Introduction

The prognosis of patients with pancreatic ductal adenocarcinoma remains extremely dismal, with the overall survival rate being less than 5% [1]. The reason for the poor prognosis of this disease is that candidates for surgical resection represent less than 20% of patients with pancreatic carcinoma, due to locoregional spread or metastatic disease at the time of diagnosis [2, 3]. Surgical resection offers the only chance of cure or long-term survival in the treatment of this disease [4]. Furthermore, the 5-year survival rate of patients with pancreatic ductal adenocarcinoma who undergo surgical resection has been reported recently to be less than 20% [5–14]. To find useful prognostic factors, many investigators have attempted to analyze the clinicopathological features of patients with pancreatic ductal adenocarcinoma after surgical resection with univariate or multivariate analysis. Potential factors include tumor size [6, 7, 9, 13, 14], surgical margin status [6–9, 12–14], and postoperative adjuvant chemotherapy [5–7, 14]. In addition, nodal involvement is one of the most important factors for predicting the long-term survival of patients with pancreatic ductal adenocarcinoma [5–7, 14]. Lymph node metastasis initially occurs in the para-pancreatic nodes and finally spreads to the para-aortic lymph nodes through the nodes along the superior mesenteric artery [15, 16].

However, to our knowledge, there have been few reports concerning the prognostic impact of para-aortic lymph node metastasis, because the para-aortic lymph nodes are not routinely dissected in America and Europe [16–20]. In our institution, dissection of the para-aortic lymph nodes as well as the regional lymph nodes has been performed routinely for patients with pancreatic ductal adenocarcinoma since 1992. The aim of the present study was to evaluate the clinicopathological features of patients with para-aortic lymph node metastasis and to clarify the prognostic impact of para-aortic lymph node metastasis in pancreatic ductal adenocarcinoma. This was achieved by using univariate and multivariate survival analysis to assess patients treated at a single institution.

Methods

Medical records for 103 consecutive patients with pancreatic ductal adenocarcinoma who underwent pancreatic resection with regional and para-aortic lymph node dissection at the Department of Surgery, Hiroshima University Hospital, between January 1992 and December 2008 were reviewed retrospectively. The 103 patients were diagnosed as having no apparent para-aortic lymph node metastasis by preoperative imaging examinations. All patients underwent tumor resection with the aim of achieving cure and had a confirmed pathological diagnosis. Patients with pancreatic ductal adenocarcinoma derived from an intraductal papillary-mucinous neoplasm (16 patients) or a mucinous cystic neoplasm (one patient) were excluded from this analysis [21, 22].

Preoperative workup included ultrasonography, computed tomography, endoscopic sonography, endoscopic retrograde pancreatocholangiography, and percutaneous transhepatic cholangiography to evaluate the local or distant extension of the tumors. Tumor resection was abandoned if distant metastasis, invasion to the celiac or superior mesenteric artery, and apparent para-aortic lymph node metastasis were found by these preoperative examinations.

Patients with carcinoma in the pancreatic head usually underwent pylorus-preserving pancreatoduodenectomy. If the tumor was close to the duodenal bulb in the superior pancreatic head, conventional pancreatoduodenectomy was performed. All patients with carcinoma in the pancreatic body or tail underwent distal pancreatectomy with splenectomy. Total pancreatectomy was performed when the tumor invaded the whole pancreas. All 103 patients underwent regional and para-aortic lymph node dissection. The para-aortic lymph nodes were dissected from the upper part of the celiac trunk to the upper part of the origin of the inferior mesenteric artery. Intraoperative pathological assessment of the proximal or distal pancreatic margins was performed on frozen tissue sections. If the pancreatic margin was positive for cancerous cells, further resection of the pancreas was performed to the maximum extent possible.

After tumor resection, hematoxylin and eosin staining was performed. All specimens were examined pathologically, and each tumor was classified as well-differentiated, moderately differentiated, or poorly differentiated adenocarcinoma according to the predominant pathological grading of differentiation. Anterior serosal invasion, retropancreatic tissue invasion, choledochal invasion, duodenal invasion, portal vein invasion, lymph node metastasis, and extrapancreatic nerve plexus invasion were all examined pathologically. Surgical margins were considered positive if infiltrating adenocarcinoma was present at the proximal or distal pancreatic transection line or in the dissected peripancreatic soft tissue margins. The final stage of pancreatic carcinoma was examined pathologically according to the TNM classification system of malignant tumors published by the International Union Against Cancer (UICC), 6th edition [23].

