Abstract
Aim: The aim of this study was to confirm the predictive/prognostic value of the preadjuvant Glasgow Prognostic Score (GPS) and carbohydrate antigen (CA) 19-9 level in pancreatic cancer patients receiving adjuvant gemcitabine (GEM) after surgery. Patients and Methods: A total of 67 resected pancreatic cancer patients, treated with adjuvant GEM, were included. The GPS and CA19-9 level were calculated prior to administration of adjuvant therapy and were found to correlate with the outcomes and rate of early recurrence. Results: An elevated preadjuvant GPS or CA19-9 level was significantly associated with a shorter overall survival (OS) (p=0.003 and p<0.001, respectively). Either an elevated GPS or CA19-9 level predicted early recurrence and the combined use of these two factors improved the ability to predict early recurrence, with a specificity and accuracy up to 0.958 and 0.821, respectively. Conclusion: Both an elevated preadjuvant GPS and CA19-9 level, when used alone, are significant predictors of poor outcomes in pancreatic cancer patients receiving adjuvant GEM. The combined use of these parameters improves the ability to predict early recurrence in such patients.
- Glasgow prognostic score
- carbohydrate antigen 19-9
- pancreatic cancer
- early recurrence
- gemcitabine monotherapy
- curative surgery
Pancreatic cancer is one of the most lethal cancers despite recent advances in cancer treatment. At present, surgery is the only curative approach. Combined treatment with postoperative gemcitabine (GEM), which represents the current gold standard treatment for resected pancreatic cancer, improves outcomes, although early recurrence (recurrence within six months) is observed at a considerable rate (1). Therefore, predicting the outcomes of patients receiving adjuvant chemotherapy remains a major challenge and identifying markers to predict early recurrence may improve the outcomes of such patients leading to better management and treatment.
The presence of an ongoing systemic inflammatory response has been recognized to be associated with poor outcomes among oncologic patients. The Glasgow Prognostic Score (GPS) is an inflammation-based prognostic score derived from the acute-phase C-reactive protein (CRP) and serum albumin concentrations (2). A number of studies have identified a relationship between the GPS and clinical outcomes in a variety of cancers, including pancreatic cancer (3-7). The GPS is predictive and/or prognostic for patients with advanced unresectable pancreatic cancer and those receiving curative surgery (3, 7, 8).
At present, the serum carbohydrate antigen 19-9 (CA19-9) level is the only biomarker approved by the Food and Drug Administration (FDA) for pancreatic cancer and is widely used to predict survival and treatment responses in such patients (9). The postoperative CA19-9 level and/or a change in the perioperative CA19-9 level, although the cut-off values vary among studies, are recognized to be the most important prognostic factors in pancreatic cancer patients undergoing surgery (10-13). A decrease in the CA19-9 level during chemotherapy for advanced pancreatic cancer is also predictive of treatment response (14). The secondary end-point results from the Radiation Therapy Oncology Group (RTOG) trial 9704 that showed the ability of the post-resection CA19-9 level to predict the clinical outcomes of pancreatic cancer patients undergoing adjuvant chemoradiation therapy (15). However, the utility of the CA19-9 level in patients receiving adjuvant GEM remains to be clarified.
Combining multiple independent prognostic biomarkers may improve the ability to predict the clinical outcomes of oncologic patients. In this way, the ability to more accurately predict early recurrence may allow physicians to provide more appropriate management and treatment of patients undergoing adjuvant therapy. A recent study reported the improved prognostic performance of the combination of CA19-9 and carcinoembryonic antigen (CEA) levels in pancreatic cancer patients undergoing radical resection (16); however, no previous studies have investigated the potential predictive performance of combining the GPS and CA19-9 level in this setting.
The aim of this study was, therefore, to confirm the predictive/prognostic value of the preadjuvant GPS and CA19-9 level in pancreatic cancer patients undergoing adjuvant GEM after curative surgery and clarify whether combining the GPS and CA19-9 level improves the ability to predict early recurrence in such patients.
Patients and Methods
Patients. A total of 67 pancreatic ductal adenocarcinoma patients (36 males and 31 females, 40-81 years of age) who received adjuvant GEM after curative surgery at the Kanagawa Cancer Center between January 2007 and December 2012 were included in this study. All patients received adjuvant chemotherapy according to one of the following protocols: GEM at a dose of 1,000 mg/m2 biweekly × 12 (6 months) or GEM at a dose of 1,000 mg/m2 on days 1, 8 and 15 every four weeks for six months. Biochemical profiles, including the serum albumin, CRP and CA19-9 levels were measured prior to the administration of adjuvant chemotherapy. Tumor staging was carried out according to the International Union against Cancer (UICC) classification guidelines (Sixth Edition, 2002). Informed consent was obtained from all patients for participation in this study according to the institutional rules of Kanagawa Cancer Center. The study protocol conformed to the ethical guidelines of the 2013 Declaration of Helsinki, as reflected in a priori approval by the Institution's human research committee (17).
