- Email: [email protected]
Preoperative pulmonary function as a prognostic factor for stage I non–small cell lung carcinoma
GENERAL THORACIC
Preoperative Pulmonary Function as a Prognostic Factor for Stage I Non–Small Cell Lung Carcinoma Toshihiko Iizasa, MD, Makoto Suzuki, MD, Kazuhiro Yasufuku, MD, Akira Iyoda, MD, Mizuto Otsuji, MD, Shigetoshi Yoshida, MD, Yasuo Sekine, MD, Kiyoshi Shibuya, MD, Yukio Saitoh, MD, Kenzo Hiroshima, MD, and Takehiko Fujisawa, MD Departments of Thoracic Surgery and Basic Pathology, Graduate School of Medicine, Chiba University, Chiba, Japan
Background. The aim of this study was to clarify preoperative lung function as a prognostic factor for the long-term survival of, and to discuss the appropriateness of lobectomy for, patients with stage I non–small cell lung carcinoma who have poor preoperative pulmonary function. Methods. The study group consisted of 402 lobectomized patients with stage I non–small cell lung carcinoma treated by complete resection from 1985 to 1997. Preoperative percent forced vital capacity [(forced vital capacity/predicted forced vital capacity) ⴛ 100], FEV1% [(forced expiratory volume in 1 second/forced vital capacity) ⴛ 100], arterial carbon dioxide tension, and smoking were statistically analyzed as prognostic factors together with other host and tumor biologic factors. Results. Multivariate analysis demonstrated that tumor size (p < 0.0001) was the most significant prognostic factor for survival from primary lung cancer. Age (p < 0.0001), sex (p ⴝ 0.0036), and FEV1% (p ⴝ 0.0046) were
found to be independent prognostic factors for survival from death by nonprimary lung cancer–related causes. Smoking was highly correlated with FEV1% (correlation coefficient ⴝ ⴚ0.511; p < 0.0001). The 100 patients with a preoperative FEV1% less than 70% included 34 patients with nonprimary lung cancer–related deaths, whereas the 302 patients with an FEV1% of 70% or greater included only 23 patients (p < 0.0001). Conclusions. Along with tumor size, FEV1% is the most significant prognostic factor for patients with stage I non–small cell lung carcinoma with regard to survival from death by other causes. Lobectomy may not be preferred as an appropriate surgical modality for patients with stage I non–small cell lung carcinoma with small peripheral nodules who exhibit poor pulmonary function, especially lowered FEV1%.
A
divided by predicted FVC times 100 (%FVC), forced expiratory volume in 1 second (FEV1) divided by FVC times 100 (FEV1%), and arterial carbon dioxide tension (Paco2) were examined to gauge pulmonary disorder preoperatively. The present retrospective study is an attempt to clarify prognostic factors related to long-term survival in surgery cases of stage I NSCLC. We discuss appropriateness of lobectomy and important considerations for following up for patients with stage I NSCLC with poor pulmonary function.
s screening with spiral computed tomography becomes more prevalent, more patients with small peripheral lesions are being detected, and a great deal of debate regarding the appropriateness of lobectomy versus limited resection has been renewed. Lobectomy has been associated with a more positive prognosis for non– small cell lung carcinoma (NSCLC) than limited operations have been [1, 2]. The disadvantage of impaired lung function as a result of a lobectomy, however, makes it a controversial option [3– 8], especially for stage I patients who can expect long-term survival. Errett and colleagues [3] reported that wedge resection could be an alternative for poor-risk patients in stage I NSCLC. However, they could not determine the most important prognostic factors because of a univariate evaluation. Few reports have considered preoperative pulmonary function as a prognostic factor for the long term [9], because these functions have no relationship to the biologic features of NSCLC. We selected eight factors that may significantly influence prognosis for stage I NSCLC. These included the patient variables of age, sex, and smoking, and the tumor size and histologic type. Also, forced vital capacity (FVC)
Accepted for publication Oct 2, 2003. Address reprint requests to Dr Fujisawa, Department of Thoracic Surgery, Graduate School of Medicine, Chiba University, 1-8-1 Inohana, Chuo-ku, Chiba 260-8670, Japan; e-mail: [email protected].
© 2004 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2004;77:1896 –903) © 2004 by The Society of Thoracic Surgeons
Material and Methods Patients Between January 1985 and December 1997, 1,131 patients underwent surgical resection for primary lung cancer in our institute. Of those, 451 patients were confirmed as having stage I NSCLC after pathologic examination of the excised materials. For this study, 11 patients with incomplete resection, 14 patients with low-grade malignant type of NSCLC (11 carcinoids, 2 mucoepidermoid carcinomas, and 1 adenocystic carcinoma), 10 patients with a central lesion from a segmental bronchus, 20 patients with nonlobectomy (10 pneumonectomies, 7 wedge resections, 2 segmentectomies, and 1 main bronchus resection), and 5 patients who died of complications 0003-4975/04/$30.00 doi:10.1016/j.athoracsur.2003.10.014
within 1 month postoperatively were excluded from the survival evaluation. Some patients have overlapped on exclusion criteria; in all, 402 cases were used in this study. The histologic types of the tumors included 136 squamous cell carcinomas, 245 adenocarcinomas, 18 large cell carcinomas, and 3 adenosquamous carcinomas. Of the patients, 147 were women and 255 were men, ranging in age from 34 to 87 years with a mean age and standard deviation of 62.9 ⫾ 9.8 years. The advanced age group was defined as those patients 70 years of age or older. Tumors were classified by pathologic findings as 253 T1 (T1; ⱕ3 cm) and 149 T2 (T2; ⬎3 cm). Surgical procedures performed were 402 lobectomies, including 27 bilobectomies. Histologic type, pathologic stage, and TNM classification were identified according to the criteria of the American Joint Committee on Cancer [10] and the Classification of Lung Cancer description by the Japan Lung Cancer Society [11]. Large cell carcinomas and adenosquamous carcinomas were classified into the nonsquamous group together with adenocarcinomas for the univariate and multivariate studies. Preoperative %FVC, FEV1%, and Paco2 were adapted as clinical prognostic factors for pulmonary disorder. The patient’s history of cigarette smoking was evaluated according to the description on the patient’s chart. This was expressed as pack-years, calculated by the number of packs of cigarettes smoked per day multiplied by the number of years the subject smoked. For the first year after surgery, patients were followed up every month, then every 3 months from 2 to 5 years, and every 6 months thereafter. Additional follow-up and survival were determined by telephone or postal contact with the patient. When distant or local disease recurrence developed, any treatment was permitted.
Survival Rate and Statistical Analysis The end point was survival from the primary lung cancer–related deaths, the nonprimary lung cancer–related deaths, and overall survival in this study, calculated between the date of operation and the time of death. Cause of death was confirmed by telephone contact with the doctor who monitored the patient. Cause of death was classified as primary lung cancer–related if the patient had a recurrence of the primary lung cancer at the time of death (distant metastasis, local recurrence, or both). Nonprimary lung cancer–related death classification was determined in patients free of recurrence of the primary lung cancer at the time of death (eg, death as a result of respiratory disease, cardiac disease, brain vascular disease, second malignancy, or other). Overall survival and survival from the primary and nonprimary lung cancer–related deaths were calculated by the KaplanMeier method [12]. We have analyzed survival from the primary lung cancer by censoring patients with nonprimary lung cancer–related deaths at the date of death and vice versa. There is, therefore, no algebraic relationship between the overall survival rate and the survival rates in the two subgroups of primary lung cancer and nonprimary lung cancer. Because of this there is not an inconsistency between the overall 5-year survival of 76.3% and
IIZASA ET AL PREOPERATIVE PULMONARY FUNCTION IN STAGE I NSCLC
1897
the subgroup rates of 84.2% and 90.6%. Clinical prognostic factors examined in this study included age, sex, smoking status, histologic type, tumor size, preoperative %FVC, preoperative FEV1%, and preoperative Paco2. Each variable for survivals was analyzed in both strata and continuous using Cox proportional hazards. Multivariate analysis was calculated in a forward stepwise procedure, according to the Cox proportional hazards model [13], using the SPSS statistical software program package (SPSS version 11.0 for Windows, SPSS Inc, Chicago, IL). A 0.10 level of probability value was the significance level used for adding and deleting a covariable from the model. The 2 test was used for contingency table analysis in the case of two-by-two or threeby-two tables. The correlation of %FVC, FEV1%, Paco2, and pack-years was evaluated with the Spearman rank correlation coefficient. Statistical analysis was regarded as statistically significant at a p value less than 0.05.
