Visceral pleural involvement in nonsmall cell lung cancer: prognostic significance

Visceral pleural involvement in nonsmall cell lung cancer: prognostic significance

Toshihiro Osaki, MD, PhD, Akira Nagashima, MD, PhD, Takashi Yoshimatsu, MD, PhD, Sosuke Yamada, MD, and Kosei Yasumoto, MD, PhD Department of Chest Su...

191KB Sizes 0 Downloads 31 Views

Toshihiro Osaki, MD, PhD, Akira Nagashima, MD, PhD, Takashi Yoshimatsu, MD, PhD, Sosuke Yamada, MD, and Kosei Yasumoto, MD, PhD Department of Chest Surgery, Kitakyushu Municipal Medical Center, and Department of Surgery II, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan

Background. A tumor of any size that invades the visceral pleura is classified in the T2 category; however, the definition of the visceral pleural involvement has remained somewhat ambiguous. It is unclear whether the T2 category includes the p2 status alone or incorporates the extent of the p1 status. Methods. We retrospectively analyzed the survival of 474 patients with T1 and T2 nonsmall cell lung cancer to evaluate the influence of the degree of visceral pleural involvement (p0, p1, and p2) on the prognosis and to clarify the definition of the visceral pleural involvement. Results. The 5-year survival rates according to the degree of visceral pleural involvement were 68.0% in p0 (n ⴝ 345), 43.9% in p1 (n ⴝ 110), and 54.9% in p2 (n ⴝ 19; p0 versus p1, p ⴝ 0.0004; p0 versus p2, p ⴝ 0.013; and p1 versus p2, p ⴝ 0.61). The degree of visceral pleural

involvement (p0 versus p1/p2) was a significant independent prognostic factor from tumor size and lymph node involvement, by multivariate analysis (relative risk ⴝ 1.47, p ⴝ 0.033). The prognosis of pN0 patients with p1 and tumor size 3 cm or less was significantly poorer than that of those with p0 and tumor size 3 cm or less (p ⴝ 0.0004), and the prognosis of patients with p1 and tumor size more than 3 cm was significantly poorer than that of those with p0 and tumor size more than 3 cm (p ⴝ 0.024). Conclusions. The degree of visceral pleural involvement (p0 versus p1/p2) is an important component of the lung cancer staging system. Tumors with p1 and p2 status should be regarded as representing visceral pleural involvement and T2 disease. (Ann Thorac Surg 2004;77:1769 –73) © 2004 by The Society of Thoracic Surgeons

V

Patients and Methods

isceral pleural involvement (VPI) of lung cancer has prognostic implications [1– 6]; therefore, tumors with VPI are considered to be T2 even when they are 3 cm or less in size, according to the TNM descriptors by Mountain [7]. The degrees of VPI are classified into the following three stages: p0, tumor with no pleural involvement or that reaches the visceral pleura but does not extend beyond its elastic layer; p1, tumor that extends beyond the elastic layer of the visceral pleura but is not exposed on the pleural surface; and p2, tumor that is exposed on the pleural surface but does not involve the parietal pleura [8]. However, the definition of VPI in the TNM descriptions has remained somewhat ambiguous: it is not clear whether a T2 tumor includes the p2 status alone or also incorporates the extent of the p1 status. The staging manual of the International Union Against Cancer (UICC) [9] also does not offer detailed advice on the diagnosis of VPI. In this study, we retrospectively analyzed the survival of 474 nonsmall cell lung cancer (NSCLC) patients with T1 and T2 tumors, to evaluate the influence of the degree of VPI (p0, p1, and p2) on the prognosis and to clarify the definition of the VPI.

Accepted for publication Oct 8, 2003. Address reprint requests to Dr Osaki, Department of Chest Surgery, Kitakyushu Municipal Medical Center, 2-1-1 Bashaku, Kokurakita-ku, Kitakyushu 802-0077, Japan; e-mail: [email protected] fukuoka.jp.

