Survival in Synchronous vs Single Lung Cancer

Survival in Synchronous vs Single Lung Cancer Upstaging Better Reflects Prognosis Marcel Th. M. van Rens, MD; Pieter Zanen, MD, PhD; Aart Brutel de la Rivie`re, MD, PhD, FCCP; Hans R. J. Elbers, MD, PhD; Henry A. van Swieten, MD, PhD; and Jules M. M. van den Bosch, MD, PhD, FCCP

Objective: To define prognostic parameters for patients with synchronous non-small cell lung cancer (NSCLC). Design: Retrospective study of period from 1970 through 1997. Patients: Patients with a single (n ⴝ 2,764) and synchronous NSCLC (n ⴝ 85) who underwent pulmonary resection. Methods: All tumors were classified postsurgically, and the tumors of the patients with synchronous lung cancer were staged separately. The most advanced tumor was used for comparison. Actuarial survival time was estimated, and risk factors influencing survival were evaluated. Patients who died within 30 days of surgery were excluded. Measurement and results: Five-year survival for single NSCLC was 41% and for synchronous lung cancer it was 19%. The relative risk of death for patients with synchronous lung cancer was 1.75, compared to that for patients with single lung cancer. The most advanced tumor in synchronous cancer was a significant predictor of survival (p < 0.005). The survival of patients with synchronous lung cancer in which the most advanced tumors were stage I (n ⴝ 40) and stage II (n ⴝ 27) was not different from that of patients with stage II (n ⴝ 834) and stage IIIA (n ⴝ 405) single lung cancer, respectively. Conclusion: The poorer survival of patients with synchronous NSCLC is confirmed and quantified. The stage of the most advanced tumor was the best predictor of prognosis. The prognosis of patients with synchronous NSCLC resembles the prognosis of patients with a single lung cancer of a higher stage. Upstaging in synchronous lung cancer is recommended on the basis of these observations. (CHEST 2000; 118:952–958) Key words: lung neoplasm; multiple primary neoplasm; neoplasm staging; non-small cell lung carcinoma; prognosis; pulmonary surgical procedures; survival Abbreviations: CI ⫽ confidence interval; NSCLC ⫽ non-small cell lung carcinoma

resection remains the most effective P ulmonary treatment for patients with non-small cell lung cancer (NSCLC) with limited disease.1 In some patients, more than one tumor is present or a second one is detected later (multiple lung cancer). Generally, multiple lung cancer is subdivided into synchronous and metachronous lung cancer,2 with synchro*From the Departments of Pulmonary Diseases (Drs. van Rens, Zanen, and van den Bosch), Pathology (Dr. Elbers), and Thoracic Surgery (Dr. van Swieten), Sint Antonius Hospital, Nieuwegein; and Department of Thoracic Surgery (Dr. Brutel de la Rivie`re), University Medical Center, Utrecht, the Netherlands. Manuscript received November 16, 1999; revision accepted May 11, 2000. Correspondence to: Marcel Th. M. van Rens, MD, Sint Antonius Hospital, Department of Pulmonary Diseases, PO Box 2500, 3430 EM Nieuwegein, The Netherlands; e-mail: [email protected] 952

nous cancer being defined as two or more tumors present or detected at the same time, and metachronous cancer being defined as tumors detected with an interval. However, diagnosis of multiple lung carcinoma with the same histology may be difficult because of the possibility of metastatic or recurrent disease.2 There have been several studies of synchronous lung cancer,3– 8 and these studies should be distinguished from those analyzing primary lung cancer with intrapulmonary metastasis.9 –13 In general, the data of patients with synchronous lung cancer are not related to therapy and survival of patients with single lung carcinoma. Therefore, it remains difficult to probe changes in survival of patients with synchronous lung cancer. Clinical Investigations

