Prognostic value of a novel FPR biomarker in patients with surgical stage II and III gastric cancer

Background Inflammation and nutrition are two main causes contributing to progression of gastric cancer (GC), and inflammatory biomarker may be presented as its valuable prognostic factor. Thus, this study was carried out to investigate the prognostic significance of preoperative circulating albumin/fibrinogen ratio (AFR), fibrinogen/pre-Albumin ratio (FPR), fibrinogen (Fib), albumin (Alb) and pre-Albumin (pAlb) in surgical GC. Materials and Methods Three hundred and sixty surgical stage II and III GC patients from June 2011 to December 2013 were enrolled in this retrospective study. X-tile software, Kaplan–Meier curve and Cox regression model were used to evaluate the prognostic role of them. A predictive nomogram was established to predict prognosis of overall survival (OS), and its accuracy was assessed by concordance index (c-index). Results Decreased Alb, pAlb, AFR and elevated FPR were significantly associated with shorter OS. FPR was identified as the most effective prognostic factor to predict 3-year’s OS by time-dependent ROC analysis. A long survival was observed in patients with low level of FPR and the prognosis of stage III FPR-low GC patients undergoing chemotherapy was significantly superior to the patients without the treatment (P=0.002). However, no difference of survival was examined in stage II subgroups stratified by FPR and high FRP of stage III patients with or not the treatment of chemotherapy. C-index of nomogram containing FPR (c-index=0.756) was high in comparison with the nomogram without FPR (c-index =0.748). Conclusion Preoperative FPR might be a feasible prognostic biomarker in surgical stage II and III GC and it could precisely distinguish stage III patients who appeared to obviously benefit from adjuvant chemotherapy. Meanwhile established nomogram based on clinical parameters and FPR could improve its predictive efficacy.


INTRODUCTION
Gastric cancer (GC), one of the most common malignancies, is the second most cause of mortality worldwide [1].Although rapid improvement of surgery and adjuvant treatment in past decade, the prognosis of GC patients remained unsatisfactory owing to recurrence or metastasis after curative resection [2].Therefore, promising prognostic biomarker which predicted its progression and survival would be helpful for management and treatment in these patients.
It has been well known that inflammation and nutrition are closely associated with progression and survival of GC [3][4][5].Anti-inflammatory treatment and nutritional care could prevent cancer progression and improve prognosis of the patients [5][6][7].Seo et al. reported that preoperative adequate albumin (Alb) and energy intake could improve therapeutic effect of the patients [5].Kim et al. demonstrated that long-term low-dose aspirin intake could reduce susceptibility to GC [7].Circulating nutritional and inflammatory mediators such as fibrinogen(Fib), Alb and pre-albumin(pAlb) are usually aberrant in these patients.Emerging evidences indicated that high level of plasma Fib were significantly associated with poor clinical outcome of GC patients [8][9][10], and preoperative low serum pAlb level and hypoalbuminemia were considered to be predictors for shorter overall survival (OS) in GC patients [11,12].A recent study reported that circulating albumin to gamma-glutamyltransferase ratio could apparently improve predictive accuracy for OS in resected intrahepatic cholangiocarcinoma patients in comparison with TNM staging systems alone [13].Thus, we speculated that circulating Alb/Fib ratio (AFR) and Fib/pAlb ratio (FPR), which reflected status of inflammation and nutrition, would be novel inflammatory biomarkers of prognostic prediction for postoperative stage II and III GC patients.Abbreviation: NA: not available; Fib: fibrinogen; Alb: albumin; pAlb: pre-Albumin; AFR: albumin/fibrinogen ratio; FPR: fibrinogen/pre-Albumin ratio (FPR); CEA: carcinoembryonic antigen; FAS: FPR and Alb Score; mFAS: modified FPR and Alb Score; OS: overall survival; We firstly compared the clinical efficacy of preoperative circulating Fib, Alb, and pAlb, either alone or pooled, for 3 years' clinical outcome in stage II and III GC patients.Our findings revealed that FPR could independently predict postoperative OS with superior accuracy compared with the other prognostic indicators and select the patients who could benefit from adjuvant chemotherapy.Additionally, a reliable prognostic nomogram based on clinical parameters and FPR could improve its predictive value of OS in the patients.

The optimal thresholds for Fib, Alb, pAlb, AFR and FPR
The optimal cut-points using X-tile program for preoperative circulating Fib, Alb, pAlb, AFR and FPR were 3.3 mg/dl, 37 g/l, 195.9 mg/l, 8.9 and 12.1, respectively (Figure 1 and Supplementary Figure 1).According to the optimal cut-points, enrolled patients were divided into lowand high-groups.The details are shown in Table 2.

The correlation of Fib, Alb, pAlb, AFR and FPR with the clinical parameters
In order to investigate associations of these factors with tumor stage, 44 stage I GC patients were enrolled in our study.We compared the high groups and low groups for these indicators and increased Fib, FPR and deceased Alb, pAlb and AFR were positively correlated with age (more than 60 years), tumor size (larger than 5cm), tumor stage (III), depth of invasion depth (T3-T4), lymph node metastasis (N1-N3) and poor OS (all P<0.001) (Table 2 and Figure 2).Compared with 30 preoperative patients, higher FPR were in the patients with recurrent GC (P=0.001) (Figure 2F).Besides, no significant association was observed among alcohol, tobacco, hypertension, diabetes, tumor differentiation and adjuvant chemotherapy in two groups.