© 2004 by The Society of Thoracic Surgeons Published by Elsevier Inc

Patients Between April 1992 and December 2001, 738 consecutive patients with NSCLC underwent surgical resections at the Department of Chest Surgery, Kitakyushu Municipal Medical Center, Japan. Among them, 474 patients were pathologic T1 and T2 NSCLC with N0 to N2 nodal status. Excluded from the analysis were 5 T2 NSCLC patients with interlobar invasion (interlobar p3), and 4 patients with tumors involving the main bronchus, 2 cm or more distal to the carina. The characteristics of the 474 patients are summarized in Table 1. The patients ranged in age from 17 to 87 years (median, 66.5), and comprised 304 men and 170 women. The histologic types were 290 adenocarcinomas, 156 squamous cell carcinomas, 11 large cell carcinomas, 7 adenosquamous carcinomas, and 10 other histologic types. The postoperative pathologic N status were N0 in 341 patients, N1 in 52 patients, and N2 in 81 patients. Surgical procedures consisted of pneumonectomy in 21 patients, bilobectomy in 27 patients, lobectomy in 393 patients, segmentectomy in 17 patients, and wedge resection in 16 patients.

Assessment of Visceral Pleural Involvement In reference to Hammar’s staging classification [8], we defined the degrees of VPI as follows; p0, tumor that does not penetrate the elastic layer of the visceral pleura; p1, tumor that penetrates the elastic layer but is not exposed 0003-4975/04/$30.00 doi:10.1016/j.athoracsur.2003.10.058

GENERAL THORACIC

Visceral Pleural Involvement in Nonsmall Cell Lung Cancer: Prognostic Significance

1770

OSAKI ET AL VISCERAL PLEURAL INVOLVEMENT IN NSCLC

Results

Table 1. Patient Characteristics Characteristics GENERAL THORACIC

Age (y) Mean Range Sex Male Female Histology Adenocarcinoma Squamous cell carcinoma Large cell carcinoma Adenosquamous carcinoma Other Pleural involvement p0 p1 p2 Tumor size ⱕ 3 cm ⬎ 3 to ⱕ 5 cm ⬎ 5 cm Tumor differentiation Well Moderate Poor Pathologic N status N0 N1 N2 Type of resection Pneumonectomy Bilobectomy Lobectomy Segmentectomy/wedge

Ann Thorac Surg 2004;77:1769 –73

No. of Patients (%) 66.5 17– 87 304 (64.1) 170 (35.9) 290 (61.2) 156 (32.9) 11 (2.3) 7 (1.5) 10 (2.1) 345 (72.8) 110 (23.2) 19 (4.0) 225 (47.5) 172 (36.3) 77 (16.2) 159 (33.5) 183 (38.6) 101 (21.3) 341 (71.9) 52 (11.0) 81 (17.1) 21 (4.4) 27 (5.7) 393 (82.9) 33 (7.0)

on the pleural surface; and p2, tumor that is exposed on the pleural surface but does not involve the parietal pleura. When it was difficult to evaluate the degree of the VPI by hematoxylin and eosin staining, an elastic stain (Verhoeff-Van Gieson [VVG] stain) was used, which helped in distinguishing between p0 and p1 (Fig 1). The degrees of visceral pleural involvement were p1 status in 345 patients (72.8%), p1 in 110 patients (23.2%), and p2 in 19 patients (4.0%).

Statistical Analyses The associations between the clinicopathologic characteristics and the degree of visceral pleural involvement were evaluated using the ␹2 test. A survival analysis for each prognostic variable on the overall survival was estimated, according to the Kaplan-Meier method. The terminal event was death attributable to cancer or noncancer causes. The statistical significance of the differences in the survival distribution among the prognostic groups was evaluated by the log-rank test. The Cox proportional hazards model was applied to the multivariate survival analysis. The statistical difference was considered to be significant if the p value was below 0.05. Data were analyzed with the use of the Abacus Concepts, Survival Tools for StatView program (Abacus Concepts, Berkeley, CA).