Postoperative adjuvant chemotherapy was administered beginning in 1999. Patients who were offered postoperative adjuvant chemotherapy had three options after surgical resection: intra-arterial chemotherapy alone, intravenous chemotherapy alone, or intravenous and oral chemotherapy. Intra-arterial chemotherapy was performed with Seldinger’s method. Briefly, a catheter was placed into the common hepatic artery with a reservoir in the lower abdomen. An intra-arterial bolus injection of 5-fluorouracil (5-Fu) 160 mg/m2 was administered bi-weekly. Intravenous chemotherapy consisted of gemcitabine 700 mg/m2 administered bi-weekly for 30 min by intravenous drip infusion. Patients who received intravenous and oral chemotherapy were given intravenous gemcitabine 700 mg/m2 on day 1 and orally administered S-1 50 mg/m2 for 7 consecutive days; this cycle was repeated every 14 days [24]. No patient received radiation therapy during the study periods.

Patients were followed regularly in outpatient clinics by undergoing computed tomography twice a year for 5 years after surgery. Information on outcomes beyond 5 years after surgery was collected by telephone or personal interview. For patients who died, survival time after surgery and cause of death were recorded. For surviving patients, postoperative survival time and status of recurrence were recorded.

Survival analyses on four clinical factors (gender, age, tumor location, and use of adjuvant chemotherapy) and 13 pathological factors (tumor size, tumor differentiation, anterior serosal invasion, retropancreatic tissue invasion, choledochal invasion, duodenal invasion, portal vein invasion, lymph node metastasis, para-aortic lymph node metastasis, extrapancreatic nerve plexus invasion, surgical margin status, UICC pT factor, and UICC stage) were performed with univariate and multivariate methods. Clinicopathological factors and survival were compared between patients with and without para-aortic lymph node metastasis.

The χ2 test or Fisher’s exact test was used for comparison among the two groups. Survival curves were constructed based on the Kaplan–Meier method, and differences in survival curves were compared with a univariate log-rank (Mantel-Cox) test. Factors found to be significant on univariate analysis were subjected to multivariate analysis with a Cox proportional hazards model. A value of p < 0.05 was considered statistically significant. Statistical analysis was performed using the Macintosh version of StatView version 5.0 (SAS Institute, Cary, NC).

Results

The 103 eligible patients included 50 men and 53 women (median age: 67 years; range: 31–83 years), and 47 patients (46%) were more than 70 years old. No 30-day postoperative deaths occurred among the 103 patients. Tumors were located mainly in the pancreatic head in 73 patients and in the body or tail in 30 patients. Pylorus-preserving pancreatoduodenectomy, conventional pancreatoduodenectomy, distal pancreatectomy with splenectomy, and total pancreatectomy were performed for 65, 4, 30, and four patients, respectively. Postoperative chemotherapy was performed for 71 patients: intra-arterial chemotherapy alone in 3 patients, intravenous chemotherapy alone in seven patients, and intravenous and oral chemotherapy in 61 patients. Overall survival rates for the 103 patients were 74% at 1 year, 41% at 2 years, and 19% at 5 years (median survival, 14.0 months; range: 1–122 months).

The mean number of dissected lymph nodes was 29 (range: 2–75), and the mean number of dissected para-aortic lymph nodes was 5 (range: 1–26). There were 72 tumors (70%) with lymph node metastasis and 31 (30%) without lymph node metastasis. The number of involved lymph nodes ranged from 1 to 40 (median 3). Para-aortic lymph node metastasis was found in 18 patients (17%). The number of involved para-aortic lymph nodes ranged from 1 to 16 (median 2).

Table 1 indicates a comparison of clinicopathological factors between patients with and without para-aortic lymph node metastasis. Extrapancreatic nerve plexus invasion (p = 0.043), surgical margin status (p = 0.002), and UICC pT factor (p = 0.039) were associated significantly with para-aortic lymph node metastasis.

