Prognostic significance of preoperative gamma-glutamyltransferase to lymphocyte ratio index in nonfunctional pancreatic neuroendocrine tumors after curative resection

Various inflammation-based prognostic scores have been associated with reduced survival in patients with nonfunctional pancreatic neuroendocrine tumor (NF-PNET). However, few studies have illuminated the relationship between the preoperative gamma-glutamyltransferase (GGT) to lymphocyte ratio index (GLRI) and the prognosis of NF-PNET. A retrospective review of 125 NF-PNET patients following curative resection was conducted. The cut-off values for the inflammation-based prognostic scores, including GLRI, were selected using receiver operating characteristic curve analysis. Univariate, multivariate and Kaplan-Meier analyses were used to calculate overall survival (OS) and disease-free survival (DFS). The optimal cut-off value for GLRI was 10.3. Multivariate analysis showed that GLRI was an independent predictor of OS (P = 0.001) and DFS (P = 0.007) for NF-PNET. Kaplan-Meier analysis also showed that preoperative GLRI had significant prognostic value in various subgroups of patients with NF-PNET. The discriminatory capability of GLRI was superior to that of other inflammation-based scores in OS prediction. Furthermore, the predictive range was expanded by incorporating GLRI into the conventional stratification systems, including AJCC staging and WHO classification. These results indicated that preoperative GLRI was an independent predictor for NF-PNET patients undergoing curative resection. The incorporation of GLRI into the existing conventional stratification systems resulted in improved predictive accuracy.

Gamma-glutamyltransferase (GGT) is a membrane bound enzyme essential to glutathione (GSH) metabolism that protects cells from reactive oxygen species 8 . GGT expression is found predominantly on the luminal surface of secretory epithelial cells, especially of the hepato-biliary tract, pancreas and kidneys 9 . Increasing evidence has suggested that an elevated serum GGT level is an independent predictor of poor survival in several cancer types, such as pancreatic cancer 10 , intrahepatic cholangiocarcinoma 11 and hepatocellular carcinoma 12 . Engelken FJ et al. reported that elevated GGT and leukocytosis predicted shorter survival for patients with unresectable pancreatic cancer 10 . Research from Diergaarde B suggested that a common variation in the GGT1 gene could affect the risk of pancreatic cancer 13 . Furthermore, recent data suggested that higher serum levels of GGT, within the normal range, was an early marker of oxidative stress and an indicator of higher cancer risk 14 .
Lymphocyte count, which can comprehensively reflect the systemic inflammatory response in patients with cancer, has been a reliable marker for predicting the survival of patients with different types of cancer 15,16 . Katz SC et al. demonstrated that lymphocyte infiltration was common in the majority of NETs, as assessed by immunohistochemistry for CD3, CD4, CD8, and CD56 17 . Tumor-infiltrating lymphocytes (TILs) have been shown to predict outcomes in numerous primary human malignancies, including pancreatic adenocarcinoma, hepatocellular carcinoma and colorectal cancer 18 . In contrast, lymphocytopenia has been reported in various cancers, but is particularly marked in patients with pancreatic cancer. Preoperative lymphocytopenia is a poor prognostic factor in patients with pancreatic cancer 15 . It has been suggested that lymphocytopenia, indicating a state of depressed immune function, may influence survival adversely due to reduced host response to the tumor cells.
In combination with the effects of GGT and lymphocytes mentioned above, the GGT to lymphocyte ratio index (GLRI) may be a potentially effective biomarker for tumor prognosis. Additionally, the GGT to platelet ratio index (GPRI), a predictor of liver fibrosis and cirrhosis 19 , was an independent predictive factor for HBV-related hepatocellular carcinoma after hepatic resection 20 . To our knowledge, the GLRI has not been used to predict the survival and tumor recurrence after curative resection for NF-PNETs. The goal of this study was to assess the prognostic value of GLRI in patients with NF-PNETs following curative resection. Further, we aimed to compare the discriminative ability of GLRI with that of other inflammation scores to determine whether the GLRI could be a useful marker for predicting patients' outcomes. Additionally, we attempted to refine the existing stratification systems by incorporating GLRI into the existing TNM staging system or WHO classification.

