Upstaging by Vessel Invasion Improves the Pathology Staging System of Non- Small Cell Lung Cancer*

Transcription

1 CHEST Original Research Upstaging by Vessel Invasion Improves the Pathology Staging System of Non- Small Cell Lung Cancer* Tomoshi Tsuchiya, MD, PhD; Satoshi Hashizume, MD; Shinji Akamine, MD, PhD; Masashi Muraoka, MD, PhD; Sumihisa Honda, PhD; Koichi Tsuji, MD, PhD; Shougo Urabe, MD, PhD; Tomayoshi Hayashi, MD, PhD; Naoya Yamasaki, MD, PhD; and Takeshi Nagayasu, MD, PhD LUNG CANCER Background: There is a need for a more complete classification system of lung cancer. To address this issue, we assessed whether the new staging could differentiate patients with early-stage cancers who have poorer prognosis and improve the unbalanced patient numbers with overlapping prognoses arising from the current TNM staging system. Methods: The study included 995 patients with pathology stages I and II non-small cell lung cancer (NSCLC) who underwent surgical resection at two institutions. We subclassified patients with stage IA and IB NSCLC based on the presence of vessel invasion (Vi). Stage IA Vi and stage IB non-vi were combined into new stage IB, as were stages IB Vi and IIA into new stage IIA. Results: The numbers of patients of stages IA, IB, IIA, and IIB were 477, 314, 55, and 149, and their 5-year survival rates were 86.0%, 66.2%, 60.7%, and 50.4%, respectively. Vi groups showed significantly poorer prognosis than non-vi groups at stage IA (p 0.011) and at stage IB (p 0.036). The numbers of patients of new stages IA, IB, and IIA were 333, 260, and 253, and their 5-year survival rates were 88.7%, 76.4%, and 61.2%, respectively. Regression analysis indicated that the new staging improved predictability of overall survival according to disease stage, and Akaike information criterion (3023.7) was significantly lower than that for current staging system (3032.5). Conclusion: Upstaging of Vi groups allows differentiation of patients with early-stage cancers with poor prognosis and improves the unbalanced numbers of patients and prediction of prognosis in cases of lung cancer. (CHEST 2007; 132: ) Key words: lung cancer; staging; vessel invasion Abbreviations: AIC Akaike information criteria; LBAC localized bronchioloalveolar carcinoma; NSCLC nonsmall cell lung cancer; UICC/AJCC International Union Against Cancer/American Joint Committee on Cancer; Vi vessel invasion Staging and classification of lung cancers are important for appropriate patient management and for estimating prognosis. In 1997, the International *From the Departments of Chest Surgery (Drs.Tsuchiya, Akamine, and Muraoka) and Pathology (Drs. Tsuji and Urabe), Oita Prefectural Hospital, Oita; Division of Surgical Oncology (Drs. Hashizume, Yamasaki, and Nagayasu), Department of Translational Medical Sciences, Department of Radiation Epidemiology, Atomic Bomb Disease Institute (Dr. Honda) and Nagasaki University Graduate School of Biomedical Sciences, Nagasaki City; and Department of Pathology (Dr. Hayashi), Nagasaki University Hospital, Nagasaki City, Japan. The authors have no conflicts of interest to disclose. Union Against Cancer/American Joint Committee on Cancer (UICC/AJCC) published the fifth edition of the TMN classification system for lung cancer. 1 Manuscript received August 9, 2006; revision accepted April 2, Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( org/misc/reprints.shtml). Correspondence to: Tomoshi Tsuchiya, MD, PhD, Division of Surgical Oncology, Department of Translational Medical Sciences, Nagasaki University Graduate School of Biomedical Sciences, Sakamoto, Nagasaki City, , Japan; tomoshi@nagasaki-u.ac.jp DOI: /chest Original Research

