Extended lymphadenectomy benefits patients with borderline resectable pancreatic head cancer—a single-center retrospective study

Background: Whether standard lymphadenectomy or extended lymphadenectomy should be performed is still under debate during pancreaticoduodenectomy (PD). We aimed to compare their morbidity and mortality rates among patients with pancreatic head cancer (PHC). Methods: In this retrospective study, a total of 322 patients were enrolled. According to the scope of intraoperative lymph node dissection, patients were divided into extended lymphadenectomy group (n=120) and standard lymphadenectomy group (n=202). Based on the resectability of the tumor, there were 198 cases of resectable PHC and 124 cases of borderline resectable PHC, respectively, in which further stratified analysis was carried out according to the extent of lymph node dissection. Results: All patients completed the operation successfully, with a perioperative morbidity rate of 27.9% and mortality rate of 0.9%. As for the overall patients, patients in the extended lymphadenectomy group had higher neutrophil-to-lymphocyte ratio (NLR), longer operation time, more intraoperative blood loss, lymph node dissection and patients with borderline resectable pancreatic head cancer (BRPHC) (P<0.05). The 1-, 2-and 3-year overall survival rates of patients with extended lymphadenectomy and standard lymphadenectomy were 71.9%, 50.6%, 30.0% and 70.0%, 32.9%, 21.5%, respectively (P=0.068). With regards to patients with BRPHC, the number of lymph node dissection in the extended lymphadenectomy group was more (P<0.05), and the 1-, 2- and 3-year overall survival rates of patients with extended lymphadenectomy and standard lymphadenectomy were 60.7%, 43.3%, 27.4% and 43.2%, 17.7%, 17.7%, respectively (P=0.007). Conclusions: Patients with BRPHC tended to have vast lymph node metastasis. Extended lymphadenectomy can improve their long-term survival. intraoperative blood loss, operation time, curative resection postoperative complication and vein invaded by the tumor, 124 patients with borderline resectability underwent allogeneic vascular replacement after invasive resection in 87 end-to-end anastomosis in and direct after wedge resection in The blood loss was 500 [IQR, was performed in 108 (33.5%), and the operation was


Introduction
Pancreatic cancer is becoming one of the most common cause of cancer mortality, whose incidence has increased in the past decade despite a decrease in incidence rates for other cancers (1). Although considerable progress has been made to improve the survival rate surgical resection remains the only potential curative method for patients with pancreatic cancer (2)(3)(4). Notably, lymphadenectomy, which is a powerful prognostic factor after pancreaticoduodenectomy (PD) that is independent of the histology, has been strongly suggested to achieve the potential cure when pancreatic cancer is suspected (5,6). At present, whether standard lymphadenectomy or extended lymphadenectomy should be performed is still under debate. Ishikawa et al. reported extended clearance of regional lymph nodes and soft tissue could significantly prolong the survival without increasing the morbidity rate (7). In contrast, Jang et al. found extended resection failed to improve oncologic efficacy or long-term survival compared with standard resection (8). Whereas, Pedrazzoli et al. did notice a trend toward longer survival after extended lymphadenectomy instead of standard lymphadenectomy in spite of similar overall survival rates between the two groups (9). Therefore, it is the surgeons' major concern to determine their efficacy among these patients.
Moreover, previous studies focused on all patients with pancreatic head cancer without distinguishing their stages, who actually had different tumor biology and lymph node status (10,11). Since it is really difficult to thoroughly assess the extent of positive lymph nodes in the preoperative and intraoperative periods due to currently limited clinical technology, we conducted a retrospective analysis in terms of resectability including resectable and borderline resectable pancreatic head cancer (BRPHC) recommended by the National Comprehensive Cancer Network (NCCN) guidelines (12). We aimed to compare the morbidity and mortality rates between extended lymphadenectomy and standard lymphadenectomy.

