Differences in the Prognostic Impact between Single-Zone and Multi-Zone N2 Node Metastasis in Patients with Station-Based Multiple N2 Non–Small Cell Lung Cancer
Article information
Abstract
Purpose
The International Association for the Study of Lung Cancer suggests further subdivision of pathologic N (pN) category in non–small-cell lung cancer (NSCLC) by incorporating the location and number of involved lymph node (LN) stations. We reclassified patients with the station-based N2b disease into single-zone and multi-zone N2b groups and compared survival outcomes between the groups.
Materials and Methods
This retrospective study included patients with pN2 NSCLC who underwent lobectomy from 2006 to 2019. The N2 disease was subdivided into four categories: single-station N2 without N1 (N2a1), single-station N2 with N1 (N2a2), multiple-station N2 with single zone involvement (single-zone N2b), and multiple-station N2 with multiple zone involvement (multi-zone N2b). LN zones included in the subdivision of N2 disease were upper mediastinal, lower mediastinal, aortopulmonary, and subcarinal.
Results
Among 996 eligible patients, 211 (21.2%), 394 (39.6%), and 391 (39.3%) were confirmed to have pN2a1, pN2a2, and pN2b disease, respectively. In multivariable analysis after adjustment for sex, age, pT category, and adjuvant chemotherapy, overall survival was significantly better with single-zone N2b disease (n=125, 12.6%) than with multi-zone N2b disease (n=266, 26.7%) (hazard ratio [HR], 0.67; 95% confidence interval [CI], 0.49 to 0.90; p=0.009) and was comparable to that of N2a2 disease (HR, 1.12; 95% CI, 0.83 to 1.49; p=0.46).
Conclusion
Prognosis of single-zone LN metastasis was better than that of multiple-zone LN metastasis in patients with N2b NSCLC. Along with the station-based N descriptors, zone-based descriptors might ensure optimal staging, enabling the most appropriate decision-making on adjuvant therapy for patients with pN2 NSCLC.
Introduction
In patients with non–small cell lung cancer (NSCLC), lymph node (LN) metastasis is considered as one of the most reliable prognostic indicators as it greatly impacts decision-making regarding medical and surgical treatment strategies [1]. According to the 8th edition of the TNM Classification of Malignant Tumors, pathologic N descriptors are divided into four categories, with no change from those included in the 7th edition of the TNM Classification of Malignant Tumors [2]. Nonetheless, the International Association for the Study of Lung Cancer (IASLC) proposed the further subdivision of pN1 and pN2 based on the number and location of metastatic LNs in patients with heterogeneous prognosis [2]. However, due to the intrinsic heterogeneity of N2 disease, the prognosis of patients in real-world clinical practice does not appear to be sufficiently explained by only three subclasses.
The 7th and 8th editions demonstrated prognostic relevance by quantifying nodal disease (subdividing the N2 category into N2a and N2b subcategories) only in the pathological stage owing to limitations in information quality. However, with the upcoming introduction of the 9th edition, which will incorporate subdivided N descriptors showing prognostic relevance in both clinical and pathological stages, the subdivided N descriptors may potentially yield even more meaningful results.
In addition to the LN station-based subclassification [3], the IASLC previously conducted an exploratory analysis to subdivided the pN1 and pN2 groups according to the LN zones for the 7th edition of TNM staging system [4]. It might be clinically valuable to incorporate the LN zone–based classification with the currently used LN station–based classification.
In this study, we adopted the LN zone–based classification to subdivide the currently used N2 descriptor, which is based on the LN stations, according to the number of involved zones in patients with pN2 NSCLC and compared long-term survival outcomes between the patients with single and multiple LN zone involvement.
Materials and Methods
1. Study population
Retrospective review of the medical records of 10,066 patients who underwent surgical resection from January 1, 2006 to December 31, 2019 at Asan Medical Center in Seoul, Korea identified a total of 1,378 patients with pN2 NSCLC. Among these 1378 patients, those with incomplete resection (R1 or R2 resection) (n=166), those who underwent exploratory thoracotomy or sublobar resection (biopsy, wedge resection, or segmentectomy) (n=68), those with distant metastasis, M1a, M1b and M1c (n=52), those who received preoperative therapy (chemo, radiation or chemo-radiation) (n=85), and those who died within 30 days after surgery (n=11) were excluded (S1 Fig.). The final cohort included 996 patients who underwent complete resection with systemic LN dissection except for those with incomplete LN dissection (< 6 resected LNs).
