Risk factors related to metastasis of para-aortic lymph nodes in pancreatic ductal adenocarcinoma: A retrospective observational study.

This study was designed to explore the risk factors related to metastasis of para-aortic lymph node (PALN).Clinicopathologic data of 241 patients with resectable or borderline resectable pancreatic cancer who underwent pancreaticoduodenectomy with extended lymphadenectomy between January 2008 and December 2015 were collected, potential factors related to metastasis of PALN were analyzed.Positive rate of PALN was 19.5% (47/241). Univariate analysis showed that back pain (P = .028), preoperative CA19-9 level (P < .001), tumor size (P < .001), portal vein (PV)/superior mesenteric vein (SMV) invasion (P < .001), superior mesenteric artery (SMA) invasion (P < .001), and diameter > 1.0 cm were in correlation with PALN involvement, multivariate analysis revealed that preoperative CA19-9 level, PV/SMV invasion, SMA invasion and diameter > 1.0 cm were independent risk factors to metastasis of PALN. Patients with LN8+ had a higher positive rate of PALN than with LN8- (38.1% vs 15.6%, P = .001), similar results could be found when LN12+ (35.8% vs 13.2%, P < .001) and LN14+ (41.2% vs 11.0%, P < .001), multivariate analysis showed that LN8+ and LN14+ were closely in correlation with PALN metastasis.Several factors were related to the status of PALN, preoperative CA19-9 level, PV/SMV invasion, SMA invasion and diameter > 1.0 cm were 4 independent risk factors to PALN metastasis. LN8+ and LN14+ were 2 strong predictors of PALN metastasis. A comprehensive analysis covering all possible risk factors related to metastasis of PALN should be given before design of treatment plan whenever involvement of PALN was suspected.

Introduction

Pancreatic cancer characterized by low resection rate and poor prognosis has become one of the leading causes of cancer-related death.[ A recent study shows that pancreatic cancer is the fourth most fatal cancer in men after lung, colorectal, and prostate cancer; similarly, it is also found to be the fourth most fatal cancer in women after breast, colorectal and lung cancer.[ Several factors including pathologic type, depth of invasion, status of lymph nodes, status of resection margin, etc. are closely in correlation with prognosis of pancreatic cancer after operation.[ Among these, para-aortic lymph node (PALN) has been confirmed as one of the independent risk factors for unfavorable prognosis.[ As an example, median survival time of patients with PALN metastasis ranged from 5.1 to 15.7 months reported by some centers.[ Involvement of PALN was the single independent factor associated with a shorter survival time confirmed by a study designed by Doi et al,[ and the similar results could be found in some other studies designed by Murakami et al,[ Sakai et al,[ Yoshida et al,[ Kanda et al,[ and so on.

It is important and necessary to predict the status of PALN before operation and to analyze factors related to metastasis due to its pivotal role for prognosis of patients with pancreatic cancer, the status of PALN is one of the critical factors for designing the therapeutic regimen. As the nonregional lymph nodes, metastasis of PALN is deemed as a contraindication of surgery by some authors although contrary opinions are held by others. A certain proportion of patients may lose the opportunity of surgery on account of the suspected metastasis of PALN when pancreatic cancer is diagnosed. Negative PALN may be found for some of these patients if pathological examination is carried out because of the false-positive results of imaging. It has important significance to evaluate the status of PALN accurately whenever pancreatic cancer is confirmed. Most frequently, CT and MRI are used to judge the status of lymph node before operation, or to evaluate the resectability of pancreatic cancer. However, imaging-based preoperative detection techniques have yielded disappointing results, the accuracy of CT or MRI is 72.5% or 88.4% for detecting PALN metastasis.[ As mentioned above, PALN has been classified as nonregional lymph node, and metastasis of PALN has been recognized as distant metastasis. Because of the very poor prognosis when PALN involved, surgical resection is not considered for pancreatic cancer.[ This is may be the pivotal limiting factor. Judgment of PALN status is mainly relied on imaging examination for patients without operation, consequently, accuracy of PALN status is unavoidably weakened.[ Risk factors related to metastasis of PALN should be analyzed carefully, which may help to improve judgment accuracy and to make better treatment plan. With the aim of exploring risk factors, extended lymphadenectomy with No. 16a2 and No. 16b1 lymph nodes included has been performing for pancreatic head cancer in our hospital recent years.

