Impact of retrieved lymph node count on short-term complications in patients with gastric cancer

Background It is well established that retrieved lymph node (RLN) counts were positively correlated with better overall survival in gastric cancer (GC). But little is known about the relationship between RLN count and short-term complications after radical surgery. Methods A total of 1487 consecutive GC patients between January 2016 and December 2018 at Nanjing Drum Tower Hospital were retrospectively analyzed. Univariate analyses were performed to elucidate the association between RLN count and postoperative complications. We further identified clinical factors that might affect the RLN count. Results Among all of the patients, postoperative complications occurred in 435 (29.3%) patients. The mean RLN count was 25.1, and 864 (58.1%) patients were diagnosed with lymph node metastasis. Univariate analyses showed no significant difference between RLN count and postoperative complications (both overall and stratified by CDC grade). Univariate and multivariate analyses further revealed that type of resection, tumor invasion, and lymph node metastasis were associated with RLN count. Conclusions The current study demonstrated that RLN count was not associated with postoperative short-term complications following gastrectomy of GC, which provided a rationale for the determination of a proper RLN count of curative gastrectomy.


Background
There are approximately one million new cases of gastric cancer (GC) each year worldwide, and half of them occur in Eastern Asia, including China, Japan, and South Korea [1]. Despite advances in early screening and comprehensive treatment of GC, it remains the third most common cause of cancer-related death in the world [2]. For advanced GC, a consensus has been reached of radical gastrectomy with D2 lymphadenectomy [3]. However, there is still controversy over the number of retrieved lymph nodes (RLNs) for accurate pathological staging.
Several studies have reported that RLN count was positively correlated with better overall survival in GC, even in lymph node-negative GC [4][5][6][7]. An RLN count of ≥ 16 has been recommended by the 8th edition TNM classification for GC to guarantee the accurate pN stage [8]. Moreover, Okajima et al. suggested an optimal RLN count of ≥ 25 for nodal staging [9]. Recently, by stratum analysis of 7620 patients, Deng et al. proposed an optimal RLN count of ≥ 16 for lymph node-negative GC and > 30 for lymph node-positive GC [10]. These above studies are all conducted by comparing the RLN count with long-term survival. But little is known about the relationship between the RLN count and short-term complications after radical surgery.
Postoperative complications of GC pose a significant impact on the length of postoperative stay and hospital charges, which further affect the quality of life [11]. Therefore, investigating the relationship between RLN count and postoperative short-term complications would provide more comprehensive evidence for selecting the appropriate RLN count.

Patients
A total of 1487 consecutive GC patients between January 2016 and December 2018 at Nanjing Drum Tower Hospital were retrospectively reviewed. All patients underwent curative (R0) gastrectomy and were histologically confirmed. The exclusion criteria were as follows: (1) multivisceral resection, (2) patients accepting preoperative radiotherapy or chemotherapy, (3) patients with previous stomach surgery, and (4) patients with incomplete clinical data. This study was approved by the Ethics Committee of Nanjing Drum Tower Hospital.

Data collection
Data for preoperative characteristics, intraoperative index, and postoperative features were extracted. Preoperative characteristics included age, gender, body mass index (BMI), comorbidities, and laboratory data. The intraoperative index involved the American Society of Anesthesiologists (ASA) grade, surgical approach, type of resection, operation time, and blood loss. Postoperative features included depth of tumor invasion, tumor site, retrieved lymph node count, lymph node metastasis, lymph node ratio (LNR), log odds of positive lymph nodes (LODDS), pTNM stage, Lauren subtype, short-term complications, postoperative stay, and total hospital charges. LNR was defined as the ratio of positive to retrieved lymph nodes. LODDS was calculated by log [(positive lymph nodes + 0.5)/(total lymph nodes − positive lymph nodes + 0.5)] [12]. The postoperative short-term complications occurring in the hospital or within 30 days were collected. All complications were evaluated according to the Clavien-Dindo classification system [13].  Postoperative stay (days) 12.0 ± 8.1 Total hospital charges (10 4 ¥) 7.5 ± 3.5 BMI body mass index, CRP C-reactive protein, ASA American Society of Anesthesiologists, RLNs retrieved lymph nodes, LNR lymph node ratio, LODDS log odds of positive lymph nodes

Statistical analysis
Statistical analyses were conducted by SPSS 19.0 (Chicago, IL, USA). Continuous variables were shown as means ± SD. Student's t test was applied for normally distributed data; Mann-Whitney U test was applied for non-normally distributed data. Categorical variable data were presented as numbers and analyzed using the chisquared test or the Fisher exact test. Univariate and multivariate analyses were performed to analyze the risk factors associated with the postoperative complications or retrieved lymph node count. The optimal cutoff values of LNR and LODDS were determined by receiver operating characteristic (ROC) analysis. All statistical tests were conducted two-sided, and statistical differences were termed as P value < 0.05.