Results Survival and Univariate Analysis Overall 5- and 10-year survival rates were 76.3% and 60.2%, respectively. This corresponded to 5- and 10-year survival rates from the primary lung cancer–related deaths of 84.2% and 74.9%, respectively, and from the nonprimary lung cancer–related deaths of 90.6% and 80.4%, respectively. For overall survival, each prognostic factor was determined as significant by univariate p value (age, tumor size, and FEV1%, p ⬍ 0.0001; sex, p ⫽ 0.0001; tumor histology, p ⫽ 0.0050; pack-year, p ⫽ 0.0012; %FVC, p ⫽ 0.4500; Paco2, p ⫽ 0.0275; Cox proportional hazards; Table 1). For survival rates from the primary lung cancer– related deaths, the p value for tumor size classification (T factor, p ⬍ 0.0001) was identified as significant with Cox proportional hazards (Table 2). For survival from the nonprimary lung cancer–related deaths, age (p ⬍ 0.0001), sex (p ⬍ 0.0001), tumor histology (p ⬍ 0.0001), pack-year (p ⬍ 0.0001), FEV1% (p ⬍ 0.0001), %FVC (p ⫽ 0.0016), and Paco2 (p ⫽ 0.0132) each displayed a significant difference with Cox proportional hazards (Table 3). We were able to demonstrate that survival rates from the primary lung cancer–related deaths were not significantly different with FEV1%, although differences in overall survival rates and survival rates from the nonprimary lung cancer– related deaths were statistically significant for this factor (p ⬍ 0.0001, Cox proportional hazards), which we divided into three groups based on mild chronic obstructive pulmonary disease in the European Respiratory Society [14] and the British Thoracic Society [15] (Fig 1).
Correlation Between Pulmonary Function and Smoking Pack-Years We also evaluated the rate of smoking for correlation with %FVC and FEV1% using a Spearman rank correlation coefficient. Our data disclosed a significant correlation between decreased FEV1% and increased smoking. The FEV1% correlated inversely with smoking pack-years (correlation coefficient ⫽ ⫺0.511; p ⬍ 0.0001; Fig 2).
GENERAL THORACIC
Ann Thorac Surg 2004;77:1896 –903
GENERAL THORACIC
1898
IIZASA ET AL PREOPERATIVE PULMONARY FUNCTION IN STAGE I NSCLC
Table 1. Univariate Analyses of Patient Characteristics and Overall Survival After Surgery for Stage I Non–Small Cell Lung Carcinoma
Characteristic Age ⱖ70 y ⬍70 y Sex Male Female Histology Sq Ad La Ad-sq Tumor size (T factor) T1 T2 Pack-years ⱖ30 ⬍30 FEV1% ⬍60% ⱖ60%, ⬍70% ⱖ70% %FVC ⬍70% ⱖ70%, ⬍80% ⱖ80% Paco2 ⬍40 mm Hg ⱖ40 mm Hg, ⬍45 mm Hg ⱖ5 mm Hg
No. of No. of Dead 5-Year 10-Year Univariate Cases Cases Survival Survival p Valuea 114 288
52 87
64.1 81.0
46.9 65.1
⬍0.0001
255 147
108 31
71.4 85.0
51.4 76.7
0.0001
136 245 18 3
59 73 7 0
68.9 81.1 64.4 100
49.2 65.7 64.4 100
0.0050b
253 149
70 69
84.4 62.6
67.2 48.7
⬍0.0001
212 190
98 41
68.0 85.7
47.9 75.0
0.0012
30 70 302
19 35 85
59.8 64.1 80.9
29.0 43.8 67.7
15 32 355
5 17 117
66.7 48.0 79.3
66.7 43.7 61.6
0.4500
157 208
70 58
67.5 83.0
51.9 66.3
0.0275
37
11
77.3
66.8
a Univariate p value with Cox proportional hazards. non-Sq.
Ann Thorac Surg 2004;77:1896 –903
tional hazards model using seven clinical prognostic factors (age, sex, histology, tumor size, FEV1%, %FVC, and Paco2). Smoking was excluded as a factor from these analyses because of its considerable statistical correlation with FEV1% and the fact that FEV1% is objectively measured, whereas smoking is self-reported. Multivariate analysis for overall survival revealed three independent prognostic factors: age (p ⫽ 0.0002), sex (p ⫽ 0.0065), and tumor size (p ⬍ 0.0001; Table 4). Multivariate analysis for survival from the primary lung cancer–related deaths demonstrated only tumor size (p ⬍ 0.0001) as an independent prognostic factor (Table 4). In addition, multivariate analysis for survival from the nonprimary lung cancer–related deaths demonstrated age (p ⬍ 0.0001), sex Table 2. Univariate Analyses of Patient Characteristics and Survival From the Primary Lung Cancer–Related Deaths After Surgery for Stage I Non–Small Cell Lung Carcinoma
Characteristic
⬍0.0001
b
Sq versus
Ad ⫽ adenocarcinoma; Ad-sq ⫽ adenosquamous carcinoma; FEV1% ⫽ percent forced expiratory volume in 1 second; %FVC ⫽ percent forced vital capacity; La ⫽ large cell carcinoma; Paco2 ⫽ arterial carbon dioxide partial pressure; Sq ⫽ squamous cell carcinoma; T1 ⫽ 3 cm or less; T2 ⫽ more than 3 cm.
Furthermore, the number of cases (85 of 100 cases) with FEV1% less than 70% for heavy smokers with 30 or more pack-years was significantly larger than that for smokers with 29 or fewer pack-years or nonsmokers, as compared with the number of cases (134 of 302) with FEV1% of 70% or greater for heavy smokers with 30 or more pack-years (p ⬍ 0.0001; 2 test). However, no correlation between %FVC and smoking or %FVC and FEV1% was found (data not shown).
Multivariate Analyses Multivariate analyses of overall survival and survival from the primary and nonprimary lung cancer–related deaths were performed according to the Cox propor-
Age ⱖ70 y ⬍70 y Sex Male Female Histology Sq Ad La Ad-sq Tumor size (T factor) T1 T2 Pack-years ⱖ30 ⬍30 FEV1% ⬍60% ⱖ60%, ⬍70% ⱖ70% %FVC ⬍70% ⱖ70%, ⬍80% ⱖ80% Paco2 ⬍40 mm Hg ⱖ40 mm Hg, ⬍45 mm Hg ⱖ45 mm Hg a
No. of No. of Dead 5-Year 10-Year Univariate Cases Cases Survival Survival p Valuea 114 288
25 56
78.6 86.2
71.6 76.3
0.1674
255 147
57 24
82.6 87.0
71.6 80.2
0.1022
136 245 18 3
24 54 3 0
84.7 84.1 81.0 100
75.8 73.9 81.0 100
0.7208b
253 149
36 45
91.4 71.6
81.3 63.9
⬍0.0001
212 190
49 32
80.4 88.2
70.7 79.3
0.4614
30 70 302
4 15 62
88.5 80.4 84.6
82.9 67.2 75.5
15 32 355
3 7 71
78.6 74.3 85.3
78.6 74.3 75.1
0.0986
157 208
36 38
79.4 87.2
73.4 76.3
0.4286
37
7
87.4
75.5
Univariate p value with Cox proportional hazards. non-Sq.
0.3732
b
Sq versus
Ad ⫽ adenocarcinoma; Ad-sq ⫽ adenosquamous carcinoma; FEV1% ⫽ percent forced expiratory volume in 1 second; %FVC ⫽ percent forced vital capacity; La ⫽ large cell carcinoma; Paco2 ⫽ arterial carbon dioxide partial pressure; Sq ⫽ squamous cell carcinoma; T1 ⫽ 3 cm or less; T2 ⫽ more than 3 cm.
IIZASA ET AL PREOPERATIVE PULMONARY FUNCTION IN STAGE I NSCLC
Table 3. Univariate Analyses of Patient Characteristics and Survival From the Nonprimary Lung Cancer–Related Deaths After Surgery for Stage I Non–Small Cell Lung Carcinoma
Characteristic Age ⱖ70 y ⬍70 y Sex Male Female Histology Sq Ad La Ad-sq Tumor size (T factor) T1 T2 Pack-years ⱖ30 ⬍30 FEV1% ⬍60% ⱖ60%, ⬍70% ⱖ70% %FVC ⬍70% ⱖ70%, ⬍80% ⱖ80% Paco2 ⬍40 mm Hg ⱖ40 mm Hg, ⬍45 mm Hg ⱖ45 mm Hg
No. of No. of Dead 5-Year 10-Year Univariate Cases Cases Survival Survival p Valuea 114 288
27 31
81.6 94.0
65.5 85.2
⬍0.0001
255 147
51 7
86.5 97.7
71.7 95.7
⬍0.0001
136 245 18 3
35 19 4 0
81.3 96.4 79.5 100
64.9 89.0 79.5 100
⬍0.0001b
253 149
34 24
92.3 87.5
82.7 76.2
0.0641
212 190
49 9
84.6 97.1
67.7 94.6
⬍0.0001
30 70 302
15 20 23
67.6 79.7 95.5
34.9 65.1 89.6
15 32 355
2 10 46
84.8 64.7 93.0
84.8 58.8 82.0
0.0016
157 208
34 20
85.0 95.2
70.8 86.9
0.0132
37
4
88.4
88.4
a Univariate p value with Cox proportional hazards. non-Sq.
