The Prognostic Value of Ratio-Based Lymph Node Staging in Resected Non Small-Cell Lung Cancer

Original Article The Prognostic Value of Ratio-Based Lymph Node Staging in Resected Non Small-Cell Lung Cancer Chen Qiu, MD,* Wei Dong, MD,* Benhua Su, MBBS, Qi Liu, MD,* and Jiajun Du, PhD Introduction: Assessment of lymph node status is a critical issue in the surgical management of non small-cell lung cancer (NSCLC). We sought to determine the prognostic value of metastatic lymph node ratio (LNR) in patients with radical surgery for NSCLC. Methods: We abstracted data from 480 consecutive patients undergoing radical surgery for NSCLC between 2006 and 2008 in our institution. Kaplan Meier estimated the survival function using the number of metastatic lymph node (MLN) and LNR as categorized variables. The prognostic value of age, sex, smoking status, location of tumor, histology, pathology T stage, pathology N stage, surgical procedure, chemotherapy, MLN, and LNR were assessed using a multivariate Cox proportional hazards model for overall survival (OS) and disease-free survival (DFS). Results: The median numbers of examined lymph nodes and MLNs were 15 and 5, respectively. Optimal cutpoints of the LNR were calculated as 0, 0 to 0.35, and greater than 0.35. Patients with higher LNR were associated with worse OS and DFS in the whole series, whereas there was no significant difference in the OS and DFS of those patients classified as pathology N2. A multivariate analysis showed that the LNR staging, smoking status, and chemotherapy were revealed to be independent prognostic factors. Conclusions: LNR is an independent predictor of survival in patients with NSCLC undergoing radical resection; the prognostic significance is more valuable in patients classified as pathology N1. Key Words: Prognostic, Metastatic lymph node, Lymph node ratio, Non small-cell lung cancer. (J Thorac Oncol. 2013;8: 429-435) Non small-cell lung cancer (NSCLC) remains a major cause of cancer death, and the 5-year survival rate is *Institute of Oncology, Provincial Hospital Affiliated to Shandong University, Shandong University, #51 Weiliu Road, Jinan 250021, P. R. China; Department of Medical Engineering, Provincial Hospital Affiliated to Shandong University, Shandong University, #324 Jingwu Road, Jinan 250021, P. R. China; and Department of thoracic Surgery, Provincial Hospital Affiliated to Shandong University, Shandong University, #324 Jingwu Road, Jinan 250021, P. R. China Disclosure: The authors declare no conflict of interest. Chen Qiu, MD, and Wei Dong, MD, contributed equally to this manuscript. Address for correspondence: Jiajun Du, Department of thoracic Surgery, Provincial Hospital Affiliated to Shandong University, Shandong University, #324 Jingwu Road, Jinan 250021, P. R. China. E-mail: dujiajun@sdu.edu.cn Copyright 2013 by the International Association for the Study of Lung Cancer ISSN: 1556-0864/13/0804-0429 Journal of Thoracic Oncology Volume 8, Number 4, April 2013 still poor despite early detection and therapeutic modalities. 1 Among the known prognostic factors, lymph node (LN) status is currently regarded as the strongest prognostic parameter for patients with NSCLC. The nodal staging in the current seventh edition of the tumor, node, metastasis (TNM) classification is defined based on the anatomical location of positive LNs, as in the previous sixth edition; however, the TNM classification system for breast, gastric, and colorectal cancers has been updated from the traditional system to include number of metastatic lymph nodes (MLN) in the N staging. 2 The prognostic value of MLNs on survival for patients with NSCLC has long been focused by several studies 3 5 ; however, the maximal number of MLNs is limited by the number of LNs sampled, and it is difficult to assess the LN status reliably without removing sufficient numbers of LNs. The lymph node ratio (LNR), defined as the ratio of positive LNs divided by the total number of retrieved LNs, reflects both the parameters and has been shown to represent a powerful independent prognosticator in several solid cancer types. 