GPS evaluation. The GPS was determined prior to the induction of adjuvant GEM. The cut-off level of CRP was set at 0.5 mg/dl according to a report that showed the mean CRP level of East Asians to be lower than half of the mean CRP level of Western individuals (18). The GPS was defined as follows: patients with both an elevated CRP level (>0.5 mg/dl) and hypoalbuminemia (<3.5 g/l) were allocated a score of 2; patients with only one of these biochemical abnormalities were allocated a score of 1; and patients with neither of these abnormalities were allocated a score of 0.
Patients' characteristics.
CA19-9 measurement. The serum CA19-9 levels were analyzed using a radioimmunoassay kit in the Kanagawa Cancer Center laboratory prior to initiation of adjuvant chemotherapy. The normal upper limit of CA19-9 is 37U/ml. We adopted a cut-off point of 180U/ml according to a previous study that prospectively investigated the ability of the post-resection CA19-9 level to predict survival in pancreatic cancer patients treated with adjuvant chemoradiation (15). The subjects were stratified into two groups: patients with a preadjuvant CA19-9 level below 180U/ml and those with a preadjuvant CA19-9 level equal to or above 180U/ml.
Statistical analysis. The cumulative disease-free survival (DFS) and overall survival (OS) rates were estimated according to the Kaplan-Meier method and compared using the log-rank test. The predictors of outcomes were assessed using univariate and multivariate analyses with a Cox proportional hazard regression model.
The differences between the two groups, with and without early recurrence, were analyzed using Fisher's exact test, while a multivariate analysis was performed using a logistic regression model. p-Values are derived from two-tailed tests. For all statistical tests, the level of significance was set at 0.05 and all statistical analyses were performed using the EZR software program (Saitama Medical Center, Jichi Medical University, Omiyaku, Saitama, Japan), a graphical user interface for R (The R Foundation for Statistical Computing, version 2.13.0) or, more precisely, a modified version of R commander (version 1.8-4) designed to add statistical functions frequently used in biostatistics (19).
Results
Patients' characteristics. The baseline clinical characteristics of the patients are shown in Table I. The median follow-up period was 14.1 months (range=3.8-43.1 months). A total of 49 patients had a primary tumor in the head of the pancreas, while 18 patients had a primary tumor in the body or tail. R0 resection was undertaken in 48 patients (71.6%). Of the 67 patients, three had pIA stage disease, 19 had pIIA stage disease and 45 had pIIB stage disease.
(a) Disease-free survival curves based on the Glasgow Prognostic Score (GPS). The 1-year disease-free survival rates of the patients with a GPS of 0 (solid line) or 1 or 2 (dotted line) were 36.8% and 14.4%, respectively. (b) Overall survival curves based on the Glasgow Prognostic Score (GPS). The 1-year overall survival rates of the patients with a GPS of 0 (solid line) or 1 or 2 (dotted line) were 89.2% and 58.0%, respectively. (c) Disease-free survival curves based on the CA19-9 level. The 1-year disease-free survival rates of the patients with a CA19-9 level below 180 U/ml (solid line) or 180 U/ml or above (dotted line) were 36.4% and 5.9% (not reached), respectively. (d) Overall survival curves based on the CA19-9 level. The 1-year overall survival rates of the patients with a CA19-9 level below 180 U/ml (solid line) or 180 U/ml or above (dotted line) were 89.1% and 40.3%, respectively.
Survival analysis. A total of 27 patients had an elevated GPS (a GPS of 1 or 2), while 40 patients had a GPS of 0. Patients with an elevated GPS (a GPS of 1 or 2) had significantly shorter DFS and OS values than the patients with a GPS of 0 (p=0.007 and p=0.003, respectively, according to the log-rank test) (Figure 1).
The CA19-9 levels were below 180 U/ml in 17 patients and equal to or above 180 U/ml in 50 patients. Patients with a CA19-9 level of ≥180 had significantly shorter DFS and OS values (p<0.001 and p<0.001, respectively, according to the log-rank test) (Figure 1).
Univariate analysis showed perineural invasion (ne of 2-3) (p=0.036), T3 (p=0.010), GPS of 1 or 2 (p=0.007), CA19-9 level of ≥180 U/ml (p<0.001) and R1 status (p=0.004) to be significant factors predicting a poor DFS. The multivariate analysis of these variables indicated (ne of 2-3) (p=0.011), GPS of 1 or 2 (p=0.042) and CA19-9 level of ≥180U/ml (p<0.001) to be independent predictors of a poor DFS (Table II).
Univariate and multivariate analyses of clinicopathological features for disease-free survival.
According to the univariate analysis, an elevated GPS of 1 or 2 (p=0.003), CA19-9 level of ≥180U/ml (p<0.001) and R1 status (p=0.027) were significant factors predicting a poor OS. Meanwhile, the multivariate analysis indicated that an elevated GPS (p<0.001) and CA19-9 level (≥180 U/ml) (p<0.001) were independent predictors of a poor OS (Table III).