2 However, there is a need to establish a new and more complete international classification system based on recent trends in patient characteristics. One of these trends is the identification of subgroups of patients with small lung cancer who show good prognosis. Noguchi and colleagues 2 reported that histopathology types A (localized bronchioloalveolar carcinoma [LBAC]) and B (LBAC with structural collapse of alveoli) should be considered in situ peripheral carcinoma without distant metastasis. However, the overall 5-year survival of patients with histopathology types C (LBAC with foci of active fibroblastic proliferation) and D (poorly differentiated adenocarcinoma) were 74.8% and 52.4%, respectively, indicating poor prognosis in some patients with small lung cancer. Several reports 3 5 indicate that some patients with stage IA disease and poor prognosis might benefit from postoperative adjuvant chemotherapy. Based on this clinical fact, we believe that there is a need to subclassify patients with stage IA disease into more than one group. The second trend is the unbalanced patient numbers for each stage. The speed of diagnostic imaging and nationwide mass screening systems results in increased number of patients with early-stage non-small cell lung cancer (NSCLC). Watanabe 6 reported that stage IA lung cancer was diagnosed in 18.9% of patients in 1991, whereas 26.5% received the same staging in Takeda and colleagues 7 also reported that 50% of patients who underwent surgery in 1991 had stage I lung cancer, whereas the percentage climbed to nearly 70% in In contrast, stage IIA NSCLC was diagnosed in only 1 to 4.4%; this proportion was extremely small compared to other stages. 1,8 11 The third trend is the overlapping prognosis of patients with stage IB and stage IIA disease. The reported survival rates of patients with stage IB disease and stage IIA disease are 57% and 55%, respectively, 1 a result supported by the study of Inoue and colleagues. 10 Based on these results, we hypothesized that upstaging of poor prognostic groups from stage IA to stage IB and from stage IB to stage IIA might compensate for such trends. The objective of this study was to assess whether restaging of NSCLC patients with a poor prognosis can improve the current staging system. We used vessel invasion as a predominant predictor to subclassify the patients with stage IA and IB disease. Patients Materials and Methods This retrospective study included 995 patients with pathology stage IA, IB, IIA, and IIB NSCLC who underwent radical surgical resection at Oita Prefectural Hospital (institution 1) from February 1978 to March 2005 (n 597) and at Nagasaki University Hospital (institution 2) from August 1988 to March 2005 (n 398). The age of the patients ranged from 31 to 91 years (median, 68 years). The clinical records of these patients were reviewed to obtain routine demographic information and to confirm the pathology staging. Clinical and pathology diagnoses from 1978 to 1997 were reclassified according to the latest TMN staging system (fifth edition of UICC/AJCC, 1997). Surgical Procedure Location of the carcinoma was recorded at the preoperative evaluation and adjusted according to the results of surgery as necessary. In agreement with the institutional policy, lobectomies or pneumonectomies with systemic lymphadenectomy were typically performed according to the methods described by Naruke and colleagues. 12 For right-side tumors, the superior mediastinal compartment containing the trachea and superior vena cava from the level of the azygos vein to the right subclavian artery was dissected (nodes 1 to 4). The hilar lymph nodes around the mainstem bronchi and lymph nodes posterior to the trachea were also routinely dissected. Subcarinal, paraesophageal, and inferior pulmonary ligament lymph nodes (nodes 7 to 9) were also removed. For left-side tumors, the subaortic compartment, containing the left pulmonary artery, aortic arch, left recurrent laryngeal, and phrenic nerve (nodes 5 and 6), was completely dissected. Thereafter, the aortopulmonary ligament of Botalli was ligated and divided to facilitate dissection of paratracheal nodes (nodes 2 to 4). For tumors in the upper and middle lobes, the dissection of paraesophageal and inferior pulmonary ligament lymph nodes was omitted. Follow-up Follow-up studies were conducted by direct patient contact every 6 months for the first 3 years after the operation and every 12 months thereafter. Time in days from the date of operation to the date of death or last follow-up was recorded. Histopathology A sample of the primary tumor from each patient was fixed in 10% formalin and embedded in paraffin. Representative hematoxylin-eosin stained sections were reviewed. Staining of elastic and connective tissue was done if necessary. In all cases, a minimum of two slides of the original tumor were examined. The diagnosis of NSCLC and categorization of the cell type were made based on World Health Organization criteria (Lyon, France, 2004). 