1899
preoperative FEV1% of less than 70% included 35 who died of nonprimary lung cancer–related causes, whereas of 302 patients with an FEV1% of 70% or more, only 23 patients with nonprimary lung cancer–related deaths were observed (p ⬍ 0.0001, 2 test; Table 5). Thirty of 35 deaths in the group with an FEV1% of less than 70%, who were also heavy smokers (⬎30 pack-years), included 13 patients with respiratory insufficiency as a result of pneumonia or chronic obstructive pulmonary disease without primary lung cancer recurrence and 10 patients with secondary malignancy in other organs.
Comment
⬍0.0001
b
Sq versus
Ad ⫽ adenocarcinoma; Ad-sq ⫽ adenosquamous carcinoma; FEV1% ⫽ percent forced expiratory volume in 1 second; %FVC ⫽ percent forced vital capacity; La ⫽ large cell carcinoma; Paco2 ⫽ arterial carbon dioxide partial pressure; Sq ⫽ squamous cell carcinoma; T1 ⫽ 3 cm or less; T2 ⫽ more than 3 cm.
(p ⫽ 0.0036), and FEV1% (p ⫽ 0.0046) as independent prognostic factors (Table 4).
Type of Recurrence and Cause of Death Of the 402 patients analyzed for survival, 139 died of a cause other than surgical complications. Eighty-one patients died of recurrent disease, and 58 died of nonrecurring diseases. Local recurrence was observed in 6 patients, distant metastases in 71, and both in 4 patients. Nonrecurring causes included 17 respiratory diseases, 7 cardiac diseases, 6 brain vascular diseases, and 17 second primary malignancies (Table 5). We further evaluated the cause in nonprimary lung cancer–related deaths according to their preoperative FEV1%. The 100 patients with a
In this study, we retrospectively investigated clinical characteristics, patterns of disease recurrence, and causes of death in surgical cases of stage I NSCLC and evaluated the statistical significance of prognostic factors. We made several noteworthy findings. First, preoperative impaired FEV1% as well as tumor size could be important prognostic factors of poor outcome in stage I NSCLC, as suggested by an increase in the number of nonprimary lung cancer–related deaths in this study. Second, a marked correlation between increased smoking and a decline in FEV1% was found in patients with stage I NSCLC. Patients with poor pulmonary function have been noted to be at a disadvantage in perioperative management. There are many reports in which the risk of postoperative complications correlated with low-grade lung function could be predicted with diverse variables [16 –18]. Most reports have discussed the risk of complications and mortality for poorly reserving lung function in a perioperative period, but few evaluate and describe the adequacy of surgical procedures for patients with poor pulmonary function from the standpoint of longterm survival [9]. In the last two decades, cure of surgical cases of NSCLC has focused on non– cancer-related death for patients with poor preoperative pulmonary function, especially those in stage I cases who can expect long-term survival and tend to have been overlooked. Many previous reports regarding prognostic factors of stage I NSCLC demonstrated that overall 5-year survival rate still remained about 60% to 80% in stage IA and IB together [19 –21], which is still worse as compared with that of other organ carcinomas [22–24]. Harpole [2] has reported that 79 patients died of other causes without any evidence of cancer, and 105 patients had died of cancer in 289 resected cases with stage I NSCLC. He demonstrated no significant difference between the numbers of primary and nonprimary lung cancer–related deaths in those resected cases with stage I NSCLC, although he did not examine and discuss the cause of nonprimary lung cancer–related deaths in detail [2]. Here we report that the cause of nonprimary lung cancer–related deaths is related to poorer FEV1% and smoking status. Death from pulmonary causes is also suspected to originate from preoperative impaired FEV1% and smoking status. Smoking would also increase the possibility of developing a second malignancy [20, 25].
GENERAL THORACIC
Ann Thorac Surg 2004;77:1896 –903
GENERAL THORACIC
1900
IIZASA ET AL PREOPERATIVE PULMONARY FUNCTION IN STAGE I NSCLC
Ann Thorac Surg 2004;77:1896 –903
Fig 1. The survival curves of the patient groups with stage I non–small cell lung cancer correlated with forced expiratory volume in 1 second divided by forced vital capacity, times 100 (FEV1%), for overall (A), primary lung cancer (B), and nonprimary lung cancer (C). (A) Overall 5and 10-year survival rates were 59.8% and 29.0%, respectively, for patients with FEV1% less than 60%; 64.1% and 43.8%, respectively, for the FEV1% 60% to 69.9% group; and 80.9%, and 67.7%, respectively, for the FEV1% 70% or greater group. (B) The 5- and 10-year survival rates from the primary lung cancer–related deaths were 88.5% and 82.9%, respectively, for the FEV1% less than 60% group; 80.4% and 67.2%, respectively, for the FEV1% 60% to 69.9% group; and 84.6%, and 75.5%, respectively, for the FEV1% 70% or greater group. (C) The 5and 10-year survival rates from the nonprimary lung cancer–related deaths were 67.6% and 34.9%, respectively, for the FEV1% less than 60% group; 79.7% and 65.1%, respectively, for the FEV1% 60% to 69.9% group; and 95.5% and 89.6%, respectively, for the FEV1% 70% or greater group.
Influences of smoking on pulmonary function have been reported since the 1960s. Dockery and associates [26] studied the correlation between long-term smoking and lung function, demonstrating that FVC and FEV1% decreased irreversibly and were inversely proportional to the amount smoked in pack-years. Our data, obtained using the Spearman rank correlation, also revealed a significant correlation between decreased FEV1% and increased smoking. Miyazawa and coworkers [27] reported that the condition of the pulmonary vascular bed after major lung resection does not improve, even in the late postoperative phase. We expect that in patients who smoke, the already impaired FEV1% might drop even further in the postoperative period compared with normal FEV1% levels. Furthermore, these patients may suffer greater morbidity related to pulmonary causes as a result of their poor pulmonary function.
In stage I cases, tumor size and histologic type are the most important factors used in determining the prognosis of stage I NSCLC for tumor progress and characteristics. A T1 designation suggested a better prognosis than T2 for survival from the primary lung cancer–related deaths in our univariate and multivariate analyses, in agreement with other reports [2, 20, 28 –30]. Most reports disclosed no significant difference in prognosis between squamous cell carcinoma and nonsquamous cell carcinoma in patients with stage I NSCLC [2, 20, 28 –30]. On the other hand, some reports using univariate analysis have noted that for T1 N0 M0 NSCLC, the squamous cell carcinoma cases have better overall survival rates than the nonsquamous cases [5]. Additionally, disease recurrence in nonsquamous cell carcinoma has been found to be more frequent than in squamous cell carcinoma in patients with T1 N0 NSCLC [19]. Our results are consis-
IIZASA ET AL PREOPERATIVE PULMONARY FUNCTION IN STAGE I NSCLC
1901
Table 5. Comparisons of Cause of Deaths in Nonprimary Lung Cancer in Patients With FEV1% ⬍ 70% Versus Those With ⱖ70%a Cause of Death
FEV1% ⬍ 70% (n ⫽ 100)
FEV1% ⱖ 70% (n ⫽ 302)
13 4 1 10 7 35
4 3 5 7 4 23b
Respiratory disease Cardiac disease Brain vascular disease Second malignancy Other Total a
Values are number of patients.
b
p value ⬍ 0.0001 with 2 test.
FEV1% ⫽ percent forced expiratory volume in 1 second.
Fig 2. Correlation between pulmonary function and smoking in pack-years. Correlation of smoking with forced expiratory volume in 1 second divided by forced vital capacity, times 100 (FEV1%), was detected at a correlation coefficient of ⫺0.515, p less than 0.0001 with Spearman rank correlation coefficient. The higher the smoking rate in pack-years, the lower the FEV1% was found to be.
tent with the former reports but contradict the latter reports. This discrepancy may be partially explained by the fact that our study included many long-term surviving patients who were identified through a health examination for elderly persons in Japan as having small adenocarcinoma. Additionally, the authors of the later reports adapted a univariate analysis that disregarded the involvement of other prognostic factors. We believe the histologic type of primary NSCLC was not regarded as a significant prognostic factor. We selected patient age and sex as primary factors that will affect survival. There are many disadvantageous Table 4. Multivariate Analyses of Prognostic Factors and Postsurgical Survival Rates Using the Cox Proportional Hazards Model Relative Riska
Variable Overall Age Sex (male vs female) Tumor size FEV1% Primary cancer related Tumor size Nonprimary lung cancer related Age Sex (female vs male) FEV1%
95% CI
p Value
1.040 1.807 1.201 0.985
1.061–1.190 0.0002 1.180 –2.786 0.0065 1.097–1.315 ⬍0.0001 0.967–1.003 0.0998
1.341
1.206 –1.490 ⬍0.0001
1.095 3.428 0.946
1.056 –1.135 ⬍0.0001 1.496 –7.857 0.0036 0.940 – 0.989 0.0046
a Relative risk of continuous variables (age, tumor size, and FEV1%) were ratios of larger value to smaller.
CI ⫽ Confidence interval; in 1 second.