In this report, the results of surgical treatment of patients with synchronous lung cancer are compared with those for patients with single lung carcinoma. Possible prognostic parameters in synchronous lung cancer are defined. Materials and Methods From 1970 through 1997, a total of 3,372 previously untreated patients underwent surgery for NSCLC at the Sint Antonius Hospital, Nieuwegein, the Netherlands. Of these, 3,086 patients underwent pulmonary resection, 2,849 of whom underwent resection for stage I, stage II, and stage IIIA NSCLC. These patients were allocated to two groups. The first group consisted of 85 patients with synchronous lung carcinoma. In 27 of these 85 patients, the second tumor was found accidentally at operation or at pathologic examination. In all 85 patients, the tumors were diagnosed at the same time and were classified as synchronous lung cancer according the criteria of Martini and Melamed 2 (Table 1). The second group consisted of 2,764 patients with single NSCLC. Recently, 2,361 of these patients with a follow-up of at least 5 years were analyzed in a study validating refinements of the staging system.14 The characteristics of the patients with a single lung cancer (n ⫽ 2,764) and synchronous lung cancer (n ⫽ 85) are presented in Table 2. Tumors were classified postsurgically, according to the 1997 staging system.1 The tumors of patients with synchronous cancer were staged separately, and the highest staged/most advanced tumor was classified as the “reference” tumor, as reported by others.4,5 Resected specimens were routinely examined pathologically and reviewed using the current criteria of the World Health Organization classification.15 Differences in frequency-based patient data were tested by the ␹2 test and normally distributed data by Student’s t test. Differences were considered significant when the p value was ⬍ 0.05. Two survival analyses were performed. One analysis used the data for both patients with single or synchronous lung cancer, to focus on differences between single and synchronous lung cancer. The other analysis used only the data for the patients with synchronous lung cancer. Survival was calculated by the KaplanMeier survival analysis method.16 Survival was estimated from date of operation. Deaths within 30 days after operation were excluded. Differences in survival between the two groups were tested by the log-rank test.17 Differences were considered significant when the p value was ⬍ 0.05. To facilitate interpretation, the hazard ratio (plus 95% confidence interval [CI]) was calculated where needed. Multivariate Cox’s regression analysis18 was used to analyze the data of the patients with synchronous lung cancer, to test the relation between survival and age, stage of tumor, histology, or tumor localization. The following variables

Table 1—Criteria for Synchronous Lung Cancer* Tumors present or diagnosed at the same time A. Tumors physically distinct and separate B. Histology Different Same, but in different segment, lobe, and lung if: Origin from carcinoma in situ No carcinoma in lymphatics common to both No extrapulmonary metastases at time of diagnosis *Adapted from Martini and Melamed.2

Table 2—Characteristics of Patients With Single and Synchronous NSCLC* Characteristics Patients, No. Sex Male Female Age, yr Median ⬎ 65 Unilateral One lobe More than one lobe Bilateral Histology Squamous Adeno Adenosquamous Large cell Resection Pneumonectomy Lobectomy Bilobectomy Limited resection‡ Side Right Left pTNM stage§ IA IB IIA IIB IIIA

Single Lung Cancer

Synchronous Lung Cancer†

2,764

85

2,528 (91.5) 236 (8.5)

76 (89.4) 9 (10.6)

64.7 1,328 (48)

65.8 49 (58) 42 (49) 16 26 43 (51)

1,821 689 98 156

(65.9) (24.9) (3.6) (5.6)

49 (57.6) 30 (35.3) 6 (7.1) 0

710 1,649 222 183

(25.7) (59.7) (8.0) (6.6)

14 53 7 11

1,483 (53.7) 1,281 (46.3) 510 954 99 781 420

(18.4) (34.5) (3.6) (28.3) (15.2)

(16.5) (62.4) (8.2) (12.9)

47 (55.3) 38 (44.7) 18 32 3 26 6

(21.2) (37.6) (3.5) (30.6) (7.1)

*Data are presented as No. (%) unless otherwise indicated. pTNM ⫽ pathologic TNM. †Details from reference tumor are presented. ‡Segmental or wedge resection. §From Mountain.1

were used as categorical variables with two classes: age (⬍ 65 years vs ⱖ 65 years); histology (identical vs different); and tumor localization (one lobe vs different lobes). The variable stage was used as a categorical variable with three classes (stages I, II, and IIIA). A forward-selection approach was used.