13 The pathology stage of each tumor was determined postoperatively according to the TNM classification system (fifth edition of UICC/AJCC, 1997). 1 Vessel invasion (Vi) was determined as described previously 14 by the identification of intravascular tumor clots in the lumen of blood or lymphatic vessels, which were often covered by endothelial cells. Blood vessels were identified by the presence of erythrocytes in the lumen, an endothelial cell lining, and the presence of elastic tissue around larger vessels. Lymphatic vessels were identified by exclusion; an endothelial cells-lined channel was considered a lymphatic vessel if the lumen was devoid of erythrocytes and contained few lymphocytes. To further help discrimination of blood vessels from lymphatic vessels, elastic-van Gieson stain was applied. In elastic-van Gieson-stained sections, the wall structure of a vein was clearly outlined by a black-stained elastic tissue and red-stained collagen (Fig 1, top left, A, and top right, B). A vein could be distinguished from a large lymphatic vessel; the latter contained rather a thin smooth muscle coat but had only a few CHEST / 132 / 1/ JULY,

3 Figure 1. Photomicrographs of blood Vi (top, A and B) and lymphatic Vi (bottom, C and D). Note that it is difficult to detect the lumen of the vessels by hematoxylin-eosin stain only (left, A and C). Elastic-van Gieson stain (original 100) shows thick black elastic fibrils at the lumen of blood vessels (top right, B). Only a few elastic fibrils with a red-stained smooth muscle coat were observed at the lumen of lymphatic vessels (bottom right, D). elastic fibrils (Fig 1, bottom right, D). Indeed, the elastic tissue in the wall of a vein was least destroyed by malignancy. At institution 2, to further support the diagnosis of microvessel invasion, immunohistochemical staining for CD34 or D2 40 was performed in some cases by the streptavidin-biotin method as described previously. 15,16 D2 40 is immunoreactive for lymphatic vessels only. 16 Two pathologists evaluated all histologic preparations at institution 2. At institution 1, all histologic preparations were evaluated by one pathologist, while another pathologist reviewed ambiguous slides and a consensus was reached. Reclassification of Stage IA and IB Disease for New Staging System To estimate the contribution of Vi to the prognosis of patients with stage IA and IB NSCLC, patients in each stage were regrouped according to the presence of such invasion (Vi group or non-vi group). Furthermore, we developed the following new stages: stage IA non-vi, which was defined as new stage IA; stage IA Vi and stage IB non-vi were combined as new stage IB; and stage IB Vi and stage IIA were combined as new stage IIA. Statistical Analysis Time to death was defined as the period from the date of surgery to death. An observation was censored at the last follow-up if the patient was alive or the patient had died of a cause other than NSCLC. Specific time-to-survival curves were plotted using the Kaplan-Meier method, while the log-rank test was used to assess the statistical significance of differences between groups. A two-tailed p value 0.05 was considered statistically significant. The appropriateness of the current and new staging systems was judged by regression analyses using a Cox proportional hazards model, and the two staging systems were compared using Akaike information criteria (AIC) statistics. 17 A smaller AIC score indicated a more appropriate model for predicting outcome. Statistical analysis was performed using statistical software (version 8.2; SAS Institute; Cary, NC). Results Table 1 shows the patient characteristics for each institution. The median observation period was longer at institution 1 than institution 2 (58.8 months vs 39.0 months). The frequency of adenocarcinoma was relatively lower and that of squamous cell carcinoma was relatively higher at institution 1 compared with institution 2. Patients with large cell neuroendocrine carcinoma were only observed at institution 2. The frequency of vessel invasion was higher at institution 2 than institution 1 (56.3% vs 45.7%). The current staging distribution is depicted in Table 2. The number of patients with stage IIA disease was extremely small compared to those with other stages at both institutions. Contrary to our expectation, the survival curves of patients with stage IIA and stage IIB disease almost overlapped at 172 Original Research

