According to the European Society of Thoracic Surgeons (ESTS), mediastinal lymph nodes are processed in 5 ways: (1) selective lymph node biopsy; (2) lymph node sampling, including systemic lymph node sampling (SLNS); (3) systematic lymph node dissection (SLND); (4) lobe-specific lymph node dissection (L-SLND); (5) extended lymph node dissection [7]. For resectable lung cancer, SLND or SLNS was able to provide accurate staging and guide the choice of treatment modality [8]. For the presence of pN2 metastasis in clinical early-stage NSCLC patients, the NCCN also recommends performing SLND or SLNS [9, 10].

The lymph node drainage pattern of stage IA NSCLC was temporarily uncertain whether it follows a lobe-specific pattern, and there were skip and multiple patterns of lymph node metastasis in NSCLC, some scholars believe that SLND remains the preferred modality in the surgical treatment of stage IA NSCLC [5]. However, several studies have shown that SLND for clinical early-stage NSCLC has similar survival to L-SLND and SLNS [2, 11]. Huang et al. suggested that early-stage NSCLC patients who had mediastinal lymph node dissection versus mediastinal lymph node sampling not only had similar overall survival but also similar rates of local recurrence and distant metastasis [12]. ACOSOG Z0030 results from this large prospective clinical trial for patients with pT1-2N0-1 disease who underwent SLND versus those who underwent SLNS showed no difference in recurrence, survival, and surgical complication rates [13]. The results above also indirectly demonstrated that the odds of early-stage NSCLC patients’ pN2 metastasis were smaller.

At present, there is no uniform standard for lymph node resection in early-stage NSCLC, and we tried to search the predictors associated with pN2 metastasis by preoperative clinical characteristics and intraoperative frozen pathology. Cheng et al. showed that the rate of cI NSCLC patients’ pN2 metastasis was 14.2%, and pN2 metastasis was correlated with tumor histology [14]. Yukinori et al. suggested that lymph node metastasis in patients with cIA stage NSCLC is associated with CEA [15]. Previous studies from our center showed that the clinically early-stage NSCLC patients’ lymph node metastasis rate was 6.8%, and the size of solid tumor component and preoperative CEA were independent predictors of lymph node metastasis, which was probably caused by the fact that most patients had ground-glass nodules [16]. Therefore, in this study, we collected data from patients with pure solid NSCLC at stage cIA and found that the rate of pN2 metastasis was 13.5%, and preoperative tumor diameter, CEA, and NSE were independent predictors of pN2 metastasis. Our research results showed that combining three preoperative independent predictive factors to form a new factor Y, pN2 metastasis was shown in only 2.1% (3 patients) of patients in the low-risk group compared with 17.8% (69 patients) of patients in the high-risk group. In this study, the pathological status of the station 10 lymph node was another independent predictor of pN2 metastasis. However, this study used the relevant data of station 10 lymph nodes after operation. If it is used to predict pN2 metastasis, it needs to clarify the situation of all station 10 lymph nodes intraoperatively. At the same time, our results suggested that the station 10 lymph node-negative group (low-risk group) had 47 (9.8%) patients with pN2 metastasis, which might not be an ideal factor to exclude patients without pN2 metastasis. Predictor Y, on the other hand, readily available without increasing the burdens of physicians and patients, might be a better predictor for excluding pN2 metastasis.

Our study found a low rate of lymph node metastasis at stations 8 and 9 in cIA pure-solid NSCLC patients, including 2 patients (0.4%) with station 8 lymph node metastasis and 6 patients (1.1%) with station 9 lymph node metastasis. Asamura and Yang et al. reported at different times that upper lobe lung cancer was prone to superior mediastinal lymph node metastasis; lower lobe lung cancer was prone to inferior mediastinal lymph node metastasis [17, 18]. Lobe-specific lymph node dissection refers to upper lobe lung cancer with resection of the superior mediastinal lymph nodes and lower lobe lung cancer with resection of the subcarinal and inferior mediastinal lymph nodes. Our data suggest that pN2 metastasis largely follows a lobe-specific pattern, with a higher rate of superior mediastinal lymph node metastasis in upper lobe lung cancer; there was a high rate of the subcarinal lymph nodes in lower lobe lung cancer; the right middle lobe lung cancer has a high rate of the subcarinal lymph node metastasis, but also to the superior mediastinal lymph nodes. We believe that the extent of lobe-specific lymph node dissection for middle lobe lung cancer should include the subcarinal and superior mediastinal lymph nodes. Postoperative survival after L-SLND in early-stage NSCLC patients was similar to SLND, and therefore L-SLND was feasible for clinical early-stage NSCLC patients with less pN2 metastasis [11]. The Y low-risk group only had 3 (2.1%) patients who presented pN2 metastasis (station 7 lymph node metastasis in 2 patients with right lower lobe, station 5 and station 6 lymph node metastasis in 1 patient with left upper lobe), and all of the patients were lobe-specific lymph node metastasis. Thus, we considered that lobe-specific lymph node sampling, or SLNS could be adopted for this group of patients. On the other hand, the Y high-risk group had 79 (17.8%) patients with pN2 metastasis, including 15 (3.9%) patients with the presence of non-lobe-specific lymph node metastasis. Thus, patients in the Y high-risk group might benefit from receiving SLND. If intraoperative frozen pathology indicates metastasis in station 10 lymph nodes, SLND would be recommended.

Intraoperative frozen pathology cannot confirm the degree of tumor differentiation and pleural invasion of each specimen in a short time. Thus, the factor cannot be studied as a potential predictor of pN2 metastasis in this test. Pulmonary lymphatic drainage is mainly composed of two major systems, namely interstitial lymphatic vessels and parenchymal lymphatic vessels. There are abundant anastomotic branches between the two major lymphatic drainage systems, and the lymph of lung segments can flow directly to mediastinal lymph nodes [19]. We divided the location of the tumor in the lobe into outer 1 / 3, middle 1 / 3, and inner 1 / 3. The results showed that pN2 metastasis was not correlated with the location of the tumor in the lobe, which did not exclude the correlation with the lymph reflux pathway. The patients included in this study did not have a clear pathological diagnosis of lung tumor before surgery, so the tumor markers related to NSCLC and small cell lung cancer (SCLC) were detected. It is well known that NSE is a relatively specific tumor marker for SCLC. In the course of our study, NSE was found to be an independent predictor of mediastinal lymph nodes in cIA stage pure-solid NSCLC. Previous studies have shown that NSE increases in patients with advanced lung cancer of different pathological types [20]. At the same time, some studies have shown that NSE is associated with the degree of pathological invasion, pathological stage, and prognosis of NSCLC [21, 22]. This phenomenon might be caused by the neuroendocrine characteristics of some patients with NSCLC [23]. More studies are needed to verify whether NSE, a tumor marker of SCLC, is associated with NSCLC.

Our findings suggested a new modality for intraoperative lymph node dissection in cIA pure-solid NSCLC patients, but this study still has some limitations. First, as a single center administrative database, we could hardly capture all the subtle factors, some of which might be critical to the results. Therefore, further multicenter randomized trials are needed to validate these results. Second, in this study, we did not incorporate the relevant data on lymph node micro-metastasis, so it still needs further investigation on whether our findings could apply to patients with lymph node micro-metastasis.