Abstract
Background/Aim: The diagnostic value of serum DUPAN-2 level has been reported; however, the relationship between preoperative DUPAN-2 level and recurrence pattern has not been fully investigated in pancreatic ductal adenocarcinoma (PDAC). Patients and Methods: We retrospectively analyzed 50 patients with PDAC who underwent pancreatectomy. The relationships between clinicopathologic factors and site-specific disease-free survival (DFS) were analyzed using Cox proportional hazard and receiver operating characteristic (ROC) curve analyses. Results: The tumor location was the pancreatic head in 31 patients and the body/tail in 19 patients. Of the 50 patients, 34 had recurrence (median DFS, 11 months). Fifteen patients had hematogenous recurrence, and 16 had locoregional recurrence. In multivariate analysis, adjuvant chemotherapy [p=0.01; odds ratio (OR)=8.10; 95% confidence interval (CI)=1.58-41.6] and venous invasion (p=0.01; OR=8.33; 95%CI=1.53-45.4) were significant factors for hematogenous recurrence-free survival, whereas the neutrophil-to-lymphocyte ratio (p=0.03; OR=2.57; 95%CI=1.10-5.98) and DUPAN-2 level (p<0.01; OR=1.00; 95%CI=1.000-1.002) were significant factors for locoregional recurrence-free survival. In ROC curve analysis, the area under the curve of DUPAN-2 level was 0.613 for hematogenous recurrence and 0.682 for locoregional recurrence. In the log-rank test, the hematogenous and locoregional recurrence-free survival rates of patients with higher DUPAN-2 levels were significantly worse than those with lower DUPAN-2 level. Conclusion: Elevation of preoperative DUPAN-2 level independently predicts locoregional recurrence after surgery. Patients with elevated preoperative DUPAN-2 level may benefit from neoadjuvant chemoradiation therapy to avoid postoperative locoregional recurrence.
Patients with pancreatic ductal adenocarcinoma (PDAC) undergoing pancreatectomy still have a poor prognosis. Up to 70% of patients with PDAC experience postoperative recurrence despite curative resection and adjuvant chemotherapy (1, 2). The sites of recurrence after pancreatectomy for PDAC have been previously reported, with the main recurrence sites being the liver, peritoneum, locoregional sites, lung, and lymph nodes (2, 3). Several studies have demonstrated that the period between surgery and recurrence and the survival duration after recurrence differ among patients with different recurrence patterns (3–5). Therefore, investigating the predictors of recurrence pattern is crucial to predict patient survival or to decide appropriate preoperative treatments.
The clinicopathologic factors predictive of recurrence patterns in patients with PDAC have been investigated. Several studies have reported predictors of locoregional recurrence, including resection margin status (5), tumor size (5, 6), positive lymph node status (6, 7), perineural invasion (7), preoperative carbohydrate antigen 19-9 (CA19-9) level (7), and neoadjuvant therapy (6, 8). In addition, several other studies have reported predictors of distant or hematogenous recurrence, including the lymph node ratio (4) and preoperative CA19-19 or duke pancreatic monoclonal antigen type 2 (DUPAN-2) levels (9). Of the recurrence patterns, only locoregional recurrence can be prevented with preoperative local therapies (i.e., chemoradiation therapy) or excellent surgical procedures.
The clinical relevance of preoperative tumor markers has been previously investigated. Many studies have demonstrated that an elevated preoperative CA 19-9 level strongly predicts poor prognosis in patients with PDAC undergoing curative resection (10). Approximately 6%-10% of patients do not secrete CA19-9 because of their Lewis negative blood phenotype. Therefore, DUPAN-2 and S-pancreas antigen-1 (SPAN-1), the expression of which is not influenced by the Lewis blood phenotype, may have important roles in the diagnosis or prognosis of PDAC. However, only a few studies have demonstrated the prognostic values of preoperative SPAN-1 and DUPAN-2 levels (9, 11–13). Thereby, the prognostic roles of these preoperative tumor markers remain unclear.