Characteristics According to Degree of VPI Table 2 shows the correlation of the degree of VPI with histologic type, tumor size, tumor differentiation, and lymph node involvement. Patients were divided into three groups according to the tumor diameter: 3 cm or smaller, 225 patients (47.5%); from 3.1 to 5 cm, 172 patients (36.3%); and larger than 5 cm, 77 patients (16.2%). The ratio of tumors with 3 cm or smaller sizes, well differentiation, and pathologic N0 status in the p0 group showed a significant increase among the three groups (p ⬍ 0.0001, ⬍ 0.0001, and 0.0046, respectively). There were no significant differences in histologic type among the three groups.

Survival Analysis The median observation period after the operations was 34.4 months (range, 0.7 to 116.3). The overall 5-year survival rate of the 474 patients was 61.9%. The overall 5-year survival rates of patients with p0, p1, and p2 status were 68.0%, 43.9%, and 54.9%, respectively (Fig 2). A significant survival difference was observed between the p0 and p1 status patients, and between the p0 and p2 status patients (p ⫽ 0.0004 and 0.013, respectively). No significant difference was found between the p1 and p2 status patients (p ⫽ 0.61). The potential prognostic factors were subjected to a univariate survival analysis (Table 3). Sex, age, histologic type, VPI (p0 versus p1/p2), tumor size, tumor differentiation, pathologic N status, and type of resection were found to be significant prognostic factors. Table 4 shows the results of the multivariate survival analysis. The degree of VPI was found to be an independent prognostic factor, along with sex, age, tumor size, and pathologic N status (relative risk ⫽ 1.47, p ⫽ 0.033). Among the patients with pathologic N0 status, according to the combination of visceral pleural involvement (p0 versus p1) and tumor size (3 cm or less versus 3 cm or more), the overall survivals were compared among the following four groups: p0 and tumor size 3 cm or less group, 149 patients; p1 and tumor size 3 cm or less group, 25 patients; p0 and tumor size more than 3 cm group, 114 patients; and p1 and tumor size more than 3 cm group, 44 patients (Fig 3). The prognosis of patients with p1 and tumor size 3 cm or less was significantly poorer than that of those with p0 and tumor size 3 cm or less (5-year survival: 46.3% versus 84.2%, p ⫽ 0.0004). The prognosis of patients with p1 and tumor size more than 3 cm was significantly poorer than that of those with p0 and tumor size more than 3 cm (5-year survival: 46.1% versus 69.6%, p ⫽ 0.024). There were no significant survival differences between the patients with p0 and tumor size more than 3 cm and those with p1 and tumor size 3 cm or less (p ⫽ 0.22).

Comment In the TNM staging system, several important questions remain to be answered, including the treatment of VPI. The degree of VPI can be classified into three subgroups, based on the elastic layer of the pleura: the p0, p1, and p2 status. Concerning the effect of the degree of VPI on the

OSAKI ET AL VISCERAL PLEURAL INVOLVEMENT IN NSCLC

1771

GENERAL THORACIC

Ann Thorac Surg 2004;77:1769 –73

Fig 1. (A) Elastic stain of a p1 status tumor. The tumor cells extend beyond the elastic layer (arrow) of the visceral pleura, but are not exposed on the pleural surface (original magnification, ⫻200). (B) Elastic stain of a p0 status tumor. The tumor cells do not extend beyond the elastic layer (arrow) of the visceral pleura (original magnification, ⫻200).