Predictive value of the combined GPS and CA19-9 level for detecting early recurrence. A total of 29 of the 67 patients died during the follow-up period, whereas 19 of the 67 patients (28.4%) displayed early recurrence within six months after surgery. The site of early recurrence was local in 3/19 patients (15.8%), in the lymph nodes in 4/19 patients (21.1%), the peritoneal region in 5/19 patients (26.3%) and distant in 12/19 patients (63.2%). The median OS of the patients with early recurrence was 10.1 months, that was significantly shorter than the 40.2 months observed in patients without early recurrence. The univariate and multivariate analyses indicated the GPS (p=0.020) and CA19-9 level (p=0.021) to be independent factors predicting early recurrence (Table IV). Patients who had both elevated GPS and CA19-9 showed poorer DFS and OS than patients who had either of these factors or patients who had neither of these factors (Figure 2). The sensitivity, specificity and accuracy of the preadjuvant GPS to predict early recurrence were 0.684, 0.708 and 0.701, respectively, while those for the CA19-9 level were 0.579, 0.879 and 0.791, respectively. Importantly, the combined use of the GPS and CA19-9 level improved the specificity and accuracy up to 0.958 and 0.821, respectively (Table V).
Univariate and multivariate analyses of clinicopathological features for overall survival.
Discussion
In this study, we found both the preadjuvant GPS and CA19-9 level to be significant predictors of DFS and OS in patients undergoing adjuvant GEM chemotherapy after curative resection for pancreatic cancer. Additionally, either the preadjuvant GPS or CA19-9 level alone was a predictive factor for early recurrence in these patients and the combined use of these two parameters improved the ability to predict early recurrence over the application of either alone.
Our findings demonstrated that an elevated preadjuvant GPS is significantly associated with shorter DFS and OS values in pancreatic cancer patients receiving adjuvant GEM, which is in line with the results of a recent study that showed an association between an elevated postoperative GPS and poor outcomes in colorectal cancer patients undergoing adjuvant chemotherapy (20). The prognostic and/or predictive value of the GPS has been demonstrated in various types of cancers, including pancreatic cancer (3-7). Furthermore, an elevated GPS is significantly associated with a poor prognosis among pancreatic cancer patients with inoperable disease and/or those undergoing surgery (3, 7, 8).
Our findings also showed that an elevated preadjuvant CA19-9 level is significantly associated with shorter DFS and OS values in pancreatic cancer patients receiving adjuvant GEM. An elevated postoperative CA19-9, although the cut-off values vary among studies, and/or the failure of CA19-9 normalization after surgery have been shown to be significantly associated with poorer survival in pancreatic cancer patients undergoing curative surgery (10-13). In addition, a decrease in the CA19-9 level under chemotherapy with GEM in patients with locally advanced or metastatic pancreatic cancer is associated with favorable outcomes (14). However, the role of the CA19-9 level in patients undergoing adjuvant therapy has not been fully investigated. The secondary endpoint of the RTOG trial 9704, a randomized phase III study comparing the use of either continuous infusion fluorouracil (FU) or gemcitabine before and after adjuvant chemoradiotherapy with FU in patients with resected pancreatic adenocarcinoma, showed that the patients with a CA 19-9 level of ≥180 U/ml had a significantly increased risk of death compared to their counterparts with a CA19-9 level below 180 U/ml (15). The present study is the first study to prospectively investigate the significance of the CA19-9 level in pancreatic cancer patients undergoing adjuvant treatment. Our results provide additional evidence of the predictive/prognostic utility of measuring the CA19-9 level in the adjuvant setting.
Relationships between early recurrence (≤6 months) and clinicopathological features.
In this study, we evaluated the accuracy of combining the preadjuvant GPS and CA19-9 level to predict early recurrence in pancreatic cancer patients receiving adjuvant GEM. Both the preadjuvant GPS and CA19-9 levels were found to be predictive for early recurrence and the combined use of these parameters improved the ability to predict early recurrence over the application of either alone. Combining multiple independent prognostic biomarkers may improve the ability to predict clinical outcomes in oncologic patients. The GPS reflects the status of the host inflammation response (2), while the CA19-9 level is a marker of tumor biology (9). Therefore, combining these two different factors, one patient-related and one tumor-related, is theoretically acceptable and may improve the ability to predict clinical outcomes in pancreatic cancer patients undergoing adjuvant chemotherapy. The combined use of routinely usable biomarkers may, thus, serve as a convenient tool for predicting clinical outcomes and guide further management in these patients.
The overall survival (OS) of the patients was significantly shorter when stratified by GPS and CA19-9 status (by log-rank test).
Accuracy for predicting early recurrence.
In summary, both an elevated preadjuvant GPS and CA19-9 level alone are significant predictors of poor outcomes in pancreatic cancer patients receiving adjuvant GEM after curative surgery and the combined use of these parameters improves the ability to predict early recurrence in such patients. However, the current study was a small, retrospective study; therefore, our results warrant further investigation in larger prospective cohorts. This report was presented, in part, at the 2014 Gastrointestinal Cancer Symposium of the American Society of Clinical Oncology as part of the General Poster Session in San Francisco, California, in January 2014.
Acknowledgments
This study was partially supported by grants from the Kanagawa Health Foundation. The Authors declare that they have no potential conflicts of interest.
- Received February 22, 2016.
- Revision received March 30, 2016.
- Accepted April 4, 2016.
- Copyright© 2016 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
