Hiroyasu Ueno, Aritoshi Hattori, Takeshi Matsunaga, Kazuya Takamochi, Shiaki Oh, Kenji Suzuki

Hiroyasu Ueno, Aritoshi Hattori, Takeshi Matsunaga, Kazuya Takamochi, Shiaki Oh, Kenji Suzuki Is Lower Zone Mediastinal Nodal Dissection Always Mandatory for Lung Cancer in the Lower Lobe? Department of General Thoracic Surgery, Juntendo University School of Medicine Address: 1-3, Hongo 3-chome, Bunkyo-ku, Tokyo, Japan, 113-8431 Kenji Suzuki, MD, Professor and Chairman, Department of General Thoracic Surgery, Juntendo University School of Medicine E-mail: kjsuzuki@juntendo.ac.jp Phone: +81-3-3813-3111 Fax: +81-3-5800-0281 Clinical Original Article 1

Abstract Purpose: Dissection of the lower zone mediastinal nodes is mandatory during systematic nodal dissection for lung cancer. However, the significance of lower zone lymph node metastasis (LZM) in lung cancer remains unclear. Therefore, we aimed to identify the predictive factors for LZM in patients with lower lobe lung cancer. Methods: A retrospective study was conducted on 257 patients with lower lobe lung cancer, in whom pulmonary resection and mediastinal nodal dissection were performed between 2009 and 2013. The radiological factors on thin-section computed tomography scans (TSCT) and several conventional clinical factors were evaluated as possible predictors of LZM. Results: Twenty (7.8%) patients exhibited LZM. The majority of the tumors were especially located in Segment 10 (50%). All patients showed a solid appearance on TSCT. In a univariate analysis, the tumor location, a solid appearance and the clinical T factor significantly predicted LZM (p=0.011, 0.005, 0.018). Furthermore, based on a multivariate analysis, the tumor location in Segment 10 significantly predicted LZM in patients with lower lobe solid lung cancer (p= 0.031). Conclusion: The appropriate surgical strategy for lower zone lymph node dissection should be selected based on the tumor location and the findings of TSCT, due to the high frequency of LZM (19.6%), especially in patients with pure-solid lung cancer in Segment 10. 2

Key words: lung cancer, lower zone lymph node, lymph node dissection, predictive factor 3

Introduction In the revision of the 7th TNM staging classification for lung cancer, the T and M descriptors were modified based on differences in survival. In contrast, there were no differences in the prognosis based on the involvement of different nodal zones within the N1 or N2 category [1]. Accurate staging of the lymph node involvement is a critical aspect of the initial management of patients with non-small cell lung cancer (NSCLC), and it influences decisions about the appropriateness and timing of treatment strategies such as surgery, radiation and systemic therapy [2]. With regard to the prognostic factors related to the lymph nodes, the number of involved nodes [3], the involved stations [3-5] and the presence of skip metastases [6-8] are of great importance [9]. Furthermore, as shown in many prospective randomized trials worldwide, the role of mediastinal lymph node dissection during the resection for NSCLC is still controversial (Table 1) [10-14]. In a recent report of the results of the American College of Surgery Oncology Group Z0030 Trial, no significant differences were observed with regard to long-term survival between mediastinal lymph node dissection and mediastinal lymph node sampling during pulmonary resection for patients with T1 or T2, and N0 or nonhilar N1 NSCLC [10]. Lung cancers that show a wide area of ground-glass opacity (GGO) are considered to have a favorable prognosis, and in most cases, their pathological features are minimally invasive. This cohort has an excellent prognosis with a recent five-year survival rate of more than 90% [15]. On the other hand, radiologically pure solid lung cancers exhibit more malignant behavior and have a dismal prognosis, with 4