FEV1% ⫽ percent forced expiratory volume
factors in patients 70 years of age or older. In our univariate analysis, overall survival rates in the elderly male group were significantly lower than those for the younger, female group, agreeing with other reports [31, 32]. However, survival from the primary lung cancer– related deaths was not significantly different between these groups when using multivariate analysis. Nonprimary lung cancer–related deaths were associated with a poorer prognosis in the overall survival of the elderly male group by multivariate analysis. Elderly men commonly have a shorter life expectancy; thus, age and sex are not prognostic factors for survival from the primary lung cancer–related deaths in stage I NSCLC, according to the Cox proportional hazard model. In conclusion, we demonstrated that non–lung cancer– related death, in particular pulmonary failure and second primary malignant disease, is sufficiently significant so as to warrant attention in stage I NSCLC patients with impaired FEV1%. The FEV1% is one of the most significant prognostic factors in stage I NSCLC patients, as is the tumor size. The FEV1% has been highly correlated with past smoking status. Damage to postoperative pulmonary function could not be disregarded in patients with poor FEV1% in stage I NSCLC and worsened their prognosis, as indicated by the increased number of non–lung cancer–related deaths. Lobectomy may not be preferred as an appropriate surgical modality. Pulmonary disease and second malignancy for any stage I NSCLC patient who has poor FEV1% should be carefully monitored so that postoperative lung function might be preserved.
References 1. Ginsberg RJ, Rubinstein LV. Randomized trial of lobectomy versus limited resection for T1 N0 non–small cell lung cancer. Lung Cancer Study Group. Ann Thorac Surg 1995; 60:615–23. 2. Harpole DH. Stage I non–small cell lung cancer. Cancer 1995;76:787–96. 3. Errett LE, Wilson J, Chiu RC, et al. Wedge resection as an alternative procedure for peripheral bronchogenic carcinoma in poor-risk patients. J Thorac Cardiovasc Surg 1985; 90:656 –61. 4. Bennet WF, Smith RA. Segmental resection for bronchogenic carcinoma: a surgical alternative for the compromised patient. Ann Thorac Surg 1979;27:169 –72.
GENERAL THORACIC
Ann Thorac Surg 2004;77:1896 –903
GENERAL THORACIC
1902
IIZASA ET AL PREOPERATIVE PULMONARY FUNCTION IN STAGE I NSCLC
5. Read RC, Yoder G, Schaeffer RC. Survival after conservative resection for T1 N0 M0 non–small cell lung cancer. Ann Thorac Surg 1990;49:391–8. 6. Miller JI, Hatcher CR Jr. Limited resection of bronchogenic carcinoma in the patient with marked impairment of pulmonary function. Ann Thorac Surg 1987;44:340 –3. 7. Hoffmann TH, Ransdell HT. Comparison of lobectomy and wedge resection for carcinoma of the lung. J Thorac Cardiovasc Surg 1980;79:211–7. 8. Pastorino U, Valente M, Bedini V, et al. Limited resection for stage I lung cancer. Eur J Surg Oncol 1991;17:42–6. 9. Ploeg AJ, Kappetein AP, van Tongeren RB, et al. Factors associated with perioperative complications and long-term results after pulmonary resection for primary carcinoma of the lung. Eur J Cardiothorac Surg 2003;23:26 –9. 10. Mountain CF. Revision in the international system for staging lung cancer. Chest 1997;111:1710 –7. 11. The Japan Lung Cancer Society. Classification of lung cancer. 5th ed. Tokyo: Kanehara, 2000. 12. Kaplan EL, Meier P. Nonparametric estimation from incomplete observation. J Am Stat Assoc 1958;53:457–81. 13. Cox DR. Regression models and life tables. J Royal Stat Soc 1972;B34:187–220. 14. Siafakas NM, Vermeire P, Pride NB, et al. Optimal assessment and management of chronic obstructive pulmonary disease (COPD). The European Respiratory Society Task Force. Eur Respir J 1995;8:1398 –420. 15. BTS guidelines for the management of chronic obstructive pulmonary disease: The COPD Guidelines Group of the Standards of Care Committee of the BTS. Thorax 1997; 52(Suppl 5):S1–28. 16. Ferguson MK, Little L, Rizzo L, et al. Diffusing capacity predicts morbidity and mortality after pulmonary resection. J Thorac Cardiovasc Surg 1988;96:894 –900. 17. Kearney DJ, Lee TH, Reilly JJ, et al. Assessment of operative risk in patients undergoing lung resection: importance of predicted pulmonary function. Chest 1994;105:753–9. 18. Pierce RJ, Copland JM, Sharpe K, et al. Preoperative risk evaluation for lung cancer resection: predicted postoperative product as a predictor of surgical mortality. Am J Respir Crit Care Med 1994;150:947–55. 19. Thomas P, Rubinstein L. Cancer recurrence after resection: T1 N0 non–small cell lung cancer. Lung Cancer Study Group. Ann Thorac Surg 1990;49:242–6. 20. Fujisawa T, Iizasa T, Saitoh Y, et al. Smoking before surgery
Ann Thorac Surg 2004;77:1896 –903
21. 22.
23.
24.
25.
26. 27. 28. 29. 30.
31. 32.
predicts poor long-term survival in patients with stage I non–small-cell lung carcinomas. J Clin Oncol 1999;17:2086 – 91. Dominioni L, Imperatori A, Rovera F, et al. Stage I nonsmall cell lung carcinoma: analysis of survival and implications for screening. Cancer 2000;89:2334 –44. van der Hage JA, van de Velde CJ, Julien JP, et al. Improved survival after one course of perioperative chemotherapy in early breast cancer patients: long-term results from the European Organization for Research and Treatment of Cancer (EORTC) Trial 10854. Eur J Cancer 2001;37:2184 –93. Lacy AM, Garcia-Valdecasas JC, Delgado S, et al. Laparoscopy-assisted colectomy versus open colectomy for treatment of non-metastatic colon cancer: a randomised trial. Lancet 2002;359:2224 –9. Barchielli A, Amorosi A, Balzi D, et al. Long-term prognosis of gastric cancer in a European country: a population-based study in Florence (Italy). 10-year survival of cases diagnosed in 1985–1987. Eur J Cancer 2001;37:1674 –80. Thomas PA Jr, Rubinstein L. Malignant disease appearing late after operation for T1 N0 non–small-cell lung cancer. The Lung Cancer Study Group. J Thorac Cardiovasc Surg 1993;106:1053–8. Dockery DW, Speizer FE, Ferris BG Jr, et al. Cumulative and reversible effects of lifetime smoking on simple tests of lung function in adults. Am Rev Respir Dis 1988;137:286 –92. Miyazawa M, Haniuda M, Nishimura H, et al. Longterm effects of pulmonary resection on cardiopulmonary function. J Am Coll Surg 1999;189:26 –33. Martini N, Bains MS, Burt ME, et al. Incidence of local recurrence and second primary tumors in resected stage I lung cancer. J Thorac Cardiovasc Surg 1995;109:120 –9. Williams DE, Pairolero PC, Davis CS, et al. Survival of patients surgically treated for stage I lung cancer. J Thorac Cardiovasc Surg 1981;82:70 –6. Gail MH, Eagan RT, Feld R, et al. Prognostic factors in patients with resected stage I non–small cell lung cancer. A report from the Lung Cancer Study Group. Cancer 1984;54: 1802–13. Ferguson MK, Wang J, Hoffman PC, et al. Sex-associated differences in survival of patients undergoing resection for lung cancer. Ann Thorac Surg 2000;69:245–50. McDuffie HH, Klaassen DJ, Dosman JA. Female-male differences in patients with primary lung cancer. Cancer 1987;59: 1825–30.
