Risk Factors for Lymph Node Metastasis in Clinical Stage IA3 Lung Adenocarcinoma Patients

Background: lymph node metastasis is a poor prognostic factor for lung cancer; however, the risk of lymph node metastasis has not been claried yet, so it is controversial to conduct systematic lymph node dissection for early lung cancer. Therefore, this study aimed to focus on analyzing the predictive factors for lymph node metastasis in patients with clinical stage IA3 lung adenocarcinoma. Methods: Our study group retrospectively analyzed all surgical patients admitted to our hospital from January 1, 2017 to June 2021, and these patients were considered having stage IA3 lung adenocarcinoma. A total of 334 patients underwent lobectomy combined with systematic lymph node dissection. Univariate and multivariate logistic regression analysis were adopted to predict the risk factors of lymph node metastasis. Results: Among the 334 patients eligible for this study, the overall mediastinal lymph node metastasis rate was 15.27%. There were 45 cases of N1 metastasis and 11 cases of N2 metastasis, 5 cases had both N1 and N2 metastasis at the same time. The patients were divided into three groups according to consolidation tumor ratio (CTR) values (<0.25, 0.25-0.5, >0.5). The lymph node metastasis rates in each CTR group were 1.8% (2/112), 11.7% (17/145) and 41.6% (32/77), respectively. The mediastinal lymph node metastasis rate in patients with carcinoembryonic antigen (CEA>5ng/ml) was 57.89% (22/38). The receiver operating characteristic curve (ROC) showed that the area under the curve (AUC) of CTR, pathological type and CEA were 0.790 [95% condence interval (CI): 0.727 – 0.853(cid:0)P<0.001]; 0.800(cid:0)95% CI:0.735–0.865(cid:0)P<0.001(cid:0);0.682(cid:0)95% CI: 0.591–0.773, P<0.001(cid:0);respectively. Multivariate regression analysis showed that these listed factors were signicantly correlated with lymph node metastasis of clinical stage IA3 lung adenocarcinoma: CEA [Odds Ratio (OR)=3.05, P=0.016], CTR 0.25 to 0.5 (OR=14.12, P<0.017), CTR>0.5 (OR=7.75, P=0.015), micropapillary adenocarcinoma (OR=15.704, P<0.001), and solid adenocarcinoma (OR=8.971, P=0.001). Conclusions: CEA (>5ng/ml), histologic subtype and CTR (>0.25) are important predictors of lymph node metastasis in clinical stage IA3 lung adenocarcinoma, systematic lymph node dissection should be the prior choice for patients with clinical stage IA3 incorporated with risk factors. The lymph node dissection method in stage IA3 should be alternative from those in stage IA1 and IA2. lymph node metastasis between EGFR gene mutation and EGFR gene wild type.


Introduction
Currently, surgery is still the main standard treatment for early lung cancer, the lymph node metastasis being known by intraoperative lymph node dissection during operation however, anatomical lobectomy combined with systematic lymph node dissection for early lung cancer treatment is controversial within the academic eld 1 . The long-term follow-up results of JCOG0804 in Japan showed that most patients do not even need systematic lymph node dissection 2 , in addition, many studies have shown that the pathological types of adenocarcinoma in situ and microinvasive adenocarcinoma will not have lymph nodes 3,4 . In particular, Stage IA has been subdivided into stage IA1, IA2 and IA3 in the 8 th Edition International Association for the Study of Lung Cancer (IASLC), and the risk of lymph node metastasis in clinical stage IA1 and IA2 lung cancer is relatively low 5 , however, the risk factors of lymph node metastasis in IA3 lung adenocarcinoma are undetermined, and there are few studies related to discuss or clarify the risk. At present, the research results of stage IA3 in JCOG1211 in Japan have not been published yet 6 . Therefore, it is necessary to predict the risk factors of lymph node metastasis of clinical stage IA3 lung adenocarcinoma, in order to important clinical reference value for the choice of lymph node dissection in stage IA3 of lung adenocarcinoma.

Patients
The retrospective study was approved by the ethics committee of Wenzhou Central Hospital (L2021-12-42). As the study was retrospective, it was not necessary to obtain the written informed consent of each patient. During the period between January 2017 and June 2021, 334 patients with clinical stage IA3 lung adenocarcinoma were included in the study.
Inclusion criteria: preoperative clinical staging was based on TNM staging (IASLC 8 th Edition); Single tumor with diameter (>2-3cm); Malignant tumor without previous chemotherapy or radiotherapy treatment; Performance status of 0 or 1; Su cient organ function; Intraoperative pathology con rmed invasive adenocarcinoma; Preoperative examination excluded distant organ metastasis. Exclusion criteria: Incomplete case records; Severe emphysema, pulmonary brosis, interstitial pneumonia; history of severe heart failure, myocardial infarction, heart disease within the past 6 months; Pleural dissemination; The short axis of mediastinal lymph nodes was greater than 1cm or positron emission tomography-computer tomography (PET-CT) considered lymph node metastasis; Preoperative neoadjuvant radiotherapy or chemotherapy.