2. Preoperative staging and postoperative management
The details of diagnosis and therapeutic strategies for NSCLC in our institution have been described elsewhere [5]. Briefly, preoperative workup included routine blood tests, such as complete blood count and electrolyte and chemistry panels, chest X-ray, chest computed tomography (CT), pulmonary function test, current, and past medical history, surgical history, and physical examination. Additionally, 18F-fluorodeoxyglucose positron emission tomography, and cranial magnetic resonance scans were routinely performed following lung cancer diagnosis to check for metastasis. In particular, mediastinal LN biopsy was selectively performed using mediastinoscopy or endobronchial ultrasound for suspicious LNs observed in preoperative CT or positron emission tomography images, and the specific treatment strategy was determined by a multidisciplinary team in patients with confirmed N2 metastasis.
In patients with biopsy-proven clinical N2 disease, upfront surgery was performed in those with primary tumors which were resectable without pneumonectomy, in those with mediastinal LN metastasis limited to a single zone, and in those without evidence of extracapsular extension on preoperative imaging. The imaging criteria for exctracapsular extension included indistinct LN margins, irregular LN enhancement, infiltration into adjacent tissue, and central necrosis. Mediastinal LN dissection involved all nodes at stations 2R, 4R, 7R, 8R, 9R, 10R, and 11R for right-sided tumors. For left-sided tumors, the dissection encompassed nodes at stations 5L, 6L, 7L, 10L, and 11L (selectively with 4L, 8L, and 9L). All LNs, whether they were dissected during the surgical procedure or collected from the specimens at a later time, underwent a pathological examination. They were categorized based on their anatomical location using the numbering system outlined in the Mountain-Dresler modification of the American Thoracic Society (MD-ATS) guidelines. The pathological staging was conducted retrospectively, following the criteria specified in the 8th edition of the American Joint Committee on Cancer (AJCC).
The multidisciplinary team recommended adjuvant chemotherapy with a platinum-based regimen for all patients with pN2 disease, except for those over 75 years of age and those in poor physical condition. The treatment consisted of four cycles for 4-6 weeks after surgery. Since 2008, patients with epidermal growth factor receptor mutations who experienced recurrence after the first adjuvant chemotherapy received tyrosine kinase inhibitor treatment. Adjuvant radiotherapy was administered at a daily dose of 1.8 Gy, up to a total dose of 50.4 Gy, in patients with complete resection. Only those patients who completed treatment were considered to have received adjuvant therapy.
All patients were evaluated for wound infection and other short-term complications, such as pneumothorax, 2 weeks after surgery. Disease status was evaluated in an outpatient setting every 3-6 months in the first 2 years and every 6-12 months thereafter, and records were collected through the data. Chest CT was also performed in patients with suspicious recurrence, and the treatment approach, and the possibility of chemotherapy were assessed in consultation with an oncologist in patients with confirmed recurrence.
3. Subdivision of N descriptors and LN zones
In the present study, pN2 was subdivided into single-station pN2 (pN2a) and multiple-station pN2 (pN2b) based to the number of included LN stations (Table 1). pN2a was further subdivided into pN2a1, defined as pN2a without pN1 involvement, and pN2a2, defined as pN2a with pN1 involvement, according to the presence of N1.
Recommended N2 subclassification
For subdivision of pN2b, LN anatomic locations were classified into the following five zones based on the report by Rusch et al. [4]: upper (level 1-4 LNs), aortopulmonary (level 5-6 LNs), subcarinal (level 7 LNs), lower (level 8-9 LNs), and hilar (level 10-11 LNs) and peripheral (level 12-14 LNs). Next, pN2b with involvement of multiple LN stations in a single zone was defined as single-zone N2b and pN2b with involvement of multiple LN stations in multiple zones was defined as multi-zone N2b.
4. Outcome measures and statistical analyses
Clinical information was collected through January 2022 by retrieving data from institutional electronic medical databases and electronic chart reviews. The outcomes of interest were mortality and recurrence. Overall survival (OS) was described as the time interval from the date of surgery to the date of death, which was recorded by reviewing death records in the Korean National Security Death Index Database. Freedom from recurrence (FFR) was calculated as the time between the date of lung cancer surgery and recurrence, whereas patients without recurrence were counted at the latest known recurrence-free time point.
