您的位置： &
p27及Ki—67对鉴别良、恶性甲状腺滤泡状肿瘤的价值
优质期刊推荐百度拇指医生
&&&普通咨询
您的网络环境存在异常，
请输入验证码
验证码输入错误，请重新输入Ki-67 immunostaining and other prognostic factors including tobacco smoking in patients with resected nonsmall cell lung carcinoma - Shiba - 2000 - Cancer - Wiley Online Library
Advertisement
BACKGROUNDTo estimate the effectiveness of expression of the tumor proliferative marker Ki-67 antigen (Ki-67) as a postoperative prognostic marker, the authors analyzed Ki-67 expression and its correlation with postoperative survival and other clinicopathologic factors, including preoperative smoking habits, in patients with resected nonsmall cell lung carcinoma (NSCLC).METHODSA total of 156 patients with resected NSCLC at the study institution were investigated. Postoperative survival rates were estimated based on demographic and clinicopathologic factors, including Ki-67 expression and preoperative tobacco smoking habits.RESULTSThe overall postoperative 5-year survival rate in patients with high Ki-67 labeling indices (& 20%) was 39.6% compared with 67.7% in patients with low Ki-67 labeling indices. This finding was significant for all resected cases and for each pathologic disease stage (P & 0.05). The postoperative 5-year survival rate in patients with a history of heavy smoking (& 30 pack-years) was 47.6% compared with 62.5% for other patients (P = 0.027). This result was especially significant in patients with International Union Against Cancer Stage I disease and in patients with nonsquamous cell carcinoma (P & 0.03). The authors also observed a positive correlation between the Ki-67 labeling index and preoperative smoking habits (P = 0.0002). Multivariate analysis demonstrated that lymph node involvement, tumor differentiation, and Ki-67 labeling index were significant prognostic factors in NSCLC (P & 0.01).CONCLUSIONSTumor Ki-67 expression is a strong prognostic factor in NSCLC, especially adenocarcinoma. It may be hypothesized that tobacco mutagenicity may play a role in the growth and extension of NSCLC, which is one of the major impediments to postoperative survival in patients with a history of heavy smoking. Cancer 7&65. & 2000 American Cancer Society.Although lung carcinoma is the leading cause of cancer death throughout the industrialized world, few advances in treatment have been achieved in the last two decades.,
Surgical outcomes for nonsmall cell lung carcinoma (NSCLC) remain unsatisfactory. Approximately 40&50% of lung carcinoma patients believed to have limited tumor disease (Stage I or II disease) will develop a recurrence within 3 years after surgery, even though these patients undergo curative resection. To identify these at-risk patients, additional biologic prognostic parameters need to be established.Many biologic prognostic markers have been reported in NSCLC. Tumor cell cycle analysis has indicated that tumors with a higher proliferation rate showed more aggressive clinical behavior. Several immunostaining studies using antibodies that recognize the Ki-67 nuclear antigen, which is associated with cell proliferation,,
have provided a reliable method with which to evaluate tumor growth in many malignant tumors including those of the lung. Ki-67 expression also is a useful prognostic marker in NSCLC patients, especially those with early stage disease. Recently we reported that preoperative smoking habit is an important clinical postoperative prognostic factor in evaluating overall long term survival in patients with primary resected Stage I NSCLC. To our knowledge several studies have reported that genetic alterations frequently occur in the bronchial mucosa of chronic smokers at chromosomal sites of tumor suppresser genes that are related to tumor proliferative activity. The purpose of the current study was to investigate the prognostic significance of Ki-67 antigen expression and preoperative smoking habits, and associations between the two factors, in a group of well characterized, surgically treated NSCLC patients from one institution. We analyzed Ki-67 antigen expression in surgically resected specimens of NSCLC by immunohistochemical staining. To examine the effectiveness of and associations between the two factors, we compared the results with clinicopathologic factors related to tumor invasion and extension, as well as postoperative survival, in patients with NSCLC.MATERIALS AND METHODSPatientsTumor specimens resected from a series of unselected patients with primary NSCLC treated at our institute from September 1989 to January 1992 were used in