Is No. 12a Lymph Node Dissection Compliance Necessary in Patients Who Undergo D2 Gastrectomy for Gastric Adenocarcinomas? A Population-Based Retrospective Propensity Score Matching Study

Simple Summary Since nodal metastasis is the main pattern for gastric cancer (GC) metastasis, lymph node (LN) dissection is essential for accurate staging and improving prognosis. However, debates exist regarding the necessity of No. 12a LN dissection (LND) in D2 gastrectomy. Moreover, the compliance rate for No. 12a LND in practice is low. To explore No. 12a LND noncompliance’s effect on long-term prognosis in GC patients after D2 gastrectomy, we performed a retrospective propensity score matching study with 2788 patients included. The results showed that patients with No. 12a LND had a significantly greater OS than those without it before and after PSM. This study is the first propensity score matching study to demonstrate the prognostic impact of No. 12a LND noncompliance on patients who undergo D2 gastrectomy. This large population-based study may provide guidance on No. 12a LND. Abstract LN dissection is essential for accurately staging and improving GC patient prognosis. However, the compliance rate for No. 12a LND in practice is low, and its necessity is controversial. Data from GC patients who underwent total gastrectomy (TG)/distal gastrectomy (DG) plus D2 lymphadenectomy between January 2000 and December 2017 at West China Hospital, Sichuan University were reviewed. No. 12a LND noncompliance’s effect on the long-term prognosis of patients with GC after D2 gastrectomy was explored. Of the 2788 patients included, No. 12a LND noncompliance occurred in 1753 patients (62.9%). Among 1035 patients with assessable LNs from station 12a, 98 (9.5%) had positive LNs detected at station 12a. No. 12a LN metastasis patients (stage IV not included) had significantly better overall survival (OS) than TNM stage IV patients (p = 0.006). Patients with No. 12a LND compliance had a significantly higher OS than those without, both before (p < 0.001) and after (p < 0.001) PSM. Cox multivariate analysis confirmed that No. 12a LND noncompliance was an independent prognostic factor before (HR 1.323, 95% CI 1.171–1.496, p < 0.001) and after (HR 1.353, 95% CI 1.173–1.560, p < 0.001) PSM. In conclusion, noncompliance with No. 12a LND compromised the long-term survival of patients who underwent D2 gastrectomy for GC.


Introduction
Gastric cancer (GC) is the fifth most common malignant tumor and the fourth leading cause of cancer-related death worldwide [1]. Surgical resection remains the mainstream treatment for GC, especially for locally advanced GC [2,3]. As nodal metastasis is the main pattern for GC metastasis, lymph node (LN) dissection is essential for accurate staging and chemotherapy used. Clinicopathologic features were classified according to the classification of JGCA (3rd English edition) [6]. Adenocarcinoma of the esophagogastric junction (EGJA) was defined according to the Siewert classification [18]. The TNM staging was classified according to the eighth edition of the Cancer Staging Manual of the AJCC [12].

Scope Definition of No. 12a LNs and Definition of No. 12a LND Compliance and Noncompliance
According to the classification of JGCA (3rd English edition), hepatoduodenal ligament LNs along the proper hepatic artery were defined as station 12a LNs [6]. Scope definition of No. 12a LNs in our institute were as follows: (1)  Procedure of No. 12a LND was as follows. First, hepatoduodenal ligament between the lower border of liver and the duodenal bulb was fully exposed. Second, hepatoduodenal ligament between the upper border of pancreas at the origin of the proper hepatic artery and the confluence of the right and left hepatic arteries was opened at the left borderline of the common bile duct. Third, continuous perivascular sheath tissues, including all the fatty and lymphatic tissues along the proper hepatic artery and covering tissues of the anterior and medial wall of the portal vein, were en bloc dissected. A short rubber band was used to stretch the proper hepatic artery to facilitate the exposure of the portal vein. After a standard No. 12a LND procedure, the proper hepatic artery was skeletonized, and the anterior and medial wall of the portal vein was displayed. For definition of No. 12a LND compliance and noncompliance, patients were divided into the No. 12a LND compliance group or the No. 12a LND noncompliance group based on whether any LNs were harvested from station 12a. If any lymph nodes were retrieved from station 12a according to the final pathologic report, patients were assigned to the No. 12a LND compliance group; otherwise, patients were assigned to the No. 12a LND noncompliance group.

Follow-Up and Clinical Endpoint
All patients were followed up periodically through outpatient visits, telephone interviews, network tools and letters. The follow-up interval was every 3 to 6 months during the first 2 years postoperatively, every 6 to 12 months during the subsequent 3 years and annually thereafter. In this study, overall survival (OS) was the endpoint of interest and was calculated from the date of surgery to the date of death from any cause or the date of latest follow-up.