6 9 For NSCLC, two recent reports from America have, respectively, demonstrated that increasing LNR is a significant prognostic indicator for a worse survival in pathology N1 (pn1) patients according to different age groups. 10,11 In this retrospective study, we aimed to further evaluate the importance of LNR as a prognostic factor on the survival of patients with resectable NSCLC and to compare its prognostic power against other methods, such as the traditional pn staging and the number of MLNs resected. PATIENTS AND METHODS Patients and Pretreatment Evaluation We reviewed the records of 532 patients with NSCLC who underwent major pulmonary resection and systematic node dissection of the hilar and mediastinal LNs at our institute between 2006 and 2008. Patients who had smallcell carcinoma, concomitant double cancer, pathological Tis disease, and stage IV disease were excluded from this study. Ultimately, a total of 480 patients were enrolled. All patients underwent physical examination; computed tomography (CT) or magnetic resonance imaging of the brain, chest, and abdomen; and emission CT of bone preoperatively to make a definite diagnosis and make sure that there was no metastasis. Positron emission tomography was rarely implemented during this study period. Bronchoscopy was routinely performed to obtain a pathological diagnosis by transbronchial lung biopsy and to evaluate endobronchial staging. 429

Qiu et al. Journal of Thoracic Oncology Volume 8, Number 4, April 2013 Treatment The enrolled patients underwent lobectomy, bilobectomy, or pneumonectomy. A complete mediastinal lymphadenectomy was routinely performed. After surgery, en bloc dissected tissues were separated into each LN precisely. All resected specimens, including the primary tumor and resected hilar and mediastinal LNs, were examined to determine both the tumor histology and the extent of LN metastases. Pathological staging was based on the current seventh edition of the TNM classification. The main adjuvant treatment that patients underwent after operation was chemotherapy or radiotherapy. A total of four to six cycles of chemotherapy was administered to 261 patients (56%).The chemotherapy was a routine program for NSCLC according to the National Comprehensive Cancer Network. Follow-Up Follow-up information was obtained from all patients through office visits or telephone interviews with the patient, with a relative, or with their primary physicians and were evaluated by chest roentgenography, chest CT scans, and bone scintigraphy. Patients were followed up every 3 months for 1 year after operation, every 6 months for 3 years, and every year thereafter, with a median follow-up period of 37 months (range, 2 65 months). A total of 480 patients were followed up until death or the last day of follow-up (June 15, 2011). The overall survival (OS) in each patient was defined as the interval between the date of the definitive resection and the date of the last follow-up or death, and disease-free survival (DFS) was defined as the time interval between the date of the definitive resection and detection of first disease recurrence, metastasis, or the date of the last follow-up. Statistical Analyses Medical records and survival data were obtained from all patients. For the MLN and LNR, the optimal cutoff values were determined using χ 2 scores, which were calculated using the Cox proportional hazards regression model. The Kaplan Meier method was used to plot the survival curves, and the log-rank test was used to evaluate the differences among the subgroups. Multivariate analysis of several prognostic factors was performed with Cox proportional hazards regression model. Factors with a p value of less than 0.1 in the univariate analysis were included in the multivariate analysis. Receiver operating characteristic (ROC) curves and the areas under the curves (AUC) were calculated for the three staging systems (LNR staging, MLN staging, and pn staging) to assess