this study. All specimens were first primary tumors and no recurrent cases were included. The Karnofsky performance status of these patients were in the range of 0&1 and the patients were considered to have operable cardiopulmonary function by routine electrocardiography and pulmonary function testing using radiospirometry. All patients were evaluated using conventional staging procedures preoperatively and were found to have operable primary tumors (TNM Stage of & T3N2M0). TNM stage was classified according to the International Union Against Cancer staging system. Informed consent was obtained from all patients or their relatives. In the current series, we used those NSCLC specimens that were amenable to immunohistochemical analysis, excluding pathologic Stage IV tumors or low grade malignancies. Consequently, we examined 156 patients whose median age at the time of surgery was 62.4 years (range, 34&80 years); 112 patients were male and 44 patients were female. Histologic types diagnosed after surgery were 92 adenocarcinomas (17 well differentiated, 53 moderately differentiated, and 22 poorly differentiated), 60 squamous cell carcinomas (1 well differentiated, 40 moderately differentiated, and 19 poorly differentiated), 2 large cell carcinomas, and 2 adenosquamous carcinomas. Of the 156 patients, 80 were classified postoperatively as having Stage I disease, 38 as having Stage II disease, and 38 as having Stage III disease. Preoperative smoking habits were assessed according to the descriptions from patient's clinical charts. Forty patients were nonsmokers, 20 patients were smokers of & 30 pack-years, and 96 patients were heavy smokers (& 30 pack-years) (Table ).Table&1.&Relation between Ki-67 Expression and Clinicopathologic Features in Patients with Resected Lung CarcinomaNo. of patients81 (51.9%)75 (48.1%)Age (yrs) (mean & SD)63.3 & 8.561.4 & 9.0NSGender&Male71 (45.5)41 (26.3)&Female10 (6.4)34 (21.8)& 0.0001Histology&Adenocarcinoma31 (19.9)61 (39.1)&Squamous cell ca47 (30.1)13 (8.3)&Large cell ca1 (0.6)1 (0.6)&Adenosquamous ca2 (1.3)0& 0.0001Differentiation&Well018 (11.5)&Moderate48 (30.8)46 (29.5)&Poor or undifferentiated33 (21.2)11 (7.1)& 0.0001T classification&T117 (10.9)41 (26.3)&T238 (24.4)26 (16.7)&T320 (12.8)5 (3.2)&T46 (3.8)3 (1.9)& 0.0001Lymph node involvement&N041 (26.3)56 (35.9)&N118 (11.5)14 (9.0)&N220 (12.8)4 (2.6)&N32 (1.3)1 (0.6)& 0.0035Pathologic stage&I29 (18.6)51 (32.7)&II23 (14.7)15 (9.6)&III29 (18.6)9 (5.8)0.0001Preoperative smoking habit&Heavy smokers (≧ 30 &pack-years)62 (39.7)34 (21.8)&Nonsmokers and light &smokers19 (12.2)41 (26.3)& 0.0001Patients in the current series did not receive any anticancer chemotherapy or thoracic irradiation preoperatively or during the postoperative period until disease recurrence was confirmed by routine postoperative follow-up. Patients were followed monthly during the first year after surgery, every 3 months for 2&5 years postoperatively, and every 6 months thereafter. Telephone or postcard contact determined additional patient follow-up and survival. The cause of death was confirmed by phone contact with the physician who followed the patient.ImmunohistochemistryRepresentative tumor samples were obtained from the paraffin embedded tumor specimen that had been fixed by neutral formalin immediately after surgery. Tumor samples of 3&4 &m thickness were cut and mounted onto silan-coated glass slides. Ki-67 antigen levels were determined using the MIB-1 monoclonal antibody (MoAb) (Immunotech SA, Marseilles, France) with the slides stained according to the streptavidin-biotin staining procedure. Briefly, after deparaffinization, specimens were heated by microwave (4 times at 100 &C) in 0.01M citrate buffer solution. After cooling gradually to room temperature, the slides were incubated for 20 minutes in 0.3% hydrogen peroxide diluted in methanol. Slides then were washed with distilled water, incubated with normal rabbit serum to block nonspecific binding, and incubated with primary antibody, diluted to 1:100 in buffered solution, overnight at 4 &C. After incubation with the bridging antibody for 30 minutes, the slides then were incubated with the avidin-biotin complex for 30 minutes. Immunostaining was visualized by diaminobenzidine (stained for 8 minutes) and counterstained with hematoxylin.Evaluation of the SpecimensAll slides were evaluated without any knowledge of clinical outcomes or other clinicopathologic data. Microscopy was performed by counting & 1000 tumor cells in randomly selected high-power