Ethics approval and consent to participate
The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by the Ethics Committee of Beijing Chaoyang Hospital (No. 2020-D.-309-2) and individual consent for this retrospective analysis was waived.

Inclusion and exclusion criteria
The data of patients with pancreatic head carcinoma (PHC) who underwent PD in the Department of Hepatobiliary surgery of Beijing Chaoyang Hospital from January 2010 to December 2019 were analyzed retrospectively. According to the relevant inclusion/exclusion criteria and NCCN guidelines, 322 patients with PHC were enrolled [including 198 patients with resectable pancreatic head cancer (RPHC) and 124 patients with BRPHC].
Inclusion criteria: (I) patients with PHC who underwent PD from January 2010 to December 2019; (II) aged 20 to 85 years old; (III) assessed as RPHC or BRPHC according to the NCCN guidelines; (IV) en-bloc resection of tumor during operation; (V) postoperative pathology confirmed pancreatic ductal adenocarcinoma; (VI) the mode of operation and treatment strategy obtained the informed consent of patients and their families.
Exclusion criteria: (I) unresectable condition or metastasis found during surgery; (II) surgical rule violation; (III) pathologic diagnosis other than conventional ductal adenocarcinoma; (IV) postoperative follow-up data were incomplete or lost to follow-up ( Figure 1).

Patients grouping and definition
According to the NCCN guidelines, the patients were divided into 198 cases of RPHC and 124 cases of BRPHC. RPHC was defined as the tumor did not invade the celiac artery or vein and had no distant metastasis and BRPHC included 45 cases of tumor invading the confluence of portal vein (2 cases contacting celiac trunk but invading <180 degrees), 42 cases of tumor invading superior mesenteric vein (3 cases contacting superior mesenteric artery but invading <180 degrees), 37 cases invading portal vein (2 cases contacting common hepatic artery but invading <180 degrees) ( Figure 2).
The lymph node grouping is named according to the Japanese Society of Hepato-Biliary-Pancreatic Surgery (13). According to the range of intraoperative lymph node dissection, the patients were divided into extended lymph node dissection group and standard lymph node dissection group, and differences in the extent of resection are summarized in Table 1. Surgical outcomes such as intraoperative blood loss, operation time, curative resection (R0), postoperative complication and so forth were collected. R0 was defined as a specimen with clear resection margins, tumor cell was not found within 1mm distance from margin. Postoperative complications were graded according to the Clavien-Dindo classification (14). Drain amylase of >3 times serum amylase after the third postoperative day, as defined by International Study Group of Pancreatic Surgery (ISGPS), was defined as pancreatic fistula (15).

Surgical techniques
Some studies have shown that enlarged dissection of retroperitoneal lymphoid tissue and nerve plexus was performed on the basis of radical PD for pancreatic head cancer. The postoperative histopathology showed frequent lymph node metastasis from celiac trunk to superior mesenteric artery. It is closely related to the prognosis of pancreatic cancer, suggesting that extended lymph node dissection can benefit the survival of patients. Therefore, on the basis of standard dissection range, we dissected the lymph nodes behind the hepatic artery [8p], around the celiac artery [9], around the proper hepatic artery [12a], behind the portal vein [12p], around the superior mesenteric artery [14a-d], and around the abdominal aorta [16a2c16b1]. The scope of dissection was up to the hepatic hilum, down to the beginning of the inferior mesenteric artery, left to the left edge of the abdominal aorta, and right to the right renal hilum to complete the dissection of lymph, nerve and connective tissue in this area.

Index analysis and follow-up strategy
The clinical and pathological data were extracted from the medical records, and the perioperative data of different groups were compared. The frequency of postoperative follow-up: 1 and 3 months after operation, once every 3 months within 2 years, and once every half a year for more than 2 years. The end points of follow-up were tumor recurrence and death. The contents of follow-up mainly included blood examination (blood routine, blood biochemistry, tumor markers, etc.) and imaging examination [abdominal enhanced computed tomography (CT), pulmonary CT, enhanced CT, etc.], follow-up treatment, tumor recurrence and survival, and compared the long-term prognosis of patients in different groups.