the accuracy of their predictive ability. A p value of less than 0.05 was considered to be statistically significant. All statistical analyses were performed using SPSS (version 17.0; SPSS Inc, Chicago, IL). RESULTS Patients Characteristics A retrospective series of 480 patients with NSCLC (352 men, 128 women; median age: 59 years; range, 34 84 years) was retrieved from the original files of the Department of Thoracic Surgery of Shandong Provincial Hospital for the study. Overall, 7307 LNs were removed from the patients, and there were 1149 involved nodes. The median number of removed nodes was 15, with the median number of positive nodes being five (Table 1). The number of the MLN was significantly correlated to the LNR (r =0.751; p < 0.001) (Fig. 1). Cutoff Values for the MLN and LNR The optimal cutoff of MLN and LNR were analyzed by the Cox proportional hazards regression model. Based on the maximal χ 2 score, the optimal cutoff value was three nodes for the number of MLNs (Table 2) and 35% for LNR (Table 3). TABLE 1. Patient Characteristics Variable no. of patients Percent Gender Male 352 73.3 Female 128 26.7 Age <60 yr 295 61.5 60 yr 185 38.5 Smoking status Smoked 303 63.1 Never smoked 177 36.9 Location of tumor Left 228 47.5 Right 252 52.5 Histological diagnosis Adenocarcinoma 232 48.3 Squamous cell 212 44.2 Large cell 9 1.9 Others 27 5.6 pt stage T1 171 35.6 T2 230 47.9 T3 79 16.4 pn stage N0 254 52.9 N1 115 24.0 N2 111 23.1 Surgical procedure Lobectomy 376 78.3 Bilobectomy 44 9.2 Pneumonectomy 60 12.5 Chemotherapy Yes 269 56.0 No 211 44.0 Metastatic lymph node 0 254 52.9 1 3 118 24.6 4 108 22.5 Lymph node ratio (%) 0 254 52.9 0 35 153 31.9 >35 73 15.2 430 Copyright 2013 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology Volume 8, Number 4, April 2013 Value of Ratio-Based Lymph Node Staging FIGURE 1. Correlation between metastatic lymph node (MLN) and lymph node ratio (LNR). Based on these results, both the MLN and LNR were placed into one of three categories in subsequent analyses (0, 1 3, or 4 and 0, 0 35% or >35%, respectively). TABLE 2. Analysis of the Number of Positive Lymph Nodes Using the Cox Proportional Hazards Regression Model Cutoff value for positive node number χ 2 score p Value 0, 1, 2 32.7 0.000 0, 1 2, 3 36.0 0.000 0, 1 3, 4 39.2 0.000 0, 1 4, 5 37.4 0.000 0, 1 5, 6 34.7 0.000 Impact of MLN Staging and pn Staging on OS and DFS after Resection The median number of MLNs was five (range, 1 51). Of the 480 assessable patients, 254 patients (52%) were TABLE 3. Analysis of the Lymph Node Ratio (LNR) Calculated by the Cox Proportional Hazards Regression Model Cutoff value for LNR (%) χ 2 score p Value 0, 0 5, >5 35.1 0.000 0, 0 10, >10 38.7 0.000 0, 0 15, >15 39.8 0.000 0, 0 20, >20 44.8 0.000 0, 0 25, >25 45.4 0.000 0, 0 30, >30 46.7 0.000 0, 0 35, >35 49.4 0.000 0, 0 40, >40 46.3 0.000 0, 0 45, >45 44.1 0.000 0, 0 50, >50 39.8 0.000 FIGURE 2. (A) Overall survival rates according to metastatic lymph node (MLN) in the whole series (log-rank χ 2 = 42.7; p < 0.001). (B) Disease-free survival rates according to MLN in the whole series (log-rank χ 2 = 52.6; p < 0.001). classified as N0, 115 patients (24%) were classified as N1, and 111 patients (23%) were classified as N2. The Kaplan Meier survival curves for OS and DFS are depicted in Figs 2 and 3. In the whole series, patients with more MLN(s) removed and more advanced pn staging have worse OS and DFS (p < 0.0001). Prognostic Value of LNR for OS and DFS after Resection Survival analysis indicated that a higher LNR group was associated with worse OS (p < 0.0001) (Fig. 4A) and DFS (p < 0.0001) (Fig. 4B). The pn classification according to the Union for International Cancer Control (UICC)/TNM, however, showed an imbalance in prognostic separation. There was no significant difference in the OS and DFS of those patients Copyright 2013 by the International Association for the Study of Lung Cancer 431