postoperative nodal involvement in approximately 20% of cases, even in clinical stage IA disease [16, 17]. According to the recent Union for International Cancer Control (UICC) and International Association for the Study of Lung Cancer (IASLC) criteria, when the lower zone lymph nodes are affected, i.e., the paraesophageal nodes defined as Station 8, and pulmonary ligament nodes defined as Station 9, the disease is defined as pathological N2. The lower zone lymph nodes are generally dissected with the pulmonary ligament or mediastinal pleura, and systematic dissection is rarely performed for the lower zone. However, because the significance of dissection for lower zone lymph nodes is still unclear, we intended to identify factors predicting lower zone lymph node metastasis (LZM) in patients with lower lobe lung cancer. Clarification of the features of tumors with LZM should help clinicians decide on a suitable intraoperative strategy for lower zone lymph node dissection. Patients and Methods Between June 2009 and March 2013, 257 consecutive patients underwent pulmonary resection with mediastinal lymph node dissection for lower lobe lung cancer at our institute. Patients who underwent pneumonectomy or bilobectomy were included in this study, and those who received limited resection, such as wide wedge resection or segmentectomy, were excluded, because they were not evaluated sufficiently for mediastinal lymph node metastasis. All patients provided their written informed 5

consent before trial enrollment. The overall characteristics of the patients are shown in Table 2. A contrast-enhanced computed tomography scan was performed to evaluate the entire lung for preoperative staging. All tumors were subsequently evaluated based on the findings of a thin-section computed tomography scan (TSCT). The radiological findings were evaluated by four authors (H.U, A.H, T.M, and K.S), who were not informed of the pathological or prognostic outcome, based on the TSCT. The size and features of tumors were determined preoperatively based on the findings of TSCT. In addition, all tumors were subsequently evaluated to estimate the extent of GGO on TSCT with 2 mm collimation. The lung was photographed with a window level of -500 to -700 H and a window depth of 1000-2000 H as a lung window. A solid component was defined as an area of increased opacification that completely obscured the underlying vascular markings. GGO was defined as an area with a slight, homogeneous increase in density that did not obscure the underlying vascular markings. According to the radiological findings on TSCT, we defined the ratio of the maximum diameter of consolidation to the maximum tumor diameter as the consolidation/tumor ratio (CTR). In the current study, pure GGO was defined as a tumor with a CTR equal to 0. GGO tumors were defined as tumors with focal nodular opacity that contained both solid and GGO components (0< CTR< 1.0), and a pure-solid nodule was defined as a tumor that only showed consolidation without GGO (CTR= 1.0). All patients had clinical stage T1-3, N0 or N1 disease as assessed by preoperative TSCT. All patients were 6

expected to receive complete resection. The medical records of each patient were reviewed to determine the age, gender, pack-years of smoking and the serum carcinoembryonic antigen level (continuous variable; nanograms per milliliter). With regard to the radiological findings, the following factors were evaluated: maximal tumor dimension, presence and extent of a solid or GGO component in the tumor, and the location of the tumor. Univariate and multivariate analyses were used to identify the clinical factors that predicted LZM. Fisher s exact test and a t-test were used to compare two factors. A multivariate analysis was performed by a logistic regression analysis using the SPSS Statistics 21 software program (IBM Corporation, Somers, NY). Forward and backward stepwise procedures were used to elucidate the significant factors that were essential for predicting lower zone metastasis. The results of a statistical analysis were considered to be significant for values of p < 0.05. Results Of the 257 eligible lower lobe lung cancer patients who underwent lobectomy or pneumonectomy with mediastinal lymph node dissection, the overall cohort consisted of 170 males and 87 females. The patients ranged in age from 33 to 81 years old, with a median age of 64 years old. Among them, 20 (7.8%) patients (12 males and eight females) showed LZM. Station 8 metastasis was found in 11 cases, station 9 metastasis was found in eight cases, and metastasis in both stations in the lower zone was found 7