INVITED COMMENTARY Lobectomy has become the standard surgical management for patients with stage I non-small cell lung cancer (NSCLC), based on a report from the Lung Cancer Study Group that demonstrated improved recurrence-free survival, when compared to patients that underwent limited resection [1]. However, there may be patients with compromised pulmonary function in whom the relative benefit of lobectomy versus limited resection is outweighed by the relative risk. Patients in whom the risk-benefit analysis may be particularly important include patients with a higher risk from surgery, such as patients with restrictive lung disease [2], or patients with small peripheral nodules, in whom the relative benefits of lobectomy may not be as significant [3]. The process of preoperative risk stratification, therefore, is central to the choice of procedure [4]. Iizasa and colleagues analyze the prognostic value of preoperative pulmonary function in a series of patients who are treated with lobectomy for stage I non-small cell © 2004 by The Society of Thoracic Surgeons Published by Elsevier Inc
lung cancer [5]. The potential for this type of risk analysis is considerable: as screening with spiral CT becomes more prevalent, more patients are detected with small peripheral lesions, and the debate regarding the appropriateness of lobectomy versus limited resection has been renewed. In this series, the authors investigate the prognostic value of age, sex, pulmonary function, and tumor size in terms of death from cancer-specific and other causes, in order to examine the role of lobectomy in patients with stage I NSCLC. The analysis of the appropriateness of lobectomy, versus “lesser” surgical procedures or non-invasive procedures, must consider the risks and benefits of the alternatives. This series analyzes the results of patients who underwent lobectomy and were found on final pathologic evaluation to have stage I (node-negative) disease. Thus, the use of limited resection in this selected population may have been appropriate; however, the status of being node-negative can only be determined 0003-4975/04/$30.00 doi:10.1016/j.athoracsur.2003.12.109
Preoperative Pulmonary Function as a Prognostic Factor for Stage I Non–Small Cell Lung Carcinoma Toshihiko Iizasa, MD, Makoto Suzuki, MD, Kazuhiro Yasufuku, MD, Akira Iyoda, MD, Mizuto Otsuji, MD, Shigetoshi Yoshida, MD, Yasuo Sekine, MD, Kiyoshi Shibuya, MD, Yukio Saitoh, MD, Kenzo Hiroshima, MD, and Takehiko Fujisawa, MD Departments of Thoracic Surgery and Basic Pathology, Graduate School of Medicine, Chiba University, Chiba, Japan
Background. The aim of this study was to clarify preoperative lung function as a prognostic factor for the long-term survival of, and to discuss the appropriateness of lobectomy for, patients with stage I non–small cell lung carcinoma who have poor preoperative pulmonary function. Methods. The study group consisted of 402 lobectomized patients with stage I non–small cell lung carcinoma treated by complete resection from 1985 to 1997. Preoperative percent forced vital capacity [(forced vital capacity/predicted forced vital capacity) ⴛ 100], FEV1% [(forced expiratory volume in 1 second/forced vital capacity) ⴛ 100], arterial carbon dioxide tension, and smoking were statistically analyzed as prognostic factors together with other host and tumor biologic factors. Results. Multivariate analysis demonstrated that tumor size (p < 0.0001) was the most significant prognostic factor for survival from primary lung cancer. Age (p < 0.0001), sex (p ⴝ 0.0036), and FEV1% (p ⴝ 0.0046) were
found to be independent prognostic factors for survival from death by nonprimary lung cancer–related causes. Smoking was highly correlated with FEV1% (correlation coefficient ⴝ ⴚ0.511; p < 0.0001). The 100 patients with a preoperative FEV1% less than 70% included 34 patients with nonprimary lung cancer–related deaths, whereas the 302 patients with an FEV1% of 70% or greater included only 23 patients (p < 0.0001). Conclusions. Along with tumor size, FEV1% is the most significant prognostic factor for patients with stage I non–small cell lung carcinoma with regard to survival from death by other causes. Lobectomy may not be preferred as an appropriate surgical modality for patients with stage I non–small cell lung carcinoma with small peripheral nodules who exhibit poor pulmonary function, especially lowered FEV1%.
A
divided by predicted FVC times 100 (%FVC), forced expiratory volume in 1 second (FEV1) divided by FVC times 100 (FEV1%), and arterial carbon dioxide tension (Paco2) were examined to gauge pulmonary disorder preoperatively. The present retrospective study is an attempt to clarify prognostic factors related to long-term survival in surgery cases of stage I NSCLC. We discuss appropriateness of lobectomy and important considerations for following up for patients with stage I NSCLC with poor pulmonary function.
s screening with spiral computed tomography becomes more prevalent, more patients with small peripheral lesions are being detected, and a great deal of debate regarding the appropriateness of lobectomy versus limited resection has been renewed. Lobectomy has been associated with a more positive prognosis for non– small cell lung carcinoma (NSCLC) than limited operations have been [1, 2]. The disadvantage of impaired lung function as a result of a lobectomy, however, makes it a controversial option [3– 8], especially for stage I patients who can expect long-term survival. Errett and colleagues [3] reported that wedge resection could be an alternative for poor-risk patients in stage I NSCLC. However, they could not determine the most important prognostic factors because of a univariate evaluation. Few reports have considered preoperative pulmonary function as a prognostic factor for the long term [9], because these functions have no relationship to the biologic features of NSCLC. We selected eight factors that may significantly influence prognosis for stage I NSCLC. These included the patient variables of age, sex, and smoking, and the tumor size and histologic type. Also, forced vital capacity (FVC)
Accepted for publication Oct 2, 2003. Address reprint requests to Dr Fujisawa, Department of Thoracic Surgery, Graduate School of Medicine, Chiba University, 1-8-1 Inohana, Chuo-ku, Chiba 260-8670, Japan; e-mail: [email protected].
© 2004 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2004;77:1896 –903) © 2004 by The Society of Thoracic Surgeons
Material and Methods Patients Between January 1985 and December 1997, 1,131 patients underwent surgical resection for primary lung cancer in our institute. Of those, 451 patients were confirmed as having stage I NSCLC after pathologic examination of the excised materials. For this study, 11 patients with incomplete resection, 14 patients with low-grade malignant type of NSCLC (11 carcinoids, 2 mucoepidermoid carcinomas, and 1 adenocystic carcinoma), 10 patients with a central lesion from a segmental bronchus, 20 patients with nonlobectomy (10 pneumonectomies, 7 wedge resections, 2 segmentectomies, and 1 main bronchus resection), and 5 patients who died of complications 0003-4975/04/$30.00 doi:10.1016/j.athoracsur.2003.10.014
within 1 month postoperatively were excluded from the survival evaluation. Some patients have overlapped on exclusion criteria; in all, 402 cases were used in this study. The histologic types of the tumors included 136 squamous cell carcinomas, 245 adenocarcinomas, 18 large cell carcinomas, and 3 adenosquamous carcinomas. Of the patients, 147 were women and 255 were men, ranging in age from 34 to 87 years with a mean age and standard deviation of 62.9 ⫾ 9.8 years. The advanced age group was defined as those patients 70 years of age or older. Tumors were classified by pathologic findings as 253 T1 (T1; ⱕ3 cm) and 149 T2 (T2; ⬎3 cm). Surgical procedures performed were 402 lobectomies, including 27 bilobectomies. Histologic type, pathologic stage, and TNM classification were identified according to the criteria of the American Joint Committee on Cancer [10] and the Classification of Lung Cancer description by the Japan Lung Cancer Society [11]. Large cell carcinomas and adenosquamous carcinomas were classified into the nonsquamous group together with adenocarcinomas for the univariate and multivariate studies. Preoperative %FVC, FEV1%, and Paco2 were adapted as clinical prognostic factors for pulmonary disorder. The patient’s history of cigarette smoking was evaluated according to the description on the patient’s chart. This was expressed as pack-years, calculated by the number of packs of cigarettes smoked per day multiplied by the number of years the subject smoked. For the first year after surgery, patients were followed up every month, then every 3 months from 2 to 5 years, and every 6 months thereafter. Additional follow-up and survival were determined by telephone or postal contact with the patient. When distant or local disease recurrence developed, any treatment was permitted.
Survival Rate and Statistical Analysis The end point was survival from the primary lung cancer–related deaths, the nonprimary lung cancer–related deaths, and overall survival in this study, calculated between the date of operation and the time of death. Cause of death was confirmed by telephone contact with the doctor who monitored the patient. Cause of death was classified as primary lung cancer–related if the patient had a recurrence of the primary lung cancer at the time of death (distant metastasis, local recurrence, or both). Nonprimary lung cancer–related death classification was determined in patients free of recurrence of the primary lung cancer at the time of death (eg, death as a result of respiratory disease, cardiac disease, brain vascular disease, second malignancy, or other). Overall survival and survival from the primary and nonprimary lung cancer–related deaths were calculated by the KaplanMeier method [12]. We have analyzed survival from the primary lung cancer by censoring patients with nonprimary lung cancer–related deaths at the date of death and vice versa. There is, therefore, no algebraic relationship between the overall survival rate and the survival rates in the two subgroups of primary lung cancer and nonprimary lung cancer. Because of this there is not an inconsistency between the overall 5-year survival of 76.3% and
IIZASA ET AL PREOPERATIVE PULMONARY FUNCTION IN STAGE I NSCLC
1897
the subgroup rates of 84.2% and 90.6%. Clinical prognostic factors examined in this study included age, sex, smoking status, histologic type, tumor size, preoperative %FVC, preoperative FEV1%, and preoperative Paco2. Each variable for survivals was analyzed in both strata and continuous using Cox proportional hazards. Multivariate analysis was calculated in a forward stepwise procedure, according to the Cox proportional hazards model [13], using the SPSS statistical software program package (SPSS version 11.0 for Windows, SPSS Inc, Chicago, IL). A 0.10 level of probability value was the significance level used for adding and deleting a covariable from the model. The 2 test was used for contingency table analysis in the case of two-by-two or threeby-two tables. The correlation of %FVC, FEV1%, Paco2, and pack-years was evaluated with the Spearman rank correlation coefficient. Statistical analysis was regarded as statistically significant at a p value less than 0.05.
