement with the recent rate observed by Van Velzen and associates [6]. Our study yields similar results and shows significant difference in survival in the two N1 subgroups. We did not perform a count of the N1 nodes involved; as stressed by Yano and associates [5] and Van Velzen and associates [6], it is often difficult, not to say impossible, to quantify the number of metastatic nodes as they are very frequently conglomerated and can be hardly divided. However, in so doing, Martini and associates [4] have observed that the difference in 5-year survival rates between patients with single-station N1 disease and those with multiple N1 nodes was nearly 15% (45% vs 31%). In his study, intralobar N1 disease represented more than 55% of cases, 50% of patients had only one node involved, and almost two-thirds of them were at intralobar levels. The fact that most of the single nodes involved are intralobar could explain Martini’s conclusions. The finding of direct extension into adjacent lymph nodes from lung cancer should theoretically provide a better prognosis, such invasion being regional disease and not metastasis. Our study does not support this hypothesis. Van Velzen and associates [15] reported in stage IIa (T1N1) a highly significant difference between patients with direct extension and those with extralobar hilar metastases, on the other hand, he found comparable survivals between patients with direct extension and those with extralobar hilar metastases in stage IIb (T2N1). Such paradoxical observations may be explained if we consider that direct extension concerns essentially intralobar N1 levels in T1 tumors, and that direct extension concerns also extralobar hilar N1 levels in T2 disease according to both size and location of the primary lung tumor. We observed such extralobar hilar N1 direct extension in 32 out of 154 patients (Table 3). While proposing this classification into 2 N1 subgroups, Naruke and associates commented that they found no significant differences in survival rates between patients without metastases to lymph nodes (N0) and patients with metastases to intrapulmonary or peribronchial lymph nodes [12]. This is in agreement with the first conclusion of our study, making intralobar N1 group a regional disease. Maggi and associates [7] and Yano and associates [5] initially noticed that extralobar hilar N1 disease appears to be similar to N2 disease, and Van Velzen and associates [6] suggested that in these patients with extralobar hilar N1 involvement, an underestimated N2 disease may be present. This underestimated N2 disease should be ruled out. Radical mediastinal lymph node dissection is mandatory when assessing the value of “N” involvement in lung cancer. Series of patients in whom lymph node sampling or dissection is not complete have a demonstrable lower surgical survival rate due to errors in assigning nodal stage [16]. In the series that included performance of radical mediastinal lymphadenectomy, the N1 5-year survival rates ranged from 40% to 49% [4, 7, 12]. Yano and associates [5] state that a 40% 5-year survival is essential to assess the prognostic factors in N1 disease. In the study of Van Velzen and associates [6], the

1576

RIQUET ET AL SIGNIFICANCE OF N1 INVOLVEMENT IN LUNG CANCER

cumulative 5-year survival rate was 37.8% and the mediastinal nodes were only controlled by biopsy samples; this can explain Van Velzen and associates’ suggestion [6], but does not mean that N2 disease is underestimated because it is included with N1 and left in the mediastinum at surgery; in addition, skipping N2 metastasis exists and is by definition unrelated to concomittant N1 [17, 18]. Yano and associates [5, 19] reported that the behavior of extralobar hilar N1 disease was similar to that of N2 disease with regard to the modality of recurrence. In the first study, they noticed that the first site of metastatic recurrence was the lung in most cases of extralobar hilar N1, whereas it was essentially the brain in intralobar N1 cases. Shortly thereafter, they published another work demonstrating similar results when comparing with N0 and N2 diseases; the mode of metastasis in intralobar N1 tended to resemble that of N0, while that of extralobar hilar N1 behaves like N2 disease. We did not observe the same pattern of metastasis in N1 disease as they did (Table 6); however, we think that Yano and associates’ observations are indirect arguments that support the results we obtained when directly comparing prognosis of extralobar hilar N1 against that of N2. These two groups have the same prognosis when single-station N2 disease is present in the ipsilateral mediastinum. In conclusion, N1 disease seems to be the compound of an intralobar group that behaves like N0, and an extralobar hilar group that has a prognosis identical to that of a single-station ipsimediastinal N2 disease. If further confirmed, this can have radical influence when selecting therapeutic schemes in N1, and offers further information for clinical and research purposes.