Methods
All patients had completed listed medical tests before operation, including contrast-enhanced thoracic CT, head magnetic resonance imaging, abdominal CT, echocardiography, pulmonary function and CEA. Some patients completed PET-CT examination. No puncture biopsy was performed before operation. Lobectomy combined with systemic lymph node dissection was performed, Systematic lymph node dissection for right-side tumors contained 2R, 4R, 7-12 stations, while 4-12 stations were required for left side tumors, the lymph node classi cation comply with the IASLC (8 th Edition). Age, sex, smoking, preoperative serum CEA level, CTR, histopathological subtypes (lepidic predominant, acinar predominant, papillary predominant, micropapillary predominant, solid predominant), epidermal growth factor receptor (EGFR) mutation, PET-CT, and spread through air spaces (STAS) are medical records and pathological data taken into account in the analysis.

Statistical analysis
Univariate analysis and multivariate logistic regression analysis were performed by SPSS (version 23.0, Inc, Chicago, IL), to identify the risk factors of lymph node metastasis. Univariate analysis used Fisher's exact test or X 2 test, and the categorical variables with univariate analysis P<0.1 were included in multivariate regression analysis, where P<0.05 was statistically signi cant.
After univariate factor analysis, the following 6 factors were identi ed as important predictors for mediastinal lymph node metastasis: CTR (p<0.001), PET-CT (p=0.013), CEA level (p<0.001), Pathologic type (p<0.001), EGFR mutation (p<0.001) and STAS (P=0.001). Age, gender and smoking were not identi ed as signi cant predictors (Table 1). Six predictive factors mentioned at the beginning were included in multivariate analysis, and the results showed that mediastinal lymph node metastasis had the following three vital risk factors: preoperative CEA level>5 (ng/ml) (OR: 3.05, 95% CI:

Discussion
Many studies have shown that mediastinal lymph node metastasis may occur even in early non-small cell lung cancer 7,8 . If lymph node metastasis can be identi ed before surgery, these patients will bene t from surgical resection after neoadjuvant therapy 9 . However, it is di cult to evaluate mediastinal lymph node metastasis by enhanced CT or PET-CT before operation. This study excluded patients with mediastinal lymph node metastasis considered by preoperative PET-CT. Nevertheless, 17.96% (46/256) of patients with mediastinal lymph node metastasis were not found by PET-CT before operation, which explained the limitations of PET-CT in evaluating mediastinal lymph node metastasis.
In this study, mediastinal lymph node metastasis occurred in any pathological subtype of IA3 lung invasive adenocarcinoma, but Ye et al 10 suggested that no mediastinal lymph node metastasis in stage IA lepidic predominant taking place, which may due to the small sample size. In our study, the total metastasis rate of mediastinal lymph nodes in IA3 stage was 15.27%, but Koike et al 11 reported that the overall mediastinal lymph node metastasis rate in patients with stage IA non-small cell lung cancer was only 7.5%, Ye et al 10 showed that the overall mediastinal lymph node metastasis rate in patients with stage IA lung adenocarcinoma was 10.6%, Ding et al 12 showed that only 8.5% of patients with stage IA had mediastinal lymph node metastasis. The latest research results in Japan show that the lymph node metastasis rate of stage IA lung adenocarcinoma is only 4.9% 13 . The above results showed that the proportion of mediastinal lymph node metastasis in IA3 stage was signi cantly higher than that in IA stage during clinical research, implying that the size of the tumor possibly affects the occurrence of mediastinal lymph node metastasis. The mediastinal lymph node metastasis rate in stage IA3 was higher than that stage IA, so the lymph node dissection method in stage IA3 should be different from that in stage IA1 and IA2, suggesting that further re nement of clinical staging may help to optimize mediastinal lymph node dissection.
Lee et al 14 have shown that the abnormal CEA level is closely related to the distant metastasis of nonsmall cell lung cancer. Our results showed that the mediastinal lymph node metastasis rate in patients with CEA (>5ng/ml) was 57.89% (22/38), however, Wang et al 15 reported mediastinal lymph node metastasis rate in patients with CEA (>5ng/ml) was 15.25%, and it was a risk factor for lymph node metastasis (OR=1.574, P=0.001). Ye et al 10 also showed that the lymph node metastasis rate of CEA (>5ng/ml) was 19.16%, and it was a risk factor for lymph node metastasis in stage IA lung adenocarcinoma (OR=3.923, P<0.001), Koike et al 11 revealed that preoperative CEA(>3.5ng/ml) was a risk factor for mediastinal lymph node metastasis of clinical stage IA non-small cell lung cancer. All the results mentioned above were consistent with our study (OR=3.05, P=0.016). In clinical stage IA3 lung adenocarcinoma, the mediastinal lymph node metastasis rate of CEA (>5ng/ml) was signi cantly higher than that in previous stage IA studies, it shows that preoperative CEA abnormality is an important risk factor for mediastinal lymph node metastasis of lung cancer, the lymph node dissection of stage IA3 lung adenocarcinoma should be alternative from those in stage IA1 and IA2, systemic lymph node dissection should be preferably choice for patients with stage IA3 lung adenocarcinoma with CEA (>5ng/ml). Previous JCOG series studies have shown that CTR will affect the long-term survival of lung cancer patients 2,16 . The results of JCOG0201 suggest that small-sized (<2cm) patients with CTR (<0.25) should be classi ed as non-invasive adenocarcinoma 16 . JCOG0804 study further showed that most patients with CTR (<0.25) did not need systematic lymph node dissection, and their 5-year Disease-free survival (DFS) reached 99.7% (95% CI: 98.3-99.9%) 2 . In our study, the rate of lymph node metastasis in patients with tumor size (<2-3cm) and CTR (<0.25) was only 1.8% (2/112). Although the results did not receive long-term follow-up, it still showed that CTR (<0.25) was not prone to mediastinal lymph node metastasis occurrences. The probability of mediastinal lymph node metastasis signi cantly increased in CTR (> .004) showed that pathological subtype was an important risk factor for mediastinal lymph node metastasis of stage IA lung adenocarcinoma, complying with the results of our study. For these two special types of lung adenocarcinoma above, systematic lymph node dissection should be the prior choice. In the ROC analyses, the pathological subtype showed the highest area value under the AUC curve of 0.800, being considered to be the most accurate predictor among the three predictors. In addition, another study found that patients with pathological subtype of micropapillary or solid adenocarcinoma should be followed with adjuvant therapy after operation 9 . As for patients with stage IA3 lung adenocarcinoma combined with micropapillary or solid adenocarcinoma, the necessity of postoperative adjuvant treatment should be evaluated by the results of long-term follow-up.
Dai et al 20 have shown that STAS has relatively little effect on the prognosis of lung adenocarcinoma (< 2cm), while STAS can cause obvious poor prognosis in patients with lung adenocarcinoma (>2-3cm). Therefore, this study included STAS in multivariate analysis and did not nd that it was a risk factor for mediastinal lymph node metastasis. The possible reason is that different evaluation methods and standards may lead to different evaluation results. Moreover, the sample capacity included in this study is small, and whether the sample is related to mediastinal lymph node metastasis needs further research.
Wang et al 21 revealed that patient with advanced lung cancer with EGFR gene mutation were more prone to distant brain metastasis, while Nie et al 22 reported that there was no signi cant difference in the rate of mediastinal lymph node metastasis between EGFR gene mutation and EGFR gene wild type.
Nonetheless, above two studies did not show statistical signi cance, for they were carried out in locally advanced lung cancer. However, there are only a few studies on whether EGFR gene mutation is related to lymph node metastasis in patients with early lung cancer. In our study, we included EFGR gene mutation into multivariate analysis and did not nd that whether EGFR gene mutation was a risk factor for mediastinal lymph node metastasis.
Our study had some limitations. First, it was a retrospective study and single-center with a relatively small number of cases, therefore, patient-selection bias may be existed. Prospective multi-institutional trials are required for a de nitive conclusion. Secondly, the long-term follow-up outcome was not known.
And the advantage of this study was that the authors performed lobectomy and mediastinal lymph node dissection in all cases, we believe that our results will be helpful in identifying the mediastinal lymph node dissection of IA3 lung adenocarcinoma.

Conclusion
Systematic lymph node dissection should be the prior choice for patients with clinical stage IA3 incorporated with risk factors such as CTR, CEA and pathological subtypes. The overall occurrence rate of mediastinal lymph node metastasis in IA3 lung adenocarcinoma is high, so the lymph node dissection method of stage IA3 should be alternative from those in stage IA1 and IA2. Funding: Tables   Table 1 clinical characteristics and univariate analysis of lymph node metastasis for patients