Results Survival and Univariate Analysis Overall 5- and 10-year survival rates were 76.3% and 60.2%, respectively. This corresponded to 5- and 10-year survival rates from the primary lung cancer–related deaths of 84.2% and 74.9%, respectively, and from the nonprimary lung cancer–related deaths of 90.6% and 80.4%, respectively. For overall survival, each prognostic factor was determined as significant by univariate p value (age, tumor size, and FEV1%, p ⬍ 0.0001; sex, p ⫽ 0.0001; tumor histology, p ⫽ 0.0050; pack-year, p ⫽ 0.0012; %FVC, p ⫽ 0.4500; Paco2, p ⫽ 0.0275; Cox proportional hazards; Table 1). For survival rates from the primary lung cancer– related deaths, the p value for tumor size classification (T factor, p ⬍ 0.0001) was identified as significant with Cox proportional hazards (Table 2). For survival from the nonprimary lung cancer–related deaths, age (p ⬍ 0.0001), sex (p ⬍ 0.0001), tumor histology (p ⬍ 0.0001), pack-year (p ⬍ 0.0001), FEV1% (p ⬍ 0.0001), %FVC (p ⫽ 0.0016), and Paco2 (p ⫽ 0.0132) each displayed a significant difference with Cox proportional hazards (Table 3). We were able to demonstrate that survival rates from the primary lung cancer–related deaths were not significantly different with FEV1%, although differences in overall survival rates and survival rates from the nonprimary lung cancer– related deaths were statistically significant for this factor (p ⬍ 0.0001, Cox proportional hazards), which we divided into three groups based on mild chronic obstructive pulmonary disease in the European Respiratory Society [14] and the British Thoracic Society [15] (Fig 1).
Correlation Between Pulmonary Function and Smoking Pack-Years We also evaluated the rate of smoking for correlation with %FVC and FEV1% using a Spearman rank correlation coefficient. Our data disclosed a significant correlation between decreased FEV1% and increased smoking. The FEV1% correlated inversely with smoking pack-years (correlation coefficient ⫽ ⫺0.511; p ⬍ 0.0001; Fig 2).
GENERAL THORACIC
Ann Thorac Surg 2004;77:1896 –903
GENERAL THORACIC
1898
IIZASA ET AL PREOPERATIVE PULMONARY FUNCTION IN STAGE I NSCLC
Table 1. Univariate Analyses of Patient Characteristics and Overall Survival After Surgery for Stage I Non–Small Cell Lung Carcinoma
Characteristic Age ⱖ70 y ⬍70 y Sex Male Female Histology Sq Ad La Ad-sq Tumor size (T factor) T1 T2 Pack-years ⱖ30 ⬍30 FEV1% ⬍60% ⱖ60%, ⬍70% ⱖ70% %FVC ⬍70% ⱖ70%, ⬍80% ⱖ80% Paco2 ⬍40 mm Hg ⱖ40 mm Hg, ⬍45 mm Hg ⱖ5 mm Hg
No. of No. of Dead 5-Year 10-Year Univariate Cases Cases Survival Survival p Valuea 114 288
52 87
64.1 81.0
46.9 65.1
⬍0.0001
255 147
108 31
71.4 85.0
51.4 76.7
0.0001
136 245 18 3
59 73 7 0
68.9 81.1 64.4 100
49.2 65.7 64.4 100
0.0050b
253 149
70 69
84.4 62.6
67.2 48.7
⬍0.0001
212 190
98 41
68.0 85.7
47.9 75.0
0.0012
30 70 302
19 35 85
59.8 64.1 80.9
29.0 43.8 67.7
15 32 355
5 17 117
66.7 48.0 79.3
66.7 43.7 61.6
0.4500
157 208
70 58
67.5 83.0
51.9 66.3
0.0275
37
11
77.3
66.8
a Univariate p value with Cox proportional hazards. non-Sq.
Ann Thorac Surg 2004;77:1896 –903
tional hazards model using seven clinical prognostic factors (age, sex, histology, tumor size, FEV1%, %FVC, and Paco2). Smoking was excluded as a factor from these analyses because of its considerable statistical correlation with FEV1% and the fact that FEV1% is objectively measured, whereas smoking is self-reported. Multivariate analysis for overall survival revealed three independent prognostic factors: age (p ⫽ 0.0002), sex (p ⫽ 0.0065), and tumor size (p ⬍ 0.0001; Table 4). Multivariate analysis for survival from the primary lung cancer–related deaths demonstrated only tumor size (p ⬍ 0.0001) as an independent prognostic factor (Table 4). In addition, multivariate analysis for survival from the nonprimary lung cancer–related deaths demonstrated age (p ⬍ 0.0001), sex Table 2. Univariate Analyses of Patient Characteristics and Survival From the Primary Lung Cancer–Related Deaths After Surgery for Stage I Non–Small Cell Lung Carcinoma
Characteristic
⬍0.0001
b
Sq versus
Ad ⫽ adenocarcinoma; Ad-sq ⫽ adenosquamous carcinoma; FEV1% ⫽ percent forced expiratory volume in 1 second; %FVC ⫽ percent forced vital capacity; La ⫽ large cell carcinoma; Paco2 ⫽ arterial carbon dioxide partial pressure; Sq ⫽ squamous cell carcinoma; T1 ⫽ 3 cm or less; T2 ⫽ more than 3 cm.
Furthermore, the number of cases (85 of 100 cases) with FEV1% less than 70% for heavy smokers with 30 or more pack-years was significantly larger than that for smokers with 29 or fewer pack-years or nonsmokers, as compared with the number of cases (134 of 302) with FEV1% of 70% or greater for heavy smokers with 30 or more pack-years (p ⬍ 0.0001; 2 test). However, no correlation between %FVC and smoking or %FVC and FEV1% was found (data not shown).
Multivariate Analyses Multivariate analyses of overall survival and survival from the primary and nonprimary lung cancer–related deaths were performed according to the Cox propor-
Age ⱖ70 y ⬍70 y Sex Male Female Histology Sq Ad La Ad-sq Tumor size (T factor) T1 T2 Pack-years ⱖ30 ⬍30 FEV1% ⬍60% ⱖ60%, ⬍70% ⱖ70% %FVC ⬍70% ⱖ70%, ⬍80% ⱖ80% Paco2 ⬍40 mm Hg ⱖ40 mm Hg, ⬍45 mm Hg ⱖ45 mm Hg a
No. of No. of Dead 5-Year 10-Year Univariate Cases Cases Survival Survival p Valuea 114 288
25 56
78.6 86.2
71.6 76.3
0.1674
255 147
57 24
82.6 87.0
71.6 80.2
0.1022
136 245 18 3
24 54 3 0
84.7 84.1 81.0 100
75.8 73.9 81.0 100
0.7208b
253 149
36 45
91.4 71.6
81.3 63.9
⬍0.0001
212 190
49 32
80.4 88.2
70.7 79.3
0.4614
30 70 302
4 15 62
88.5 80.4 84.6
82.9 67.2 75.5
15 32 355
3 7 71
78.6 74.3 85.3
78.6 74.3 75.1
0.0986
157 208
36 38
79.4 87.2
73.4 76.3
0.4286
37
7
87.4
75.5
Univariate p value with Cox proportional hazards. non-Sq.
0.3732
b
Sq versus
Ad ⫽ adenocarcinoma; Ad-sq ⫽ adenosquamous carcinoma; FEV1% ⫽ percent forced expiratory volume in 1 second; %FVC ⫽ percent forced vital capacity; La ⫽ large cell carcinoma; Paco2 ⫽ arterial carbon dioxide partial pressure; Sq ⫽ squamous cell carcinoma; T1 ⫽ 3 cm or less; T2 ⫽ more than 3 cm.
IIZASA ET AL PREOPERATIVE PULMONARY FUNCTION IN STAGE I NSCLC
Table 3. Univariate Analyses of Patient Characteristics and Survival From the Nonprimary Lung Cancer–Related Deaths After Surgery for Stage I Non–Small Cell Lung Carcinoma
Characteristic Age ⱖ70 y ⬍70 y Sex Male Female Histology Sq Ad La Ad-sq Tumor size (T factor) T1 T2 Pack-years ⱖ30 ⬍30 FEV1% ⬍60% ⱖ60%, ⬍70% ⱖ70% %FVC ⬍70% ⱖ70%, ⬍80% ⱖ80% Paco2 ⬍40 mm Hg ⱖ40 mm Hg, ⬍45 mm Hg ⱖ45 mm Hg
No. of No. of Dead 5-Year 10-Year Univariate Cases Cases Survival Survival p Valuea 114 288
27 31
81.6 94.0
65.5 85.2
⬍0.0001
255 147
51 7
86.5 97.7
71.7 95.7
⬍0.0001
136 245 18 3
35 19 4 0
81.3 96.4 79.5 100
64.9 89.0 79.5 100
⬍0.0001b
253 149
34 24
92.3 87.5
82.7 76.2
0.0641
212 190
49 9
84.6 97.1
67.7 94.6
⬍0.0001
30 70 302
15 20 23
67.6 79.7 95.5
34.9 65.1 89.6
15 32 355
2 10 46
84.8 64.7 93.0
84.8 58.8 82.0
0.0016
157 208
34 20
85.0 95.2
70.8 86.9
0.0132
37
4
88.4
88.4
a Univariate p value with Cox proportional hazards. non-Sq.
1899
preoperative FEV1% of less than 70% included 35 who died of nonprimary lung cancer–related causes, whereas of 302 patients with an FEV1% of 70% or more, only 23 patients with nonprimary lung cancer–related deaths were observed (p ⬍ 0.0001, 2 test; Table 5). Thirty of 35 deaths in the group with an FEV1% of less than 70%, who were also heavy smokers (⬎30 pack-years), included 13 patients with respiratory insufficiency as a result of pneumonia or chronic obstructive pulmonary disease without primary lung cancer recurrence and 10 patients with secondary malignancy in other organs.