field (10 & 100) from different representative parts of the tumor. Typically, more than eight different areas of the tumor were counted. The median value of positive tumor cells was 25.2% in the current series. With reference to previous reports, the Ki-67 labeling index (LI) was defined as high if & 20% of the tumor nuclei stained positively with the MoAb, and was considered low if staining was & 20%.Statistical AnalysisThe demographic and clinicopathologic factors examined for statistical associations with postoperative survival were age, gender, histology, differentiation, T classification, lymph node involvement, preoperative smoking habits, and Ki-67 LI. These eight factors, including quantitative factors, were divided into two or three categories based on clinical accounts to yield the most significant results in univariate analysis. Associations between Ki-67 protein expression and the clinicopathologic factors were analyzed using the Student t test for unpaired data and the chi-square test. Postoperative survival rates in dichotomized factor categories of the variables were estimated using the method of Kaplan and Meier and survival distributions were compared between the dichotomized factor categories using the log rank test to determine influential prognostic factors. Using these factors as explanatory variables, the Cox proportional hazards model was applied to evaluate joint influences on postoperative survival. Statistical analyses were performed using the StatView statistical software (Abacus Concepts, Inc., Berkeley, CA). The level of significance was set at 0.05.RESULTSExpression of Ki-67 in NSCLCThe relation between Ki-67 expression and clinicopathologic features in the current series of resected NSCLCs is shown in Table . High LIs (& 20%) were observed in 81 of 156 tumors (51.9%). Among the 81 high LI tumors, 30 showed indices of & 40%. There was no statistically significant difference in age between the high LI and low LI groups. However, tumors in male patients showed high LIs more frequently than tumors in female patients. With regard to histologic type, squamous cell carcinomas showed high LIs in 78% of tumors, whereas a high LI was observed in only 34% of adenocarcinomas. With regard to histologic differentiation, poorly differentiated tumors showed high Ki-67 LIs, especially in adenocarcinomas. These differences all were statistically significant (P & 0.0001). Comparison of pathologic TNM stage demonstrated that T1 or N0 tumors showed low Ki-67 LIs more frequently than tumors of other grades, and that early stage tumors more frequently showed lower Ki-67 LIs compared with more advanced stage tumors. There was a significant correlation between T classification, N classification, TNM stage, and Ki-67 LI. Comparison of preoperative smoking habits and Ki-67 LI demonstrated that tumors with higher LIs were more frequent in heavy smokers. We observed a statistically significant correlation between the Ki-67 LI and preoperative smoking habits.The relation between preoperative smoking habits and clinicopathologic features in the current series of patients with resected NSCLC also was examined (Table ). Heavy smokers were more frequently male and had squamous cell carcinoma with poorly differentiated tumors and T2-4 tumors (P & 0.05). However, associations between preoperative smoking habits and N classification or pathologic stage were not observed. We also analyzed distribution of Ki-67 LIs for each patient characteristic stratified by preoperative smoking habit (Table ). In heavy smokers, tumors with high LIs were more frequent in older patients, in males, in patients with squamous cell carcinomas, in patients with poorly differentiated tumors, and in patients with advanced disease compared with the other characteristics. In nonsmokers or light smokers (those who smoked & 30 pack-years) tumors with low LIs were more frequent in females and in patients with nonsquamous cell carcinoma.Table&2.&Relation between Preoperative Smoking Habit and Clinicopathologic Features in Patients with Resected Lung CarcinomaNo. of patients96 (61.5%)60 (38.5%)Age (yrs) (mean & SD)63.7 & 7.860.3 & 9.8NSGender&Male92 (59.0)20 (12.8)&Female4 (2.6)40 (25.6)& 0.0001Histology&Adenocarcinoma43 (27.6)49 (31.4)&Squamous cell ca52 (33.3)8 (5.1)&Large cell ca1 (0.6)1 (0.6)&Adenosquamous ca02 (1.3)& 0.0001Differentiation&Well4 (2.6)14 (9.0)&Moderate58 (37.2)36 (23.1)&Poor/undifferentiated34 (21.8)10 (6.4)0.0003T classification&T127 (17.3)31 (19.9)&T244 (28.2)20 (12.8)&T320 (12.8)5 (3.2)&T45 (3.2)4 (2.6)0.0138Lymph node involvement&N060 (38.5)37 (23.7)&N119 (12.2)13 (8.3)&N216 (10.3)8 (5.1)&N31 (0.6)2 (1.3)0.7191Pathologic stage&I47 (30.1)33 (21.2)&II26 (16.7)15 (9.6)&III23 (14.7)12 (7.7)0.6992Table&3.