Clinicopathologic predictors of metastasis of different regional lymph nodes in patients intraoperatively diagnosed with stage-I non-small cell lung cancer

Background Selection of the best lymph node for dissection is a controversial topic in clinical stage-I non-small cell lung cancer (NSCLC). Here, we sought to identify the clinicopathologic predictors of regional lymph node metastasis in patients intraoperatively diagnosed with stage-I NSCLC. Methods A retrospective review of 595 patients intraoperatively diagnosed as stage I non-small-cell lung cancer who underwent lobectomy with complete lymph node dissection was performed. Univariate and multivariable logistic regression analysis was performed to determine the independent predictors of regional lymph node metastasis. Results Univariate logistic regression and multivariable analysis revealed three independent predictors of the presence of metastatic hilar lymph nodes, five independent predictors for lobe specific mediastinal lymph nodes, two independent predictors for lobe nonspecific mediastinal lymph nodes and two independent predictors for skipping mediastinal lymph nodes. Conclusions A complete mediastinal lymph node dissection may be considered for patients suspected of nerve invasion and albumin (> 43.1 g/L) or nerve and vascular invasions. Lobe-specific lymph node dissection should probably be performed for patients suspected of pulmonary membrane invasion, vascular invasion, CEA (> 2.21 ng/mL), and tumor (> 1.6 cm) in the right lower lobe or mixed lobes. Hilar lymph node dissection should probably be performed for patients suspected of having bronchial mucosa and cartilage invasion, vascular invasion, and CEA (> 2.21 ng/mL). Electronic supplementary material The online version of this article (10.1186/s12885-019-5632-2) contains supplementary material, which is available to authorized users.


Background
Patients diagnosed with stage-I non-small cell lung cancer (NSCLC) are most likely to be cured by surgical radical resection. Lymph node dissection is an important part of this procedure that can improve the prognosis of the patients in stage-I [1]. However, selection of the best strategy for lymph node dissection remains controversial.
In general, lymph nodes with short-axis diameters of > 1 cm as seen on CT scan are considered by the radiologists to represent metastasis when other reasons causing lymph node enlargement, such as chronic inflammation and tuberculosis, are excluded. Unfortunately, the diagnostic accuracy of CT scan for the preoperative lymph node stage is only 45-79% [2][3][4][5][6]. Also, 12-17% of patients pathologically diagnosed as N2 are preoperatively considered as N0. Skipping metastasis is also found in a part of these patients, as their CT scan results revealed lymph nodes with short-axis diameters of < 1 cm [4,5,7]. With technological developments, many methods, such as positron emission tomography, mediastinoscopy, and endoscopic ultrasound-guided fine-needle aspiration, are used to make accurate diagnosis when the surgeon ambiguously confirms preoperative lymph node metastasis. However, owing to the invasive nature of the procedure and the associated expenses, these diagnostic methods could not be routinely used for screening patients with clinical stage-I disease. Also, these procedures yield a considerable number of false-negative results and complications [8][9][10].
Although systematic nodal dissection can guarantee an accurate pathological nodal (pN) staging with a sufficient quantity of lymphatic tissue, the occurrence of postoperative complications will increase. For early stage patients, getting accurate patterns of lymph node dissection will decrease the postoperative complications and speed-up the patient recovery.
Regional lymph nodes attract attention for less invasive intraoperative lymph node dissection in early stage patients. We can classify the lymph nodes into four regional lymph nodes: Interlobar lymph node, Hilar lymph nodes, Lobe-specific mediastinal lymph nodes and Lobe non-specific mediastinal lymph nodes. Skipping mediastinal metastasis is defined as the metastasis of lobe nonspecific mediastinal lymph nodes, and it is confirmed pathologically by the absence of lobe specific mediastinal lymph node metastasis. The concept of lobe specific mediastinal lymph node is based on the lobe specificity of the lymphatic spread, and the characteristic lymph nodal metastasis patterns could be derived from different primary tumor locations [11,12]. Surgeons can design an accurate strategy for lymph node dissection according to the regulation of regional lymph node metastasis. However, each patient exhibits different clinical and pathological characteristics. Several studies have demonstrated that the incidence of lymph node metastasis differs according to individual clinical parameters and histologic components within the tumor [13][14][15][16][17]. This patient heterogeneity finally affects the pattern of regional lymph node metastasis in early stage lung cancer.