Results One hundred twelve patients (4.1%) with single lung cancer (n ⫽ 2,764) and 12 patients (14.1%) with synchronous lung cancer (n ⫽ 85) died within 30 days after operation. Eleven patients who underwent resection for synchronous lung cancer died because of cardiorespiratory complications. Postoperative mortality was especially high (6 of 28) among patients with bilateral lesions who underwent resection in one session. However, the characteristics and the causes of postoperative death did not differ from those of the other patients with synchronous lung CHEST / 118 / 4 / OCTOBER, 2000

953

Table 3—Characteristics of Evaluable Patients With Synchronous Lung Cancers (n ⴝ 73)* Characteristics Histology15 Identical Different Stage† I–I II–I II–II IIIA–ⱕ II Resection Pneumonectomy (Bi)Lobectomy Limited resection Tumor in first resection

No.

Reference Tumor, No.

Second Tumor, No.

50 23 40 22 5 6 13 53 7

19 29 25

*Deaths within 30 days are excluded. †Reference tumor vs second lesion.

cancer. Eight patients with single lung cancer were unavailable for follow-up, and therefore the data for 2,717 patients (2,644 single and 73 synchronous lesions cancer) were included for detailed survival analysis. The characteristics of the 73 patients with synchronous lung cancer included in the analysis are given in Table 3, but were not different from those of the total group of 85 patients who underwent surgery for synchronous lung cancer. As of January 1, 2000, 9 of the 73 patients with synchronous lung cancer were alive, with survival times of 2.1 to 6.4 years (median, 4.6 years) and were free of cancer. Thirty-five of the 64 patients who died, died of recurrence or metastasis of lung cancer, and 20 died of noncancer-related causes; in 9 pa-

Figure 2. Top: Estimated 5-year survival of patients with synchronous lung cancer with reference tumor stage I (n ⫽ 40; broken line), compared to survival of patients with stage II single NSCLC (n ⫽ 834; solid line; no survival difference; p ⫽ 0.48). ⫹ ⫽ censored cases. Bottom: Estimated 5-year survival of patients with synchronous lung cancer with reference tumor stage II (n ⫽ 27; broken line), compared to survival of patients with stage IIIA single NSCLC (n ⫽ 405; solid line; no survival difference; p ⫽ 0.52). ⫹ ⫽ censored cases.

tients, the cause of death was not recorded. Recurrence or metastases was diagnosed at a median time of 12 months (range, 2 to 143 months) after resection. Single and Synchronous Lung Cancer

Figure 1. Estimated survival after resection of patients with single (n ⫽ 2,644; solid line) and synchronous (n ⫽ 73; broken line) NSCLC (p ⬍ 0.0001). ⫹ ⫽ censored cases. 954

The demographic characteristics of the two groups did not differ. The patients with synchronous lung cancer were older (median age difference, 1.1 year), however not significantly (p ⫽ 0.35), and the stage distribution of the single and reference tumor was not different (p ⫽ 0.36). Overall survival was significantly higher (p ⬍ 0.0001) in patients with a single lung cancer. Five-year survival after surgery was 41% Clinical Investigations

Table 4 —Median Survival and 5-yr Survival Rates After Resection for Single and Synchronous Lung Cancer* Single Stage† I IA IB II IIA IIB IIIA ⱖ II

Synchronous

n

Median, yr

5-yr Survival, %

n

Median, yr

5-yr Survival, %

1,405 495 910 834 96 738 405 1,239

5.35 6.88 4.39 2.52 4.96 2.39 1.37 1.97

52 61 46 35 49 33 18 29

40 16 24 27 3 24 6 33

2.78 3.73 2.46 1.45 1.45 1.46 0.94 1.31

23 20 25 16 0 19 0 13

*Reference tumor used as stage marker for comparison. †From Mountain.1

for patients with single lung cancer and 19% for patients with synchronous lung cancer (Fig 1). Survival after surgery in single lung cancer was significantly different between stage I and stage II disease (p ⬍ 0.0001), stage II and stage IIIA disease (p ⬍ 0.0001), and between subsets IA and IB (p ⬍ 0.0001), and IIA and IIB (p ⬍ 0.04). Using the reference tumor as stage marker in patients with synchronous lung cancer, there was no significant difference between subsets IA and IB (p ⫽ 0.42; hazard ratio, 0.73; 95% CI, 0.34 to 1.55) and between subsets IIA and IIB (p ⫽ 0.80; hazard ratio, 1.17; 95% CI, 0.34 to 4.00). Because of this observation and the small numbers of synchronous lung cancer patients, comparative analyses of subsets of patients with single and synchronous lung cancer were not done. The survival of patients with stage I synchronous lung cancer (n ⫽ 40) was not significantly different (p ⫽ 0.48; hazard ratio, 0.88; 95% CI, 0.61 to 1.26) from that of patients with stage II single lung cancer (n ⫽ 834; Fig 2, top), and neither was the survival of patients with stage II synchronous lung cancer (n ⫽ 27) significantly different (p ⫽ 0.52; hazard ratio, 0.88; 95% CI, 0.58 to 1.31) from patients with stage IIIA single lung cancer (n ⫽ 405; Fig 2, bottom; Table 4). The relative risk of death was 1.75 higher in patients with synchronous lung cancer than in patients with single lung cancer,