Statistical analysis
Results are presented by mean ± standard deviation in accordance with normal distribution and by median [interquartile range (IQR)] in non-normal distribution. Nominal data were compared using χ 2 tests and continuous variables using Student's t-tests. Survival outcomes were calculated using the Kaplan-Meier method and compared using the log-rank test. Only variables statistically significant by univariate analysis were included in the multivariate analysis, which was performed using a Cox proportional hazards regression model. All statistical analyses were performed using SPSS, version 22.0 (IBM), with two-sided P values less than 0.05 considered statistically significant.

Clinical characteristics and outcomes
A total of 322 patients with pancreatic head cancer were identified. Three patients died in perioperative period, with a perioperative mortality rate of 0.9%, and the reason for the death was grade C pancreatic fistula complicated with abdominal hemorrhage. One-hundred and twenty patients with extended dissection and 202 patients with standard dissection were performed according to the extent of lymph node dissection. According to the location of the vein invaded by the tumor, 124 patients with borderline resectability underwent allogeneic vascular replacement after invasive vascular resection in 87 cases, end-to-end anastomosis in 25 cases, and direct suture after wedge resection in 12 cases. The intraoperative blood loss was 500 [IQR, 400, 1,000] mL, blood transfusion was performed in 108 patients (33.5%), and the operation time was 9.8±2.9 h.

Perioperative condition
Postoperative complications occurred in 90 cases, with a complication rate of 27.9%. Biochemical fistula occurred in 29 cases (9.0%), B pancreatic fistula in 8 cases

Survival data and recurrence
The follow-up period was up to June 2020, with a median follow-up period of 51 months. During this period, 107 patients (33.2%) received adjuvant chemotherapy for 1 to 8 cycles. The overall median disease-free survival time of the patients was 15 months ( Figure 3A) and the overall median survival time of the patients was 19 months ( Figure 3B).

Comparison of patients with different lymph node dissection groups
The comparison of perioperative general data between the extended dissection group and the standard dissection group was shown in Table 2. It can be seen that the patients in the extended dissection group had higher neutrophilto-lymphocyte ratio (NLR), longer operation time, more intraoperative blood loss, more lymph node dissection and more patients with borderline resectability (P<0.05). The comparison of postoperative morbidity between two groups was shown in Table 3. It can be seen that the incidence of diarrhea in the extended dissection group was higher than that in the standard dissection group (P<0.05) and there were no statistically significant differences in other perioperative complication rates and mortality rate. (P>0.05) The median disease-free survival time of patients with extended lymph node dissection and standard lymph node dissection was 16 and 15 months, respectively. The 1-, 2and 3-year disease-free survival rates were 61.4%, 40.7%, 29.0% and 57.2%, 29.3% and 20.8%, respectively (P=0.125, Figure 4A). And the overall median survival time of patients with extended lymph node dissection and standard lymph node dissection was 25 and 18 months, respectively. The 1-, 2-and 3-year overall survival rates were 71.9%, 50.6%, 30.0% and 70.0%, 32.9%, 21.5%, respectively (P=0.068, Figure 4B).

Comparison of patients with RPHC by different lymph node dissection groups
The comparison of perioperative general data between the extended dissection group and the standard dissection group was shown in Table 4. It can be seen that the NLR value of the extended dissection group was higher, the operation time was longer, and the number of lymph node dissection was more than that of the standard dissection group (P<0.05). The comparison of postoperative complications between two groups was shown in Table 5. It can be seen that the incidence of diarrhea in the extended lymph node  dissection group was also higher than that in the standard lymph node dissection group (P<0.05). There was no significant difference in the incidence and mortality of other perioperative complications. (P>0.05). The median disease-free survival time of patients with extended lymph node dissection and standard lymph node dissection was 20 and 17 months, respectively, and the 1-, 2-and 3-year disease-free survival rates were 79.4%, 41.6%, 32.1% and 65.7%, 32.9%, and 22.7%, respectively (P=0.205, Figure 5A). The overall median survival time of patients with extended lymph node dissection and standard lymph node dissection was 25 and 21 months, respectively, and the 1-, 2-and 3-year overall survival rates were 87.7%, 60.9%, 33.8% and 79.1%, 38.9%, 24.6%, respectively (P=0.146, Figure 5B).

