Preoperative Plasma Fibrinogen Level Represents an Independent Prognostic Factor in a Chinese Cohort of Patients with Upper Tract Urothelial Carcinoma

Background Increased plasma fibrinogen is thought to contribute to tumor progression and metastasis. The association of plasma fibrinogen with clinicopathological characteristics, and the optimal cutoff with an ideal predictive value has not been fully determined in patients with upper tract urothelial carcinoma (UTUC). We aimed to investigate the clinical significance of this parameter in a Chinese cohort of patients with UTUC. Methods A retrospective study was conducted to analyze the clinical data of 184 operable UTUC patients in a Chinese cohort with a high incidence of chronic kidney disease (CKD). An optimal cutoff was set for further analysis according to validated web-based software. The associations of plasma fibrinogen with clinicopathological characteristics and survival were assessed. Multivariate analyses were performed to determine the independent prognostic factors. Results Elevated plasma fibrinogen was significantly associated with tumor necrosis, lymph node involvement, and a higher preoperative CKD stage, pathological tumor stage and grade (all P < 0.05). Kaplan-Meier analysis showed that plasma fibrinogen ≥ 3.54 g/L predicted a poorer overall and cancer-specific survival than < 3.54 g/L (P < 0.001 for both). Multivariate analyses revealed that elevated preoperative plasma fibrinogen was an independent negative prognostic factor for overall survival (HR = 2.026; 95% CI: 1.226–3.349; P = 0.006) and cancer-specific survival (HR = 1.886; 95% CI: 1.019–3.490; P = 0.043). Conclusions Increased plasma fibrinogen was an independent prognostic risk factor for poor outcomes in UTUC. This parameter may serve as an effective biomarker with easy accessibility for evaluating prognosis for patients with UTUC.


Methods
A retrospective study was conducted to analyze the clinical data of 184 operable UTUC patients in a Chinese cohort with a high incidence of chronic kidney disease (CKD). An optimal cutoff was set for further analysis according to validated web-based software. The associations of plasma fibrinogen with clinicopathological characteristics and survival were assessed. Multivariate analyses were performed to determine the independent prognostic factors.

Results
Elevated plasma fibrinogen was significantly associated with tumor necrosis, lymph node involvement, and a higher preoperative CKD stage, pathological tumor stage and grade (all P < 0.05). Kaplan-Meier analysis showed that plasma fibrinogen 3.54 g/L predicted a poorer overall and cancer-specific survival than < 3.54 g/L (P < 0.001 for both). Multivariate analyses revealed that elevated preoperative plasma fibrinogen was an independent negative prognostic factor for overall survival (HR = 2.026; 95% CI: 1.226-3.349; P = 0.006) and cancer-specific survival (HR = 1.886; 95% CI: 1.019-3.490; P = 0.043).

Introduction
Upper tract urothelial carcinoma (UTUC) is a rare type of urothelial tumor. It accounts for only 5% of urinary tract urothelial carcinomas in western countries [1,2]. Nevertheless, UTUC is relatively more common and is often complicated by chronic kidney disease (CKD) in the Chinese population [3][4][5]. This phenomenon is possibly attributed to widely used Chinese herbs that contain aristolochic acid [6,7]. Despite non-metastatic UTUC patients not receiving a radical nephroureterectomy (RNU) with bladder cuff excision as a standard therapeutic approach, an adjuvant treatment might be indispensable for a high-risk subset. In addition to pathological T stage and tumor grade as established prognostic indicators of great importance, several other pathological parameters, such as tumor necrosis and lymphovascular invasion, were suggested to predict prognosis in UTUC [8][9][10]. Because of the impaired preoperative renal function and further damage after RNU, neo-adjuvant chemotherapy presented notable advantages in view of the limited opportunity to receive postoperative cisplatin-based chemotherapy [11]. Thus, this highlights the demand for discovering low cost, highly effective and easily accessible preoperative prognostic predictors.
Increasing evidence shows that there is an intimate connection between hemostatic factors and tumor biology. Fibrinogen, as a type of plasma glycoprotein, plays a vital part in clot formation, wound healing and supporting platelet aggregation. It is mainly produced by liver epithelium, and inflammatory stimuli could also promote fibrinogen synthesis in the lung and intestinal epithelium. Elevated plasma fibrinogen was reported to predict tumor progression, distant metastasis and poor oncological outcome in various malignancies [12][13][14][15]. Additionally, previous studies have shown plasma fibrinogen levels were related to certain pathological characteristics such as T stage, lymph node involvement, tumor grade, tumor size, and lymphovascular invasion in several cancers [16][17][18]. Two prior studies have also proven the prognostic value of plasma fibrinogen in UTUC in the Japanese and European populations [17,18], whose carcinogenesis is quite different from the Chinese. Nevertheless, the potential connection between plasma fibrinogen level and inflammatory biomarkers has not yet been determined in UTUC.
We conducted a retrospective study to validate the prognostic impact of preoperative plasma fibrinogen level in a Chinese cohort of patients with UTUC and explore its potential association with clinicopathologic characteristics and systemic inflammation.