Categorical variables were presented as percentages and frequencies, and continuous variables were expressed as medians with interquartile range. OS and FFR related to the N variables were estimated using the Kaplan-Meier method and evaluated using the log-rank test. Prognostic factors affecting OS and FFR were identified using the Cox proportional hazards model and evaluated with univariate and multivariate analyses. The initial multivariate Cox model included independent variables that had a p-value of 0.05 or less in the univariate analysis. The final multivariate model was constructed using stepwise model selection. Cox proportional hazard analysis was employed to account for initial factors (such as age, sex, histological type, pathological tumor factor, and adjuvant therapy) and calculate hazard ratios (HRs) to compare the risk between consecutive N categories.
All reported p-values were two-sided, and a p-value of < 0.05 was considered statistically significant. All statistical analyses were performed using R statistical software ver. 3.4.00 (R Foundation, Vienna, Austria; http://www.R-project.org/).
Results
1. Baseline characteristics
The clinicopathologic characteristics of 996 patients enrolled in the study are summarized in Table 2. The median age was 62 years (interquartile range, 55 to 69 years), and there were 607 males (60.9%). Pathologically, 747 (75.0%) and 219 (22.0%) patients were diagnosed with adenocarcinoma and squamous cell carcinoma, respectively. Lobectomy was the most common surgical approach, employed in 888 patients (89.2%), followed by bilobectomy in 65 patients (6.5%) and total pneumonectomy in 43 patients (4.3%). In the overall cohort, 211 (21.2%), 394 (39.6%), and 391 (39.3%) patients had pN2a1, pN2a2, and pN2b disease, respectively. Further, 125 (12.6%) and 266 (26.7%) of the 391 patients with pN2b disease had single-zone and multi-zone pN2b disease, respectively. After surgery, 854 patients (85.74%) received adjuvant chemotherapy (Table 2). The distribution of clinical N category for each pN category subgroup is also presented (S2 Table).
Clinicopathologic characteristics of patients (n=996)
Table 3 shows the comparison between the patients with single-zone and multi-zone pN2b disease. Briefly, age, sex, pT status, number of comorbidities, surgical approach, histologic diagnosis, postoperative adjuvant therapy, and time variable were not significantly different between the groups. Lobectomy was performed significantly less frequently in patients with multi-zone pN2b disease than in those with single-zone pN2b disease (86.1% vs. 96.8%, p=0.002).
Clinicopathologic characteristics of patients with multiple-station N2 involvement (n=391)
2. Analysis of OS and FFR based on pathologic N descriptors
The median follow-up was 38 months (interquartile range, 18 to 64 months). At the end of follow-up period, 521 of the 996 patients (52.3%) had died whereas and the remaining 475 patients (47.7%) were alive or censored.
The analysis of patients according to the N descriptors revealed significant differences in 5-year OS between the N2a1 and multi-zone N2b groups (60.3% vs. 34.6%, p < 0.001) and between the N2a2 and multi-zone N2b groups (51.0% vs. 34.6%, p < 0.001). Further, there was a significant difference in 5-year OS between the single-zone and multi-zone N2b groups (47.0% and 34.6%, respectively; p=0.007) (Fig. 1A).
Kaplan-Meier estimates of overall survival (A) and freedom from recurrence (B) in patients with pathologic N2 subdivided into N2a1, N2a2, multi-zone N2b, and single-zone N2b. CI, confidence interval; MST, median survival time.
The FFR of the N2a1 group (47.4%) was significantly different from that of the N2a2 (36.2%), single-zone N2b (35.5%), and multi-zone N2b (17.7%) groups (p=0.012, p=0.015, and p < 0.001, respectively). Additionally, the recurrence rate was significantly different between the N2a2 and multi-zone N2b groups (36.2% vs. 17.7%, p < 0.001) and between the single-zone and multi-zone N2b groups (35.5% vs. 17.7%, p=0.001) (Fig. 1B).