Comparison of patients with BRPHC by different lymph node dissection groups
The comparison of perioperative general data between the extended dissection group and the standard dissection group was shown in Table 6, which shows that the number of lymph node dissection in the extended dissection group was more than that in the standard dissection group (P<0.05). The comparison of postoperative complications between the two groups was shown in Table 7. It can be seen that the incidence of diarrhea in the extended dissection group was higher than that in the standard dissection group (P<0.05). There was no significant difference in the incidence and mortality of other perioperative complications. (P>0.05). The median disease-free survival time of patients with extended lymph node dissection and standard lymph node dissection was 12 and 8 months, respectively, and the 1-, 2and 3-year disease-free survival rates were 48.5%, 40.3%, 26.2% and 30.4%, 15.6%, 15.6%, respectively (P=0.026, Figure 6A). The overall median survival time of patients with extended lymph node dissection and standard lymph node dissection was 18 and 12 months, respectively. The 1-, 2-and 3-year overall survival rates were 60.7%, 43.3%, 27.4% and 43.2%, 17.7%, 17.7%, respectively (P=0.007, Figure 6B).

Discussion
Lymph node status is a powerful prognostic factor in patients with pancreatic head cancer after resection as lymph node invasion occurs in a high percentage of these patients indicating poor survival (16)(17)(18)(19). Therefore, lymphadenectomy has been recognized as an important and fundamental step during PD. A growing body of evidence shows local recurrence after conventional PD results   CA19-9, carbohydrate antigen 199; CEA, carcinoma embryonic antigen; IQR, interquartile range; LN, lymph node; MLNR, metastatic lymph node ratio; NLR, neutrophil to lymphocyte ratio; OP, operation; PBD, preoperative biliary drainage; TB, total bilirubin; SD, standard deviation; SEM, standard error of the mean; Y, yes; N, no.   from incomplete clearance of the lymph nodes suggesting extended lymphadenectomy may improve survival outcomes theoretically. For example, the survival rates after extended lymphadenectomy were reported to be greatly enhanced compared with those after standard lymphadenectomy (7,(20)(21)(22). However, recent studies found that extended resection did not have survival advantages over standard resection (8,(23)(24)(25). Consequently, it is uncertain whether extended lymphadenectomy is superior to standard lymphadenectomy in PD.
In our study, we found extended lymphadenectomy did not improve the survival in all patients with pancreatic head cancer but to cause an increase in the difficulty of surgery (prolonged operating time) as well as in the postoperative complication (high incidence of diarrhea) due to its nature. Accordingly, the number of harvested lymph nodes was larger after extended lymphadenectomy. When we further analyzed the data merely from the subgroup of patients with resectable disease, there were no difference in survival rates between extended lymphadenectomy and standard lymphadenectomy. Therefore, comparison of the number of metastatic lymph nodes did not favor extended lymphadenectomy in patients with resectable disease.
In contrast, distinguished results with respect to the survival rates were observed in patients with BRPHC. Extended lymphadenectomy improved the disease-free  survival and the overall survival greatly in patients with BRPHC compared with standard lymphadenectomy. BRPHC, an intermediate stage between resectable and unresectable disease, invades the mesenteric-portal or arterial axis (26). In case of superior mesenteric-portal vein involvement, venous resection and reconstruction during PD can contribute to long-term survival of these patients (27,28). There has been increasing evidence indicating that patients with BRPHC could reach the survival similar to patients with resectable disease following surgery (29)(30)(31). However, few studies are reported concerning lymphadenectomy in patients with BRPHC. Actually, the reasonable extent of surgery is based on the range of lymph node invasion. The number of metastatic lymph nodes increased greatly in addition to the number of total lymph nodes in this study, which meant patients with BRPHC might have a wide range of lymph node invasion. Therefore, extended lymphadenectomy might benefit these selected patients. Since vascular resection and reconstruction can significantly affect the process, operating time was similar between the subgroups.