Patients and clinicopathologic evaluation
We retrospectively reviewed clinicopathologic and follow-up data of 211 patients with UTUC who received RNU from January 2006 to December 2008 in our center. To clarify, a pending related manuscript using the same cohort of patients was also submitted to PLOS ONE (PONE-D-15-26646). All patients underwent routine hematologic examination, ultrasound, cystoscopy, computer tomography/magnetic resonance imaging, and/or ureteroscopy with tissue biopsy before RNU. Patients with the following conditions were excluded from this analysis: distant metastasis at the initial diagnosis; no data on preoperative plasma fibrinogen levels; blood coagulation disorders; severe hypertension; autoimmune disease; liver disease; conservative surgery instead of RNU; neo-adjuvant chemotherapy. Finally, a total of 184 consecutive patients with pathologically diagnosed UTUC were enrolled in this study. Data regarding patients' demographic and clinicopathologic characteristics, therapeutic regimens, and followup information were collected from a comprehensive database containing medical/pathological records of patients with UTUC. Glomerular filtration rate (GFR) was calculated using a CKD-EPI equation [19]. Preoperative CKD stages were classified on the basis of GFR (mL/ min/1.73 m 2 ) measured before RNU: stage 5 (GFR <15 or renal replacement therapy), stage 4 (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29), stage 3 (30-59), stage 2 (60-89) and stage 1 (90). Tumors were staged according to the 2002 Union for International Cancer Control (UICC) TNM classification system. Tumor grading was assessed based on the World Health Organization (WHO) 1973 guidelines. This study was approved by the institutional ethics committee of Peking University First Hospital. As a retrospective analysis of routine data, a waiver of written informed consent was granted from the ethics committee. Patient records/information was anonymized and de-identified prior to analysis.

Treatment and follow-up
All patients received open or laparoscopic RNU in this cohort. Based on a preoperative radiographic evaluation or intraoperative findings, a regional lymphadenectomy was performed in patients with enlarged lymph nodes. An extended lymphadenectomy was not routinely performed. Patients were assessed every 3 months for the first 2 years after the RNU, and annually thereafter. Evaluation items included routine blood tests, biochemical tests, urinalysis, cystoscopy, chest x-ray, and ultrasound/computed tomography/magnetic resonance imaging. The co-primary endpoints of this study were all-cause and cancer-specific deaths. Death causes were determined on the basis of death certificates or by patients' treating clinicians. Overall survival (OS) and cancer-specific survival (CSS) periods were defined as the time from the date of RNU to the date of all-cause and cancer-specific deaths.

Fibrinogen measurement
Plasma fibrinogen levels were routinely measured before surgery or diagnostic interventions as a part of a routine evaluation of coagulation function. For the plasma fibrinogen analysis, 3 mL of peripheral venous blood was collected on the day of admission to the hospital. Fibrinogen was measured based on the Clauss method as previously described [20]. The inter-assay coefficient of variation was <3% and <7.5% for the plasma control in the normal and pathological range, respectively.

Statistical analysis
Continuous variables were reported as medians (IQR), and categorical variables were expressed as frequencies and percentages. The optimal cutoff value for plasma fibrinogen was determined using validated web-based software [21], which allowed us to treat it as a binary variable. In short, every possible cutoff point was examined for survival analysis, and the most significant (log-rank test) threshold was used throughout all further analyses. The associations of preoperative plasma fibrinogen levels with patients' clinicopathological features were assessed by a nonparametric chi-squared test or Mann-Whitney U test. The relationships between plasma fibrinogen and inflammatory parameters (namely leukocyte, neutrophil and platelet counts, neutrophil-lymphocyte ratio and lymphocyte-monocyte ratio) were evaluated using a correlation analysis. OS and CSS were analyzed using the Kaplan-Meier method. A univariate analysis using the long rank test was performed to screen variables that may potentially predict prognosis. Then, those statistically significant variables were included in a multivariate Cox regression model to determine the independent prognostic risk factors. The optimal cutoff was determined using software based on R package, and the other statistical analyses were performed using the IBM Statistical Package for Social Sciences (SPSS) version 20.0. A result was considered statistically significant with P value of <0.05.