3. Multivariable Cox regression analysis to compare the prognostic value of N descriptors based on LN zones
Further analysis was performed to evaluate the prognostic value of the zone-based pN2a1, pN2a2, single-zone pN2b, and multi-zone pN2b subdivisions. Univariable and multivariable Cox proportional hazards analyses were performed to identify independent prognostic factors in patients with pN2 NSCLC. In univariable analysis, the sex, age, pathologic node factor, histological structure, pathologic tumor factor and adjuvant therapy was significant prognostic factors. (Table 4). The prognostic impact of these subdivided descriptors was evaluated with multivariate analysis, with adjustment of variables such as age, sex, pT factor, and adjuvant therapy (Fig. 2). Within the pN2a group, the OS was significantly different between the pN2a1 and pN2a2 groups (HR, 1.34; 95% confidence interval [CI], 1.04 to 1.72; p=0.023) (Fig. 2A). Additionally, the OS was significantly different between the single-zone and multi-zone pN2b groups among the patients with pN2b disease (HR, 0.67; 95% CI, 0.49 to 0.90; p=0.009) (Fig. 2F). Finally, the OS was not significantly different between the N2a2 and single-zone N2b groups (HR, 1.12; 95% CI, 0.83 to 1.49; p=0.46) (Fig. 2D).
Univariate and multivariate Cox analysis in patients with pN2 NSCLC
Multivariate Cox proportional hazards regression adjusted for age, sex, bovine serum albumin, and adjuvant therapy in patients with pathologic N2 subdivided N2a1, N2a2, multi-zone N2b, and single-zone N2b. (A) N2a1 vs. N2a2. (B) N2a1 vs. single-zone N2b. (C) N2a1 vs. multi-zone N2b. (D) N2a2 vs. single-zone N2b. (E) N2a2 vs. multi-zone N2b. (F) Single-zone N2b vs. multi-zone N2b. CI, confidence interval; HR, hazard ratio.
Discussion
In a recent study comparing the discrimination ability of N descriptors subdivided based on LN stations and zones, we were the first to report that the prognosis of station-based pN2b disease differed depending on whether multiple N2 nodes were confined within a single zone [5]. However, this finding was based on analyses conducted without adjustment for significant covariates such as age, sex, and tumor stage. Moreover, considering the aim and scope of the study, in-depth analysis for this finding was skipped. The current study incorporating rigorous adjustment for significant covariates provided further evidence that the prognosis was better in patients with single-zone LN metastasis than in those with multi-zone LN metastasis among those in the same station-based N2b disease category according to the 8th edition of TNM staging system.
It should be noted that two main factors might affect this findings. First, inaccurate LN harvest contributes to the difference in prognosis between single- and multi-zone LN metastasis within the station-based N2b disease. Previously, the IASLC proposed an LN map, which provided definite anatomical landmarks for all LN stations, to achieve uniformity, and to promote future analyses of a planned prospective international database [6]. However, in real-world settings, anatomical ambiguity continues to exist during LN dissection and labeling. For example, during systematic LN dissection, some LNs are often bundled together with surrounding adipose tissue and harvested en bloc. In such cases, the specimens might be labeled as LN 2R&4R, LN 5&6, or LN 8&9. Otherwise, these specimens might be arbitrarily resected and separately labeled without consideration of the anatomical landmarks. In addition, incorrect labeling might occur during the harvest of LNs that are on or across anatomical boundaries. Thus, N2 metastasis involving a single station might be overdiagnosed as multi-station N2 metastasis whereas two ambiguous metastatic N2 nodes within the same zone might not be diagnosed as multi-station N2 metastasis.
Second, the pattern of lymphatic drainage in lungs might contribute to the observed differences in the prognosis. Lung cancer, especially NSCLC, exhibits a lymphophilic behavior [7]. In lungs, lymphatic drainage toward the hilar LNs and eventually to the mediastinal LNs occurs via parenchymal lymphatic vessels through a well-characterized peribronchial pathway [8-10], which involves the segmental, or intersegmental drainage of lung lymphatics through the same peribronchial pathway [11]. Several studies have reported that the LN metastasis pattern is affected by the primary tumor location, which serves as the basis for the lymphadenectomy approach that is specific to each lobe, i.e., lobe-specific LN dissection [12-14]. While tumors in the upper lobe tend to drain into upper mediastinal LNs (stations 2-6), those in the middle and lower lobes tend to drain into lower mediastinal LNs (stations 7-9) [15]. Therefore, it is possible that LNs located in the same zone are influenced by the same lymphatic drainage flow and that LN metastasis in different zones might indicate a more advanced disease status than that is occurring within the same zone. However, it was not feasible to test this theory in patients with station-based multi-zone N2 NSCLC.