4 Table 1 Patient Characteristics* Characteristics Institution 1 Institution 2 Total Patients, No Median observation period, mo Age range, yr Median Gender, No. (%) Male 392 (65.7) 255 (64.1) 647 (65.0) Female 205 (34.3) 143 (35.9) 348 (35.0) Histology, No. (%) Adenocarcinoma 356 (59.6) 269 (67.6) 625 (62.8) Squamous cell carcinoma 297 (49.7) 95 (23.9) 392 (39.4) Adenosquamous cell carcinoma 7 (1.2) 9 (2.3) 16 (1.6) Large cell carcinoma 27 (4.5) 10 (2.5) 37 (3.7) Large cell neuroendocline carcinoma 0 15 (3.7) 15 (1.5) Vi 273 (45.7) 224 (56.3) 497 (49.9) *Institution 1, Oita Prefectural Hospital; institution 2, Nagasaki University Hospital. institution 1. However, at institution 2 the survival rate of patients with stage IIA cancers was close to that of patients with stage IB disease. The current staging distribution for all patients from two institutions showed that the overall 5-year survival rates of patients with stage IA, IB, IIA, and IIB disease were 86.0%, 66.2%, 60.7%, and 50.4%, respectively (Fig 2). Significant differences in survival were observed between patients with stage IA and IB disease (p 0.001). The differences in survival rates were not statistically significant between patients with stage IB and IIA disease (p 0.251) and stage IIA and IIB disease (p 0.598). We reclassified the patients in stage IA and IB groups according to the presence or absence of Vi to estimate the contribution of this factor to prognosis. The overall 5-year survival rates of all patients with stage IA non-vi, the stage IA Vi group, the stage IB non-vi group, and the stage IB Vi group were 88.7%, 78.7%, 73.6%, and 61.4%, respectively (Fig 3). As predicted, the invasive nature of the cancer significantly affected survival rates for patients with stage IA disease (p 0.011) and stage IB disease (p 0.036) [Fig 3]. In addition, there was no significant difference in the survival rates of patients with stage IA Vi disease and those with stage IB non-vi disease (p 0.467), and those of patients with stage IB Vi and stage IIA disease (p 0.251). Based on these results, we regrouped the patients based on our new staging system. In the system, stage IA non-vi was defined as new-stage IA. Stage IA Vi and stage IB non-vi were combined as new stage IB. Stage IB Vi and stage IIA were combined as new stage IIA. Figure 4 shows the new staging distribution for all patients from the two institutions. The 5-year survival rates of all patients with new Table 2 Distribution of Current and New Staging System, With Hazard Ratios and 95% Confidence Intervals for Death in Relation to Stage IA* Institution 1 (n 597) Institution 2 (n 398) Total (n 995) Stages Patients, No. 5-yr Survival, % HR (95% CI) p Value Patients, No. 5-yr Survival, % HR (95% CI) p Value Patients, No. 5-yr Survival, % HR (95% CI) p Value Current staging Stage IA Stage IB ( ) ( ) ( ) Stage IIA ( ) ( ) ( ) Stage IIB ( ) ( ) ( ) AIC 1, ,032.5 New staging New stage IA New stage IB ( ) ( ) ( ) New stage IIA ( ) ( ) ( ) New stage IIE ( ) ( ) ( ) AIC 1, ,023.7 *Institution 1, Oita Prefectural Hospital; institution 2, Nagasaki University Hospital; HR hazard ratio; CI confidence interval. CHEST / 132 / 1/ JULY,

5 Figure 2. Overall 5-year survival according to the current TNM staging system. *p 0.005, log-rank test. stage IA, IB, and IIA disease were 88.7%, 76.4%, and 61.2%, respectively (Table 2). Significant differences in survival were observed between patients with new stage IA and IB disease (p 0.001) and between patients with new stage IB and IIA disease (p 0.001). The survival rates of patients with new stage IIA and those with stage IIB disease tended to be different (p 0.063). Cox proportional hazards model analysis of overall survival of all patients from two institutions showed Figure 3. Overall 5-year survival of the stage IA non-vi group, stage IA Vi group, stage IB non-vi group, and stage IB Vi group, and patients with stage IIA NSCLC. *p and **p 0.036, log-rank test. 174 Original Research