Qiu et al. Journal of Thoracic Oncology Volume 8, Number 4, April 2013 FIGURE 3. (A) Overall survival rates according to pn in the whole series (log-rank χ 2 = 47.2; p < 0.001). (B) Disease-free survival rates according to pn in the whole series (log-rank χ 2 = 60; p < 0.001). classified as pn2 as a function of LNR (Fig. 5C,D), whereas in patients classified as pn1, OS and DFS were significantly worse with higher LNR (p < 0.0001) (Fig. 5A,B). Multivariate Analysis of Prognostic Factors The number of MLNs (with the three-grade category) and the metastatic LNR (with the three-grade category) were evaluated, along with other potential prognostic factors (including sex, age, smoking status, tumor location, histological type, pt category, pn category, surgical procedure, and chemotherapy), for the prognostic significance in the multivariate analysis using Cox proportional regression hazard model. The results of multivariate analyses are presented in Table 4. pn stage was one of the significant prognostic factors, when the LNR was not included in the analysis (data not shown). However, when the LNR-based classification was FIGURE 4. (A) Overall survival rates according to lymph node ratio (LNR) in the whole series (log-rank χ 2 = 54.1; p < 0.001). (B) Disease-free survival rates according to LNR in the whole series (log-rank χ 2 = 59.7; p < 0.001). included in the model as a covariate, LNR, smoking status, and chemotherapy were revealed to be independent prognostic factors for OS and DFS, and pn stage lost its significance. Comparison of the Predictive Ability of the Three Staging Systems Based on LN The ROC curves were used to compare the three aforementioned N staging systems, and the AUC were calculated to assess the accuracy of each system s predictive ability for patients with NSCLC who underwent radical resection (Fig. 6). The AUC was 0.678 for LNR staging, 0.665 for MLN staging, and 0.672 for the seventh edition of UICC N staging. The 95% confidence interval for the three AUCs are crossed, suggesting that there was no statistical significance among the three staging methods. 12 432 Copyright 2013 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology Volume 8, Number 4, April 2013 Value of Ratio-Based Lymph Node Staging FIGURE 5. (A and B) Overall survival (OS) and disease-free survival (DFS) rates according to lymph node ratio (LNR) in pn1 NSCLC. (C and D) OS rate and DFS rates according to LNR in pn2 NSCLC. NSCLC, non small-cell lung cancer. DISCUSSION Adequate assessment of the nodal status is a critical issue in the surgical management of patients with lung cancer. In the latest TNM classification, 2 the nodal system in lung cancer is still based on the anatomical location of involved LNs. However, there is also growing evidence establishing the prognostic role of the number of MLN and LNR in NSCLC. 5,11,13 15 TABLE 4. Prognostic Factors for OS and DFS Retained in a Multivariate Analysis of 480 Patients with NSCLC Variables OS DFS HR (95% CI) p Value HR (95% CI) p Value LNR 0.000 0.000 0 1 1 0 0.35 0.249 (0.159 0.389) 0.318 (0.212 0.477) >0.35 0.422 (0.277 0.643) 0.570 (0.390 0.833) Smoking 1.614 (1.111 2.346) 0.012 1.347 (0.977 1.857) 0.069 Chemotherapy 0.542 (0.376 0.781) 0.001 0.444 (0.321 0.615) 0.000 CI, confidence interval; DFS, disease-free survival; HR, hazard ratio; LNR, lymph node ratio; OS, overall survival. Copyright 2013 by the International Association for the Study of Lung Cancer 433

Qiu et al. Journal of Thoracic Oncology Volume 8, Number 4, April 2013 FIGURE 6. Receiver operating characteristics curve of the seventh edition of the N staging systems, number of metastatic lymph node staging, and metastatic lymph node ratio staging for predicting survival of patients with NSCLC with radical resection. Fukui et al. 4 first analyzed the number of MLNs in resected NSCLC and suggested that the MLNs is a strong independent prognostic factor in NSCLC; they revealed the 5-year survival of patients with seven or more metastatic LNs to be 6%, which was significantly worse than n 4 6 patients (5-year survival: 42%), n 1 3 patients (5-year survival: 58%), or n 0 patients (5-year survival: 