irrespective of the stage of the disease. When the stage of the disease was taken into account, the relative risk of death ranged from 1.73 to 1.93 (Table 5). Synchronous Lung Cancer With the reference tumor as stage marker, survival was found to be significantly different for the different disease stages (stages I, II, and IIIA; p ⬍ 0.005). Survival differed significantly between stage I and stage II disease (p ⬍ 0.05) or stage IIIA disease (p ⬍ 0.004), but not between stage II and stage IIIA disease (p ⫽ 0.15; hazard ratio, 0.50; 95% CI, 0.20 to 1.28; Fig 3; Table 6). After pooling, the patients with stage II and stage IIIA disease had a significantly lower survival (p ⬍ 0.01) than the patients with stage I disease. Five-year survival was 13% vs 23%, respectively (Fig 4; Table 6). According to the new staging system,1 synchronous lesions with the same histology are classified as stage IIIB and stage IV. In the present survival analysis, 50 patients had synchronous lesions with the same histology. Although the numbers are small, a significant difference (p ⬍ 0.03) in survival was found between patients with tumors located within one lobe (n ⫽ 7; stage IIIB) and tumors located in different lobes (n ⫽ 43; stage IV). The same result was found when the data of all the patients with

Table 5—Cox’s Proportional Hazards Model of Factors Associated With Postoperative Survival After Resection in Patients With Single and Synchronous Lung Cancer Variables

Hazard Ratio

95% CI

␤

SE

Synchronous vs single Stage I* Stage II* Stage IIIA* Stage ⱖ II*

1.75 1.89 1.87 1.93 1.73

1.36–2.25 1.32–2.71 1.26–2.80 0.86–4.33† 1.21–2.48

0.56 0.64 0.63 0.66 0.55

0.13 0.18 0.20 0.41 0.18

*Synchronous vs single; in synchronous lung cancer, reference tumor is used as stage marker. †No significant difference. CHEST / 118 / 4 / OCTOBER, 2000

955

Figure 3. Estimated 5-year survival after tumor resection of patients with synchronous lung cancer with reference tumor stage I (n ⫽ 40; solid line), stage II (n ⫽ 27; intermediate broken line), and stage IIIA (n ⫽ 6; broken line). ⫹ ⫽ censored cases.

synchronous cancer (n ⫽ 73) were analyzed (p ⬍ 0.03; hazard ratio, 0.44; 95% CI, 0.22 to 0.90). Five-year survival was 29% and 16% for single and multiple lobe disease, respectively. Further analysis (Cox’s regression analysis) demonstrated that this difference was related to the stage of the reference tumor. Kaplan-Meier analysis demonstrated that tumor histology did not affect survival (p ⫽ 0.79). When the cases were grouped by histologic subtype, there was no significant difference in survival. Furthermore, the median survival was poorer in patients treated by a surgical procedure, including limited resection (such as a wedge or segmental resection; n ⫽ 32), than it was in patients treated by pneumonectomy or lobectomy(ies) (n ⫽ 41); however, this difference was not statistically significant (p ⫽ 0.16; hazard ratio, 1.44; 95% CI, 0.86 to 2.40). Cox’s regression analysis showed that the stage of the reference tumor was the only significant predictor of survival in patients with synchronous lung cancer (p ⬍ 0.005); age (p ⫽ 0.06), histology (p ⫽ 0.90) and localization of tumors (p ⫽ 0.11)

Figure 4. Estimated 5-year survival after tumor resection of patients with synchronous lung cancer treated by pneumonectomy or lobectomy(ies) (n ⫽ 41; broken line), compared to surgical procedures including limited resections (n ⫽ 32; solid line). ⫹ ⫽ censored cases.

