Prognostic significance of preoperative plasma fibrinogen levels in primary gastrointestinal stromal tumours: a retrospective cohort study

Improved prediction of prognosis for gastrointestinal stromal tumours (GISTs) has become increasingly important since the introduction of targeted therapy. Here, we aimed to evaluate the prognostic significance of preoperative plasma fibrinogen (Fib) levels in patients with primary GISTs and to analyse their correlations with clinicopathological characteristics. A total of 201 previously untreated patients with primary GISTs who had undergone radical surgery at our institution between October 2004 and July 2018 were enrolled. The optimal cut-off value for Fib levels was calculated using time-dependent receiver-operating characteristic curve analysis. RFS, the primary endpoint, was calculated by the Kaplan–Meier method and compared by the log-rank test. Univariate and multivariate Cox regression models were calculated. High preoperative plasma Fib levels were detected as an independent adverse prognostic factor (p = 0.008, hazard ratio 3.136, 95% CI 1.356‒7.256). Furthermore, high preoperative plasma Fib levels also indicated a poor prognosis within the modified National Institutes of Health (mNIH) high-risk subgroup (p = 0.041). In addition, preoperative plasma Fib levels showed a positive correlation with several prognostic factors and even a linear relationship with tumour size (Spearman correlation coefficient [r] = 0.411, p < 0.001). Our results suggest that high preoperative plasma Fib levels may indicate a poor prognosis in patients with primary GISTs. As a cost-effective biomarker, preoperative assessment of plasma Fib levels may help to further risk stratify patients with mNIH high-risk GISTs and instruct the application of targeted therapy.


Abstract
Background Improved prediction of prognosis for gastrointestinal stromal tumours (GISTs) has become increasingly important since the introduction of small molecule tyrosine kinase inhibitors. Here, we aimed to evaluate the prognostic significance of preoperative plasma fibrinogen (Fib) levels in patients with primary GISTs and to analyse their correlations with clinicopathological characteristics. Methods A total of 201 previously untreated patients with primary GISTs who had undergone radical surgery at our institution between October 2004 and July 2018 were enrolled. Patient demographics, clinicopathological characteristics, preoperative plasma Fib levels and recurrence-free survival (RFS) were analysed. The optimal cut-off value for Fib levels was calculated using time-dependent receiver operating characteristic curve analysis. RFS, the primary endpoint, was calculated by the Kaplan-Meier method and compared by the log-rank test.
Univariate and multivariate Cox regression models were calculated. Results Patients in the high Fib group had a shorter RFS than those in the low Fib group (P < 0.001). In multivariate analysis, high preoperative plasma Fib levels were detected as an independent adverse prognostic factor (P = 0.008, hazard ratio 3.136, 95% CI 1.356-7.256). Furthermore, high preoperative plasma Fib levels also indicated a poor prognosis within the modified National Institutes of Health (mNIH) high-risk subgroup (P = 0.041). In addition, preoperative plasma Fib levels showed a positive correlation with several prognostic factors and even a linear relationship with tumour size (Spearman correlation coefficient [ r ] = 0.411, P < 0.001). Conclusions High preoperative plasma Fib levels may indicate a poor prognosis in patients with primary GISTs. As a cost-effective biomarker, preoperative assessment of plasma Fib levels may help to further risk stratify patients with mNIH high-risk GISTs and instruct the application of targeted therapy. 3 Background Gastrointestinal stromal tumours (GISTs) are rare but the most common mesenchymal neoplasms in the gastrointestinal (GI) tract, with an annual incidence of approximately 10 per million people [1,2]. The malignant potential of GISTs ranges from benign tumours to rapidly progressing sarcomas [3].
The only well-recognized curative treatment for primary GISTs is radical resection [4].
However, the recurrence rate remains high after curative resection alone, especially for high-risk patients, who have a greater than 50% chance of recurrence at 2 years [5].
Fortunately, the introduction of small molecule tyrosine kinase inhibitors (TKIs) revolutionized the treatment strategies and greatly improved the prognosis of patients with GISTs [4,6]. The well-recognized indications for TKIs are inoperable or metastatic GISTs and GISTs with a moderate or high risk of recurrence after resection [4]. However, all the current popularized risk-stratification schemes are based only on tumour-specific factors after resection [7][8][9][10][11]. Therefore, in this era of adjuvant TKI therapy, it is urgent to find some preoperative factors of prognosis and perfect the current risk-stratification schemes to improve the prediction of prognosis and instruct the application of targeted therapy early in the treatment setting.
In addition to tumour-specific factors, hypercoagulation is thought to be associated with several developmental processes of tumours, such as tumour angiogenesis, invasion, progression, and metastasis [12,13]. Fibrinogen (Fib) is one of the most significant indicators of coagulation [12,14]. In the last few years, high preoperative Fib levels have been found to be associated with poor prognosis in various solid tumours [15][16][17][18][19][20][21].
However, to our knowledge, there are few studies on the relationship between Fib and the prognosis of patients with GISTs, and their results have been controversial [22,23].
Furthermore, no studies have explored whether Fib correlates with prognosis within different recurrent risk subgroups. Hence, there is limited evidence confirming the prognostic ability of Fib in GISTs.
In the present study, we analysed the preoperative plasma Fib levels in 201 patients with primary GISTs who had undergone radical surgery, seeking to evaluate their prognostic significance in the overall cohort and subgroups, and further analysed their correlation with clinicopathological prognostic factors.