Results
Patients' clinicopathological characteristics. The clinicopathological characteristics are provided in Table 1. This study included 64 male patients (51.2%) and 61 female patients (48.8%). These patients were diagnosed at a mean age of 53.0 ± 12.73 years and were evaluated over a mean follow-up period of 45.76 ± 37.01 months. In this study, 54 patients were identified incidentally during health examinations. The most common presentation of symptomatic PNETs was abdominal pain in 56 (44.8%) patients, followed by abdominal discomfort in 8 (6.4%), obstructive jaundice in 5 (4%) and diarrhea in 2 (1.6%). Only two patients had hepatitis B and were receiving antiviral therapy with entecavir. According to the inclusion criteria, no patients suffered from the cholangitis. The numbers of patients classified into AJCC stages I, II, III and IV were 88, 18, 4 and 15, respectively. The numbers of patients classified into grades 1, 2 and 3 were 41, 62 and 22, respectively. The 1-, 3-, and 5-year OS and DFS rates were 98%, 89%, and 76% and 79%, 66%, and 63%, respectively.

Relationships between GLRI and clinicopathological characteristics. The relationships between
preoperative GLRI and the clinicopathological variables of patients with NF-PNET were investigated, and the data showed that preoperative GLRI was correlated with gender (P < 0.001), presence of symptoms (P = 0.005), alkaline phosphatase (AKT) (P = 0.003), T-stage (P = 0.03), distant metastasis (P = 0.022), perineural invasion (P = 0.047), and AJCC stage (P = 0.005). Furthermore, there were no significanct differences between preoperative GLRI and other clinicopathological parameters, such as age, albumin, tumor size, tumor location, LN metastasis and WHO classification (all P > 0.05, Table 1). Multivariate analysis revealed that the independent parameters associated with an elevated GLRI were gender (P < 0.001), AKT (P = 0.003) and AJCC stage (P = 0.028), which suggested that male gender, high AKT, and advanced AJCC stage were associated with elevated GLRI.
Comparative performance of GLRI and other predictive models. To further evaluate the prognostic value of GLRI, other inflammation-based scores and conventional stratification systems, ROC analysis was performed, and AUC values were compared. The GLRI had a higher AUC value (0.682; P = 0.001) than GNRI, GPRI, PLR, NLR and PNI (Fig. 4). However, the conventional staging systems, including AJCC staging and WHO classification, were superior to GLRI in OS prediction for NF-PNET (Table 4). In addition, the predictive ability of the AJCC staging systems was superior to the WHO classification in our cohort (AUC value: 0.738 vs 0.701). The model integrating the GLRI and the AJCC stage for OS prediction yielded higher AUC values than the AJCC stage alone (0.814 vs 0.738), while the predictive ability of the model integrating the GLRI and the WHO classification was superior to the WHO classification alone (0.806 vs 0.701). In addition, the model including the AJCC stage and WHO classification had the highest AUC values, followed by the model integrating the GLRI and AJCC stage (Table 4).

Discussion
The present study showed that preoperative GLRI, a novel and easily accessible inflammation-based score derived from serum GGT and lymphocytes,was a predictor of survival in patients with NF-PNET following curative resection. Furthermore, we observed that elevated preoperative GLRI was associated with advanced tumor stages. Finally, we showed that GLRI outperformed other inflammation-based scores in terms of discriminatory capacity. The predictive models incorporating GLRI and conventional stratification systems, including the WHO classification and AJCC stage, showed improved predictive power, compared to stratification systems alone.
Recently, increasing evidence has confirmed that tumor development is associated with inflammation and immunity. Inflammation plays an important role in tumor growth [21][22][23] . GGT was reported to play a prooxidant role, and the subsequent production of reactive oxygen species (ROS) could promote certain intracellular and extracellular molecular signals 24 . ROS have been reported to promote an epithelial-to-mesenchymal transition via the Snail-E-cadherin pathway and to induce inflammation via the nuclear factor kappa B pathway [25][26][27] . Therefore, the GGT level has been characterized as a biomarker for oxidative stress and has been shown to be correlated with inflammation in the extracellular tissue microenvironment. It has been found that GGT plays a pivotal role in tumor progression and aggressiveness, and it acts as a significant prognostic biomarker in several cancer entities 9-11 .   Table 2. Univariate Cox proportional hazards regression models of prognostic factors associated with OS and DFS. OS, overall survival; DFS, disease-free survival; AKT, alkaline phosphatase; NLR, neutrophil to lymphocyte ratio; PLR, platelet to lymphocyte ratio; GLRI, gamma-glutamyltransferase to lymphocyte ratio index; GNRI, gamma-glutamyltransferase to neutrophil ratio index; GPRI, gamma-glutamyltransferase to platelet ratio index; LN, lymph node; AJCC, American Joint Committee on Cancer; PNI, prognostic nutritional index. P-value < 0.05, marked in bold font, shows statistical significance.