Time-dependent ROC analysis
To further evaluate the prognostic value of inflammation-based prognostic factors, time-dependent ROC analysis was performed.The result of timedependent ROC analysis presented that the lower area under the receiver operating characteristic curve (AUC) for FPR in the early period (<6 months) and the higher AUC therefore (>6 months) among these prognostic indicators including Fib, Alb, pAlb, AFR, CEA and CA199 (Figure 3).

FPR and clinical adjuvant chemotherapy
We compared the prognosis of stage II and III GC patients receiving or not adjuvant chemotherapy in the subgroups stratified by FPR.Survivals of stage II and III GC patients were significantly longer in low FPR subgroup than them in high FPR subgroup (P=0.007 and P=0.002, respectively).Low level of FPR (adjusted HR=5.851, 95%CI=2.147-15.949)were significantly associated with reduced survival in the III stage patients without chemotherapy comparing to the patients undergoing chemotherapy.However, no difference of survival was examined in stage II subgroups stratified by FPR and high FRP of stage III subgroup receiving or not the treatment of adjuvant chemotherapy (Table 4 and Figure 4).

DISCUSSION
Most of the GC patients are diagnosed in an advanced stage and the survival rates of them are relatively low, therefore, promising prognostic biomarkers that enable to identify the patients who could obviously benefit from chemotherapy and predict survival of them are crucial [14,15].In this study, we found that evaluated FPR was significantly associated with T3-4 invasion, node metastasis and larger tumor size and was superior to other biomarkers to independently predict poor survival both within stage II-III, II and III subgroups; moreover, clinical outcome of III stage patients with low FPR appeared to obviously benefit from adjuvant chemotherapy in comparison with high FPR stage III patients, and the biomarker could improve the predicted efficacy of nomogram for stage II-III GC.
To date, some researchers have reported that high level of Fib, low level of Alb and pAlb were recognized as important prognostic factors influencing cancer progression [10,11,16], which were consistent with our findings.Due to few patients died from the disease within 6 months after surgical resection, low AUC of FPR was observed in the early period, and the AUC was gradually increased and higher than the other biomarkers, indicating that FPR was superior to these biomarkers to apparently improve predictive efficacy of prognosis within II-III stage GC patients.In addition, it could precisely classify stage III GC patients who appeared to benefit from adjuvant chemotherapy obviously.The following causes might be accounted for our findings.Firstly, it had been shown that Fib acts as a bridging molecule between GC cells and surrounding microenvironment.Adams et al. demonstrated that it as a ligand for integrin and intercellular adhesion molecule presented on malignant cell surface to mediate coagulation, inflammation and immunity [17].Secondly, Fib enhanced b3-integrin-mediated vascular endothelial adhesion of platelets to tumor cells, and platelets in turn promoted more Fib to aggregate around tumor cells by forming thrombin.They facilitated each other to protect tumor cells escaping from cytotoxicity of nature killer cells [18].Palumbo et al. reported that lymphatic metastasis, but not primary tumor growth or angiogenesis, was diminished in fibrinogen-deficient mice, suggesting that Fib was a critical determinant of the metastatic potential by impeding elimination of tumor cell by natural killer cell [19,20].Thirdly, serum Alb was one of the most widely used markers for reflecting nutritional status and hypoproteinmia was reported as a crucial parameter of malnutrition and directly influenced prognosis of GC; low levels of Alb and pAlb levels have an impact on determinant of immune responses and malnutrition, which could impair immune system defending against GC [21,22].
This study, to best of our knowledge, is the first to investigate prognostic role of AFR and FPR in GC.Certain advantages and limitations should address to explain our results.To some extent, hypoalbuminemia has been considered to be an inflammatory indicator in GC, rather than only a factor indicates malnutrition [4].Therefore, single clinical blood marker is limited and unstable to predict prognosis of GC.Our results did firstly find that FPR is a superior prognostic indicator compared to Fib, Alb, or pAlb alone, for they reflected not only inflammation but also nutritional status of GC patients.Besides, circulating Fib to pAlb ratio will expand prognostic range to avoid a single indicator causing false negative or positive results.Finally, we figured out the visual nomogram based on FPR, which could predict prognosis in postoperative stage II and III GC patients within 3 years more accurately.Therefore, preoperative calculation of FPR may help to predict 3 years' OS in surgical GC patients.However, we acknowledge some potential limitations in our study.Since the results of our study may be affected by a short follow-up period, singleinstitution design and a small sample size retrospective study, larger patients with GC are required to confirm our findings.

Figure 1 :
Figure 1: The optimal cut-off of preoperative circulating FPR in 360 surgically resected GC patients using X-tile software.The optimal cut-point of FPR ratio in the panels is shown on the histogram and corresponding populations are displayed on the Kaplan-Meier curve.

Figure 5 :
Figure 5: Postoperative nomogram estimated by clinical characteristics and FPR for 3-years' OS in 360 GC patients.(A) without FPR; (B) with FPR.