T status, the TNM staging system only states that a tumor of any size that invades the visceral pleura is classified as a T2 tumor [7, 9]. It is not clear whether the T2 tumor category includes the p2 status alone or incorporates the extent of the p1 status. In this study, the prognosis of patients with p1 and p2 status was significantly poorer than that of those with p0 status. In addition, the VPI (p0 versus p1/p2 status) was a potentially significant and independent prognostic factor, along with tumor size and lymph node involvement. Given this prognostic implication, it is important to clarify the definition of the VPI: tumors with both p1 and p2 status are regarded as having VPI and T2 disease. According to the general rules of the Japan Lung Cancer Society [10], the detailed definition concerning the degree of VPI and its relation to T status is as follows: T1, tumor with a size of 3 cm or less that extends beyond

the elastic layer of the visceral pleura but is not exposed on the pleural surface (p1); and T2, tumor of any size that is exposed on the pleural surface but does not involve the parietal pleura (p2). However, in many studies [1, 3, 6, 11–13] tumors with p1 status are grouped as having VPI and with T2 tumors, whereas other studies grouped them as not having VPI and with T1 tumors [4, 5]. Thus, discrepancies exist concerning the relation between the VPI and the T status. The degree of VPI correlated with tumor size, tumor differentiation, and lymph nodal involvement. There is an apparent trend toward large tumor size, moderate or poor differentiation, and nodal involvement in patients with p1 status, as compared with those with p0 status, indicating that a sharp distinction can be made between the p0 and p1 status. Kondo and associates [14] reported a close correlation between the VPI and the malignant

Table 2. Characteristics According to Degree of Visceral Pleural Involvement Visceral Pleural Involvement Characteristics Histology Adenocarcinoma Squamous cell carcinoma Tumor size ⱕ 3 cm ⬎ 3 to ⱕ 5 cm ⬎ 5 cm Tumor differentiation Well Moderate or poor Pathologic N status N0 N1 N2 a

Numbers in parentheses are percentages.

p0 (n ⫽ 345)

p1 (n ⫽ 110)

p2 (n ⫽ 19)

213 (61.7)a 113 (32.8)

62 (56.4) 39 (35.5)

15 (78.9) 4 (21.1)

0.33

179 (51.9) 121 (35.1) 45 (13.0)

37 (33.6) 41 (37.3) 32 (29.1)

9 (47.4) 10 (52.6) 0 (0.0)

⬍ 0.0001

137 (39.7) 186 (53.9)

19 (17.3) 82 (74.5)

3 (15.8) 16 (84.2)

⬍ 0.0001

263 (76.2) 35 (10.1) 47 (13.6)

69 (62.7) 14 (12.7) 27 (24.5)

9 (47.4) 3 (15.8) 7 (36.8)

0.0046

p Value

1772

OSAKI ET AL VISCERAL PLEURAL INVOLVEMENT IN NSCLC

Ann Thorac Surg 2004;77:1769 –73

Table 4. Multivariate Cox Regression Analysis of Overall Survival Variable GENERAL THORACIC Fig 2. Survival curves according to the degree of visceral pleural involvement.

pleural lavage cytology. The pleural lavage cytology was positive in 3 of 221 (1.3%), 13 of 96 (14%), and 15 of 41 (37%) of patients in the p0, p1, and p2 groups, respec-

Overall Sex Male Female Agea ⬍ 67 years ⱖ 67 years Histology Squamous cell carcinoma Nonsquamous cell carcinoma Pleural involvement p0 p1 or p2 Tumor size ⱕ 3 cm ⬎ 3 cm Tumor differentiation Well Moderate or poor Pathologic N status N0 N1 or N2 Type of resection Pneumonectomy or bilobectomy Lobectomy Segmentectomy or wedge

No. of 5-Year Patients Survival (%) p Value 474

61.9

304 170

55.7 73.0

⬍ 0.0001

218 256

70.0 54.8

0.0033

156 318

53.0 66.5

0.0010

345 129

68.0 45.7

⬍ 0.0001

225 249

73.5 51.7

⬍0.0001

159 284

76.9 55.0

0.0001

341 133

71.0 40.1

⬍ 0.0001

48

42.0

⬍ 0.0001b

393 33

65.5 36.4

The mean age of the 474 patients was 66.5 ⫾ 9.9 years. Therefore, the patients were divided into two groups: patients aged 67 years or over, and b patients aged less than 67 years. The p value was calculated between the lobectomy group and the pneumonectomy or bilobectomy group. There were no significant differences between the lobectomy group and the segmentectomy or wedge group (p ⫽ 0.22), and between the pneumonectomy or bilobectomy group and the segmentectomy or wedge group (p ⫽ 0.12). a