&Distribution of Ki-67 Labeling Indices for Each Patient Characteristic Stratified by Preoperative Smoking HabitAge (yrs)&≧ 70 (n = 38)19 (50)8 (21)3 (8)8 (21)&≧ 60, & 70 (n = 67)26 (39)16 (24)10 (15)15 (22)&& 59 (n = 51)17 (33)10 (20)6 (12)18 (35)Gender&Male (n = 112)59 (53)33 (29)12 (11)8 (7)&Female (n = 44)3 (7)1 (2)7 (16)33 (75)Histologic type&Squamous cell ca (n = 60)41 (68)11 (18)6 (10)2 (3)&Nonsq. cell ca (n = 96)21 (22)23 (24)13 (14)39 (41)Differentiation&Well or moderate (n = 112)36 (32)26 (23)12 (11)38 (34)&Poor or undiff (n = 44)26 (59)8 (18)7 (16)3 (7)p stage&Stage I (n = 80)22 (28)25 (31)7 (9)26 (33)&Stage II and III (n = 76)40 (53)9 (12)12 (16)15 (20)Postoperative Survival of Patients and Prognostic FactorsAmong the 156 patients, 71 died within 5 years postoperatively. The cause of death was lung carcinoma recurrence in 58 patients, other cancer death in 3 patients, and other diseases in 10 patients. We excluded 3 surgical death patients who died within 30 days postoperatively in the current survival analysis. Estimated postoperative 5-year survival rates in the dichotomized categories for 8 factors and the results of the log rank test are shown in Table . The mean and median follow-up times for the 156 patients were 4.1 years (49 months) and 5.7 years (68.4 months), respectively. Statistically significant differences in survival were observed in T1 tumors versus T2-4 tumors for pathologic T classification (P & 0.0001), negative versus positive lymph node involvement (P & 0.0001), grade of tumor differentiation (well or moderate vs. poor or undifferentiated) (P & 0.0001), low versus high Ki-67 LI in resected tumor tissue (P = 0.0043), and heavy smokers who smoked & 30 pack-years versus others (P = 0.0272). We obtained P values & 0.05 for comparisons of survival distributions with the stratified categories of age (stratified into three groups) and histology type (dichotomized into squamous and nonsquamous cell carcinomas). The Kaplan&Meier survival curve with respect to the Ki-67 LI of tumor tissues revealed the low Ki-67 labeling group (& 20%) showed the highest postoperative survival compared with the other 2 this difference was more prominent in nonsquamous cell carcinoma (Fig. ). This trend also was significant in nonsquamous cell carcinoma stratified by pathologic stage. The low Ki-67 LI group showed a markedly higher postoperative survival rate in patients with advanced stage disease compared with patients with early stage tumors (Fig. ). Statistically significant differences in the Ki-67 LI also were observed between the heavy smokers and nonsmokers or light smokers. This result was more significant in Stage I NSCLC. (P = 0.006) (Fig. ). By histologic stratification, heavy smokers showed poorer postoperative survival curves compared with others in nonsquamous cell carcinoma patients (P = 0.024) but did not in squamous cell carcinoma patients (P = 0.26) (Fig. ).Table&4.&Postoperative 5-Year Survival Rates Using the Kaplan&Meier Method and Results of the Log Rank Test in All PatientsAge (yrs)&≧ 7037 (24.2)45.8%&≧ 60, & 7065 (42.5)50.5%&& 6051 (33.3)62.7%0.0674Gender&Male109 (71.2)49.3%&Female44 (28.8)63.3%0.0601Histologic type&Squamous cell ca58 (37.9)55.0%&Nonsquamous cell ca95 (62.1)52.3%0.777Differentiation&Well or moderate111 (72.5)65.5%&Poor42 (27.5)21.2%& 0.0001T classification&T158 (37.9)81.0%&T2&495 (62.1)36.2%& 0.0001Lymph node involvement&Negative96 (62.7)65.3%&Positive57 (37.3)33.2%& 0.0001Ki-67 LI&Low (LI & 20.0%)75 (49.0)67.7%&High (LI &#x%)78 (51.0)39.6%0.0043Preoperative smoking habit&Nonsmokers or smokers && 30 pack-years59 (38.6)62.5%&Heavy smokers (≧ 30 &pack-years)94 (61.4)47.6%0.0272Figure&1. Overall postoperative (postop) survival stratified by Ki-67 labeling index for (a) all cases and (b) nonsquamous cell carcinoma (ca) cases.Figure&2. Overall postoperative (postop) survival for nonsquamous cell carcinoma (ca) cases stratified by Ki-67 labeling index for (a) Stage I cases and (b) Stage II and III cases.Figure&3. Over-all postoperative (postop) survival for nonsquamous cell carcinoma cases stratified by preoperative smoking habit for (a) all cases and (b) Stage I cases.Figure&4. Overall postoperative (postop) survival for nonsquamous cell carcinoma (ca) cases stratified by preoperative smoking