6 Figure 4. Overall 5-year survival using the proposed new TNM staging system based on Vi. *p and **p 0.063, log-rank test. that the hazard risks of patients with new stage IB disease, new IIA disease, and new IIB disease were 2.09, 3.99, and 5.31 times higher than that of patients with new IA disease, respectively (Table 2). Although we could not detect statistical differences in survival rates between patients with new stage IIA and stage IIB (p 0.063), the overall 5-year survival rates and hazard ratios under the new system were more clearly divided than with the current staging system. Analysis of system suitability according to the AIC score also indicated a significant improvement in ability to predict survival using the new staging classification compared to the current system ( vs ). The patient numbers of each stage were more equally distributed at institution 2. However, the AIC score at institution 2 could not detect statistical improvement of the new staging system, possibly because of pathology procedure. Discussion This study aimed to establish a new staging system to resolve some of the problems with the current TNM staging system, in particular the imbalance in patient numbers among stages and overlapping of survival curves. To validate the appropriateness of this approach, we analyzed the combination data and reviewed the data of two institutions. Under the current staging system, the largest proportion of patients had stage IA disease, while the stage IIA group numbered the least (Table 1). Only 4.8% of patients at institution 1 and 6.5% of those at institution 2 in our cohort were categorized as stage IIA, which was consistent with the 1 to 4.4% reported in other surgical series. 1,8 11,18 The results of our study at institution 2 and several previous studies 1,10 indicated similar survival rates for patients with stage IB and stage IIA disease. Surprisingly, the 5-year survival of patients with current stage IIA and stage IB disease did not overlap at institution 1, although an overlap was apparent for those with stage IIA and IIB disease. A similar finding was also reported by Jassem and colleagues. 11 Because stage IIA was represented by only 29 patients at institution 1 and 26 patients at institution 2, it is difficult to verify the appropriateness of this modification. Subclassifying all patients of stage I disease by Vi demonstrated a significant difference in overall 5-year survival between patients of stage IA non-vi and stage IA Vi, but not between those of stage IA Vi and stage IB non-vi. In addition, the overall 5-year survival of patients of stage IB non-vi was significantly different from that of patients of stage IB Vi. Furthermore, the overall 5-year survival rates of CHEST / 132 / 1/ JULY,

7 patients of stage IB Vi group and those of stage IIA disease were similar. These findings indicate that stage IA Vi patients should be upstaged to stage IB, and those with stage IB Vi should be upstaged to stage IIA. We thus modified the current staging system by combining stage IA Vi and stage IB non-vi, as well as stage IB Vi and stage IIA. Predictably, the patient numbers distributed among stages were more balanced and survival curves of patients were divided at both institutions (Table 2). Furthermore, at institution 1, AIC statistical analysis suggested that our modified staging system is more suitable than the current system, and appeared to improve the current UICC/AJCC TNM staging system. However, at institution 2, AIC statistic did not show improvement of the new staging system because the survival rates of patients of stage IA and stage IB became similar under the new staging system. In this institution, pathologists scrutinized vessel invasion using immunohistochemistry, and the frequency of vessel invasion was 11 percentage points higher than that of institution 