Materials and methods

Patients enrolled in this study

Clinicopathological data of patients with ductal adenocarcinoma of pancreatic head confirmed by pathological examination postoperatively who underwent standard or extended pancreaticoduodenectomy with extended lymphadenectomy between January 2008 and December 2015 were collected and analyzed. Inclusion criteria: resectable or borderline resectable pancreatic head cancer based on the definition of NCCN Guidelines 2018; pancreatic ductal adenocarcinoma was confirmed by postoperative pathology. Exclusion criteria: carcinomas of lower end of common bile duct, ampulla, duodenal papilla, and uncinate process; neuroendocrine neoplasm; acinar cell carcinomas; benign tumors; patients who underwent neoadjuvant therapy. The protocol and procedures employed were reviewed and approved by the institutional review committee of Beijing Chaoyang Hospital.

Definitions

PALNs were dissected from the level of the celiac trunk down to the root of the inferior mesenteric artery, including stations No.16 a2 and No.16 b1. In this study, standard or extended pancreaticoduodenectomy was consistent with the definitions recommended by International Study Group for Pancreatic Surgery (ISGPS).[ Lymph nodes of stations 5, 6, 8a, 8p, 9, 12a, 12p, 12b1, 12b2, 12c, 13a-b, 14a-d, 16a2, 16b1, and 17a–b were removed, which was defined as extended lymphadenectomy.[

Evaluation before operation

Either contrast-enhanced CT or MRI was used for evaluation of resectability and judgment of lymph node status, especially the status of nonregional lymph node including PALN. Metastasis of lymph node was suspected when the greatest diameter was >1.0 cm.

Adjuvant therapy after operation

Patients with ECOG 0-1 were suggested to receive adjuvant therapy within 8 to 12 weeks after operation. Regimes including gemcitabine, gemcitabine plus capecitabine, capecitabine, and FOLFIRINOX (leucovorin and fluorouracil plus irinotecan and oxaliplatin) were selected. Radiotherapy was recommended for patients with positive margin.

Factors for analysis

The following potential risk factors for PALN metastasis were analyzed: general factors included age at diagnosis, gender, back pain or not, preoperative CA19-9 level; imaging factor mainly included the diameter of PALN observed in computed tomography; pathologic factors covered tumor size, tumor differentiation, portal or superior mesenteric vein (PV/SMV) invasion, superior mesenteric arterial (SMA) invasion. Meanwhile, we also analyzed the station of lymph node (LN) which was in close relationship with the status of PALN.

Statistical analysis

Sample size was calculated by using the formula before conducting this study and no less than 206 patients should be enrolled. SPSS 16.0 (IBM, Chicago, IL) was used for data analysis. A P < .05 was considered to be statistically significant. Survival analysis was carried out using the Kaplan–Meier method with the log-rank test. Categorical variables were analyzed by Chi-square test. A multivariable analysis accomplished by a binary logistic regression model was used to test for independent risk factors associated with metastasis of PALN.

Results

General parameters

A total of 241 patients including 108 males and 133 females with ductal adenocarcinoma of pancreatic head were enrolled in this study. The age of patients at diagnosis ranged from 40 to 79 years, with a median age of 57 years. The CA19-9 level before operation ranged from 17 to 1848 U/mL, with a median of 156 U/mL. Twenty-five patients suffered from back pain preoperatively. Tumor size ranged from 1.4 to 6.4 cm, with a median of 2.5 cm. As classification of the 8th edition AJCC/UICC TNM staging system, T1 was found in 43 patients, T2 in 169 patients and T3 in 29 patients, respectively. PV/SMV invasion was found in 47 patients, SMA invasion was found in 24 patients. Swollen PALN (diameter > 1.0 cm) observed in CT scan was found in 44 patients. Among these patients, 213 patients received chemotherapy: gemcitabine for 46 cases, gemcitabine + capecitabine for 83 cases, capecitabine for 55 patients, and FOLFIRINOX for 29 patients. Of the 213 patients, 21 patients received chemotherapy associated with radiotherapy.

Lymph nodes

A total of 5431 lymph nodes were detected, and the number of lymph nodes retrieved ranged from 6 to 45 with a median number of 21 per patient. Five hundred eighty-one PALNs were retrieved (ranged from 1 to 5 with a median of 2). Lymph node involvement was detected in 164 patients (68.0%), number of patients at different station of lymph node is shown in Fig. 1, and PALN metastasis was confirmed in 47 patients (19.5%). As classification of the 8th edition AJCC/UICC TNM staging system, N0 was found in 77 patients, N1 in 52 patients and N2 in 112 patients. The greatest diameter of positive PALN ranged from 0.5 to 1.6 cm with the median diameter of 1.1 cm, and 0.3 to 1.2 cm with the median diameter of 0.6 cm was examined in patients with negative PALN.