were not. Survival was poorer when the reference tumor was higher than stage I (Table 6). Discussion This study confirms and quantifies the poorer postsurgical survival of patients with synchronous NSCLC, compared with that of patients with single NSCLC. The risk of dying of patients with synchronous lung cancer was almost double that of patients with single lung cancer. The stage of the most advanced tumor was the best predictor of survival in patients with synchronous lung cancer. Our results as well as the results of others show that surgery is a treatment option for patients with synchronous lung cancer.3– 8 Survival after resection in the patients with synchronous lung cancer was comparable with that reported in the literature.3,6,8,19 For patients with stage IA single NSCLC, survival is poorer after limited resection.20 However, in the present study, this observation was not confirmed clearly, possibly because several resections were performed. Never-

Table 6 —Cox’s Proportional Hazards Model of Factors Associated With Postoperative Survival After Resection in Patients With Synchronous Lung Cancer Variables

Hazard Ratio

95% CI

␤

SE

Stage II reference tumor* Stage ⱖ II reference tumor* Stage IIIA reference tumor*

1.73 1.92 3.84

1.01–2.96 1.15–3.20 1.53–9.64

0.55 0.65 1.35

0.27 0.26 0.47

*Related to stage I reference tumor. 956

Clinical Investigations

theless, the use of limited resections for synchronous lesions is not supported and should be considered as a viable “compromise” surgical treatment.20 In the literature, two separate definitions are used to define and stage synchronous malignant lesions in the lung. The time-dependent diagnosis (ie, synchronous vs metachronous) is rather definite. Malignant lesions have to grow before they can be detected. If CT and other scanning techniques had been available at time of diagnosis of all the patients, it is possible that more patients would have been diagnosed with synchronous cancer. However, we aimed to include only patients whose disease met the strict definitions used by Martini and Melamed.2 These authors excluded synchronous tumors affecting common lymphatics if these had the same histology, because of possible metastatic disease. In contrast to former definitions, the new staging system limits the staging of multiple synchronous tumors to lesions with the same histology and possible metastatic disease.1 Synchronous satellite pulmonary nodules situated in the same lobe are considered to be locally advanced disease and are staged as T4 (stage IIIB). Tumors located in different lobes are staged as metastatic disease (stage IV).1 The definitions used by Martini and Melamed2 and in the 1997 staging system1 are debatable. Grossly, both definitions fail to implement the multiple field cancerization theory21 and to deal with the possibility of metastatic disease. This is really a problem when the diagnosis is made on the basis of histology according to World Health Organization criteria10 and the histology of both tumors is identical. In the future, new molecular techniques will make it possible to distinguish metastatic lesions from separate primary tumors.22,23 Cancer staging is designed to determine therapy and prognosis.1 In the new staging system, primary lung cancer with secondary pulmonary nodules is staged by upstaging, but there is no proposal for the staging of synchronous lung cancers. In order to analyze the meaning of staging and upstaging in synchronous disease, we first staged all tumors separately. The most advanced tumor was defined as the reference tumor. Because the stage of the reference tumor proved to be the main predictor of survival, we then compared patients with single and synchronous lung cancer according to the stage of the (reference) tumor. The survival of patients with synchronous lung cancer with stage I reference tumor appeared to be comparable to that of patients with stage II single lung cancer, and the survival of patients with synchronous lung cancer with reference tumor stage II appeared comparable to that of patients with stage IIIA single NSCLC. Although our study included relatively a small number of patients, our findings may help to stage synchronous

multiple NSCLC and to determine prognosis. We propose that the reference tumor should be upstaged by one stage. Thus, patients with synchronous lung cancer with reference tumor of stage I and stage II are eligible for surgery and have a survival comparable to that of patients with primary stage II and stage IIIA NSCLC, in whom surgery is performed. In conclusion, the prognosis of patients with synchronous NSCLC is worse than that of patients with single NSCLC. After resection, the risk of dying is almost double that of patients with single NSCLC. The stage of the most advanced tumor is the best predictor of prognosis, and survival is significantly better in patients with a stage I reference tumor than it is in patients with a higher-staged reference tumor. The survival of patients with synchronous NSCLC with stage I and stage II reference tumors was not different from that of patients with single lung cancer stage II and stage IIIA, respectively. These results indicate that synchronous cancer should be upstaged. ACKNOWLEDGMENT: The authors thank Jane Sykes for her advice in editing the manuscript.