Association of preoperative plasma fibrinogen level with clinical characteristics
Based on the software mentioned above, the optimal cutoff value of 3.54 g/L was set, which best discriminated between survival and all-cause/cancer-specific death. Therefore, the cohort was divided into 2 groups. Of all the patients, 97 (52.7%) had a plasma fibrinogen level of less than 3.54 g/L, and 112 (67.1%) had a level of 3.54 g/L or above. There were no statistically significant differences in patient age, gender, smoking history, and lymphovascular invasion between the two groups (P > 0.05). Nevertheless, a high plasma fibrinogen level was significantly associated with tumor necrosis (P = 0.020), lymph node involvement (P = 0.028), and higher preoperative CKD stage (P = 0.002), pathological T stage (P = 0.010) and tumor grade (P = 0.011), which is shown in Table 1. Furthermore, Spearman's correlation analysis showed that plasma fibrinogen level was positively correlated with neutrophil-lymphocyte ratio, leukocyte, neutrophil and platelet counts and was negatively correlated with lymphocyte-monocyte ratio (Fig 1).

Association of preoperative plasma fibrinogen level with survival
During the median follow-up of 78 (34-92) months, 82 (44.6%) died and 53 (28.8%) died of cancer-specific causes. In the entire cohort, the 5-year OS and CSS rate were 67.4±3.5% and 74.8±3.3%, respectively. The Kaplan-Meier survival analysis revealed that a plasma fibrinogen 3.54 g/L predicted significantly poorer OS and CSS than < 3.54 g/L (P < 0.001 for both), which is shown in Figs 2 and 3. In the patients with plasma fibrinogen < 3.54 and 3.54, the 5-year OS rates were 79.4±4.1% and 54.0±5.3%, and the 5-year CSS rates were 84.1±3.8% and 63.9±5.4%, respectively.
The univariate analysis identified a preoperative plasma fibrinogen level as a statistically significant predictor for both OS and CSS. To clarify the independent prognostic significance of preoperative plasma fibrinogen level for OS and CSS, multivariate analyses using a Cox proportional hazards regression model were performed to adjust for other prognostic indicators. The results of multivariate analyses revealed that preoperative plasma fibrinogen level was an independent risk factor in patients with UTUC. Patients with plasma fibrinogen 3.54 g/L had a higher risk of all-cause death (P = 0.006) and cancer-specific death (P = 0.043) than those with plasma fibrinogen < 3.54 g/L. Gender and pathological T stage were also independent prognostic indicators for OS and CSS. Older age ( 70) and preoperative CKD stage 4/5 were also identified as independent risk factors for OS. Tumorous ureter involvement was statistically significantly associated with poor CSS (Table 2).