Figure 1

Number of patients with positive node and negative node at different station of lymph node.

Relationship between factors and status of PALN

Univariate analysis showed that age at surgery, gender, and BMI were not the risk factors to metastasis of PALN. Patients with back pain had a higher positive rate of PALN metastasis compared to patients without (P = .028); preoperative CA19-9 level was another potential risk factor, a significantly higher positive rate of PALN metastasis was found when CA19-9 was more than 1000 U/mL (P < .001). Several pathologic factors including tumor size, PV/SMV invasion, and SMA invasion were in correlation with metastasis of PALN. As classification of the 8th edition AJCC/UICC TNM staging system, tumor size ≤2 cm, > 2 and ≤4 cm, >4 cm was defined as T1, T2, and T3, respectively. Patients were divided into 3 groups according to the current classification, different positive rates of PALN were revealed. Patients had the highest positive rate of 58.6% when tumor size was >4 cm compared to 7.0% when ≤2 cm. Patients with PV/SMV invasion or SMA invasion had the higher positive rate than without, diameter > 1 cm was also correlated to high metastasis rate of PALN, as shown in Table 1. Multivariate analysis showed that preoperative CA19-9 level, PV/SMV invasion, SMA invasion and diameter > 1 cm were 4 independent risk factors to metastasis of PALN, which are listed in Table 2.

Table 1

Univariate analysis for the positive rates of PALN in patients with different factors.

Table 2

Multivariate analysis for risk factors to metastasis of PALN.

Relationship between different stations of LNs and metastasis of PALN

As univariate analysis shown in Table 3, much more patients with positive PALN when patients had metastasis of the 8th station of LN compared with nonmetastasis of LN8 (38.1% vs 15.6%, P = .001), similar results were found when metastasis was confirmed in LN12, LN13, and LN14. Whereas no significantly different proportion of patients with positive PALN was detected in LN5, LN6, LN9, and LN17, similar proportions of patients with PALN involvement were found regardless of the status of these stations. Multivariate analysis showed that LN8+ and LN14+ were 2 strong predictors to metastasis of PALN.

Table 3

Univariate and multivariate analysis for the relationship between different stations of LN and metastasis of PALN.

Different survival time of patients with or without PALN metastasis

As shown in Table 4, patients with PALN involvement had a much poorer prognosis compared to patients without. The 1-year overall survival rate (OS) was 79.4% in patients with negative PALN, but only 52.7% in patients with positive PALN. A 5-year OS of 22.6% was calculated in patients without PALN involvement, whereas no 3-year survival was found in patients with PALN involvement, let alone 5-year survival (Fig. 2).

Table 4

The survival rates of patients with or without PALN metastasis.

Figure 2

Comparison of survival time between patients with para-aortic lymph node metastasis and without.

Discussion

Among all malignant solid tumors, pancreatic cancer is characterized by delayed diagnosis, low curative rate, and high mortality.[ Cancer Statistics in China which is completed by National Cancer Center of China shows that the estimated new cancer cases and deaths of pancreas are 90.1 and 79.4 thousand in 2015.[ With the increasing incidence and also with the high mortality, pancreatic cancer has been becoming a major public problem. Despite decades of effort, pancreatic cancer remains one of the most aggressive and lethal malignancies, and its 5-year survival rate remains at only ∼5%.[ Poor prognosis is in correlation with several factors, a key factor confirmed by centers is metastasis of PALN.[ As the nonregional lymph nodes, metastasis of PALN has been recognized as distant metastasis, that is to say, a large proportion of patients with PALN involvement will lose the opportunity to receive operation. Before the design of treatment plan, estimation of clinical staging is an essential process. The status of PALN is one of focus of attention, and judgment is mainly based on CT or MRI. Metastasis is highly suspected when diameter of lymph node is more than 1.0 cm, and patient with diameter of PALN > 1.0 cm will lose the opportunity of operation. However, false positive of imaging is unavoidable, and several other factors may help us to improve the accuracy of judgment. Thus, it is necessary to explore the high-risk factors related to metastasis of PALN.