1 (Table 1). We speculate that extremely mild Vi such as a few cells in one capillary, which was hardly detected at institution 1, was identified at institution 2. Consequently, patients with stage IA Vi and stage IB Vi disease at institution 2 might include patients with good prognosis. Because there is no criterion for Vi in the current pathology staging system, each pathologist has own subjective diagnostic criteria including presence or absence of vessel invasion, which makes the impact of vessel invasion on survival uncertain. The early standardization of diagnostic criterion of vessel invasion is necessary. We propose that obvious vessel invasion should be adopted as vessel invasion for the new staging system. The subdivision of early-stage NSCLC according to prognostic factors, including smoking history, serum carcinoembryonic antigen level, extent of operation, tumor size, pleural involvement, and Vi has also been proposed. 4,19 23 Vi is one of the most reliable predictors, although some studies 4,14,23,24 showed that only lymphatic Vi was significantly related to decreased survival. However, other authors indicated that cells from small cancers metastasize preferentially via the blood vessels. Our preliminary study indicated that both blood Vi and lymphatic Vi were significant poor prognostic factors in patients with stage IA NSCLC (data not shown). Goldstein and colleagues 4 also demonstrated that the nonpulmonary metastasis-free survival advantage for patients without Vi was significant; 78% of patients were disease free at 5 years when Vi was absent, compared to only 35% when Vi was present. 4 Therefore, we adopted the diagnostic criteria of Vi, defined by a combination of lymphatic and blood Vi, in order to simplify the substaging. Several studies 14,23 have demonstrated a significant correlation between Vi and larger tumor size. However, we speculate that tumor size is not directly correlated with tumor aggressiveness, although Vi has been linked to tumor aggressiveness 29 and is crucial in the earliest stages of tumor metastasis in a variety of human cancers. 30,31 The literature therefore supports our proposal that Vi should be incorporated into an improved TNM staging system of lung cancer, as it is for cancer of the liver. Problem with the classification system still exist even with the modifications proposed by this study. The upgrading system by Vi only applies to pathology staging, and discrepancies will no doubt arise between clinical and pathology staging systems. However, we believe that it is more important to identify the poor prognostic group in patients with stage I NSCLC. Our previous study demonstrated that patients with stage IA NSCLC benefited significantly from postoperative oral uracil-tegafur chemotherapy when Vi was present (data not shown). We therefore consider that the upgrading of stage IA Vi patients to stage IB may expand the use of adjuvant chemotherapy to cover those of stage IA group with poor prognosis. In conclusion, Vi can be used as a predictor of prognosis for patients with stage IA and IB NSCLC. Upstaging of patients with Vi from stage IA to stage IB and from stage IB to stage IIA allows identification of patients with stage I NSCLC who have poor prognosis and compensates for unbalanced patient numbers and overlapped prognosis in the current staging system. References 1 Mountain CF. Revisions in the International System for Staging Lung Cancer. Chest 1997; 111: Noguchi M, Morikawa A, Kawasaki M, et al. Small adenocarcinoma of the lung: histologic characteristics and prognosis. Cancer 1995; 75: Reynolds RD, Pajak TF, Bateman JR, et al. Considerations in designing and analyzing surgical adjuvant study in resected stage I and II carcinoma of the lung. Cancer 1979; 44: Goldstein NS, Mani A, Chmielewski G, et al. Prognostic factors in T1N0M0 adenocarcinomas and bronchioloalveolar carcinomas of the lung. Am J Clin Pathol 1999; 112: Kato H, Ichinose Y, Ohta M, et al. A randomized trial of adjuvant chemotherapy with uracil-tegafur. N Engl J Med 2004; 350: Watanabe Y. Substaging of stage I: A commentary. Lung Cancer 2003; 42: Takeda S, Fukai S, Komatsu H, et al. Impact of large tumor size on survival after resection of pathologically node negative (pn0) non-small cell lung cancer. Ann Thorac Surg 2005; 79: Original Research