Currently, the risk factors for extensive lymph node metastasis remain limited. In this study, we determined large tumor size, poor tumor differentiation and BRPHC were the independent predictors favoring extended lymphadenectomy. Tumor size has been repeatedly reported to be a crucial prognostic factor for patients with pancreatic cancer (32)(33)(34)(35)(36), in relation to positive peritoneal lavage cytology (37) and recurrence (38,39). It is thought that cells within the tumor usually acquire mutations, and finally, a sub-clone that is capable of lymph node metastasis may evolve as the primary tumor grows. Feng et al. reported nuclear-enriched abundant transcript 1 (NEAT1) could promote pancreatic cancer cell proliferation and metastasis both in vitro and in vivo associating with tumor size and lymph node (40). Oshima et al. found mutations in KRAS, p53 and SMAD4 were significantly associated with tumor size and lymph node metastasis (41). It is possible the extent of lymph node metastases tends to increase with tumor size (42,43).
Another expected risk factor for extensive lymph node metastasis was the poor tumor differentiation which influenced the clinical outcome of patients. In essence, poor degree of tumor differentiation reflects rapid tumor progression (44), whereas there are rich lymphatic networks around the pancreas including the internal pancreatic lymphatics, extrapancreatic lymphatics and peripancreatic lymphatics, which make lymph node metastasis easier. Actually, differentiation of tumor cells is controlled by complex regulatory networks, which can be explained with molecular mechanism of pancreatic cancer differentiation. Xie et al. reported dysregulated KLF4 expression associated with poor differentiation of pancreatic cancer (45). Milan et al. determined FOXA2 regulated well-and poorly differentiated pancreatic ductal adenocarcinoma cells via interactions with transcription factors (46). Liang et al. found patients with poor tumor differentiation had higher MAP4K4 expression, and consequently increased number of positive lymph nodes (47). All these data support a positive association between lymph node metastasis and poor tumor differentiation.
As borderline resectability is noted as a high risk for a margin-positive status in patients with BRPHC vascular resection and reconstruction is usually required in these patients. Nevertheless, the survival still decreases with the severity of vascular involvement (25). Surprisingly, the disease-free survival and the overall survival improved in patients with BRPHC undergoing extended lymphadenectomy in our study due to a wide range of lymph node invasion. There are two possible reasons for this phenomenon. On one hand, the superior mesentericportal vein invasion might occur in patients with higher rates of poor tumor differentiation (48) as well as in patients with a larger tumor size (49,50), which was also confirmed in our study; on the other hand, invasions of the plexuses around the arteries are the main causes of local recurrence after resection in these patients (51). All these suggest a higher rate of lymph node metastasis. Nevertheless, local tumor control by extended lymphadenectomy cannot overcome the negative aspects of pre-existing lymph node metastasis as well as superior mesenteric-portal vein involvement, which might explain why patients with BRPHC still have lower survival rates.
Taken together, we found that patients with BRPHC tended to have vast lymph node metastasis compared with patients with resectable disease. Extensive surgery for pancreatic carcinoma should be entertained in selected patients as it can improve their long-term survival.
The main limitation of this study is that it represented the experience of a single center. The number of patients in each subgroup is relatively small, which may limit the accuracy of our assessment. Future studies, preferably random clinical trials from multi-centers, are needed to further confirm our preliminary outcomes.