References 1. Mountain CF. Revisions in the International System for Staging Lung Cancer. Chest 1997;111:1710–7. 2. Riquet M, Manac’h D, Le Pimpec Barthes F, et al. Facteurs pronostiques du cancer bronchique non a` petites cellules ope´re´. Etude critique du TNM. Rev Pneumol Clin 1996;52: 181–7. 3. Mountain CF, Dresler CM. Regional lymph node classification for lung cancer staging. Chest 1997;111:1718–23. 4. Martini N, Burt ME, Bains MS, McCormack PM, Rusch VW, Ginsberg RJ. Survival after resection of stage II non-small cell lung cancer. Ann Thorac Surg 1992;54:960– 6.

Ann Thorac Surg 1999;67:1572– 6

5. Yano T, Yokoyama H, Inoue T, Asoh H, Tayama K, Ichinose Y. Surgical results and prognostic factors of pathologic N1 disease in non-small-cell carcinoma of the lung. Significance of N1 level: lobar or hilar nodes. J Thorac Cardiovasc Surg 1994;107:1398 – 402. 6. Van Velzen E, Snijder RJ, Brutel de la Rivie`re A, Elbers HRJ, van den Bosch JMM. Lymph node type as a prognostic factor for survival in T2 N1 M0 non-small cell lung carcinoma. Ann Thorac Surg 1997;63:1436– 40. 7. Maggi G, Casadio C, Mancuso M, Oliaro A, Cianci R, Ruffini E. Resection and radical lymphadenectomy for lung cancer: prognostic significance of lymphatic metastases. Int Surg 1990;75:17–21. 8. Martini N, Flehinger BJ. The role of surgery in N2 lung cancer. Surg Clin North Am 1987;65:1037– 49. 9. Naruke T, Suemasu K, Ishikawa S. Lymph node mapping and curability at various levels of metastasis in resected lung cancer. J Thorac Cardiovasc Surg 1978;76:832–9. 10. Kaplan EL, Meier P. Non-parametric estimation form incomplete observations. J Am Stat Assoc 1958;53:45– 81. 11. Peto R, Peto J. Asymptomatically efficient rank invariant test procedures (with discussion). JR Stat Soc [A] 1972;135:185– 207. 12. Naruke T, Goya T, Tsuchiya R, Suemasu K. Prognosis and survival in resected lung carcinoma based on the new international staging system. J Thorac Cardiovasc Surg 1988; 96:440–7. 13. Shields TW. The location of station 11 pulmonary lymph nodes. J Thorac Cardiovasc Surg 1995;109:402–3. 14. Riquet M. Anatomic basis of lymphatic spread from carcinoma of the lung to the mediastinum: surgical and prognostic implications. Surg Radiol Anat 1993;115:271–7. 15. Van Velzen E, Snijder RJ, Brutel de la Rivie`re A, Elbers HJJ, van den Bosch JMM. Type of lymph node involvement influences survival rates in T1 N1 MO non-small cell lung carcinoma. Lymph node involvement by direct extension compared with lobar and hilar node metastases. Chest 1996; 110:1469–73. 16. Nesbitt JC, Putnam JB, Walsh GL, Roth JA, Mountain CF. Survival in early-stage non-small cell lung cancer. Ann Thorac Surg 1995;60:466–72. 17. Martini N, Flehinger BJ, Zahan MB, Beattie EJ. Results of resection in non oat cell carcinoma of the lung with mediastinal lymphnode metastases. Ann Surg 1983;198:386–97. 18. Riquet M, Manac’h D, Dupont P, Dujon A, Hidden G, Debesse B. Anatomic basis of lymphatic spread of lung carcinoma to the mediastinum: anatomo-clinical correlations. Surg Radiol Anat 1994;16:229–38. 19. Yano T, Yokoyama H, Inoue T, et al. The first site of recurrence after complete resection in non-small-cell carcinoma of the lung. Comparisons between pN0 disease and pN2 disease. J Thorac Cardiovasc Surg 1994;108:680–3.