Table 1 Comparison of clinicopathological factors of patients with pancreatic ductal adenocarcinoma who did or did not have para-aortic nodal involvement
Full size table

Seventeen clinicopathological factors were investigated to determine their prognostic significance. The results of the log-rank test are shown in Table 2. Gender, age, tumor location, tumor size, anterior serosal invasion, choledochal invasion, and duodenal invasion did not influence postoperative survival. Univariate analysis revealed that postoperative adjuvant chemotherapy (p < 0.001), tumor differentiation (p = 0.037), retroperitoneal tissue invasion (p = 0.040), portal or splenic vein invasion (p = 0.002), extrapancreatic nerve plexus invasion (p < 0.001), lymph node metastasis (p < 0.001), para-aortic lymph node metastasis (p < 0.001), surgical margin status (p < 0.001), UICC pT factor (p = 0.003), and UICC stage (p < 0.001) were associated significantly with increased survival. These factors were entered into multivariate analysis with a Cox proportional hazards model, and use of postoperative adjuvant chemotherapy (p < 0.001) and absence of extrapancreatic nerve plexus invasion (p = 0.041) remained associated independently with longer survival (Table 3). In contrast, para-aortic lymph node metastasis (p = 0.078) was not associated significantly with survival in the final multivariate model. To avoid confounding due to nodal status and other pathological factors, neither UICC pT factor nor UICC stage were used as dependent variables in the multivariate survival analysis.

Table 2 Univariate survival analysis of prognostic factors for patients with pancreatic ductal adenocarcinoma
Full size table
Table 3 Multivariate survival analysis of prognostic factors for patients with pancreatic ductal adenocarcinoma
Full size table

Overall survival rates for patients without para-aortic lymph node metastasis were 79% at 1 year, 49% at 2 years, 29% at 3 years, and 23% at 5 years, whereas those for patients with para-aortic lymph node metastasis were 53% at 1 year, 12% at 2 years, and 0% at 3 years. All patients who exhibited para-aortic lymph node metastasis except for one died of recurrence within 2 years after surgery (Fig. 1). The 5-year survival rate of patients with para-aortic lymph node metastasis was significantly lower than that of patients without nodal involvement (p < 0.001). In addition, a significant difference in the survival rates of patients with or without para-aortic lymph node metastasis was found among patients who had nodal involvement (p = 0.012, Fig. 2). Figure 3 indicates survival curves of patients with para-aortic lymph node metastasis who did or did not receive postoperative adjuvant chemotherapy. Survival time of patients who received postoperative adjuvant chemotherapy was significantly longer than that of patients who did not (median survival time, 14.1 months versus 10.1 months; p = 0.025).

Fig. 1
figure 1

Comparison of postoperative survival in patients with or without para-aortic lymph node involvement following resection for pancreatic ductal adenocarcinoma (p < 0.001)

Full size image
Fig. 2
figure 2

Comparison of postoperative survival in patients who had no lymph node involvement, lymph node involvement without para-aortic involvement, and para-aortic lymph node involvement after resection for pancreatic ductal adenocarcinoma. N(−) versus N(+), para-A N (−) p = 0.006; N(−) versus para-A N (+), p < 0.001; N(+), para-A N (−) versus para-A N (+) p = 0.012. N lymph node involvement, Para-A N para-aortic lymph node involvement

Full size image
Fig. 3
figure 3

Comparison of postoperative survival in patients with para-aortic lymph node involvement who did or did not receive postoperative adjuvant chemotherapy following resection for pancreatic ductal adenocarcinoma (p = 0.025)

Full size image

Discussion

Lymph node involvement has been reported to range from 50 to 78% in cases of resected pancreatic ductal adenocarcinoma [5–14]. However, reports concerning para-aortic lymph node metastasis rates are few, because para-aortic lymph node dissection is performed routinely only in Japanese institutions. According to these reports, the rate of para-aortic lymph node metastasis ranged from 13 to 26% (Table 4) [16–20]. In the present study the rates of nodal involvement and para-aortic lymph node involvement were 70 and 17%, respectively. These rates were similar to those in previous reports.