77%). The present results also indicate that improved survival outcomes are associated with a less number of MLNs for various tumor histologies. In support of our study, Lee et al. 15 reviewed 1081 patients who underwent major pulmonary resection between 1990 and 2006 and found the number of MLNs proved to be a good prognostic factor in NSCLC. According to the study by Wei et al. 5, MLN shown to be a better prognostic determinant than the location-based pn stage. Along with the number of MLN, another related measure is LNR, which has been identified as a significant prognostic value in breast cancer, 8 pancreatic cancer, 16 gastric cancer, 17 and colon cancer. 6 In an examination of 651 Japanese patients with NSCLC, Matsuguma et al. 13 showed that LNR was a significant prognostic factor and showed superiority to the UICC classification with respect to the location of metastatic LNs. In another American study, Nwogu et al. 18 analyzed 25,887 patients with NSCLC from the SEER database to determine the influence of LNs examined and LNR on survival and found more LNs resected and lower LNR are associated with better patient survival independent of age, sex, grade, tumor size, and stage of disease. There is, however, no clear consensus about the cutoff point of the staging classification. In a few articles related to our study, Matsuguma et al. divided the LNRs into three categories (0, 0.01 0.12, or 0.12) according to the median number of 11.1% for node-positive patients. In another study considering the N1 NSCLC, they divided the LNR into three categories (<0.15, 0.15 0.5, and >0.5) based on previous studies on other types of tumors and on the distribution of the LNR in the Surveillance, Epidemiology and End Results database. The current study used χ 2 scores calculated by the Cox proportional hazards regression model and thus determined the cutoff value to be three positive nodes for MLN and 0.35 for LNR. Based on the three categories, we found that a higher LNR group was associated with a worse OS and DFS in the whole series, which was also shown to be an independent factor for OS and DFS. We also investigated the impact of the LNR on survival rates according to pn staging system, the trend mentioned above can only be seen in pn1 patients. The benefits of adjuvant chemoradiotherapy for early-stage NSCLC remain controversial. Despite the significant role of staging in predicting survival, the prognosis may differ according to the number, level, or type of LNs involved. Patients with N1 NSCLC are a heterogeneous group in terms of survival and often do not get adjuvant chemotherapy despite an estimated 33 to 60% risk of disease-related death within 5 years of diagnosis. 19 21 Our results showed that patients with a higher LNR seem to be at an increased risk of recurrence, which is also consistent with the findings of two recent studies evaluating the relationship between the LNR and survival for N1 patients with NSCLC. 3,10 These data could provide useful prognostic information to physicians and patients, allowing the selection of patients for more aggressive postoperative therapy or closer follow-up strategy. To date, a number of studies have shown that LNR staging can be an alternative to the pn staging in breast, pancreatic, and stage III colon cancers 8,16,22. The multivariate analysis in our study also showed that when the LNR-based classification was included, pn stage lost its significance, and LNR, smoking status, and chemotherapy were revealed to be independent prognostic factors for OS and DFS. From the ROC curve, we still cannot reach the conclusion that LNR staging has superiority than the pn staging in NSCLC. The combination of the LNR and pn status provides a valuable help with prognosis. In summary, the current study demonstrates that LNR is an independent predictor of survival in patients with NSCLC. Moreover, we found that LNR is more informative toward predicting survival in pn1 stage patients. Acknowledgments The work in this study was supported by Provincial Natural Science Foundation of Shandong (ZR2010HM067) and Provincial Science and Technology Development Planning of Shandong (2011GGH21819 and 2012G0021836). REFERENCES 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin 2011;61:69 90. 