Pancreatic cancer is classified into resectable, locally advanced or borderline resectable (LA/BR), and nonresectable tumors. Even among patients with resectable PDAC, approximately 60%-70% experience recurrence after pancreatectomy with a curative intent (4, 14, 15). In patients with LA/BR PDAC, neoadjuvant chemotherapy and chemoradiation therapy [total neoadjuvant therapy (TNT)] are recommended to improve surgical resectability and postoperative prognosis (7, 16, 17). TNT involves the administration of chemoradiation therapy to control locoregional tumor invasion and systemic chemotherapy to prevent distant metastases (17). Meanwhile, upfront surgery followed by adjuvant chemotherapy is the universally accepted standard treatment for resectable pancreatic cancer (18). However, recent studies have recommended TNT for patients with resectable PDAC, although the survival benefit of neoadjuvant therapy remains controversial (8). Therefore, the selection of candidates for TNT among patients with resectable PDAC is an important issue for improving patient survival or reducing the unnecessary adverse effects of therapies.
In this study, we retrospectively investigated the recurrence patterns after pancreatectomy and their predictive factors in patients with PDAC, with a focus on the preoperative serum levels of tumor markers such as carcinoembryonic antigen (CEA), CA19-9, SPAN-1, and DUPAN-2.
Patients and Methods
Patients. From October 2004 to July 2020, a total of 89 patients with PDAC underwent pancreatectomy at Oita Prefectural Hospital, Oita, Japan. All patients had histologically confirmed ductal adenocarcinoma. Of these 89 patients, 6 had histologically confirmed liver metastasis or were positive for para-aortic lymph node metastases, 3 had adenosquamous carcinoma or anaplastic carcinoma, and 2 had operative death. These 11 patients were excluded from the study. Of the remaining 78 patients, 52 had preoperative DUPAN-2 level measurements. Two of the 52 patients were lost to follow-up and thus excluded. Finally, 50 patients who underwent pancreatectomy with a curative intent were enrolled in this study.
Clinicopathologic factors and recurrence pattern. We investigated the following clinicopathologic factors to identify the risk factors for recurrence pattern after pancreatectomy: sex, body mass index, pathologic factors (tumor size, serosal invasion, retropancreatic invasion, portal invasion, plexus invasion, nodal metastasis, lymphatic invasion, venous invasion, and neural invasion), preoperative nutritional factors [C-reactive protein-to-albumin ratio, neutrophil-to-lymphocyte ratio (NLR), and prognostic nutritional index (PNI)], preoperative tumor markers (CEA, CA19-9, SPAN-1, and DUPAN-2), postoperative complications, and adjuvant chemotherapy. All biochemical and tumor marker measurements were obtained before the initial therapies. Recurrent tumor was confirmed with radiologic examination (computed tomography, magnetic resonance imaging, or positron emission tomography). The first confirmed site of recurrence was recorded as the recurrence site for each patient. In this study, the recurrence pattern after pancreatectomy was defined as either hematogenous recurrence (recurrent tumor in the liver or lung) or locoregional recurrence (recurrent tumor in the retropancreatic tissue around the pancreatic stump or in the nerve plexus of the celiac or superior mesenteric arteries).
Measurement of tumor markers. The serum levels of tumor markers were evaluated using a chemiluminescent enzyme immunoassay (Lumipulse® Presto®, Fujirebio Inc. Tokyo, Japan) for CA19-9 (normal range, ≤37.0 U/ml) and CEA (normal range, ≤5 ng/ml), immunoradiometric assay for SPAN-1 (normal range, ≤30 U/ml; outsourced to BML Inc., Tokyo, Japan), and an enzyme immunoassay for DUPAN-2 (normal range, ≤150 U/ml; outsourced to SRL Inc., Tokyo, Japan). The chemiluminescent enzyme immunoassay for CA19-9 and CEA were performed according to the manufacturers’ instructions. For SPAN-1 and DUPAN-2, when the measured value was in the normal range, the patients’ serum levels were considered to be 30 U/ml for SPAN-1and 150 U/ml for DUPAN-2.