Statistical Methods
Continuous variables that fit a normal distribution were expressed as the mean ± SD and were compared using Student's t test; otherwise, variables were expressed as the median [IQR] and were compared using the rank sum test. Categorical variables are expressed as numbers (%) and were compared using chi-squared tests. Kaplan-Meier curves of cumulative survival were compared using the log rank test. Logistic regression analysis was used to analyze independent predictors associated with No. 12a LN metastasis. Cox proportional hazard regression analysis was performed to identify independent predictors associated with overall survival. Those variables with a univariable p < 0.05 were entered into the multivariable regression model using backward stepwise variable selection. Nomograms and calibration curves of the Cox regression model and the logistic regression model were generated using the "rms" package of R version 4.1.0. The performance of the models was evaluated and validated using the concordance index (C-index) and calibrated using 1000 bootstrap samples. Calibration was evaluated using the Hosmer-Lemeshow test.
Propensity scores were calculated using a logistic regression model based on covariates, including age, sex, comorbidity, tumor location, macroscopic type, tumor size, extent of gastrectomy, operative approaches, concomitant organ resection, number of harvested LNs, histology type, perineural invasion, lymphovascular invasion, venous invasion, cancer nodules, TNM stage and adjuvant chemotherapy. A 1:1 nearest neighbor matching, without replacement, was performed with a caliper width of 5% of the standard deviation of the logit of propensity score by using the "MatchIt" package of R version 4.1.0 [19]. After PSM, the absolute standardized mean difference (SMD) was used to measure covariate balance, and an SMD threshold of 0.1 was considered substantial imbalance.

Comparison of Clinicopathologic Findings before and after PSM
As shown in Table 1, the lowest compliance rate was reported for the dissection of station 12a LNs. The No. 12a LND compliance rate of the whole study cohort was 37.1% (1035/2788). The comparison of clinicopathological parameters between the No. 12a LND compliance group and the No. 12a LND noncompliance group showed that the longitudinal location of the tumor (p < 0.001), distribution of macroscopic types (p < 0.001), tumor size (p < 0.001), number of harvested LNs (p < 0.001), perineural invasion rate (p < 0.001), vascular invasion rate (p < 0.001), venous invasion rate (p < 0.001), T stage distribution (p < 0.001), TNM stage distribution (p = 0.011) and proportion of patients who received adjuvant chemotherapy (p = 0.003) were significantly different between groups. After a 1:1 matching based on propensity scores, there were 916 patients left in each group, and the clinicopathological parameters were well balanced between the two groups (Table 2 and Figure S1).

Subgroup Analysis of No. 12a LND Compliance and Noncompliance
Subgroup analysis was performed in the matched cohort according to clinicopathologic parameters. It showed that compliance with No. 12a LND could bring a survival benefit for most subgroups, especially for patients with tumors involving the middle/lower third of the stomach (p = 0.012), with EGJA Siewert type III (p = 0.034), who had a tumor with or without lesser curvature involvement (both p < 0.05), who underwent distal gastrectomy (p = 0.010), who had a tumor size of <5 and ≥5 cm (both p < 0.05), with pT3 (p = 0.001) and pT4 (p = 0.005), with pN3a (p = 0.048), with stage II (p = 0.007) and stage III (p = 0.002), and who received adjuvant chemotherapy (p = 0.004). The detailed results of the subgroup analysis are depicted in a forest plot (Figure 3).

Net Survival Benefit of No. 12a LND Compliance
To evaluate the net survival benefit of No. 12a LND compliance, two nomograms were established for No. 12a LND compliant patients and No. 12a LND noncompliant patients based on independent prognostic factors confirmed in the entire study population. These two nomograms were compared, and the difference between the two estimates was the expected net survival benefit from No. 12a LND compliance ( Figure 4A,B). Bootstrapping with 1000 resamples demonstrated good predictive performance of the nomograms, with C-indexes of 0.732 (95% CI 0.707-0.757) for No. 12a LND compliant patients and 0.718 (95% CI 0.702-0.735) for No. 12a LND noncompliant patients. The calibration curves to predict 3-and 5-year survival probabilities among No. 12a LND compliant patients and No. 12a LND noncompliant patients also showed good consistency with the ideal predictive curves (all p > 0.05, Figure 4C,D).