The goal of this study was to identify the clinicopathologic characteristics that could predict the differences in metastasis among the various regional lymph nodes and to discuss the patterns for patients intraoperatively diagnosed with stage-I NSCLC. These clinicopathologic predictors will probably provide surgeons with useful information to select the appropriate lymph node dissections, especially for early stage patients.

Patient selection
A total of 595 patients who consecutively underwent surgical resection for primary lung cancer at our hospital from January 2015 to December 2017 were reviewed retrospectively. The records of patients intraoperatively diagnosed with stage-I NSCLC who underwent lobectomy or segmentectomy with complete lymph node dissection as per the nomenclature were selected for this study. All patients met the criteria for stage-I NSCLC based on the new International Staging System for NSCLC (National Comprehensive Cancer Network (NCCN) Guidelines Version 3.2014: Staging Non-Small Cell Lung Cancer) [12]. We excluded patients who exhibited any one of the following conditions: 1) preoperative tumor size > 4 cm and lymph node > 1 cm at the largest diameter on CT imaging or evidence of distant metastasis; 2) preoperative chemotherapy or radiotherapy; 3) previous or coexistent tuberculosis or malignant disease; 4) complete lymph node dissection that did not meet the current standards (i.e., all lymph node stations, including right-hand stations 2-4 and 7-9 and left-hand stations 2-9); 5) synchronous lung cancers, or 8) intraoperative frozen rapid pathological results depicting tumor size > 4 cm in the largest diameter.
Patients were preoperatively assessed with chest X-ray, chest and upper abdominal CT scan, brain magnetic resonance imaging, and bone scintigraphy. CT scan was used for preoperative N-staging. The approach for primary lung cancer resection was video-assisted thoracic surgery.
Tissue specimens contained pulmonary nodules and lymph nodes. Pulmonary nodules were analyzed using rapid frozen section in the pathology department of our medical center. The remaining nodules and lymph nodes were fixed using 10% formalin, and conventional formalin-fixed, paraffin-embedded pathological tests were performed.

Statistical analysis
The baseline patient characteristics were summarized in percent for categorical variables. The significance of associations with the outcome of lymph nodal metastases was first evaluated using a univariate logistic analysis (P < 0.20). These significant variables were further analyzed by multivariable analysis as independent predictors for lymph node metastasis (P < 0.05). Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated. Statistical analyses were performed using STATA software, release 13.
Multivariable analysis of clinicopathologic characteristics associated with metastasis of different regional lymph nodes For the interlobar lymph node, multivariate analysis of the 8 risk factors obtained from univariate analysis suggested that none of them were significant predictors of interlobar lymph node metastasis ( Table 2).
For the lobe specific mediastinal lymph node, multivariate analysis of the 10 risk factors resulting from univariate analysis indicated that only the maximum diameter of the tumor (≤1.6 cm vs. > 1.6 cm, OR = 3.18,      (Table 4). For the lobe nonspecific mediastinal lymph node, multivariate analysis of the 4 risk factors obtained from univariate analysis revealed that only vascular (absent vs. present, OR = 4.89, 95% CI 1.78-13.40; P = 0.002) and nerve invasions (absent vs. present, OR = 4.73, 95% CI 1.05-21.35; P = 0.043) were the 2 independent predictors associated with the presence of metastasis (Table 5).