in one case. The tumor was found in Segment 6 in three (15%), Segment 8 in four (20%), Segment 9 in three (15%), and Segment 10 in 10 cases (50%). In patients with LZM, the number of tumors in Segment 10 was significantly higher than that in any other segment (p=0.011). Furthermore, all patients with LZM showed a pure solid appearance on TSCT (p=0.005). The radiological maximal tumor diameter ranged from 18 to 83 mm, with a mean of 33.9 ± 18.2 mm. In patients with more than clinical T2a lung cancer, LZM was observed at a higher frequency compared with patients with less than clinical T1b lung cancer (p=0.018). According to a univariate analysis, the tumor location, i.e., segment 10 or not, consolidation status and clinical T factor were significant predictors of LZM (p=0.011, 0.005 and 0.018, respectively) (Table 3). On the other hand, LZM was never found in patients with a GGO component on TSCT. Therefore, we evaluated 197 patients with lower lobe lung cancer, especially with a radiologically solid appearance on TSCT. Among them, a tumor location in segment 10 was the only significant predictor of LZM as determined by a multivariate analysis (p=0.031) (Table 4). Based on these results, when we combined these prognostic factors, LZM was observed in approximately 19.6% of patients with a solid tumor located in segment 10, whereas the frequency of LZM in the overall patients with lower lobe lung cancer was about 7.8%. Discussion 8

Intraoperative lymph node assessment is important for the accurate staging of NSCLC. Compared to systematic mediastinal lymph node sampling, systematic mediastinal lymph node dissection could improve the intraoperative staging and can be expected to provide better locoregional control [11]. In contrast, more extensive mediastinal dissection might be associated with longer hospitalization and potential damage to mediastinal structures [11]. However, as presented in Table 1, the extent and impact of mediastinal node removal are still controversial, despite several prospective trials [10-14]. It is difficult to evaluate the clinical characteristics of tumors that show LZM. Furthermore, the postoperative lymph node involvement clinically depends on the radiological findings of the tumor. Lung cancers that show a wide area of GGO are considered to have a good prognosis, and in most cases, their pathological features are minimally invasive. However, postoperative nodal involvement is found in approximately 20% of cases, even in patients with clinical stage IA disease [16, 17]. Furthermore, postoperative nodal involvement is often found in patients with radiologically pure solid lung cancer [18]. Thus, a consensus regarding the proper treatment strategies for lower nodal involvement has not yet been reached [19]. According to our present results, all tumors with LZM showed a radiologically pure solid appearance. Moreover, among radiologically pure solid tumors located in segment 10, the incidence of LZM (19.6%) was higher than that of tumors located in other segments of the lower lobe (6.9%). Thus, lower zone lymph node dissection is mandatory for such cases, and in solid tumors located in segment 6 to 9, lower zone lymph node dissection is recommended. In contrast, in patients with tumors that show a 9

GGO component on TSCT, lower zone metastasis is extremely rare. Therefore, selective dissection for the lower zone lymph node could be an appropriate surgical strategy, i.e., lower zone lymph node sampling is sufficient for GGO tumors. Several authors have insisted that lung adenocarcinoma with a large area of GGO on TSCT has a much better prognosis than conventional adenocarcinoma of the lung, regardless of the maximum tumor dimension [20-24]. Suzuki and colleagues reported that the consolidation status based on TSCT was a significant factor predicting nodal involvement, regardless of the hilar or mediastinal lymph node status in patients with small-sized lung cancer [16]. In our study, the consolidation status was also found to be a significant factor predicting the presence of lower zone lymph node metastasis according to a univariate analysis (p= 0.005). The mediastinal lymph nodes that play important roles in the lymphatic drainage of the lungs can be divided into four distinct but interconnected groups: the anterior (prevascular) lymph nodes in the anterior mediastinal component, the tracheobronchial lymph nodes, the paratracheal lymph nodes and the posterior lymph nodes in the posterior area of the visceral compartment of the mediastinum. Furthermore, the posterior lymph nodes may be separated into groups: the paraesophageal nodes and those located in either pulmonary ligament [25]. These posterior nodes are less commonly identified in the superior portion than in the inferior portion of the mediastinum [25]. The dominant flow of lymphatic drainage occurs from basal segments of the right lower lobe into the subcarinal nodes through the 10