Statistical analysis. The disease-free survival (DFS) rates were calculated using the Kaplan–Meier method, and survival curve analysis was performed using the log-rank test. The relationship between DFS according to recurrence pattern and each clinicopathologic factor was compared using univariate Cox proportional hazard models. Variables with p<0.1 in univariate analyses were included in multivariate analyses. Values of p<0.05 were considered significant in all analyses. The cutoff value of preoperative DUPAN-2 level according to recurrence pattern was defined using receiver operating characteristic (ROC) curve analysis and measurements of area under the curve (AUC). All statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), which is a graphical user interface for R (version 3.3.3; The R Foundation for Statistical Computing, Vienna, Austria). More precisely, it is a modified version of R Commander (version 1.3-6) designed to add statistical functions frequently used in biostatistics (19).
Results
Patient characteristics. The 50 patients included in the study comprised 24 men (48%) and 26 women (52%) with a mean age of 69.2 years. The location of the tumor was the pancreatic head in 31 patients (62%) and the body/tail in 19 (38%). Accordingly, 31 patients underwent pancreatoduodenectomy and 19 distal pancreatectomy. The median follow-up period after surgery was 29 months. Adjuvant chemotherapy with oral S1 was administered in 34 (68%) patients, followed by intravenous infusion therapy with gemcitabine hydrochloride in 3 (6%) patients. Neoadjuvant chemotherapy was used in 7 (14%) patients, most of whom received albumin-bound paclitaxel plus gemcitabine therapy.
Recurrence patterns. Of the 50 patients, 34 (68.0%) developed recurrence after pancreatectomy with a median DFS of 11 months. The recurrence sites after pancreatectomy are listed in Table I. The first recurrence site was local only in 7 patients (14.0%), the liver only in 4 (8.0%), the lung only in 2 (4%), the liver and lung in 1 (2.0%), the peritoneum only in 4 (8.0%), and the lymph nodes only in 4 (8.0%). The remaining 12 patients with postoperative recurrence had recurrent tumors at multiple sites as follows: local and distant sites in 8 patients and multiple distant sites in 4 patients. The patients were classified according to recurrence pattern as follows: 15 patients had hematogenous recurrence, 16 had local recurrence, and 8 had other recurrence. The clinicopathologic characteristics of the patients according to recurrence pattern are listed in Table II. Patients with local recurrence had higher levels of tumor markers (CA19-9, SPAN-1, and DUPAN-2) than those with other recurrence or with no recurrence.
Recurrence sites after pancreatectomy in patients with pancreatic ductal adenocarcinoma.
Clinicopathologic characteristics according to recurrence pattern in patients with pancreatic ductal adenocarcinoma who underwent pancreatectomy.
Univariate and multivariate Cox proportional hazard regression analyses according to recurrence pattern. The results of analyses using univariate and multivariate Cox proportional hazard models for hematogenous recurrence-free survival are summarized in Table III. In univariate analyses, the presence of nodal metastasis, lymphatic invasion, venous invasion, higher preoperative PNI, higher DUPAN-2 level, and absence of adjuvant chemotherapy were significant (p<0.1) predictive factors for hematogenous recurrence. In multivariate analyses, the presence of venous invasion [p=0.01; odds ratio (OR)=8.33; 95% confidence interval (CI)=1.53-45.4] and the absence of adjuvant chemotherapy (p=0.01; OR=8.10; 95% CI=1.58-41.6) were significant (p<0.05) predictive factors for hematogenous recurrence after pancreatectomy. The preoperative serum DUPAN-2 level was not a significant (p = 0.73) factor in multivariate analyses.
Univariate and multivariate Cox proportional hazard regression analyses of predictors of hematogenous recurrence after pancreatectomy for pancreatic ductal adenocarcinoma.