References 1 Mountain CF. Revisions in the international system for staging lung cancer. Chest 1997; 111:1710 –1717 2 Martini N, Melamed MR. Multiple primary lung cancers. J Thorac Cardiovasc Surg 1975; 70:606 – 611 3 Verhagen AF, Van De Wal HJ, Cox AL, et al. Surgical treatment of multiple lung cancers. Thorac Cardiovasc Surg 1989; 37:107–111 4 Ferguson MK, DeMeester TR, DesLauriers J, et al. Diagnosis and management of synchronous lung cancers. J Thorac Cardiovasc Surg 1985; 89:378 –385 5 Okada M, Tsubota N, Yoshimura M, et al. Operative approach for multiple primary lung carcinomas. J Thorac Cardiovasc Surg 1998; 115:836 – 840 6 Rosengart TK, Martini N, Ghosn P, et al. Multiple primary lung carcinomas: prognosis and treatment. Ann Thorac Surg 1991; 52:773–779 7 Ribet M, Dambron P. Multiple primary lung cancers. Eur J Cardiothorac Surg 1995; 9:231–236 8 Deschamps C, Pairolero PC, Trastek VF, et al. Multiple primary lung cancers. J Thorac Cardiovasc Surg 1990; 99: 769 –778 9 Yano M, Arai T, Inagaki K, et al. Intrapulmonary satellite nodule of lung cancer as a T factor. Chest 1998; 114:1305– 1308 10 DesLauriers J, Brisson J, Cartier R, et al. Carcinoma of the lung: evaluation of satellite nodules as a factor influencing prognosis after resection. J Thorac Cardiovasc Surg 1989; 97:504 –512 11 Nakajima J, Furuse A, Oka T, et al. Excellent survival in a subgroup of patients with intrapulmonary metastasis of lung cancer. Ann Thorac Surg 1996; 61:158 –163 12 Yoshino I, Nakanishi R, Osaki T, et al. Postoperative prognosis in patients with non-small cell lung cancer with synchronous ipsilateral intrapulmonary metastasis. Ann Thorac Surg 1997; 64:809 – 813 CHEST / 118 / 4 / OCTOBER, 2000

957

13 Okada M, Tsubota N, Yoshimura M, et al. Evaluation of TMN classification for lung carcinoma with ipsilateral intrapulmonary metastasis. Ann Thorac Surg 1999; 68:326 –331 14 Van Rens MThM, Brutel de la Rivie`re A, Elbers HRJ, et al. Prognostic assessment of 2,361 patients who underwent pulmonary resection for non-small cell lung cancer stage I, II and IIIA. Chest 2000; 117:374 –379 15 Histologic typing of lung tumors. 2nd ed. Geneva, Switzerland: World Health Organization, 1981 16 Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53:457– 481 17 Peto R, Peto J. Asymptotically efficient rank invariant test procedures. J Stat Soc (series A) 1972; 135:185–198 18 Cox DR. Regression models and life tables. J R Stat Soc B 1972; 34:187–220 19 Pommier RF, Vetto JT, Lee JT, et al. Synchronous non-small

958

cell lung cancers. Am J Surg 1996; 171:521–524 20 Landreneau RJ, Sugarbaker DJ, Mack MJ, et al. Wedge resection versus lobectomy for stage I (T1 N0 M0) non-small cell lung cancer. J Thorac Cardiovasc Surg 1997; 113:691– 698 21 Slaughter DP, Southwick HW, Smejkal W. Field cancerization in oral stratified squamous epithelium: clinical implications of multicentric origin. Cancer 1953; 6:963–968 22 Leong PP, Rezai B, Kock WM, et al. Distinguishing second primary tumors from lung metastases in patients with head and neck squamous cell carcinoma. J Natl Cancer Inst 1998; 90:972–977 23 Kropveld A, Rozemuller EH, Leppers FGJ, et al. Sequencing analysis of RNA and DNA of exons 1 through 11 shows p53 gene alterations to be present in almost 100% of head and neck squamous cell cancers. Lab Invest 1999; 79:347–353

Clinical Investigations