8 8 Yang P, Allen MS, Aubry MC, et al. Clinical features of 5,628 primary lung cancer patients: experience at Mayo Clinic from 1997 to Chest 2005; 128: Naruke T, Goya T, Tsuchiya R, et al. Prognosis and survival in resected lung carcinoma based on the new international staging system. J Thorac Cardiovasc Surg 1988; 96: Inoue K, Sato M, Fujimura S, et al. Prognostic assessment of 1310 patients with non-small-cell lung cancer who underwent complete resection from 1980 to J Thorac Cardiovasc Surg 1998; 116: Jassem J, Skokowski J, Dziadziuszko R, et al. Results of surgical treatment of non-small cell lung cancer: validation of the new postoperative pathologic TNM classification. J Thorac Cardiovasc Surg 2000; 119: Naruke T, Suemasu K, Ishikawa S. Surgical treatment for lung cancer with metastasis to mediastinal lymph nodes. J Thorac Cardiovasc Surg 1976; 71: Travis WD, Brambilla E, Müller-Hermelink HK, et al, eds. WHO classification of tumours: pathology and genetics of tumours of the lung, pleura, thymus and heart. Lyon, France: IARC Press, Brechot JM, Chevret S, Charpentier MC, et al. Blood vessel and lymphatic vessel invasion in resected non-small cell lung carcinoma: correlation with TNM stage and disease free and overall survival. Cancer 1996; 78: Mineo TC, Ambrogi V, Baldi A, et al. Prognostic impact of VEGF, CD31, CD34, and CD105 expression and tumour vessel invasion after radical surgery for IB-IIA non-small cell lung cancer. J Clin Pathol 2004; 57: Renyi-Vamos F, Tovari J, Fillinger J, et al. Lymphangiogenesis correlates with lymph node metastasis, prognosis, and angiogenic phenotype in human non-small cell lung cancer. Clin Cancer Res 2005; 11: Akaike H. A new look at the statistical model identification. IEEE Trans Automatic Control 1974; 19: Rami-Porta R. Reflections on the revisions in the international system for staging lung cancer. Chest 1998; 113: Kawai H, Tada A, Kawahara M, et al. Smoking history before surgery and prognosis in patients with stage IA non-small-cell lung cancer: a multicenter study. Lung Cancer 2005; 49: Sakao Y, Tomimitsu S, Takeda Y, et al. Carcinoembryonic antigen as a predictive factor for postoperative tumor relapse in early-stage lung adenocarcinoma. Eur J Cardiothorac Surg 2004; 25: 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: ; discussion Gajra A, Newman N, Gamble GP, et al. Impact of tumor size on survival in stage IA non-small cell lung cancer: a case for subdividing stage IA disease. Lung Cancer 2003; 42: Suzuki K, Nagai K, Yoshida J, et al. Predictors of lymph node and intrapulmonary metastasis in clinical stage IA non-small cell lung carcinoma. Ann Thorac Surg 2001; 72: Shields TW. Prognostic significance of parenchymal lymphatic vessel and blood vessel invasion in carcinoma of the lung. Surg Gynecol Obstet 1983; 157: Collier FC, Blakemore WS, Kyle RH, et al. Carcinoma of the lung: factors which influence five year survival with special reference to blood vessel invasion. Ann Surg 1957; 146: Macchiarini P, Fontanini G, Hardin MJ, et al. Blood vessel invasion by tumor cells predicts recurrence in completely resected T1 N0 M0 non-small-cell lung cancer. J Thorac Cardiovasc Surg 1993; 106: Gabor S, Renner H, Popper H, et al. Invasion of blood vessels as significant prognostic factor in radically resected T1 3N0M0 non-small-cell lung cancer. Eur J Cardiothorac Surg 2004; 25: Pechet TT, Carr SR, Collins JE, et al. Arterial invasion predicts early mortality in stage I non-small cell lung cancer. Ann Thorac Surg 2004; 78: Macchiarini P, Fontanini G, Hardin JM, et al. Most peripheral, node-negative, non-small-cell lung cancers have low proliferative rates and no intratumoral and peritumoral blood and lymphatic vessel invasion: rationale for treatment with wedge resection alone. J Thorac Cardiovasc Surg 1992; 104: Kopfstein L, Christofori G. Metastasis: cell-autonomous mechanisms versus contributions by the tumor microenvironment. Cell Mol Life Sci 2006; 63: Schoppmann SF. Lymphangiogenesis, inflammation and metastasis. Anticancer Res 2005; 25: CHEST / 132 / 1/ JULY,