bronchopulmonary nodes. The most important route of drainage from the basal segmental bronchi of the left lower lobe is the route that runs under the left main bronchus to the subcarinal nodes. This route extends to the right superior tracheobronchial nodes or through the lower pretracheal nodes to the right upper paratracheal nodes. Thus, the route of the posterior lymph which flows into the subcarinal nodes may be related to the fact that, in our study, half of the tumors that showed LZM were located in segment 10. In other words, the lymphatic route from segment 10 might flow into the lower zone lymph nodes directly, rather than into the hilar nodes. When we consider this possibility, segment 10 may have minor variations in lymphatic flow toward the lower zone lymph nodes. These minor lymphatic variations lead to lymphatic flow into the subcarinal nodes via the lower zone lymph nodes. Some patients demonstrate tumor involvement only in the N2 or N3 nodes, and none in the N1 nodes, and this phenomenon is called skip metastasis [6-8]. With respect to anatomy, patients who have N2 disease and N1 negative nodes may have lymphatic channels that bypass the N1 nodes [7]. In the 20 patients in this study who showed LZM, five (25%) patients exhibited skip metastasis to a lower zone lymph node (Table 5). All of these cases of skip metastasis were found in segment 10 in the left lung. This finding supports the notion that segment 10 may possess minor variations in lymphatic flow that lead it to flow toward the lower zone lymph nodes directly, especially in the left lung. Nomori et al. noted that the lymphatic flow went not only to resected segmental nodes but also to the segmental nodes of posterior segments, probably because the lobar bronchi were located in the posterior part of the chest cavity [26]. Imai et al. reported that the 11

rate of direct subpleural lymph flow to the mediastinum without passing through the hilum pulmonis was 21.4%, which is consistent with the reported range of skip metastasis of approximately 20-40 % [27]. The sentinel nodes are sometimes found in different segments from the resected segment; therefore, one of the sentinel nodes which belongs to segment 10 may exist in the lower zone lymph nodes. One of the important objectives of this study was to determine the proper strategies for treating cancer in the lower zone lymph nodes. As indicated in our study, the radiological findings based on TSCT correlated well with the postoperative nodal involvement, even in lower zone lymph nodes. Several imaging modalities, such as CT, magnetic resonance imaging (MRI), positron emission tomography (PET) and PET-CT play central roles in the detection and staging of the disease [28]. In the future, these solid tumors located in the lower lobe may be divided into several groups by PET. If a solid tumor shows a positive result by PET, it may be associated with a higher frequency of lymph node metastasis and a poorer prognosis [16, 17]. Further studies on this matter are warranted to provide an economic perspective and to determine the best combination of tests that provides the greatest impact on the clinical outcomes [28]. One of the limitations of this study is its retrospective nature, and there may be some degree of selection bias. In the future, a prospective randomized study of systematic lymph nodal dissection versus lobe-specific lymph node dissection during pulmonary resection in patients with radiologically invasive lung cancer will be planned in Japan. However, our present results provide useful information for 12

determining the appropriate surgical strategies for lower lobe lung cancers based on TSCT. In conclusion, among patients with lower lobe lung cancer, the radiological findings on TSCT and tumor location were significant factors predicting LZM. Therefore, the surgical strategy for lower zone dissection should be appropriately selected based on the tumor location and the findings of TSCT, due to the high frequency of LZM (19.6%) in patients with pure solid lung cancer in segment 10, and the extremely low frequency of LZM in patients with GGO tumors. This paper was presented at the IASLC 15th World Conference on Lung Cancer, Sydney, Australia, October 27-30, 2013 13

Acknowledgements This study was supported in part by a Grant-in-Aid for Cancer Research from the Ministry of Health, Labour and Welfare of Japan and the Smoking Research Foundation. Kenji Suzuki has no conflicts of interest to declare in association with this article. 14