Recently, judging the status of PALN are mainly depended on imaging examinations including abdominal CT, MRI, or PET-CT if operation is avoided. While, imaging-based preoperative detection techniques have yielded disappointing results.[ A study designed by Imai et al[ showed the 0 sensitivity of CT, MRI, and PET-CT for detecting PALN metastasis, and the accuracy was 72.5%, 88.4%, and 90.0%, respectively. Meanwhile, their study confirmed that no significant difference was found for longer diameter between metastatic and nonmetastatic PALN, same results could also be found for shorter diameter and long/short ratio although diameter of PALN > 1 cm was confirmed as an independent risk factor in this study. However, metastasis of PALN could not be ruled out when diameter was no more than 1.0 cm. If potential risk factors to metastasis of PALN can be confirmed, it may help to improve the accuracy of preoperative diagnosis. There are a few of studies about high-risk factors related to metastasis of PALN. A study designed by Komo et al[ showed that PV/SMV invasion, tumor size > 3 cm and regional lymph node metastasis were associated with high involvement rate of PALN. Our study showed that back pain, high CA19-9 level (more than 500 U/mL, especially more than 1000 U/mL), larger than 4 cm of tumor size, PV/SMV invasion, and SMA invasion were potentially had the higher positive rate of PALN, and multivariate analysis revealed that preoperative CA19-9 level, PV/SMV invasion, and SMA invasion were independent risk factors to metastasis of PALN. Based on our results, patients who had extremely high CA19-9 level, or PV/SMV and/or SMA invasion detected on imaging examination at diagnosis should be highly suspected to have PALN metastasis although no swollen lymph nodes were found in the para-aortic area.

Which stations of LN should be involved when PALN metastasis happened? Several researches in this area have been completed, as an example, using the technique of injecting activated carbon particles or 111In colloid, Nagakawa et al[ concluded that passing stations 13 and 14 were the main lymphatic route to PALN. Kayahara et al[ showed that the main lymphatic pathway from the head of pancreas to PALN was via the station 14, a study designed by Kanda et al[ found that LN12+, LN14+, and LN17+ were independent risk factors to metastasis of PALN, thus station 14 was deemed as “junctional LN” by some centers. In the present series, a higher positive rate of PALN was found when LN8, LN12, LN13, or LN14 was involved, and multivariate analysis showed that involvement of LN8 and LN14 were independent relative factors to metastasis of PALN. Owing to the close correlation between the 2 stations, we want to know whether LN16 should be removed if metastasis of LN14 is highly suspected during operation procedure although extended lymphadenectomy is not recommended for pancreatic cancer.

Most of studies confirmed that patients with PALN involvement have decreased survival time.[ As an example, a study designed by Murakami et al[ showed that the 1-, 2-, 3-, and 5-year OS rates were 79%, 49%, 29%, and 23% in patients without metastasis of PALN, whereas 1-, 2-, and 3-year OS rates were 53%, 12%, and 0% in patients with metastasis of PALN. In this study, patients with PALN involvement had a poorer prognosis compared with patients without.

The main drawback of this study is the small sample which collected from our single center, meanwhile, several important influence factors have not been included in this study. With the aim to drawing a definitive conclusion, a big sample from multicenter which included more comprehensive risk factors is an urgent need.

In conclusions, patients with metastasis of PALN have poorer prognosis. Several factors are related to metastasis of PALN, besides the diameter > 1.0 cm, preoperative CA19-9 level, PV/SMV invasion, and SMA invasion are also 3 independent risk factors. Status of LN8 and LN14 are in correlation with PALN, a higher positive rate can be found when metastasis is confirmed in LN8 or LN14. A comprehensive analysis covering all possible risk factors related to metastasis of PALN should be given before design of treatment plan whenever involvement of PALN was suspected.

Acknowledgment

All authors thank Zhang Xin-xue and Dai Yang for supporting data, their supports are the key factor for completion of this manuscript.

Author contributions

Conceptualization: Xingmao Zhang, Jie Zhang, Qiang He.

Data curation: Xingmao Zhang, Jie Zhang, Hua Fan, Yu Liu, Qiang He.

Formal analysis: Xingmao Zhang, Jie Zhang, Hua Fan, Qiang He.

Investigation: Yu Liu.

Methodology: Hua Fan.

Software: Hua Fan, Yu Liu.

Supervision: Hua Fan, Yu Liu, Qiang He.

Writing—original draft: Xingmao Zhang, Jie Zhang, Qiang He.