95% CI

1.90

1.27–2.84

0.0017

1.66

1.16 –2.39

0.0061

1.19

0.82–1.73

0.35

1.47

1.03–2.10

0.033

1.63

1.13–2.36

0.0095

1.46

0.95–2.25

0.087

2.69

1.89 –3.83 ⬍ 0.0001

1.23

0.82–1.86

p Value

0.32

The mean age of the 474 patients was 66.5 ⫾ 9.9 years. Therefore, the patients were divided into two groups: patients aged 67 years or over, and patients aged less than 67 years.

a

CI ⫽ confidence interval;

Table 3. Five-Year Survival According To Potential Prognostic Factors Variable

Sex (female versus male) Agea (⬍ 67 years versus ⱖ 67 years) Histology (non-SCC versus SCC) Pleural involvement (p0 versus p1 or p2) Tumor size (ⱕ 3 cm versus ⬎ 3 cm) Tumor differentiation (well versus moderate or poor) Pathologic N status (N0 versus N1 or N2) Type of resection (lobectomy versus other types)

Relative Risk of Death

SCC ⫽ squamous cell carcinoma.

tively. Riquet and associates [12] also reported that among 63 patients with p0 disease, all had negative pleural lavage cytology. Their findings may also support the proposal that tumors with both p1 and p2 status should be regarded as having VPI. We should make a clear distinction between the p0 and p1 status. When it is difficult to determine the degree of the VPI by hematoxylin and eosin staining, an elastic stain should be employed. The use of an elastic stain is potentially important for assessing VPI, and especially when distinguishing between the p0 and p1 status, as noted by Bunker and associates [11]. They also reported that the elastic stain results changed the pathologic stage in approximately 4% of lung cancer resections overall and in 10% of the cases that were indeterminate for pleural involvement by hematoxylin and eosin staining. Among the 474 T1-2 NSCLC patients in this study, the T status criteria based on the VPI, whether the T2 tumor includes the p2 status alone or incorporates the extent of the p1 status, affected the status of 37 patients (7.8%) with tumor size less than 3cm and p1 status. In 249 patients, the T status was T2 based on tumor size (⬎ 3 cm); therefore, it was unchanged regardless of the degree of VPI. In 9 patients, the T status was T2 because the tumor was p2 status. In 179 patients, the T status was T1 based on both the tumor size (ⱕ 3 cm) and p0 status. In the TNM staging system, the tumor and nodal status have been considered to be the most important prognostic factors, although recent advances in molecular biology and genetics have created new diagnostic possibilities. We do not wish to further complicate the TNM staging system by using additional prognostic factors and excluding the conventional TNM descriptions. The degree of VPI (p0 versus p1/p2 status) is an important component of the lung cancer staging system. In a future revision of the TNM staging system, we believe that

OSAKI ET AL VISCERAL PLEURAL INVOLVEMENT IN NSCLC

4.

5.

6. 7. 8. 9. Fig 3. Survival curves of patients with pN0 status, according to the degree of visceral pleural involvement and tumor size.

tumors with p1 and p2 status should be regarded as representing VPI and T2 disease, and this statement should be included in the TNM descriptions.

10. 11.

12.

References 1. Manac’h D, Riquet M, Medioni J, Le Pimpec-Barthes F, Dujon A, Danel C. Visceral pleura invasion by non-small cell lung cancer: an underrated bad prognostic factor. Ann Thorac Surg 2001;71:1088 –93. 2. Harpole DH, Herndon JE, Young WG, Wolfe WG, Sabiston DC. Stage I nonsmall cell lung cancer: a multivariate analysis of treatment methods and patterns of recurrence. Cancer 1995;76:787–96. 3. Ichinose Y, Yano T, Asoh H, Yokoyama H, Yoshino I, Katsuda Y. Prognostic factors obtained by a pathologic

13.

14.