References 1. Detterbeck FC, Postmus PE, Tanoue LT. The stage classification of lung cancer, Diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. CHEST. 2013; 143(5) (suppl): e191s-e210s 2. Rusch VW, Crowley J, Giroux DJ, Goldstraw P, Im J, Tsuboi M, et al. The IASLC lung cancer staging project: Proposals for the revision of the N descriptors in the forthcoming seventh edition of the TNM classification for lung cancer. J Thorac Oncol. 2007; 2: 603-612. 3. Fukui T, Mori S, Yokoi K, Mitsudomi T. Significance of the number of positive lymph nodes in resected non-small cell lung cancer. J Thorac Oncol. 2006; 1: 120-125. 4. Keller SM, Vangel MG, Wagner H, Schiller JH, Herskovic A, Komaki R, et al. Prolonged survival in patients with resected non-small cell lung cancer and single-level N2 disease. J Thorac Cardiovasc Surg. 2004; 128: 130-137. 5. Misthos P, Sepsas E, Kokotsakis J, Skottis I, Lioulias A. The significance of one-station N2 disease in the prognosis of patients with nonsmall-cell lung cancer. Ann Thorac Surg. 2008; 86: 1626-1630. 6. Prenzel KL, Baldus SE, Monig SP, Tack D, Sinning JM, Gutschow CA, et al. Skip matastasis in nonsmall cell lung carcinoma: predictive markers and isolated tumor cells in N1 lymph nodes. Cancer. 2004; 100: 1909-1917. 7. Riquet M, Assouad J, Bagan P, Foucault C, Barthes FLP, Dujon A, et al. Skip mediastinal lymph 15

node metastasis and lung cancer: a particular N2 subgroup with a better prognosis. Ann Thorac Surg. 2005; 79: 225-233. 8. Tsubota N, Yoshimura M. Skip metastasis and hidden N2 disease in lung cancer: how successful is mediastinal dissection? Surg Today. 1996; 26: 169-172. 9. Takamochi K, Oh S, Suzuki K. Prognostic evaluation of nodal staging based on the new IASLC lymph node map for lung cancer. Thorac Cardiov Surg. 2010; 58: 345-349. 10. Darling GE, Allen MS, Decker PA, Ballman K, Malthaner RA, Inculet RI, et al. Randomized trial of mediastinal lymph node sampling versus complete lymphadenectomy during pulmonary resection in the patient with N0 or N1 (less than hilar) non-small cell carcinoma: results of the American College of Surgery Oncology Group Z0030 Trial. J Thorac Cardiovasc Surg. 2011; 141: 662-670. 11. Lardinois D, Suter H, Hakki H, Rousson V, Betticher D, Ris H. Morbidity, survival, and site of recurrence after mediastinal lymph-node dissection versus systematic sampling after complete resection for non-small cell lung cancer. Ann Thorac Surg. 2005; 80: 268-275. 12. Izbicki JR, Passlick B, Pantel K, Pichlmeier U, Hosch SB, Karg O, et al. Effectiveness of radical systematic madiastinal lymphadenectomy in patients with resectable non-small cell lung cancer. Ann Surg. 1998; 227: 138-144. 13. Wu Y, Huang Z, Wang S, Yang X, Ou W. A randomized trial of systematic nodal dissection in resectable non-small cell lung cancer. Lung Cancer. 2002; 36: 1-6. 16

14. Keller SM, Adak S, Wagner H, Johnson DH. Mediastinal lymph node dissection improves survival in patients with stages II and IIIa non-small cell lung cancer. Ann Thorac Surg. 2000; 70: 358-366. 15. Suzuki K, Koike T, Asakawa T, Kusumoto M, Asamura H, Nagai K, et al. A prospective radiological study of thin-section computed tomography to predict pathological noninvasiveness in peripheral clinical IA lung cancer (Japan Clinical Oncology Group 0201). J Thorac Oncol. 2011; 6: 751-756. 16. Suzuki K, Kusumoto M, Watanabe S, Tsuchiya R, Asamura H. Radiologic classification of small adenocarcinoma of the lung: radiologic-pathologic correlation and its prognostic impact. Ann Thorac Surg. 2006; 81: 413-420. 17. Hattori A, Suzuki K, Matsunaga T, Fukui M, Kitamura Y, Miyasaka Y, et al. Is limited resection appropriate for radiologically solid tumors in small lung cancers? Ann Thorac Surg. 2012; 94: 212-215. 18. Hattori A, Suzuki K, Maeyashiki T, Fukui M, Kitamura Y, Matsunaga T, et al. The presence of air bronchogram is a novel predictor of negative nodal involvement in radiologically pure-solid lung cancer. Eur J Cardiothorac Surg. 2013. doi: 10. 1093/ejcts/ezt467. 19. Miyoshi S, Shien K, Toyooka S, Miyoshi K, Yamamoto K, Sugimoto S, et al. Validity of using lobe-specific regional lymph node stations to assist navigation during lymph node dissection in early stage non-small cell lung cancer patients. Surg Today, 2014, 44: 2028-2036. 20. Aoki T, Tomoda Y, Watanabe H, Nakata H, Kasai T, Hashimoto H, et al. Peripheral lung 17