In addition, lymphocytes, a marker of patient immune status, play a central role in the systemic inflammatory response. A number of studies have demonstrated a relationship between the lymphocyte count or neutrophil-lymphocyte ratio and the prognosis of cancer patients [28][29][30] .
This association can be explained by CD4+ T lymphocyte cells acting as a sensor in detecting precancerous cells and then regulating their eradication 31 . The loss of CD4+ T lymphocytes accompanied by the impaired activation of CD8+ T lymphocyte cells can cause the insufficient secretion of cytotoxin performing anti-carcinogenic functions in neoplastic microenvironments 32 . Therefore, GLRI, a novel inflammatory marker that reflects the status of oxidative stress and immune function, could be a potential predictive marker.
In our study, we first identified the cut-off value of preoperative GLRI according to the ROC curve, and 10.3 appeared to be the optimal cut-off value for GLRI with maximum joint sensitivity and specificity. Notably, concerning the correlations between GLRI and clinical characteristics, we found that an elevated GLRI was positively correlated with gender, presence of symptoms, AKT, T-stage, perineural invasion, and AJCC stage. Moreover, patients with elevated GLRI were more inclined to have a higher rate of distant metastasis. All of these data showed that GLRI could reflect the tumor burden and tumor progression. With further analysis, we found that elevated GLRI was identified as an independent risk factor for OS and DFS in this cohort. The 1-, 3-and 5-year OS rates and DFS rates of patients with high levels of GLRI were markedly lower than those in the low level group. For the subgroup of patients with stage I/II, we also found that a preoperative GLRI > 10.3 was predictive of significantly worse survival. Thus, the preoperative GLRI might be able to predict poor prognosis in patients with early NF-PNET. In addition, in the subgroup with tumor size >2 cm or in patients without perineural invasion,  Table 3. Independent prognostic factors for OS and DFS identified by the Cox multivariate proportional hazards regression model. OS, overall survival; DFS, disease-free survival; LN, lymph node; AKT: alkaline phosphatase; GLRI, gamma-glutamyltransferase to lymphocyte ratio index; NLR, neutrophil to lymphocyte ratio; AJCC, American Joint Committee on Cancer; NA, not available. P-value < 0.05, marked in bold font, shows statistical significance.
preoperative GLRI > 10.3 also showed prognostic value in predicting poorer OS and DFS. All of these data provided further evidence that preoperative GLRI could act as a potential prognostic marker to predict survival in NF-PNET patients undergoing curative resection, in accordance with our above hypothesis. TNM stage, grade or markers of systemic inflammation, such as the NLR and PLR, have been emerged as prognostic indicators in PNET 1,2,30 , similar to our study. To the best of our knowledge, the existing stratification systems and predictive models for PNET, including the above mentioned TNM staging system, WHO classification and two nomograms 33,34 , lacked indicators of systemic inflammation, which could offer additional information in prognostic evaluation. Han X et al. reported that an effective nomogram including the chromogranin A level, liver metastases tumor burden, and Ki-67, predicted overall survival in well-moderate NF-PNETs with liver metastases, and the nomogram showed fitting calibration with a C-index of 0.87 (95% confidence interval, 0.82-0.92) 33 . Furthermore, the research from Ellison TA et al. had confirmed that a simple prognostic nomogram, including continuous Ki-67 labelling, sex, and age at surgery (≤63, >63) could be used to predict survival for NF-PNETs with a Harrel's c-index value of 0.74 34 . Herein, we incorporated GLRI, an inflammation-based biomarker, into AJCC staging and WHO classification and showed that the predictive abilities of models integrating the GLRI and the stratification systems for OS was superior to the stratification systems alone. The results were supportive of the integration of GLRI into conventional stratification systems for improved discriminative ability.