Patient characteristics
The background characteristics of the patients enrolled in this study were presented in Table 1 Association between perioperative characteristics and postoperative complications As presented in Table 2, univariate and multivariate analyses indicated that postoperative short-term complications were significantly correlated with age, gender, level of preoperative serum albumin, and operation time.
Stratified analyses by type of resection revealed that complications occurred frequently in proximal gastrectomy compared with total gastrectomy, while there was no significant difference between distal gastrectomy and total gastrectomy. No significant association was observed between RLN count and overall postoperative complications.

Factors associated with RLN count
We further explored the potential factors associated with RLN count. Univariate analyses revealed that preoperative serum albumin, type of resection, tumor invasion, lymph node metastasis, and pTNM stage were associated with RLN count (P < 0.05; Table 4). Stratification by type of resection showed that RLN count in either distal gastrectomy or proximal gastrectomy was significantly lower than that in total gastrectomy. Multivariate analyses further indicated that type of resection, tumor invasion, and lymph node metastasis were still significantly associated with RLN count (P < 0.05; Table 4).

Discussion
Nodal involvement significantly affected the prognosis of GC patients because it is the major root of tumor relapse after surgery [14,15]. Thus, standardized lymph node dissection is the basic requirement for curative (R0) gastrectomy. Curative gastrectomy with D2 lymphadenectomy has been considered as the standard fashion for decades in Eastern Asia, especially in Japan [16,17]. This procedure has been gradually accepted by Western countries in recent years [18,19]. As for the RLN count, the 8th edition TNM classification for GC recommended dissecting at least 16 lymph nodes. Moreover, emerging evidence revealed the positive correlations between RLN count and overall survival of GC patients [4,5,20].  [20]. Additionally, LNR and LODDS were also reported to be associated with GC prognosis [21][22][23]. These above studies mainly focused on the relationship between RLN count and long-term prognosis. However, little is known about its effects on postoperative short-term complications.
In this study, we concentrated on the association between RLN count and short-term prognosis. Univariate analyses showed no significant difference between RLN count and postoperative complications (both overall and stratified by CDC grade). Therefore, more lymph nodes were encouraged to be dissected from the perspective of short-term prognosis.
Although curative gastrectomy with D2 lymphadenectomy is considered a pivotal strategy for advanced GC, there are international and institutional differences in the number of RLN count [24,25]. Various factors were reported to influence the RLN count, including the confidence and enthusiasm of doctors (both surgeons and pathologists), surgical situation, and innate lymph node count in each patient [7,9]. In our study, we concluded that RLN count was related to the type of resection, tumor invasion, and lymph node metastasis. Of note, RLN count was positively correlated with the lymph node metastasis rate, which underlined the importance of RLN count for accurate staging.
Actually, for a thorough pathological examination, RLNs should be individually divided from a complete tissue sample after surgery. Owing to much time and effort was required during this procedure, it has not been widely implemented clinically. Therefore, the examined lymph node count by pathologists might be lower than the dissected lymph node count. Multiple attempts have been conducted to improve the detection rate of lymph nodes [26][27][28]. Li et al. elucidated that the mean number of RLNs could be significantly elevated by injecting carbon nanoparticles before surgery compared with controls (38.33 vs 28.27) [26]. Markl and colleagues reported a twofold lymph node pick up rate utilizing methylene blue staining than unstained groups (35 vs 17) [27]. Several dye materials were also used to increase the number of lymph nodes dissected during surgery, such as fluorescent indocyanine green (ICG) and 5-aminolevulinic acid (5-ALA) [29,30].
We acknowledge that this study had some potential limitations. First, it was a retrospective, single-center study, so the results might be flawed because of residual confounding factors. Second, the RLN count was closely related to the quality of surgeons and pathologists. The perioperative variables might differ in different doctors. Therefore, multicenter studies are needed to confirm our results.

Conclusions
In conclusion, the current study demonstrated that RLNs\ count was not associated with postoperative short-term complications following gastrectomy of GC. Therefore, our analysis encouraged more lymph nodes to be dissected for accurate pathologic staging.