Discussion
The role of No. 12a LN metastasis and No. 12a LND in gastric cancer is debated and is still not fully elucidated. The 5-year survival of patients with No. 12a LN metastasis has been reported to range from 5.6% to 54.4% [9,14,[20][21][22]. There were inconsistencies between different guidelines in terms of whether hepatoduodenal LNs should be classified as regional LNs and be routinely removed during surgery [6,8,12,15]. The present study showed that the 5-year OS rate or No. 12a LN metastasis patients (stage IV not included) was 51.8%, which was significantly better than that of stage IV patients. Therefore, the results of this study are in accordance with the opinion that No. 12a LN metastasis should be considered regional metastasis.
In the present study, the overall No. 12a LN metastasis rate was 9.5%. The metastasis rate or No. 12a LNs varied between previous reports due to inevitable differences in clinicopathological features between studies. In fact, it has been reported that tumor location, tumor size, tumor stage, soft tissue invasion, nerve invasion, intravascular cancer emboli, macroscopic type and histological type are possible predictors of No. 12a LN metastasis [9,21,23,24]. Logistic regression in this study revealed that an older age, the tumor longitudinal location (middle or lower third involved), the tumor cross-sectional location (lesser curvature involved), the undifferentiated type, venous invasion, cancer nodules, pT4 and distant metastasis were independent predictors of station 12a LN metastasis.
The No. 12a LND compliance rate in the present study was 37.1%, which is in accordance with other studies [14,25]. This result indicated that the compliance rate or No. 12a LND is unsatisfactory in actual practice. Chen et al. [16] analyzed data from 2401 patients who underwent D2 radical gastrectomy and found that the tumor site, BMI, range of gastrectomy, previous abdominal surgery and surgery type were independent predictive factors for noncompliance with D2 lymphadenectomy. It has been reported that the endoscopic submucosal injection of carbon nanoparticle suspension or indocyanine green 1 day before surgery could improve the number of lymph node dissections performed at station 12a [26,27]. Thus, such techniques should be selectively used in patients with a high risk of No. 12a LN metastasis or a high risk of noncompliance with D2 lymphadenectomy to improve the compliance rate of station 12a LND.
In the present study, the survival analysis showed that patients with compliance with No. 12a LND achieved better OS than those without it, both before and after PSM. Further Cox regression analysis showed that noncompliance with No. 12a LND was an independent prognostic factor for OS in the entire study cohort and in the matched cohort. Based on these results, noncompliance with No. 12a LND compromised the long-term survival of patients who received TG/DG plus D2 lymphadenectomy. However, there was an apparent contradiction between the results of the present study and the results of the real-world study reported by Seo et al. [14]. A wide gap in the proportion of early-stage patients included between studies may account for this discrepancy.
To further elucidate patients who may benefit from No. 12a LND, subgroup analysis was performed in the matched cohort. It is not surprising that stage II-III patients obtained survival benefit from the compliance of No. 12a LND. Lin et al. [25] also reported that the survival of stage II and stage III LND compliant patients was significantly superior to that of LND noncompliant patients. Moreover, the compliance of No. 12a LND provided a survival benefit for patients with tumors located at the lower or middle third or the lesser curvature involved. These results matched with the reported independent predictors of station 12a LN metastasis and may be mainly due to the right gastric artery serving as a lymphatic drainage route for these tumors [9].
Another finding in the subgroup analysis was that compliance with No. 12a LND brings survival benefits to patients with tumors located at the esophagogastric junction. Patients with EGJA were further divided into Siewert type II and type III, and the results showed that patients with Siewert type III EGJA obtained survival benefits from compliance with No. 12a LND. Galizia et al. [13] analyzed data from 73 patients who underwent radical surgery for GC and reported that modified D2 (D1/D1+) lymphadenectomy conferred the same oncologic adequacy as standard D2 lymphadenectomy for tumors located in the upper or middle third of the stomach. However, the sample size in their study seems too small to draw a reliable conclusion. Yura et al. [28] surveyed 202 patients diagnosed with T2/T3 gastric cancer exclusively located in the upper third of the stomach and reported that the station 12a LN metastasis rate was 0.6% (1/162). However, another recent study from de Jongh et al. [29] reported a No. 12a LN metastasis rate of 22.7% (5/22) in patients with T3/T4 gastric cancer located in the proximal third of the stomach. In the present study, the No. 12a LN metastasis rate of patients with tumors located at the esophagogastric junction was 2.7% (5/182). A high proportion (over 40%) of pT4 tumors included in this study may explain the result that patients with Siewert type III EGJA obtained survival benefit from the compliance of No. 12a LND. Tumor size may be another risk factor for No. 12a LN metastasis in EJGA. A previous study from our institute noted that a tumor size exceeding 5.0 cm was an independent risk factor for lower perigastric lymph node metastasis [22]. Thus, further studies are warranted for the lymph node metastasis patterns in large advanced (T3/T4) gastric cancer located at the upper third of the stomach and the esophagogastric junction. According to the results of the current study, No. 12a LND should not be neglected in patients with EGJA.
There were limitations to this study. First, although PSM was used to eliminate possible confounders, selection bias, detection bias and statistical bias are unavoidable because of the inherent nature of retrospective studies. Second, the statistical results in the present study were based on data from a single center, and external validation was lacking. Third, the results of subgroup analysis may not be robust because of the small sample size of particular groups. Fourth, overall survival was the only endpoint of interest in this study due to the lack of recurrent data in the WCH-SGCPR.
In conclusion, noncompliance with station 12a LND does compromise long-term survival in patients who underwent D2 gastrectomy for gastric adenocarcinomas. The No. 12a LND procedure should be performed more carefully to improve the compliance rate.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cancers15030749/s1, Figure S1: Standardized mean differences of clinicopathological parameters before and after PSM; Figure S2: Nomogram of the logistic regression model for No. 12a LN metastasis; Figure S3: The overall survival curves of patients with and without No. 12a LN metastasis and No. 12a LND noncompliant patients before PSM; Table S1: Univariate and multivariate survival analysis in this study by Cox proportion hazard model after PSM.  Informed Consent Statement: Written informed consent was waived because this was a retrospective study and all participants were anonymous.

Data Availability Statement:
The raw data and the original procedure video from this study are available from the corresponding author upon request.