Our study had several limitations that must be considered. First, the present study was retrospective in nature and a single-centre experience. Second, only patients who underwent curative resection were included in the study. Furthermore, due to the limited number of patients, external validation was not performed. Therefore, future studies should be performed to evaluate the prognostic significance of GLRI in patients with advanced clinical stages and different treatment modalities.
In conclusion, as a novel and easily accessible inflammation-based biomarker, preoperative GLRI was an independent predictor of OS and DFS for NF-PNET patients undergoing curative resection. Furthermore, we confirmed that prognostic models incorporating GLRI into the TNM staging system or WHO classification provided improved predictive accuracy, compared with the stratification systems alone. Therefore, we recommend that surgeons devise the treatment plans considering not only TNM stage but also these prognosis related serum biomarkers. Only in this manner can we acquire better personalised therapy for patients with NF-PNET.   (2) were treated for recurrent disease; (3) received preoperative radiochemotherapy prior to surgery; (4) underwent an R1 or R2 resection; (5) had a history of cancer of any type; and (6) did not consent to the use of their medical records for research purposes. We included in the study only those patients who had survived for at least 60 days after surgery in the study to exclude perioperative mortality-related bias. The study was approved by the Ethics Committee of the First Affiliated Hospital of Zhejiang University School of Medicine and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all of the participants before the commencement of the study. All of the methods and research activities were performed in accordance with the guidelines and regulations.
Laboratory tests included routine blood testing, albumin, GGT and AKT, which were routinely performed within 7 days before the surgical resection and 7 days after the operation. The TNM stage of each PNET was determined based on the American Joint Committee on Cancer (AJCC) TNM classification, while the grade of each PNET was determined according to the 2010 WHO classification of NETs of the GEP system. Six inflammatory factors, including GLRI, GPRI, GNRI, NLR, PLR and PNI, were included in this analysis. The definitions of the six inflammatory factors are as follows: NLR = neutrophil count/lymphocyte count; PLR = platelet count/lymphocyte count; PNI = serum albumin levels (g/dl) × 10 + total lymphocyte count (per mm 3 ) × 0.005; GLRI = (GGT value/lymphocyte count) × 10 9 /U; GPRI = (GGT value/platelet count) × 10 9 /U; GNRI = (GGT value/neutrophil count) × 10 9 /U. Follow-up. Patient follow-up was performed by reviewing hospital records or contacting patient family members. Overall survival (OS) was defined as the time span extending from the date of the initial diagnosis until the date of death from any cause or the date of last known contact. Disease-free survival (DFS) was defined as the time extending from the date of surgery to the date of PNET recurrence. Patients who did not have evidence of local recurrence or metastasis at the last follow-up and patients who died of diseases unrelated to PNET were censored from the analysis of DFS. Our department follows up with patients every 6 months for the first 5 years after surgery and then yearly thereafter. The following postoperative follow-up data were collected for each patient: clinical symptoms and signs; laboratory test results; and radiological examination results.
Statistical analysis. All of the statistical analyses were performed using SPSS software, version 16.0 for Windows (SPSS, Chicago, IL, USA). Area under the curve (AUC) values, obtained from receiver operating characteristic (ROC) curve analysis, were used to compare the predictive efficacies of GLRI and the afore mentioned inflammatory factors. Differences between groups were analysed using Pearson's chi-square test, Fisher's exact  Table 4. Areas under the ROC curve for conventional staging systems and inflammation-based prognostic scores for predicting OS in NF-PNET undergoing radical resection. ROC, receiver operating characteristic; OS, overall survival; NF-PNET, nonfunctional pancreatic neuroendocrine tumor; AJCC, American Joint Committee on Cancer; GLRI, gamma-glutamyltransferase to lymphocyte ratio index; GNRI, gamma-glutamyltransferase to neutrophil ratio index; GPRI, gamma-glutamyltransferase to platelet ratio index; NLR, neutrophil to lymphocyte ratio; PLR, platelet to lymphocyte ratio; PNI, prognostic nutritional index. P-values < 0.05, marked in bold font, indicate statistical significance.
test or the Mann-Whitney U test as appropriate. The Kaplan-Meier method and the log-rank test were used to calculate OS and DFS, respectively. Prognostic analysis was performed using univariate and multivariate Cox regression models. A P-value < 0.05 was considered statistically significant. All of the data generated or analysed during this study are included with this published article.