The results of analyses using univariate and multivariate Cox proportional hazard models for locoregional recurrence-free survival are summarized in Table IV. In univariate analyses, higher preoperative NLR and CA-19-9, SPAN-1, DUPAN-2 levels, and absence of adjuvant chemotherapy were significant (p<0.1) predictors of locoregional recurrence. In multivariate analyses, higher preoperative NLR (p=0.03; OR=2.57; 95%CI=1.10-5.98) and preoperative DUPAN-2 level (p<0.01; OR=1.00, 95% CI=1.000-1.002) were significant (p<0.05) factors to predicting locoregional recurrence after pancreatectomy.
Univariate and multivariate Cox proportional hazard regression analyses of predictors of locoregional recurrence after pancreatectomy for pancreatic ductal adenocarcinoma.
ROC curve analysis. The ROC curves for hematogenous and locoregional recurrence according to preoperative DUPAN-2 level are shown in Figure 1. The cutoff value of preoperative DUPAN-2 level was 150 U/ml for hematogenous recurrence and 200 U/ml for locoregional recurrence. The AUC was 0.613 for hematogenous recurrence (Figure 1A) and 0.682 for locoregional recurrence (Figure 1B).
Receiver operating characteristic curves of preoperative duke pancreatic monoclonal antigen type 2 level according to recurrence pattern in patients with pancreatic ductal adenocarcinoma. The area under the curve was 0.613 for hematogenous recurrence (A) and 0.682 for locoregional recurrence (B). AUC: Area under the curve.
Survival curve analysis. The recurrence site-specific DFS curves are shown in Figure 2. The locoregional recurrence-free survival rate was lower in patients with DUPAN-2 ≥200 U/ml than those with DUPAN-2 <200 U/ml (Figure 2A; p<0.01, log-rank test). Similarly, the hematogenous recurrence-free survival rate was lower in patients with DUPAN-2 ≥150 U/ml than in those with DUPAN-2 <150 U/ml (Figure 2B; p=0.04).
Hematogenous and local recurrence-free survival curves of patients with pancreatic ductal adenocarcinoma. The cutoff value of DUPAN-2 level was defined as 150 U/ml for hematogenous recurrence and 200 U/ml for locoregional recurrence using receiver operating characteristic curve analysis. Both hematogenous (A) and locoregional (B) recurrence-free survival after pancreatectomy were significantly worse in patients with higher DUPAN-2 level than in those with lower DUPAN-2 level (p=0.04 for hematogenous recurrence and p<0.01 for locoregional recurrence, log-rank test). DUPAN-2: Duke pancreatic monocle antigen type 2.
Discussion
Previous studies have investigated the sites of recurrence after pancreatectomy in patients with PDAC. In the surgery alone era, the most frequent sites of recurrence were the liver (approximately 60%), peritoneum (approximately 30%), lung (approximately 15%), and locoregional sites (approximately 70%) (14, 20). After the introduction of adjuvant or neoadjuvant therapies, the postoperative recurrence sites changed. In the setting of neoadjuvant chemotherapy, the rate of recurrence according to location in patients with resectable or LA/BR PDAC was reported to be approximately 15% for locoregional sites (8, 21) and approximately 40%-50% for distant organs (15, 21). Meanwhile, neoadjuvant TNT has improved the resectability of the tumor or the local recurrence rate after surgery in patients with LA/BR PDAC (16, 22). In the present study, the recurrence rate according to location was 20.6% (7/34) for locoregional sites only, 20.6% (7/34) for hematogenous organs, 11.8% (4/34) for peritoneum only, and 35.3% (12/34) for multiple sites. In total, local recurrence occurred in 44.1% (15/34) of patients who developed recurrence. In this study, adjuvant chemotherapy was used in 74% (37/50) of patients, whereas neoadjuvant chemotherapy was used in only 7 (14%). This might explain the high rate of locoregional recurrence.