habits for (a) nonsquamous cell carcinoma cases and (b) squamous cell carcinoma cases.Smoking cessation after surgery also was investigated in 70 Stage I patients. Of these, there were 36 patients considered to be heavy smokers in the preoperative period. Among these patients, 28 stopped smoking and 8 patients began smoking again after surgery. There were 33 patients considered to be light or nonsmokers and no patients began smoking after surgery. The postoperative 5-year survival rate in the 28 patients who stopped smoking was 67.7% and was 62.5% in the 8 patients who began smoking again after surgery. No significant difference in survival was observed by the log rank test (P = 0.3433).Significant Prognostic FactorsSeven factors with P values & 0.1 (according to the log rank test) were used as explanatory variables in the Cox proportional hazards regression model and were applied to evaluate joint influences on survival. For all patients, statistically significant factors influencing survival were lymph node involvement, tumor differentiation, and T classification. In the 95 nonsquamous cell carcinoma cases, the regression coefficients of lymph node involvement, tumor differentiation, and Ki-67 LI were statistically significant (P values of 0.6, and 0.0087, respectively) (Table ). The relative risk of cases with high Ki-67 LIs was 2.39 compared with cases with low LIs, and was 2.0 for heavy smokers compared with nonsmokers or light smokers in nonsquamous cell carcinoma cases. Log cumulative hazards curves for each stratified category of the seven factors were calculated and confirmed the assumptions of proportions of the proportional hazards by inspection.Table&5.&Relative Risks and 95% CI for Each Variable in Nonsquamous Cell Carcinoma Using the Cox Proportional Hazards Regression Model (n = 95)Lymph node involvement3.4001.830&6.3610.0001Differentiation3.2811.664&6.4670.0006Ki-67 LI2.3931.246&4.5940.0087Smoking habit2.0030.799&5.0610.1384Age1.9220.900&4.1050.1199Gender1.8600.679&5.0940.2273T classification1.5570.794&3.0560.1977We also analyzed prognostic factors using disease specific survival in multivariate analysis. However, we were unable to find any remarkable differences between overall and disease specific survival in this series.DISCUSSIONRecent progress in tumor cell biology has led to the description of various biologic markers and genes connected to cell proliferation or tumor progression, and the analysis of these markers as useful postoperative prognostic factors in patients with NSCLC. Examples include DNA ploidy,ras mutation and expression, p53 mutation, increased c-erbB-2 expression,bcl-2 expression, and other cell proliferative factors. Tumor proliferation rates have been considered predictive of clinical course and studies using flow cytometry, proliferating cell nuclear antigen (PCNA), or bromodeoxyuridine (BrdU) uptake have indicated a correlation between high proliferation rates and tumor agressiveness.The Ki-67 nuclear antigen is associated with cell proliferation and found in the nucleus of cycling (G1, S-, G2, and M-phases) cells but is absent in resting (G0 phase) cells. It is believed that biologically active tumors express high levels of Ki-67 nuclear antigen.,
We selected the MIB-1 MoAb to detect Ki-67 antigen expression because it allowed stable immunodetection even in the paraffin embedded specimens and it was used to calculate Ki-67 LIs in a series of tumor specimens. In the current study, the LIs of MIB-1 were higher in squamous cell carcinoma samples than in adenocarcinoma specimens, whereas within the adenocarcinoma group, the poorly differentiated tumors tended to show higher LIs compared with the more differentiated tumors. These results are in agreement with previous studies using BrdU labeling of S-phase cells or PCNA in G1&S-phase cells.,
Growth fraction analysis using MIB-1 indicated lower postoperative survival rates in the group with higher LIs (& 20%) for all patients, adenocarcinoma patients, Stage I patients, and Stage IIIA patients. By multivariate analysis, the Ki-67 LI using MIB-1 was a significant prognostic factor in nonsquamous cell carcinoma. These results suggest that MIB-1 is a reliable prognostic marker in postoperative NSCLC patients, especially those with adenocarcinoma.Several investigators have studied the role of Ki-67 expression in predicting the surgical outcome of patients with lung carcinoma. Scagliotti et al. reported that disease free survival was significantly lower in patients with resectable