Discussion
Evaluation of regional lymph node metastasis is important for surgeons to determine the treatment and prognosis [18]. Accordingly, regional lymph node maps have been created to standardize the assessment of metastasis. In these maps, lymph nodes are labeled using a system of numerical levels and assigned names based on their anatomical location [19,20]. The International Association for the Study of Lung Cancer (IASLC) lymph node map is employed in the eighth edition of the TNM staging system [21]. According to the sequence of the lymph node map, lung cancer cells initially spread to the ipsilateral interlobar lymph nodes, then to the hilar lymph nodes, and finally to the mediastinal lymph nodes.
The concept of lobe specific mediastinal lymph nodes is based on the lobe specificity of the lymphatic spread [22]. In literature, lobe specific MLNs have been defined as 2R, 3, and 4R for the right upper lobe; 3, 7, and 8 for the right lower lobe; 4 L, 5, and 7 for the left upper lobe; and 4 L, 7, and 8 for the left lower lobe. However, Kotoulasa et al. and Shapiro et al. proposed a simpler pattern. Right upper lobe tumors mainly metastasize to 4R, right middle lobe to 4R and 7, right lower lobe to 7, left upper lobe to 5, and left lower lobe to 7 and 9 [11,12]. An analysis of the recent literature led to our definitions of lobe specific lymph nodes: 2, 3, and 4 for the right upper lobe; 4 and 7 for the right middle lobe; 7 and 8 for the right lower lobe; 5 and 7 for the left upper lobe; and 7, 8, and 9 for the left lower lobe.
A complete mediastinal lymph node dissection which removes all ipsilateral mediastinal lymph nodes [23], can provide more accurate pathological staging and improved clinical outcomes for some patients. This approach is considered a standard surgical treatment for patients diagnosed preoperatively with mediastinal lymph node metastases. However, complete mediastinal lymph node dissection is not considered a routine surgical treatment for patients with stage-I NSCLC because of the increased incidence of postoperative complications including increase in blood loss, median operative time, total chest-tube drainage and occurrence rate of chylothorax. The rapid pathological results would help surgeons to make decisions about which patterns should be performed; wedge resection, segmentectomy, or lobectomy. However, surgeons do not know which pattern should be chosen for lymph node dissection and we need some guidance from clinical research. Mark Shapiro et al. further demonstrated the importance of lobe specific MLN regarded as sentinel lymph nodes in mediastinal position in the surgical treatment of early stage lung cancer [12]. Each patient exhibits different clinicopathologic characteristics that determine the risk for regional lymph node metastasis in early stage lung cancer. We attempted to identify the risk factors to predict lymph node metastasis and allow surgeons to make appropriate decisions on the extent of the dissection. For some early patients, surgeon can remove regional lymph nodes such as lobe specific MLN that are most likely to contain metastases and avoid unnecessary systemic complete lymph nodes dissection in order to accelerate patients' postoperative recovery.
First, we used univariate analysis to ascertain the associations between clinicopathologic factors and regional lymph node metastasis. The results disclosed that sex (male patients), maximum diameter of the tumor (> 1.6 cm), position (right lower lobe, left lower lobe, and bilateral mixed lobes), pulmonary membrane invasion, bronchial mucosa and cartilage invasion, vascular invasion, CEA (> 2.21 ng/mL), and albumin (≤43.1 g/L) were the 8 significant risk factors associated with the presence of metastatic interlobar lymph nodes.
Sex (male patients), maximum diameter of the tumor (> 1.6 cm), position (right lower lobe, right middle lobe, and ipsilateral mixed lobes), tumor differentiation, pulmonary membrane invasion, bronchial mucosa and cartilage invasion, vascular invasion, nerve invasion, CEA (> 2.21 ng/mL), and albumin (≤43.1 g/L) were the 10 significant risk factors associated with the presence of metastatic hilar lymph nodes.
Sex (male patients), maximum diameter of the tumor (> 1.6 cm), position (right lower lobe, and right middle lobe), tumor differentiation, pulmonary membrane invasion, bronchial mucosa and cartilage invasion, vascular invasion, nerve invasion, CEA (> 2.21 ng/mL), and albumin (≤43.1 g/L, 11.5%) were the 10 significant risk factors associated with the presence of metastatic lobe specific mediastinal lymph nodes.