Discussion
Despite the developments in identifying biomarkers to predict survival, the conventional prognostic risk stratification heavily depends on postoperative parameters such as pathologic T stage, tumor grade, and N classification in patients with UTUC. However, more recommendations of neo-adjuvant chemotherapy due to decreased renal function after RNU highlights the significance of exploring effects and simple preoperative prognostic indicators in this cohort.
To the best of our knowledge, to date only two studies have revealed the prognostic value in UTUC in Japanese and Austrian populations [17,18], whose pathogenesis is quite different from that in China. Before generally applying plasma fibrinogen for prognostic prediction,  external validation in different ethnic subgroups is of great importance. Additionally, potential associations between plasma fibrinogen level and inflammatory biomarkers were not assessed in the two above-mentioned studies. Considering that the threshold of defining hyperfibrinogenemia does not necessarily best discriminate a poor and good oncological outcome, the optimal cutoff of plasma fibrinogen with an ideal predictive value remained to be determined in patients with UTUC.
Compared to two prior studies on the prognostic value of plasma fibrinogen in UTUC [17,18], there were some differences in the cohort of this current study, which are illustrated in Table 3. In the cohort of our study, there were more female patients (54.3% vs. 26.6% and 39.5%), more tumors with localized pathological T stage (pTa-2: 78.3% vs. 46.8% and 61.3%) and lower tumor grades (G1-2: 63.6% vs. 39.6% and 54.4%) than in the Japanese and European cohorts. This was in accordance with the previous conclusion that Chinese UTUC patients are predominantly female with preferable pathological characteristics and relatively good oncological outcomes compared with those of western and Japanese populations. Similarly, our findings also revealed an association of high plasma fibrinogen with high pathological T stage and tumor grade. However, no statistically significant relationship between plasma fibrinogen and lymphovascular invasion was noted, which is inconsistent with what Tanaka et al. reported [18]. Importantly, we observed that elevated preoperative plasma fibrinogen was significantly associated with lymph node involvement and tumor necrosis. The relationship between increased plasma fibrinogen and lymph node metastasis was also observed in another study on gastric cancer [16].
Patients with malignancy suffered an abnormal state of hypercoagulability that contributes to tumor invasion, distant metastasis and other poor clinical outcomes [22,23]. On the other hand, anticoagulants, including heparins and warfarin, have been reported to have antitumor and anti-metastatic effects [24]. Several possible mechanisms through which fibrinogen promotes tumor progression and leads to poor oncological outcome were raised. As an important molecular bridge, fibrinogen might facilitate stable adhesion between tumor cells, platelets and endothelial cells [25]. It has been reported that fibrinogen enhanced metastatic potential in part by impeding natural killer cell-mediated elimination of tumor cells [26]. It has been also suggested that fibrinogen and platelets facilitate each other in protecting tumor cells from natural killer cytotoxicity by forming thrombin, which depended on β3-integrins expressed on  human cancer cells [27]. Endogenously synthesized fibrinogen interacts with several growth factors (including VEGF and FGF-2), which promotes cellular adhesion, proliferation and migration of tumor cells and neoplastic angiogenesis [28,29]. In addition, plasma fibrinogen also plays a crucial role in the coagulation cascade, and elevated preoperative fibrinogen might also result in poor clinical outcome by inducing thromboembolism. It has been reported that high preoperative fibrinogen levels were related to elevated risk of pulmonary embolism, leading to a shortened overall survival [30]. In the Chinese population, patients with UTUC often also have CKD, which complicates the case and leads to a state of chronic systemic inflammation [3]. Growing evidence has proved the adverse impact of systemic inflammation on cancers [31,32], including UTUC [33]. Importantly, genetic and pharmacologic studies have clarified the vital significance of fibrinogen in determining the extent of systemic inflammation, and its proinflammatory role has been reported in several types of malignancies [34]. Furthermore, certain inflammatory proteins, such as interleukin-6 or C-reactive protein, promote the production of fibrinogen. Our findings showed that plasma fibrinogen level is greatly associated with several inflammatory parameters, namely leukocyte count, neutrophil count, platelets count, NLR and LMR. Therefore, in addition to its direct adverse influence on oncological outcome, plasma fibrinogen reflects the degree of systemic inflammation that contributes to poor survival in patients with UTUC.
This current study has limitations: First, it was a single-center, small-sample and retrospective research, making certain unknown bias inevitable. Secondly, levels of some specific inflammatory proteins, such as interleukin-6 or C-reactive protein, were unavailable. Therefore, their association with plasma fibrinogen level was not evaluated in our study. Moreover, detailed information about steroid use that might affect plasma fibrinogen levels might be insufficient. However, even considering these limitations, our findings clearly suggest that an elevated plasma fibrinogen level is an independent risk factor for poor OS and CSS. We validated its Meanwhile, it was also associated with tumor length 30mm (P = 0.022), higher T stage (P = 0.007) and tumor grade (P < 0.001).
This current study 184 Chinese UTUC patients (54.3% female) treated with RNU 3.54g/L was set according to a cutoff optimization tool [21].
CSS, cancer-specific survival; OS, overall survival; UTUC, upper tract urothelial carcinoma. doi:10.1371/journal.pone.0150193.t003 prognostic value in a Chinese cohort of patients with UTUC. We also elucidated the association of fibrinogen with several inflammatory parameters, and underlined its significance in systemic inflammation in UTUC for the first time. This parameter should be applied to improve prognostic risk stratification models and assist clinicians in therapeutic strategy decisionmaking.
In conclusion, elevated preoperative plasma fibrinogen level is an independent risk factor of OS and CSS in patients with UTUC. As a low-cost and effective biomarker with easy accessibility, it should be monitored and applied in prognostic risk stratification methods in the future, which will assist clinician in therapeutic decision-making and improve treatment outcomes.