Table 4 Reports on survival of patients with para-aortic lymph node metastasis in pancreatic ductal adenocarcinoma
Full size table

The prognosis of patients with para-aortic lymph node involvement has been reported to be extremely poor. The reported 2-year survival rates of patients with para-aortic lymph node involvement ranged from 0 to 18%, with few 3-year survivors (Table 4) [16–20]. Doi et al. [20] reported that metastasis to the para-aortic lymph nodes was the single independent factor associated with a shorter survival time by multivariate analysis (median survival time, 5.1 months), and they concluded that surgical resection should be abandoned and alternative treatment strategies including chemotherapy or radiotherapy should be considered if para-aortic lymph node metastasis is detected. Shimada et al. [19] reported that the median survival time in 26 patients with para-aortic lymph node involvement was 13 months versus 30 months in 106 patients without para-aortic lymph node involvement (p < 0.001), and they concluded that radical pancreatectomy including extended lymphadenectomy should not be recommended when para-aortic lymph node involvement is confirmed by intraoperative frozen diagnosis. In our study, the survival time of patients with para-aortic lymph node involvement was significantly shorter than that of patients without para-aortic lymph node involvement by univariate analysis, but para-aortic lymph node involvement was not found to be an independent prognostic factor by multivariate analysis. Moreover, adjuvant chemotherapy with new anticancer drugs significantly improved the survival of patients with para-aortic lymph node involvement. Further studies on larger numbers of patients are needed to determine whether surgical resection should be abandoned for patients who are found to have para-aortic lymph node involvement by pathological examination despite no apparent para-aortic lymph node metastasis by preoperative imaging examinations.

Reports concerning the survival benefits of postoperative adjuvant chemotherapy for patients with para-aortic lymph node metastasis in pancreatic ductal adenocarcinoma are scarce. It was reported that application of adjuvant systemic chemotherapy rather than radiotherapy might be suitable in para-aortic node-positive patients, because the major cause of failure in these patients was distant metastases [19]. Recently, postoperative gemcitabine chemotherapy, compared with surgery alone, significantly delayed the development of recurrent disease after complete resection of pancreatic ductal adenocarcinoma in a multi-center randomized controlled phase III trial (CONKO-001) [25]. Moreover, gemcitabine plus S-1 therapy has been associated with an excellent survival benefit in patients with unresectable [26] or resected [24, 27] pancreatic ductal adenocarcinoma. In the present series, we selected mainly new anticancer drugs including gemcitabine or S-1, as an adjuvant chemotherapy regimen. As a result, adjuvant chemotherapy significantly improved the survival of patients with para-aortic node involvement compared with surgery alone. We believe that adjuvant chemotherapy with new anticancer drugs may provide a survival benefit for para-aortic lymph node-positive patients in pancreatic ducta adenocarcinoma, although our result is based on a small number of patients.

In this study, para-aortic lymph node metastasis was associated significantly with extrapancreatic nerve plexus invasion, UICC pT factor, and surgical margin status. No patients who had pT1 or pT2 tumors exhibited para-aortic lymph node involvement. Sakai et al. [18] reported that no patients with a tumor diameter of less than 2 cm had para-aortic lymph node involvement, and Shimada et al. [19] reported that para-aortic lymph node involvement was associated with elevated CA19-9 a month after operation, larger tumor size, and a positive surgical margin. In addition, Connor et al. [28] described that para-aortic lymph node metastasis was found to be associated with a positive resection margin. Based on our results and these reports, patients with para-aortic lymph node involvement tend to exhibit extensive locoregional invasion of the tumor. For this reason, margin-negative resection cannot be performed for most patients with para-aortic lymph node involvement.

In the present study, several node-positive patients without para-aortic lymph node metastasis had longer survival and one patient who has survived for more than 5 years had only one involved node in the regional lymph nodes. There have been several previous reports concerning relationship between the number of involved lymph nodes and survival in pancreatic ductal adenocarcinoma. Riediger et al. [29] reported that patients with a single metastatic node had the same survival as node-negative patients and that survival of patients with no more than one metastatic node was significantly better than that of patients with two or more involved nodes. Another group of investigators reported that more than three lymph nodes positive for metastasis was an independent predictor of survival in a consecutive series of 77 patients: they established that the cutoff of three nodes was useful in identifying subsets of patients with worst prognosis among those with positive nodes [30]. We routinely performed dissection of the regional lymph nodes in our series. We believe that lymph node dissection may contribute to longer survival for patients with pancreatic ductal adenocarcinoma, especially for node-positive patients without para-aortic lymph node metastasis.

In conclusion, the prognosis of patients with para-aortic lymph node involvement is extremely poor compared with that of patients without para-aortic lymph node involvement. However, postoperative adjuvant chemotherapy with new anticancer drugs may improve the survival of patients with para-aortic lymph node involvement. Further studies are needed to clarify the effectiveness of surgical resection for patients with para-aortic lymph node involvement.