1773

examination in completely resected non-small-cell lung cancer. An analysis in each pathologic stage. J Thorac Cardiovasc Surg 1995;110:601–5. Carbone E, Asamura H, Takei H, et al. T2 tumors larger than five centimeters in diameter can be upgraded to T3 in non-small cell lung cancer. J Thorac Cardiovasc Surg 2001; 122:907–12. Maeshima AM, Niki T, Maeshima A, Yamada T, Kondo H, Matsuno Y. Modified scar grade. A prognostic indicator in small peripheral lung adenocarcinoma. Cancer 2002;95: 2546 –54. 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 –80. Mountain CF. Revisions in the international staging system for lung cancer. Chest 1997;111:1710 –7. Hammar SP. Common neoplasms. In: Dail DH, Hammar SP, eds. Pulmonary pathology. New York: Springer Verlag, 1994: 1123–278. UICC. International Union Against Cancer. In: Sobin LH, Wittekind Ch, eds. TNM Classification of malignant tumours. 5th ed. New York: Wiley-Liss, 1997:93–7. The Japan Lung Cancer Society. General rules for clinical and pathological record of lung cancer. 5th ed. Tokyo: Kanehara, 1999:77–89. Bunker ML, Raab SS, Landreneau RJ, Silverman JF. The diagnosis and significance of visceral pleural invasion in lung carcinoma. Histologic predictors and the role of elastic stains. Am J Clin Pathol 1999;112:777–83. Riquet M, Badoual C, Le Pimpec Barthes F, et al. Visceral pleura invasion and pleural lavage tumor cytology by lung cancer: a prospective appraisal. Ann Thorac Surg 2003;75: 353–5. Suzuki K, Nagai K, Yoshida J, et al. Conventional clinicopathologic prognostic factors in surgically resected nonsmall cell lung carcinoma. A comparison of prognostic factors for each pathologic TNM stage based on multivariate analyses. Cancer 1999;86:1976 –84. Kondo H, Asamura H, Suemasu K, et al. Prognostic significance of pleural lavage cytology immediately after thoracotomy in patients with lung cancer. J Thorac Cardiovasc Surg 1993;106:1092–7.

INVITED COMMENTARY Detailed studies of both the healthy and diseased pleura have been somewhat neglected. The pleura is an interesting and important set of membranes that actively transports both fluids and cells; and pleural involvement with lung cancer has a significant effect on prognosis. It has long been known that there are three histologically identifiable components that comprise the pleura: (1) a mesothelial membrane, lined by (2) an elastic membrane, and (3) underlying connective tissue. From the earliest publication (1974) of the TNM Lung Cancer Staging System, it was appreciated that any involvement with the visceral pleura was a significant adverse prognostic finding. This study provides objective evidence in support of that conclusion and, importantly, draws attention to the fact that the definition of “pleural invasion” was never provided in the staging rules. The authors demonstrate that malignant cells extending beyond the elastic layer (Hammar’s p1) have an essentially equal impact on prognosis, whether or not they are on the mesothelial surface (p2). Accordingly, this study supports the current staging rules. © 2004 by The Society of Thoracic Surgeons Published by Elsevier Inc

It should be noted that the International Staging System is derived from clinical observations alone, ie physical findings, biochemical tests, imaging studies, endoscopic findings, and biopsies. Therefore, it is doubtful that any accurate evaluation of the p1–p2 layers of involvement could be clinically made in most cases. Such evaluation, therefore, cannot be a part of clinical staging. Of interest also in this study is the additional data on the prognostic significance of other associated factors such as age, gender, histology, and tumor differentiation. I would encourage the authors to address the additional issue of interlobar pleural involvement. Clifton F. Mountain, MD Cardiothoracic Surgery University of California, San Diego 1150 Silverado St Suite 110 La Jolla, CA 92037 e-mail: [email protected] 0003-4975/04/$30.00 doi:10.1016/j.athoracsur.2003.12.089

GENERAL THORACIC

Ann Thorac Surg 2004;77:1769 –73