Comment
⬍0.0001
b
Sq versus
Ad ⫽ adenocarcinoma; Ad-sq ⫽ adenosquamous carcinoma; FEV1% ⫽ percent forced expiratory volume in 1 second; %FVC ⫽ percent forced vital capacity; La ⫽ large cell carcinoma; Paco2 ⫽ arterial carbon dioxide partial pressure; Sq ⫽ squamous cell carcinoma; T1 ⫽ 3 cm or less; T2 ⫽ more than 3 cm.
(p ⫽ 0.0036), and FEV1% (p ⫽ 0.0046) as independent prognostic factors (Table 4).
Type of Recurrence and Cause of Death Of the 402 patients analyzed for survival, 139 died of a cause other than surgical complications. Eighty-one patients died of recurrent disease, and 58 died of nonrecurring diseases. Local recurrence was observed in 6 patients, distant metastases in 71, and both in 4 patients. Nonrecurring causes included 17 respiratory diseases, 7 cardiac diseases, 6 brain vascular diseases, and 17 second primary malignancies (Table 5). We further evaluated the cause in nonprimary lung cancer–related deaths according to their preoperative FEV1%. The 100 patients with a
In this study, we retrospectively investigated clinical characteristics, patterns of disease recurrence, and causes of death in surgical cases of stage I NSCLC and evaluated the statistical significance of prognostic factors. We made several noteworthy findings. First, preoperative impaired FEV1% as well as tumor size could be important prognostic factors of poor outcome in stage I NSCLC, as suggested by an increase in the number of nonprimary lung cancer–related deaths in this study. Second, a marked correlation between increased smoking and a decline in FEV1% was found in patients with stage I NSCLC. Patients with poor pulmonary function have been noted to be at a disadvantage in perioperative management. There are many reports in which the risk of postoperative complications correlated with low-grade lung function could be predicted with diverse variables [16 –18]. Most reports have discussed the risk of complications and mortality for poorly reserving lung function in a perioperative period, but few evaluate and describe the adequacy of surgical procedures for patients with poor pulmonary function from the standpoint of longterm survival [9]. In the last two decades, cure of surgical cases of NSCLC has focused on non– cancer-related death for patients with poor preoperative pulmonary function, especially those in stage I cases who can expect long-term survival and tend to have been overlooked. Many previous reports regarding prognostic factors of stage I NSCLC demonstrated that overall 5-year survival rate still remained about 60% to 80% in stage IA and IB together [19 –21], which is still worse as compared with that of other organ carcinomas [22–24]. Harpole [2] has reported that 79 patients died of other causes without any evidence of cancer, and 105 patients had died of cancer in 289 resected cases with stage I NSCLC. He demonstrated no significant difference between the numbers of primary and nonprimary lung cancer–related deaths in those resected cases with stage I NSCLC, although he did not examine and discuss the cause of nonprimary lung cancer–related deaths in detail [2]. Here we report that the cause of nonprimary lung cancer–related deaths is related to poorer FEV1% and smoking status. Death from pulmonary causes is also suspected to originate from preoperative impaired FEV1% and smoking status. Smoking would also increase the possibility of developing a second malignancy [20, 25].
GENERAL THORACIC
Ann Thorac Surg 2004;77:1896 –903
GENERAL THORACIC
1900
IIZASA ET AL PREOPERATIVE PULMONARY FUNCTION IN STAGE I NSCLC
Ann Thorac Surg 2004;77:1896 –903
Fig 1. The survival curves of the patient groups with stage I non–small cell lung cancer correlated with forced expiratory volume in 1 second divided by forced vital capacity, times 100 (FEV1%), for overall (A), primary lung cancer (B), and nonprimary lung cancer (C). (A) Overall 5and 10-year survival rates were 59.8% and 29.0%, respectively, for patients with FEV1% less than 60%; 64.1% and 43.8%, respectively, for the FEV1% 60% to 69.9% group; and 80.9%, and 67.7%, respectively, for the FEV1% 70% or greater group. (B) The 5- and 10-year survival rates from the primary lung cancer–related deaths were 88.5% and 82.9%, respectively, for the FEV1% less than 60% group; 80.4% and 67.2%, respectively, for the FEV1% 60% to 69.9% group; and 84.6%, and 75.5%, respectively, for the FEV1% 70% or greater group. (C) The 5and 10-year survival rates from the nonprimary lung cancer–related deaths were 67.6% and 34.9%, respectively, for the FEV1% less than 60% group; 79.7% and 65.1%, respectively, for the FEV1% 60% to 69.9% group; and 95.5% and 89.6%, respectively, for the FEV1% 70% or greater group.
Influences of smoking on pulmonary function have been reported since the 1960s. Dockery and associates [26] studied the correlation between long-term smoking and lung function, demonstrating that FVC and FEV1% decreased irreversibly and were inversely proportional to the amount smoked in pack-years. Our data, obtained using the Spearman rank correlation, also revealed a significant correlation between decreased FEV1% and increased smoking. Miyazawa and coworkers [27] reported that the condition of the pulmonary vascular bed after major lung resection does not improve, even in the late postoperative phase. We expect that in patients who smoke, the already impaired FEV1% might drop even further in the postoperative period compared with normal FEV1% levels. Furthermore, these patients may suffer greater morbidity related to pulmonary causes as a result of their poor pulmonary function.
In stage I cases, tumor size and histologic type are the most important factors used in determining the prognosis of stage I NSCLC for tumor progress and characteristics. A T1 designation suggested a better prognosis than T2 for survival from the primary lung cancer–related deaths in our univariate and multivariate analyses, in agreement with other reports [2, 20, 28 –30]. Most reports disclosed no significant difference in prognosis between squamous cell carcinoma and nonsquamous cell carcinoma in patients with stage I NSCLC [2, 20, 28 –30]. On the other hand, some reports using univariate analysis have noted that for T1 N0 M0 NSCLC, the squamous cell carcinoma cases have better overall survival rates than the nonsquamous cases [5]. Additionally, disease recurrence in nonsquamous cell carcinoma has been found to be more frequent than in squamous cell carcinoma in patients with T1 N0 NSCLC [19]. Our results are consis-
IIZASA ET AL PREOPERATIVE PULMONARY FUNCTION IN STAGE I NSCLC
1901
Table 5. Comparisons of Cause of Deaths in Nonprimary Lung Cancer in Patients With FEV1% ⬍ 70% Versus Those With ⱖ70%a Cause of Death
FEV1% ⬍ 70% (n ⫽ 100)
FEV1% ⱖ 70% (n ⫽ 302)
13 4 1 10 7 35
4 3 5 7 4 23b
Respiratory disease Cardiac disease Brain vascular disease Second malignancy Other Total a
Values are number of patients.
b
p value ⬍ 0.0001 with 2 test.
FEV1% ⫽ percent forced expiratory volume in 1 second.
Fig 2. Correlation between pulmonary function and smoking in pack-years. Correlation of smoking with forced expiratory volume in 1 second divided by forced vital capacity, times 100 (FEV1%), was detected at a correlation coefficient of ⫺0.515, p less than 0.0001 with Spearman rank correlation coefficient. The higher the smoking rate in pack-years, the lower the FEV1% was found to be.
tent with the former reports but contradict the latter reports. This discrepancy may be partially explained by the fact that our study included many long-term surviving patients who were identified through a health examination for elderly persons in Japan as having small adenocarcinoma. Additionally, the authors of the later reports adapted a univariate analysis that disregarded the involvement of other prognostic factors. We believe the histologic type of primary NSCLC was not regarded as a significant prognostic factor. We selected patient age and sex as primary factors that will affect survival. There are many disadvantageous Table 4. Multivariate Analyses of Prognostic Factors and Postsurgical Survival Rates Using the Cox Proportional Hazards Model Relative Riska
Variable Overall Age Sex (male vs female) Tumor size FEV1% Primary cancer related Tumor size Nonprimary lung cancer related Age Sex (female vs male) FEV1%
95% CI
p Value
1.040 1.807 1.201 0.985
1.061–1.190 0.0002 1.180 –2.786 0.0065 1.097–1.315 ⬍0.0001 0.967–1.003 0.0998
1.341
1.206 –1.490 ⬍0.0001
1.095 3.428 0.946
1.056 –1.135 ⬍0.0001 1.496 –7.857 0.0036 0.940 – 0.989 0.0046
a Relative risk of continuous variables (age, tumor size, and FEV1%) were ratios of larger value to smaller.
CI ⫽ Confidence interval; in 1 second.