adenocarcinoma: correlation of thin-section CT findings with histologic prognostic factors and survival. Radiology. 2001; 220: 803-809. 21. Matsuguma H, Yokoi K, Anraku M, Kondo T, Kamiyama Y, Mori K, et al. Proportion of ground-glass opacity on high-resolution computed tomography in clinical T1 N0 M0 adenocarcinoma of the lung: a predictor of lymph node matastasis. J Thorac Cardiovasc Surg. 2002; 124: 278-284. 22. Suzuki K, Asamura H, Kusumoto M, Kondo H, Tsuchiya R. Early peripheral lung cancer: prognostic significance of ground glass opacity on thin-section computed tomographic scan. Ann Thorac Surg. 2002; 74: 1635-1639. 23. Takamochi K, Nagai K, Yoshida J, Suzuki K, Ohde Y, Nishimura M, et al. Pathologic N0 status in pulmonary adenocarcinoma is predictable by combining serum carcinoembryonic antigen level and computed tomographic findings. J Thorac Cardiovasc Surg. 2001; 122: 325-330. 24. Maeyashiki T, Suzuki K, Hattori A, Matsunaga T, Takamochi K, Oh S. The size of consolidation on thin-section computed tomography is a better predictor of survival than the maximum tumour dimension in resectable lung cancer. Eur J Cardiothorac Surg. 2013; 43: 915-918. 25. Shields TW. Chapter 6 Lymphatics of the lungs, General Thoracic Surgery, 7th ed. Philadelphia: Lippincott Williams & Wilkins, 2009: 87-101. 26. Nomori H, Kohno M, Izumi Y, Ohtsuka T, Asakura K, Nakayama T. Sentinel nodes in lung cancer: review of our 10- year experience. Surg Today, 2011; 41: 889-895. 18

27. Imai K, Minamiya Y, Saito H, Nakagawa T, Ito M, Ono T, et al. Detection of pleural lymph flow using indocyanine green fluorescence imaging in non-small cell lung cancer surgery: a preliminary study. Surg Today, 2013, 43: 249-254. 28. Imai K, Minamiya Y, Saito H, Motoyama S, Sato Y, Ito A, et al. Diagnostic imaging in the preoperative management of lung cancer. Surg Today, 2014, 44: 1197-1206. 19

Tables and Legends TABLE 1 The Previous Literatures Regarding the Differences of Mediastinal Lymph Node Dissection and Sampling in the Treatment of Non-Small Cell Lung Cancer Author/ Year Phase Patients Outcome Darling JTCVS 2011 [10] Phase III c- T1-2N0M0 Dissection Sampling Overall Survival Izbicki Ann Surg 1998 [12] Phase III c- T1-4N0-1M0 Dissection Sampling Disease-Free Survival, Overall Survival Wu Lung Cancer 2002 [13] Phase III c- Stage I-IIIA Dissection Sampling Overall Survival Keller ATS 2000 [14] Phase III p- T1-3N1-2M0 Dissection Sampling Overall Survival in Right Lung Cancer Lardinois ATS 2005 [11] Retrospective c- T1-3N0-1M0 Dissection Sampling Disease-Free Survival 20