NSCLC with high Ki-67 LIs (& 25% positive cells). Harpole et al. also showed that a higher Ki-67 proliferation index (& 5%) was a significant predictor of early recurrence and death from disease in patients with Stage I NSCLC by univariate but not multivariate analysis. Tungekar et al. demonstrated initial differences in survival between 3 grades of Ki-67 LIs (0&10%, 11&40%, and & 40%), but after 5 years these differences largely had disappeared. Because a significant correlation in Ki-67 LIs between preoperative biopsy and postoperative surgically resected specimens also has been reported, our method could be useful in preoperative evaluation. When deciding whether to perform surgery, the Ki-67 LI can be used along with the N classification, especially in nonsquamous cell carcinoma patients.In the current study we also demonstrated by univariate analysis that a patient's preoperative smoking habit had a significant influence on postoperative long term survival in NSCLC patients. The discrepancy found between the univariate and multivariate analyses indicates that preoperative smoking habits had some relevance to the other strong prognostic factors demonstrated (to a certain degree) in multivariate analysis. Several studies have reported smoking to be one of the major causes of lung carcinoma. The majority of lung carcinoma patients who undergo surgical resection had a smoking habit before surgery and a correlation between this smoking habit and postoperative complications also has been reported. However, to our knowledge few studies have reported a preoperative smoking habit to be an influential postoperative prognostic factor in patients with lung carcinoma patients. Johnston-Early et al. reported that nonsmokers and former smokers with small cell carcinoma had a better prognosis, whereas Sobue et al. reported that nonsmokers and former smokers with surgically resected adenocarcinoma had a better prognosis. However, with lung carcinoma patients as a whole, the majority of reports investigating smoking history were unable to show that it had any significant influence on patient survival. Although we could not demonstrate a statistically significant difference using multivariate analysis, univariate analysis of preoperative smoking habits suggested an influence on the survival of patients with NSCLC, especially in those with early stage disease and those with adenocarcinoma. Furthermore, a significant association between the preoperative smoking habits of the patients and the Ki-67 LI of the tumor also was demonstrated. To our knowledge the current study is the first to report that heavy smokers (those who smoked & 30 pack-years) developed tumors with high Ki-67 LIs. Many reports have been published showing an association between smoking and DNA alteration. Westra and Slebos reported that codon 12 of the K-ras oncogene may be a specific target of the mutagenic activity of tobacco smoke. Suzuki et al. reported that p53 mutation was associated closely with lifetime cigarette consumption. It also was reported that tumors with p53 gene alterations showed higher growth fraction or drug resistance.,
Thus, the current study data suggest that smoking may have some influence on the biologic behavior of NSCLC and that this may be due to genetic alterations in the tumor induced by the mutagenic activity of tobacco smoke.In the current study, the postoperative survival rate of patients who were heavy smokers (& 30 pack-years) was found to be significantly lower than that of nonsmokers or light smokers (& 30 pack-years), especially in patients with early stage disease or those with adenocarcinoma (P & 0.05). In patients with tumors showing a high Ki-67 LI (& 20%), poorer postoperative survival rates were observed in all surgically resected cases and for each pathologic stage, especially adenocarcinoma (P & 0.02). A positive correlation between immunohistochemical expression of Ki-67 and preoperative smoking habit was observed (P & 0.01). This correlation was especially high for males, those patients with squamous cell carcinomas, those patients with poorly differentiated tumors, and in patients with advanced disease. Because the results of the current study did not find a preoperative smoking habit to be a significant predictor of prognosis by multivariate analyses, we suggest that tobacco mutagenicity may be related to tumor growth and extension, which is one of the major impediments to postoperative survival in patients with a history of heavy smoking.
Advertisement