Nowadays surgical resection with lymph node dissection remains the only curative treatment for AGC patients. Some papers reported that there was approximately 10% of AGC patients who have developed NO.10 LN metastasis[9-11]. Some surgeons performed standard D2 lymphadenectomy in order to complete NO.10 LN dissection, but the complications have disadvantaged them to rethink this surgery type[12, 13]. Moreover patients with NO.10 LN metastasis had a worse survival than patients without them, but they could not gain survival benefits from NO.10 LN dissection[8, 14]. Thinking the safety of surgery and patients’ quality of life, however, there is no clear consensus as to whether AGC patients without metastasis to NO.10 LN preoperatively could avoid prophylactic NO.10 lymph node dissection. Thus, the aim of this study was to evaluate the impact of prophylactic NO.10 lymph node clearance on the prognosis and postoperative complications of AGC patients.
As is known to us, location was a key factor correlated with NO.10 LN especially with those in upper and greater curvature of the stomach[15, 16]. There was a reasonable explanation to this. The lymphatic flow of greater curvature located along upper body, is drained to splenic hilar lymph nodes via the short gastric artery, left gastroepiploic artery and posterior gastric artery. Tammaro reported that NO.10 LN dissection should be considered for those which tumor localized in the upper two thirds of the stomach, which was in good accordance with the current study[17]. Kusano et al. revealed that the rate of positive NO.10 LN locating at the greater curvature was 17.0%, higher than that at the lesser curvature(10%)[18]. Meanwhile, our paper revealed that the NO.10 LN rate locating at the greater curvature was 11.1%, higher than that at the lesser curvature(1.1%). The results suggested that NO.10 LN easily occurred in the gastric cancer at the greater curvature. Obviously, tumor location is absolutely a predictor for AGC patients.
To date, some researchers believed there existed closely negative correlation between differentiation degree and NO.10 LN metastasis[19, 20] In general, the poorer differentiation degree was, the higher rate of NO.10 LN would be. Our study showed that the rate of NO.10 LN in the group with well/moderately differentiation(2.8%) was lower than that in the group with poor differentiation(10.2%). Wu et al. explained that the positive NO.10 LN was correlated with higher level of Matrix Metalloproteinase3(MMP-3) expression which was induced by poorer differentiation of AGC patients[21]. Consequently, the lower differentiation should be a criterion for AGC patients who could benefit from NO.10 LN resection.
It was well-known that depth of invasion and tumor size can predict tumor staging in a certain degree. Some studies reported that the above two were potential NO.10 LN influencing factors [22, 23]. In our study, the positive rate of NO.10 LN in tumor size equal or larger than 5 cm was 10.2% while the rate in tumor size less than 5 cm was 2.8%. In our study, the NO.10 rate of T4 stage was 10.2%, higher than that T2 and T3 stage(2.8%). Some reported that NO.10 LN metastasis was observed in 3.4% patients with tumors smaller than 5 cm and in 21.3% patients with tumors larger than 5cm[24]. However there was no statistic difference between depth of invasion and tumor size on the NO.10 LN metastasis in our paper. Aoyagi et al. also reported that T stage was not associated with splenic hilar lymph node metastasis[14].
In addition, in the present study it showed that NO.10 LN was not associated with Borrmann type. The NO.10 LN rate of AGC patients with the type I/II and III/IV were 3.7% and 7.4% respectively. Some other studies reported that NO.10 LN metastasis more easily occurred in the infiltrative AGC patients compared with the local AGC patients, maybe because of tumor growth pattern[25]. But in our paper we failed to confirm the correlation between NO.10 LN and Borrmann type.
In our study we also found that when NO.4 lymph nodes was positive, NO.10 lymph node positive detection rate was as high as 26.7%; when NO.4 lymph nodes was negative, NO.10 lymph node positive detection rate was only 3.2%. The difference between them was statistically significant (P<0.001). Therefore, we can take NO.4 positive lymph nodes as the sentinel lymph nodes. This can be explained as followed: The NO.10 lymph node locates near the greater curvature and it will occur through NO.4 lymph nodes located along the greater curvature via lymphatic drainage[26]. Bian et al. have reported that for the certain group of AGC patients avoiding unnecessary NO.10 lymphadenectomy can lead to less invasive trauma and tissue damage[27]. As for the cases of NO.4 LN negative, NO.10 LN was still positive. The cancer cells will have to pass through lesser curvature lymph nodes, then NO.7, 9, and 11 lymph nodes, finally reach the NO.10 LNs. This is one of the possible mechanisms of explaining the phenomenon of "skip metastasis", but the rate is too low to confirm. And the incidence of this kind of skip metastasis was only 3.2 % in our study and was also very low in other reports[28].
Moreover, in our paper multivariate analysis showed that differentiation, depth of invasion and lymph node metastasis were independent prognostic factors(P<0.05). The patients with those characteristics have poor survival time. The results were basically consistent with other statistical reports[29-31]. As for the lymphadenectomy extent, Sano et al. enrolled 505 patients in his research showed no survival difference between SD2 group and MD2 group in AGC patients[32]. In 2014 some scholars analysed 8 RCTs and found that there was not a significant difference in the overall 5-year survival rate between the two groups[33]. In our study, we also found no significant difference in the survival rates between the two groups; the median survival time was 72.23 months and 68.56 months for the SD2 and MD2 groups, respectively(P=0.635). Although patients in MD2 group had a worse survival than patients in SD2 group, there is no statistical significance. A recent study also showed there was no statistically significant difference in 5-year survival for the patients with and without splenic hilar lymph node metastasis in the greater curvature group[34]. Huang et al. retrospectively compared patients who underwent and did not undergo NO.10 lymph node dissection in the spleen-preserving surgery group and found the 3-year disease-free survival time was significantly better in the dissection group but the overall survival time has no statistical significance[35].
The mortality and morbidity rates in the two groups were summarized in Table 4. The results in our study were similar to those previously reported. The overall postoperative morbidity in the SD2 group was higher, and the difference was statistically significant(P=0.022). Postoperative major morbidity and mortality rates were 37.96% and 3.70% in the SD2 group, and 23.64% and 1.82% in the MD2 group, respectively. We speculated that the fragile texture of the spleen and special anatomical location might increase the risk of postoperative complications.
Several limitations inherent in this study should be addressed. First, as a retrospective study, our study with a relative small number of patients from a single center included possible selection bias. Therefore, further randomized studies are required to compare the outcomes of SD2 versus MD2. Second, there were two false-negative cases in our study, namely, patients with negative NO.4 LN but positive NO.10 LN. So, we should find the sensitivity and specificity of better indicators of NO.10 LN. And large well-designed studies would also be needed to explore the role of NO.10 lymph node clearance in the AGC patients. But the unique nature of this paper provides a new insight into the lymphadenectomy extent for AGC patients.