Tumor differentiation, bronchial mucosa and cartilage invasion, vascular invasion, and nerve invasion were the 4 significant risk factors associated with the presence of metastatic lobe nonspecific mediastinal lymph nodes.
Age (≤60), sex (female patients), nerve invasion, and albumin (> 43.1 g/L) were the 4 significant risk factors associated with the presence of metastatic skipping mediastinal lymph nodes.
Furthermore, multivariate analysis of these factors identified using univariate analysis suggested that all the risk factors were not significant predictors of interlobar lymph node metastasis.
Only bronchial mucosa and cartilage invasion, vascular invasion, and CEA (> 2.21 ng/mL) were the three independent predictors associated with the presence of metastatic hilar lymph nodes. Therefore, when patients are suspected of having bronchial mucosa and cartilage invasion, vascular invasion, and CEA (> 2.21 ng/mL), hilar lymph node dissection should probably be performed.
Only maximum diameter of the tumor (> 1.6 cm), position (right lower lobe, ipsilateral mixed lobes, and bilateral mixed lobes), pulmonary membrane invasion, vascular invasion, and CEA (> 2.21 ng/mL) were the 5 independent predictors associated with the presence of metastatic lobe specific mediastinal lymph nodes. Hence, when patients are suspected of pulmonary membrane invasion, vascular invasion, CEA (> 2.21 ng/mL), and tumor (> 1.6 cm) in the right lower lobe or mixed lobes, lobe specific lymph node dissection should probably be performed.
Only nerve and vascular invasions were the two independent predictors associated with the presence of metastatic lobe nonspecific mediastinal lymph nodes. Hence, when patients are suspected of nerve and vascular invasions, complete mediastinal lymph node dissection should probably be performed.
Only nerve invasion and albumin (> 43.1 g/L) were the two independent predictors associated with the presence of metastatic skipping mediastinal lymph nodes. Therefore, when patients are suspected of nerve invasion and albumin (> 43.1 g/L), complete mediastinal lymph node dissection should probably be performed.
These results demonstrate the possibility of changes to lymph node metastasis when a tumor invades different tissues. In early-stage metastasis, the tumor invades bronchial mucosa and cartilage, pulmonary membranes, and vascular tissue only. During this stage, the CEA level (> 2.21 ng/mL) is likely to be an important predictor indicating that the tumor began to metastasize from the lymphatic system. Therefore, hilar and lobe-specific mediastinal lymph nodes, which are most likely to become the first metastatic stations, should be surgically removed. In later stages of lymph node metastasis, when the tumor begins to invade vascular and neural tissues, the albumin level (> 43.1 g/L) is likely to be an important predictor that indicates and promotes skip metastasis. When there is an increased possibility of broad mediastinal metastases, a complete mediastinal lymph-node dissection is required to ensure that all suspected metastatic lymph nodes are removed.
However, our study has some limitations. This study was conducted at a single institution with retrospective methods and demonstrated the necessity of further prospective study. Further prospective study with multicenter trial should be performed to comprehensively evaluate clinicopathologic predictors of metastasis of different regional lymph nodes in patients intraoperatively diagnosed with stage-I non-small cell lung cancer.

Conclusions
After a comprehensive analysis of results concerning the different clinicopathologic factors, we conclude that complete mediastinal lymph node dissection should probably be performed for patients suspected of nerve invasion and albumin (> 43.1 g/L) or nerve and vascular invasions; lobe specific lymph node dissection should probably be performed for patients suspected of pulmonary membrane invasion, vascular invasion, CEA (> 2.21 ng/mL), and tumor (> 1.6 cm) in the right lower lobe or mixed lobes; hilar lymph node dissection should probably be performed for patients suspected of having bronchial mucosa and cartilage invasion, vascular invasion, and CEA (> 2.21 ng/mL).