Patients
We retrospectively reviewed the medical records of 237 consecutive patients with primary GISTs who had undergone surgery at Beijing Hospital between October 2004 and July 2018. A total of 201 patients were enrolled in this study. This retrospective study was approved by the institutional review board of Beijing Hospital.
Participants in the present study met the following inclusion criteria: 1) postoperatively confirmed GIST; 2) R0 resection; 3) available plasma Fib level data within 7 days before operation; and 4) age > 18 years. The exclusion criteria were as follows: 1) 6 months or less of follow-up data; 2) preoperative imatinib treatment; 3) coexistent haematological disorders, including splenectomy, thromboembolism, or anticoagulant therapy within 3 months before operation; 4) blood transfusion or inflammatory disorders within 2 weeks before operation; 5) synchronous metastasis or other malignancies; and 6) acute or chronic liver disease.

Plasma fibrinogen
Data on preoperative plasma Fib levels were retrospectively obtained from blood coagulation analysis before breakfast within 7 days prior to surgery. Plasma Fib levels were measured by the Clauss clotting method using the automatic blood coagulation analyser ACL TOP TM 700 (Instrumentation Laboratory, Werfen Group, America) according to the manufacturer's instructions, with its matching thrombin reagent.

Pathological assessment and tumour grading
The diagnosis of GISTs was well recognized by histopathologic examination of resection specimens according to the Protocol for the Examination of Specimens From Patients With Gastrointestinal Stromal Tumour [24]. Tumour location, size, rupture, and mitotic index (number of mitoses per 50 high-power fields [HPFs]) were recorded. The malignant potential for each GIST was based on the modified National Institutes of Health (mNIH) risk classification [7].

Statistical analysis
The optimal cut-off value for preoperative Fib levels was calculated using time-dependent receiver operating characteristic (ROC) curve analysis. Continuous values were assessed using Student's t-test or the Mann-Whitney U-test, and categorical data were compared by the χ 2 test, Fisher's exact test or the Mann-Whitney U-test, as appropriate. Spearman's rank correlation coefficients were used to examine associations between two continuous variables. Recurrence-free survival (RFS) was defined as a composite endpoint of local recurrence, distant metastasis, or death from any cause, whichever came first. RFS curves were calculated by the Kaplan-Meier product limit method and then compared using the log-rank test. Univariate and multivariate Cox proportional hazard regression models were constructed to identify associations with RFS. All tests were two-sided, and P ˂ 0.05 was considered statistically significant. Statistical analyses were performed using SPSS 20.0 and R software.

ROC curve analysis
A ROC curve for preoperative plasma Fib levels and the prediction of 5-year RFS is shown in Figure 1. The area under the ROC curve of Fib was 0.665. Based on the 5-year RFS, the optimal cut-off value of 3.48 g/L had the highest sensitivity (56.5%) and specificity (74.0%). Patients were categorized into groups of H-Fib (> 3.48 g/L) and L-Fib (≤ 3.48 g/L) according to the cut-off value.