Clinicopathologic factors predicting the recurrence pattern have been investigated in patients with PDAC. The main predictors of locoregional recurrence were reported to be advanced tumor status (larger tumor size, positive lymph node status), aggressive tumor nature (poorly differentiated tumor, perineural invasion, positive margin status), higher preoperative CA19-9 level, and absence of neoadjuvant therapy (5–7). Meanwhile, for distant recurrence, female sex and poorly differentiated tumor were reported as predictors of hepatic recurrence (3). Of these factors, only CA19-9 level is a preoperative factor. In this study, we found that preoperative NLR (p=0.03; OR=2.57; 95%CI=1.10-5.98) and DUPAN-2 level (p<0.01; OR=1.00; 95%CI=1.000-1.002) were significant predictors of locoregional recurrence, both of which can be evaluated before surgery. Patients with higher preoperative NLR and DUPAN-2 are expected to develop locoregional recurrence after pancreatectomy. Therefore, neoadjuvant TNT may be highly recommended for these patients.
The clinical relevance of preoperative DUPAN-2 level has not be fully investigated in patients with PDAC undergoing surgical resection. Shibata et al. (13) reported that an increase in both preoperative CA19-9 and DUPAN-2 levels predicted patient survival after pancreatectomy. Kurahara et al. (9) demonstrated that serum CA19-9 and DUPAN-2 levels are useful predictors of early hematogenous metastasis after surgery. Sunagawa et al. (12) observed that during the initial systemic therapy, normalized DUPAN-2 level independently predicted prolonged survival in patients with resected PDAC. Only a few studies have investigated prognostic factors including preoperative DUPAN-2 level, and they found that preoperative DUPAN-2 level had no prognostic value in multivariate analysis (11, 23, 24). All previous studies that included preoperative DUPAN-2 level measurements were from Japan. In the present study, preoperative DUPAN-2 level was more highly associated with locoregional recurrence than with hematogenous recurrence after resection. We believe that preoperative serum DUPAN-2 level is useful for predicting patient survival or postoperative recurrence pattern after pancreatectomy in patients with PDAC. Therefore, routine measurement of preoperative serum DUPAN-2 level may be recommended for patients with PDAC.
For patients with LA/BR pancreatic cancer, TNT is recommended because it improves resectability and postoperative prognosis (7, 16, 17). Truty et al. (17) investigated 194 patients with PDAC who underwent surgery after TNT and reported excellent survival results as follows: 3-year DFS and overall survival rates of 32% (median, 23.5 months) and 62% (median, 58.8 months), respectively. Moreover, they reported a low rate of locoregional recurrence (6% for locoregional recurrence, 13% for peritoneal recurrence, 25% for distant recurrence, and 8% for multisite recurrence). Several randomized trials on the efficacy of neoadjuvant therapy in patients with LA/BR PDAC have been reported and demonstrated that neoadjuvant therapy improved the locoregional recurrence and overall survival rates compared with upfront surgery (16). Versteijne et al. (22) demonstrated that preoperative chemoradiotherapy was associated with better DFS, locoregional recurrence-free survival, and R0 resection rates, but did not show a significant overall survival benefit in patients with resectable or borderline resectable PDAC. Although the efficacy of neoadjuvant chemoradiotherapy in patients with resectable or LA/BR PDAC is controversial, neoadjuvant chemoradiotherapy seemed to be effective for local control after resection.
In conclusion, elevated preoperative DUPAN-2 level predicts locoregional recurrence, rather than hematogenous recurrence, after surgery in patients with PDAC. Neoadjuvant TNT may be strongly recommended for patients with elevated preoperative DUPAN-2 level to avoid locoregional recurrence after surgery.
Acknowledgements
The Authors thank the professional editors at Enago (www.enago.jp/) for editing a draft of this manuscript.
Footnotes
Authors’ Contributions
The conception and design of the study was performed by AS and TU. Acquisition of data was performed by AS, SI, ST, TF, and TT. The analysis and interpretation of the data was performed by AS, SI, MI and TB. The writing, review, and/or revision of the manuscript were performed by AS and TU. The study supervision was performed by TU.
Conflicts of Interest
The Authors state that they have no conflicts of interest and that they did not receive any financial or material support for this study.
- Received February 18, 2022.
- Revision received March 3, 2022.
- Accepted March 4, 2022.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.