TABLE 2 The Clinical and Radiological Characteristics of Patients with Lower Lobe Lung Cancer Characteristics No. (%) Gender male / female 170 (66) / 87 (34) Extent of lung resection lobectomy bilobectomy pneumonectomy 222 (86) 21 (8) 14 (5) Tumor side left / right 148 (58) / 109 (42) Histology adenocarcinoma squamous cell carcinoma large cell carcinoma adenosquamous carcinoma others 171 (66) 63 (25) 4 (2) 8 (3) 11 (4) Clinical T factor T1a /1b T2a / 2b T3 73 (28) / 63 (25) 88 (34) / 27 (11) 6 (2) Clinical TNM stage IA / IB IIA / IIB 122 (47) / 62 (24) 54 (21) / 17 (7) 21

IIIA 2 (1) Pathological N factor N0 N1 N2 162 (63) 32 (12) 63 (25) Pathological TNM stage IA / IB IIA / IIB IIIA 101 (39) / 39 (15) 40 (16) / 11 (4) 66 (26) 22

TABLE 3 The Factors Predicting Lower Zone Metastasis in Patients with Lower Lobe Lung Cancer: The Results of a Univariate Analysis Factor LZM Present (n=20) LZM Absent (n=237) P-value Age (years) Range 33-81 33-89 Mean 64 68 70 7 128 <70 13 109 0. 110 a Gender Male 12 158 Female 8 79 0.624 a Pack-years smoked 20 12 157 <20 8 80 0.626 a CEA (ng/ml) Abnormal 11 141 Normal 9 96 0.813 a Maximum tumor diameter (mm) Range 18-83 2-105 Mean 41.2 33.7 0.082 b Consolidation status Pure -solid 20 177 GGO tumor 0 60 0.005 a Side Left Right 13 7 96 141 0.057 a S6 3 (0 / 3) 84 (34 / 50) 0.084 a Tumor location (left / right) S7 0 (0 / 0) 3 (0 / 3) >0.999 a S8 4 (4 / 0) 42 (16 / 26) 0.764 a S9 3 (1 / 2) 56 (18 / 38) 0.580 a 23

S10 basal segment 10 (8 / 2) 17 (13 / 4) 52 (28 / 24) 150 (62 / 88) 0.011 a 0.054 a Clinical T factor T1a / T1b 5 / 2 68 / 61 T2a / T2b / T3 12 / 1 / 0 76 / 26 / 6 0.018 a Clinical N factor 0 15 185 1 5 52 0.780 a LZM = lower zone metastasis; CEA = serum carcinoembryonic antigen; abnormal 3.0 ng/ml, normal < 3.0 ng/ ml; a p-value in Fisher s exact test; b p-value in the t-test; Ground glass opacity (GGO) tumors are tumors having GGO on thin section CT scans. 24

TABLE 4 The Factors Predicting Lower Zone Metastasis in Patients with Lower Lobe Lung Cancer with a Radiologically Solid Appearance: the Results of a Multivariate Analysis Variant Odds Ratio 95% CI P-value* Tumor location (Segment 10 or not) 2.902 1.099-7.664 0.031 CEA ( 3.0ng/ml) 0.530 0.200-1.407 0.202 Side (Left / Right) 2.251 0.823-6.152 0.114 Clinical T factor (T2a / T2b / T3 or not) 1.664 0.597-4.638 0.330 Clinical N factor (N0 or N1) 0.915 0.294-2.848 0.878 * p-value in the logistic regression analysis 25

TABLE 5 The Characteristics of the Patients with Skip Metastasis No. Side Location Clinical stage Surgical procedure Pathological N status 1 Left S10 T1aN0M0 IA Lobectomy Station 8, 9 2 Left S10 T2aN0M0 IB Lobectomy Station 8 3 Left S10 T2aN0M0 IB Lobectomy Station 8 4 Left S10 T2aN1M0 IIA Lobectomy Station 8 5 left S10 T2bN0M0 IIA lobectomy Station 9 26