Patient demographics and clinicopathological features
Patient demographics and clinicopathological features are summarized in Table 1 According to the mNIH risk classification [7], 99 patients (49.25%) were classified in the very low/low-risk group, and 102 patients (50.75%) were classified in the moderate/highrisk group. Among the 102 patients, 36 patients (35.29%) received adjuvant imatinib treatment following surgery.

Associations between preoperative plasma fibrinogen levels and clinicopathological factors
According to the cut-off value of 3.48 g/L, patients were divided into two groups: 141 7 patients were in the H-Fib group, and 60 patients were in the L-Fib group (Table 1).
Significantly, patients in the H-Fib group were more likely to be older (P = 0.002) and have a larger tumour size (P ˂ 0.001), a greater mitotic index (P ˂ 0.001) and a higher mNIH risk classification (P ˂ 0.001) than those in the L-Fib group. Other features, including sex, cardiovascular diseases (CVDs), diabetes, GI bleeding, body mass index (BMI), plasma albumin (Alb) levels, Eastern Cooperative Oncology Group (ECOG) performance status, adjuvant imatinib treatment and tumour location, were similarly distributed between both groups. The preoperative plasma Fib levels showed a positive correlation with tumour size

Discussion
The present study demonstrated that high preoperative plasma Fib levels were significantly associated with poor RFS in patients with primary GISTs who underwent radical surgery and explored its possible cut-off value (3.48 g/L) to predict RFS. To the best of our knowledge, this study represents the largest dedicated series published, focusing on the prognostic significance of preoperative plasma Fib levels in patients with primary GISTs. Furthermore, we found, for the first time, that high preoperative plasma Fib levels still indicated a poor RFS in patients with mNIH high-risk GISTs. In addition, our study showed a positive correlation between preoperative plasma Fib levels and wellrecognized prognostic factors, including tumour size, mitotic index, and mNIH risk classification.
In our study, the correlation between high preoperative plasma Fib levels and poor prognosis was further verified, which is consistent with previous studies on GISTs [22,23].
However, the cuff-off value (3.48 g/L) is inconsistent with the study by Cai et al.  [15][16][17][18][19][20][21]. Some researchers found that Fib was positively correlated with the neutrophil-lymphocyte ratio (NLR), and others found that the combination of Fib and NLR could better predict prognosis, but the specific mechanisms remain unclear[19, 25,26]. Moreover, some animal experiments have suggested that Fib is an important factor in the metastatic potential of tumour cells [27,28]. Some studies have indicated that, in addition to antithrombotic functions, heparins and derivatives also exert critical antimetastatic effects by interfering with P-selectin-mediated cell binding [29,30]. As a biomarker of nutritional status, Alb has been recognized as a prognostic factor in some kinds of tumours [31][32][33][34]. However, in the present study, although preoperative Alb levels were found to be associated with RFS in the univariate analysis, they were not detected as an independent predictive factor in the multivariate analysis. In the present study, tumour size was not observed to be an independent prognostic factor, which may be due to the limited sample size, the exclusion of patients who received preoperative imatinib treatment or non-radical resection, or other factors obscuring the prognostic significance.
Since the survival analyses were conducted in the overall population, prognostic benefits of adjuvant imatinib treatment were not observed. However, in our study, adjuvant imatinib treatment significantly improved the prognosis of patients with mNIH moderate/high-risk GISTs (P = 0.015), which is consistent with previous studies [4,6].
In the mNIH high-risk subgroup analysis, for the first time, preoperative plasma Fib levels were detected to be a significant prognostic factor, which would help to further risk stratify patients with mNIH high-risk GISTs and instruct the application of targeted therapy. However, correlations were not significant within the very-low/low/moderate-risk subgroup analysis, which may be due to the limited number of endpoint events and the shorter follow-up time. Although there was no significant correlation between Fib and RFS in subgroups of different tumour sizes, a trend was observed that high Fib indicated poor RFS, which may require a larger sample size and a continuous follow-up time to further determine. Preoperative plasma Fib levels were a significant prognostic factor for GISTs in both stomach and non-stomach, which may indicate that plasma Fib levels were not significantly associated with the tumour locations. For patients without adjuvant imatinib treatment, high preoperative plasma Fib levels indicated a significantly poor prognosis.