2. Sobin LH Gospodarwoicz MK, Wittekind C. TNM Classification of Malignant Tumours, 7th Ed. New York: Wiley, 2009. 3. Osaki T, Nagashima A, Yoshimatsu T, Tashima Y, Yasumoto K. Survival and characteristics of lymph node involvement in patients with N1 nonsmall cell lung cancer. Lung Cancer 2004;43:151 157. 4. 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. 434 Copyright 2013 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology Volume 8, Number 4, April 2013 Value of Ratio-Based Lymph Node Staging 5. Wei S, Asamura H, Kawachi R, Sakurai H, Watanabe S. Which is the better prognostic factor for resected non-small cell lung cancer: the number of metastatic lymph nodes or the currently used nodal stage classification? J Thorac Oncol 2011;6:310 318. 6. Berger AC, Sigurdson ER, LeVoyer T, et al. Colon cancer survival is associated with decreasing ratio of metastatic to examined lymph nodes. J Clin Oncol 2005;23:8706 8712. 7. Partelli S, Fernandez-Del Castillo C, Bassi C, et al. Invasive intraductal papillary mucinous carcinomas of the pancreas: predictors of survival and the role of lymph node ratio. Ann Surg 2010;251:477 482. 8. Vinh-Hung V, Verkooijen HM, Fioretta G, et al. Lymph node ratio as an alternative to pn staging in node-positive breast cancer. J Clin Oncol 2009;27:1062 1068. 9. Kunisaki C, Makino H, Akiyama H, et al. Clinical significance of the metastatic lymph-node ratio in early gastric cancer. J Gastrointest Surg 2008;12:542 549. 10. Jonnalagadda S, Arcinega J, Smith C, Wisnivesky JP. Validation of the lymph node ratio as a prognostic factor in patients with N1 nonsmall cell lung cancer. Cancer 2011;117:4724 4731. 11. Wisnivesky JP, Arciniega J, Mhango G, Mandeli J, Halm EA. Lymph node ratio as a prognostic factor in elderly patients with pathological N1 nonsmall cell lung cancer. Thorax 2011;66:287 293. 12. Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839 843. 13. Matsuguma H, Oki I, Nakahara R, et al. Proposal of new nodal classifications for non-small-cell lung cancer based on the number and ratio of metastatic lymph nodes. Eur J Cardiothorac Surg 2012;41:19 24. 14. Wang CL, Li Y, Yue DS, Zhang LM, Zhang ZF, Sun BS. Value of the metastatic lymph node ratio for predicting the prognosis of non-small-cell lung cancer patients. World J Surg 2012;36:455 462. 15. Lee JG, Lee CY, Park IK, et al. Number of metastatic lymph nodes in resected non-small cell lung cancer predicts patient survival. Ann Thorac Surg 2008;85:211 215. 16. Riediger H, Keck T, Wellner U, et al. The lymph node ratio is the strongest prognostic factor after resection of pancreatic cancer. J Gastrointest Surg 2009;13:1337 1344. 17. Inoue K, Nakane Y, Iiyama H, et al. The superiority of ratio-based lymph node staging in gastric carcinoma. Ann Surg Oncol 2002;9:27 34. 18. Nwogu CE, Groman A, Fahey D, et al. Number of lymph nodes and metastatic lymph node ratio are associated with survival in lung cancer. Ann Thorac Surg 2012;93:1614 1619; discussion 1619. 19. Kelsey CR, Marks LB, Hollis D, et al. Local recurrence after surgery for early stage lung cancer: an 11-year experience with 975 patients. Cancer 2009;115:5218 5227. 20. Hoffman PC, Mauer AM, Vokes EE. Lung cancer. Lancet 2000;355:479 485. 21. Arriagada R, Bergman B, Dunant A, Le Chevalier T, Pignon JP, Vansteenkiste J; International Adjuvant Lung Cancer Trial Collaborative Group. Cisplatin-based adjuvant chemotherapy in patients with completely resected non-small-cell lung cancer. N Engl J Med 2004;350:351 360. 22. Chin CC, Wang JY, Yeh CY, Kuo YH, Huang WS, Yeh CH. Metastatic lymph node ratio is a more precise predictor of prognosis than number of lymph node metastases in stage III colon cancer. Int J Colorectal Dis 2009;24:1297 1302. Copyright 2013 by the International Association for the Study of Lung Cancer 435