On the other hand, the subcarinal LN zone contains only LN No. 7, which might have led to a bias in the differences in prognostic outcomes found between the LN zones in the present study. Metastasis of LN No. 7 cannot be classified as multiple-zone N2b NSCLC. The subcarinal LN zone has been reported as a negative prognostic factor in patients with pN2 NSCLC [16-19], and it is possible that the difference in prognosis between single-zone and multi-zone N2b NSCLC might be due to the metastasis of LN No. 7. Further comparison of patients with N2 disease categorized according to the presence of LN No. 7 revealed no significant differences in prognosis among the patients with N2a1 (p=0.07), N2a2 (p=0.44) and multi-zone N2b (p=0.93) NSCLC (S3 Fig.). Additionally, there was no significant difference in prognosis when the patients were categorized depending on the location of N2 metastasis, irrespective of the N1 metastasis (p=0.32 for N2a1 and p=0.08 for N2a2) (S4 Fig.).
Particularly for patients with N2, who exhibit a less favorable prognosis, the results regarding the effectiveness of postoperative radiotherapy (PORT) are inconsistent. While it has a significant effect on local control, it is known that there is no significant difference in OS due to cardiogenic and pulmonary toxicity [20]. However, a consensus from various recently published review articles and meta-analyses suggests that selectively administering PORT to subgroups in N2 cases can be beneficial. In this context, N2b is considered a more favorable subgroup for PORT compared to N2a [21]. Based on the results of this study, we cautiously suggest that PORT could be more strongly recommended for multi-zone N2b patients. However, additional analysis was not possible in this study due to the limited number of related patients, and this should be validated through large-scale studies in the future.
The present study has several limitations. First, this was a retrospective, single-center study inherently contains selection bias. Second, the study cohort included only patients who underwent surgical treatment, and the study findings might not be applicable to all patients with NSCLC. Third, this study includes patients who underwent surgery over a long period of 14 years. There were no significant changes in preoperative and postoperative strategies and surgical techniques. However, due to the absence of a standardized protocol, the decision to administer adjuvant therapy was made on an individual basis through a multidisciplinary team approach, which could introduce selection bias. Therefore, we conducted a multivariable Cox analysis to account for potential bias, considering several factors. This is also a limitation of retrospective research. Forth, we aim to enhance our study by including additional details on the primary tumor, such as the maximum standardized uptake value (SUVmax) from positron emission tomography (PET)/CT scans, the consolidation-to-tumor ratio, and epidermal growth factor receptor (EGFR)/anaplastic lymphoma kinase (ALK) mutation status. However, obtaining this information has been challenging. Many PET scans were performed at external hospitals or the results from our reports were missing, making retrospective data collection difficult. Additionally, among N2 patients with most tumors being solid masses, a few percentage were T1a or T1b (n=79, 7.9%), so we did not emphasize the consolidation-to-tumor ratio. Regarding EGFR/ALK mutations, driver mutation tests were not conducted before 2012, and after that, testing was primarily done for patients who experienced recurrence due to insurance issues, resulting in many patients lacking this information. Overall, these factors hindered our ability to include the desired details in the study. Finally, our study, conducted between 2006 and 2019, predates the recent introduction of systemic endobronchial ultrasound (EBUS), which significantly impacts the accuracy of invasive mediastinal staging. The importance of EBUS has been increasingly emphasized in recent years, allowing for systematic evaluation with the same bronchoscope, thereby improving disease detection rates [22]. In this respect, the considerable number of initial patients included in this study have limitations due to incomplete clinical evaluation.
We demonstrate that single-zone N2b node metastasis is a favorable prognostic factor in patients with pN2b NSCLC. This finding could help clinicians enact proper treatment strategies, such as PORT, as an aggressive treatment option for these patients. In addition, reflecting this finding would enable a more precise evaluation of treatment effects when conducting prospective studies in these patients.
Electronic Supplementary Material
Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).
Notes
Ethical Statement
The need for informed consent from patients was waived by institutional review board from Asan Medical Center (approval no. 2022-1322, approval date : September 22, 2022).
Author Contributions
Conceived and designed the analysis: Kim S, Lee GD, Choi S, Kim HR, Kim YH, Kim DK, Park SI, Yun JK.
Collected the data: Kim S.
Contributed data or analysis tools: Kim S, Lee GD, Choi S, Kim HR, Kim YH, Kim DK, Park SI, Yun JK.
Performed the analysis: Kim S.
Wrote the paper: Kim S, Yun JK.
Supervision: Yun JK.
Conflict of Interest
Conflict of interest relevant to this article was not reported.