However, in the subgroup with adjuvant imatinib treatment, a correlation was not observed between Fib and RFS, which may be due to the prognostic benefits from imatinib treatment or the limited sample size.
Our study showed a positive correlation between the preoperative plasma Fib levels and several prognostic factors, including tumour size, mitotic index, and mNIH risk classification, which is similar to the study by Cai et al [23]. Furthermore, we even observed a linear correlation between preoperative plasma fibrinogen levels and tumour size, which indirectly indicates Fib's prognostic significance. Previous studies have reported that the plasma concentrations of some coagulation factors, such as Fib, increase progressively with age [35]. We also found that patients in the H-Fib group were older than patients in the L-Fib group, which was consistent with previous studies [23 , 35].
However, in our study, for patients with GISTs, the correlation between age and the preoperative plasma Fib levels was not strong (Spearman correlation coefficient [r] = 0.198, P = 0.005). Furthermore, in our study, regardless of age, high preoperative plasma Fib levels indicated a poor prognosis for patients with GISTs. Therefore, the impact of ageing could not cover up the correlation between preoperative plasma Fib levels and the malignant degree of GISTs.
The molecular mechanisms by which tumour cells interact with the haemostatic system are yet to be uncovered. Several possible mechanisms were proposed to explain the complex correlation. On the one hand, tumour cells activate the haemostatic system in multiple ways. Tumour cells not only directly activate the coagulation cascade by producing many procoagulant proteins (tissue factor, heparanase, cancer procoagulant, and tissue factor-positive microparticles) but also stimulate the procoagulant properties of the host's haemostatic cells (endothelial cells, platelets, and leukocytes), thereby increasing plasma Fib levels [36,37]. We also cannot exclude that the pathophysiological mechanism of hypercoagulation may be secondary to the tumour-derived systemic inflammatory response and/or intra-abdominal infectious disease [36][37][38][39]. Indeed, all of the procoagulant mechanisms elicited by tumour tissues, as well as the patient's general and clinical thrombotic risk factors, contribute to the occurrence of hypercoagulation in patients with cancer [37]. On the other hand, Fib could also promote tumour progression in return. In the tumour microenvironment, Fib could influence the development of tumours through complex interactions with multiple integrin or non-integrin Fib receptors (e.g., cadherins, αIIbβ3, αVβ3, αXβ2, αMβ2, α5β1, αVβ1, and Toll-like receptors), which mediate innate immune cell function, tumour cellular proliferation, migration, and apoptosis [40][41][42][43][44][45][46]. For example, Fib has been suggested to be a bridging molecule between tumour cells and vascular endothelial growth factor, which could stimulate angiogenesis and promote tumour proliferation [46]. All of these possible mechanisms 12 promote a positive feedback loop between tumour progression and hypercoagulation.
As a cost-effective biomarker, Fib is easily detected with conventional coagulation analysis before surgery. Accordingly, the evaluation of Fib levels would be clinically useful for indicating the malignant potential and prognosis in combination with imaging and pathological features.
There are several limitations in the present study. First, a selection bias cannot be excluded because it was a retrospective study of a single institution. The exclusion of patients who did not undergo radical surgery, as well as the inclusion of patients with adjuvant imatinib treatment, had an effect on the prognosis of the overall cohort. Second, due to the limited follow-up time, the inclusion of patients with adjuvant imatinib treatment and the high survival of patients with GISTs, we did not analyse OS as the endpoint of this study. Third, genetic mutation analysis was not conducted in most patients because of the high cost, which limits further studies.

Conclusions
High preoperative plasma Fib levels may indicate poor prognosis in patients with primary GISTs who underwent radical surgery. As a cost-effective biomarker, preoperative assessment of plasma Fib levels may help to further risk stratify patients with mNIH highrisk GISTs and instruct the application of targeted therapy. However, larger prospective studies and further molecular biological experiments are warranted to confirm our results.