FEV1% ⫽ percent forced expiratory volume
factors in patients 70 years of age or older. In our univariate analysis, overall survival rates in the elderly male group were significantly lower than those for the younger, female group, agreeing with other reports [31, 32]. However, survival from the primary lung cancer– related deaths was not significantly different between these groups when using multivariate analysis. Nonprimary lung cancer–related deaths were associated with a poorer prognosis in the overall survival of the elderly male group by multivariate analysis. Elderly men commonly have a shorter life expectancy; thus, age and sex are not prognostic factors for survival from the primary lung cancer–related deaths in stage I NSCLC, according to the Cox proportional hazard model. In conclusion, we demonstrated that non–lung cancer– related death, in particular pulmonary failure and second primary malignant disease, is sufficiently significant so as to warrant attention in stage I NSCLC patients with impaired FEV1%. The FEV1% is one of the most significant prognostic factors in stage I NSCLC patients, as is the tumor size. The FEV1% has been highly correlated with past smoking status. Damage to postoperative pulmonary function could not be disregarded in patients with poor FEV1% in stage I NSCLC and worsened their prognosis, as indicated by the increased number of non–lung cancer–related deaths. Lobectomy may not be preferred as an appropriate surgical modality. Pulmonary disease and second malignancy for any stage I NSCLC patient who has poor FEV1% should be carefully monitored so that postoperative lung function might be preserved.
References 1. Ginsberg RJ, Rubinstein LV. Randomized trial of lobectomy versus limited resection for T1 N0 non–small cell lung cancer. Lung Cancer Study Group. Ann Thorac Surg 1995; 60:615–23. 2. Harpole DH. Stage I non–small cell lung cancer. Cancer 1995;76:787–96. 3. Errett LE, Wilson J, Chiu RC, et al. Wedge resection as an alternative procedure for peripheral bronchogenic carcinoma in poor-risk patients. J Thorac Cardiovasc Surg 1985; 90:656 –61. 4. Bennet WF, Smith RA. Segmental resection for bronchogenic carcinoma: a surgical alternative for the compromised patient. Ann Thorac Surg 1979;27:169 –72.
GENERAL THORACIC
Ann Thorac Surg 2004;77:1896 –903
GENERAL THORACIC
1902
IIZASA ET AL PREOPERATIVE PULMONARY FUNCTION IN STAGE I NSCLC
5. Read RC, Yoder G, Schaeffer RC. Survival after conservative resection for T1 N0 M0 non–small cell lung cancer. Ann Thorac Surg 1990;49:391–8. 6. Miller JI, Hatcher CR Jr. Limited resection of bronchogenic carcinoma in the patient with marked impairment of pulmonary function. Ann Thorac Surg 1987;44:340 –3. 7. Hoffmann TH, Ransdell HT. Comparison of lobectomy and wedge resection for carcinoma of the lung. J Thorac Cardiovasc Surg 1980;79:211–7. 8. Pastorino U, Valente M, Bedini V, et al. Limited resection for stage I lung cancer. Eur J Surg Oncol 1991;17:42–6. 9. Ploeg AJ, Kappetein AP, van Tongeren RB, et al. Factors associated with perioperative complications and long-term results after pulmonary resection for primary carcinoma of the lung. Eur J Cardiothorac Surg 2003;23:26 –9. 10. Mountain CF. Revision in the international system for staging lung cancer. Chest 1997;111:1710 –7. 11. The Japan Lung Cancer Society. Classification of lung cancer. 5th ed. Tokyo: Kanehara, 2000. 12. Kaplan EL, Meier P. Nonparametric estimation from incomplete observation. J Am Stat Assoc 1958;53:457–81. 13. Cox DR. Regression models and life tables. J Royal Stat Soc 1972;B34:187–220. 14. Siafakas NM, Vermeire P, Pride NB, et al. Optimal assessment and management of chronic obstructive pulmonary disease (COPD). The European Respiratory Society Task Force. Eur Respir J 1995;8:1398 –420. 15. BTS guidelines for the management of chronic obstructive pulmonary disease: The COPD Guidelines Group of the Standards of Care Committee of the BTS. Thorax 1997; 52(Suppl 5):S1–28. 16. Ferguson MK, Little L, Rizzo L, et al. Diffusing capacity predicts morbidity and mortality after pulmonary resection. J Thorac Cardiovasc Surg 1988;96:894 –900. 17. Kearney DJ, Lee TH, Reilly JJ, et al. Assessment of operative risk in patients undergoing lung resection: importance of predicted pulmonary function. Chest 1994;105:753–9. 18. Pierce RJ, Copland JM, Sharpe K, et al. Preoperative risk evaluation for lung cancer resection: predicted postoperative product as a predictor of surgical mortality. Am J Respir Crit Care Med 1994;150:947–55. 19. Thomas P, Rubinstein L. Cancer recurrence after resection: T1 N0 non–small cell lung cancer. Lung Cancer Study Group. Ann Thorac Surg 1990;49:242–6. 20. Fujisawa T, Iizasa T, Saitoh Y, et al. Smoking before surgery
Ann Thorac Surg 2004;77:1896 –903
21. 22.
23.
24.
25.
26. 27. 28. 29. 30.
31. 32.
predicts poor long-term survival in patients with stage I non–small-cell lung carcinomas. J Clin Oncol 1999;17:2086 – 91. Dominioni L, Imperatori A, Rovera F, et al. Stage I nonsmall cell lung carcinoma: analysis of survival and implications for screening. Cancer 2000;89:2334 –44. van der Hage JA, van de Velde CJ, Julien JP, et al. Improved survival after one course of perioperative chemotherapy in early breast cancer patients: long-term results from the European Organization for Research and Treatment of Cancer (EORTC) Trial 10854. Eur J Cancer 2001;37:2184 –93. Lacy AM, Garcia-Valdecasas JC, Delgado S, et al. Laparoscopy-assisted colectomy versus open colectomy for treatment of non-metastatic colon cancer: a randomised trial. Lancet 2002;359:2224 –9. Barchielli A, Amorosi A, Balzi D, et al. Long-term prognosis of gastric cancer in a European country: a population-based study in Florence (Italy). 10-year survival of cases diagnosed in 1985–1987. Eur J Cancer 2001;37:1674 –80. Thomas PA Jr, Rubinstein L. Malignant disease appearing late after operation for T1 N0 non–small-cell lung cancer. The Lung Cancer Study Group. J Thorac Cardiovasc Surg 1993;106:1053–8. Dockery DW, Speizer FE, Ferris BG Jr, et al. Cumulative and reversible effects of lifetime smoking on simple tests of lung function in adults. Am Rev Respir Dis 1988;137:286 –92. Miyazawa M, Haniuda M, Nishimura H, et al. Longterm effects of pulmonary resection on cardiopulmonary function. J Am Coll Surg 1999;189:26 –33. Martini N, Bains MS, Burt ME, et al. Incidence of local recurrence and second primary tumors in resected stage I lung cancer. J Thorac Cardiovasc Surg 1995;109:120 –9. Williams DE, Pairolero PC, Davis CS, et al. Survival of patients surgically treated for stage I lung cancer. J Thorac Cardiovasc Surg 1981;82:70 –6. Gail MH, Eagan RT, Feld R, et al. Prognostic factors in patients with resected stage I non–small cell lung cancer. A report from the Lung Cancer Study Group. Cancer 1984;54: 1802–13. Ferguson MK, Wang J, Hoffman PC, et al. Sex-associated differences in survival of patients undergoing resection for lung cancer. Ann Thorac Surg 2000;69:245–50. McDuffie HH, Klaassen DJ, Dosman JA. Female-male differences in patients with primary lung cancer. Cancer 1987;59: 1825–30.
INVITED COMMENTARY Lobectomy has become the standard surgical management for patients with stage I non-small cell lung cancer (NSCLC), based on a report from the Lung Cancer Study Group that demonstrated improved recurrence-free survival, when compared to patients that underwent limited resection [1]. However, there may be patients with compromised pulmonary function in whom the relative benefit of lobectomy versus limited resection is outweighed by the relative risk. Patients in whom the risk-benefit analysis may be particularly important include patients with a higher risk from surgery, such as patients with restrictive lung disease [2], or patients with small peripheral nodules, in whom the relative benefits of lobectomy may not be as significant [3]. The process of preoperative risk stratification, therefore, is central to the choice of procedure [4]. Iizasa and colleagues analyze the prognostic value of preoperative pulmonary function in a series of patients who are treated with lobectomy for stage I non-small cell © 2004 by The Society of Thoracic Surgeons Published by Elsevier Inc
lung cancer [5]. The potential for this type of risk analysis is considerable: as screening with spiral CT becomes more prevalent, more patients are detected with small peripheral lesions, and the debate regarding the appropriateness of lobectomy versus limited resection has been renewed. In this series, the authors investigate the prognostic value of age, sex, pulmonary function, and tumor size in terms of death from cancer-specific and other causes, in order to examine the role of lobectomy in patients with stage I NSCLC. The analysis of the appropriateness of lobectomy, versus “lesser” surgical procedures or non-invasive procedures, must consider the risks and benefits of the alternatives. This series analyzes the results of patients who underwent lobectomy and were found on final pathologic evaluation to have stage I (node-negative) disease. Thus, the use of limited resection in this selected population may have been appropriate; however, the status of being node-negative can only be determined 0003-4975/04/$30.00 doi:10.1016/j.athoracsur.2003.12.109