Hypercoagulation after neoadjuvant immunochemotherapy as a new prognostic indicator in patients with locally advanced gastric cancer undergoing surgery

Abstract

BACKGROUND

Coagulation status is closely related to the progression of malignant tumors. In the era of neoadjuvant immunochemotherapy (NICT), the prognostic utility of coagulation indicators in patients with locally advanced gastric cancer (LAGC) undergoing new treatments remains to be determined.

AIM

To determine whether hypercoagulation is an effective prognostic indicator in patients with LAGC who underwent radical resection after NICT.

METHODS

A retrospective analysis of clinical data from 104 patients with LAGC, who underwent radical resection after NICT between 2020 and 2023, was performed. D-dimer and fibrinogen concentrations were measured one week before NICT, and again one week before surgery, to analyze the association between these two indicators and their combined indices [non-hypercoagulation (D-dimer and fibrinogen concentrations within the upper limit of normal) vs hypercoagulation (D-dimer or fibrinogen concentrations above the upper limit of normal)] with prognosis. After radical resection, patients were followed-up periodically. The median follow-up duration was 21 months.

RESULTS

Data collected after NICT revealed that the three-year overall survival (OS) and disease-free survival (DFS) rates the non-hypercoagulation group were significantly better than those in the hypercoagulation group [94.4% vs 78.0% (P = 0.019) and 87.0% vs 68.0% (P = 0.027), respectively]. Multivariate analysis indicated that hypercoagulation after NICT was an independent factor for poor postoperative OS [hazard ratio (HR) 4.436, P = 0.023] and DFS (HR 2.551, P = 0.039). Pre-NICT data demonstrated no statistically significant difference in three-year OS between the non-hypercoagulation and hypercoagulation groups (88.3% vs 84.1%, respectively; P = 0.443).

CONCLUSION

Hypercoagulation after NICT is an effective prognostic indicator in patients with LAGC undergoing radical gastrectomy.

Key Words: Locally advanced gastric cancer; Coagulation; Neoadjuvant immunochemotherapy; Prognosis; Radical gastrectomy

Core Tip: This retrospective study assessed the correlation between hypercoagulation after neoadjuvant immunochemotherapy (NICT) and the prognosis of patients with locally advanced gastric cancer (LAGC). Results indicated that the non-hypercoagulation group exhibited significantly superior overall survival (OS) and disease-free survival (DFS) rates compared with the hypercoagulation group. Furthermore, multivariate analysis revealed that hypercoagulation after NICT was an independent risk factor for adverse postoperative OS and DFS. This suggests that hypercoagulation after NICT is an independent prognostic indicator of LAGC and may be used to evaluate therapeutic effects.

INTRODUCTION

Gastric cancer (GC) is a malignant tumor that threatens human health globally, with the highest reported incidence and mortality rates in East Asia[1]. For patients with locally advanced GC (LAGC), a multitude of convincing clinical trials have consistently validated radical gastrectomy followed by neoadjuvant chemotherapy as the gold-standard treatment[2-5]. However, pathological regression and overall survival (OS) after neoadjuvant chemotherapy remain unsatisfactory, with approximately 40% of treated patients experiencing recurrence and metastasis within 3 years after surgery[6-8]. Studies have shown that immunotherapy has become a significant development in recent years and helps improve the treatment landscape of patients with LAGC[9-13]. Therefore, an increasing number of researchers have become interested in the effects of neoadjuvant immunochemotherapy (NICT) on patients diagnosed with LAGC. However, a subset of patients exhibit insensitivity to NICT, indicating an urgent need for an indicator that can accurately predict the therapeutic effects of NICT.

Since the discovery of the complex relationship between the coagulation system and tumor progression, research investigating coagulation function in patients diagnosed with cancer has burgeoned[14-18]. Many coagulation markers, such as fibrinogen, D-dimer, and platelet count, can predict cancer progression[19-25]. Recently, a hypercoagulation index comprising 2 coagulation markers-fibrinogen and D-dimer-was established and validated as a better prognostic indicator of esophageal squamous cell carcinoma[24]. However, there is still a significant lack of research in this area within LAGC, especially regarding the analysis of multiple combined coagulation markers, which has rarely been reported. In addition, although NICT has a significant impact on the coagulation system, it remains unclear whether the values of coagulation markers before and after NICT are associated with prognosis.

In this study, we analyzed the prognostic correlation between plasma D-dimer and fibrinogen concentrations before and after NICT, and the utility of their combination in assessing the prognosis of patients with LAGC undergoing radical gastrectomy after NICT.

MATERIALS AND METHODS

Patient selection

Clinical data from patients with LAGC, who underwent NICT followed by radical excision between 2020 and 2023, were retrospectively analyzed. The inclusion criteria were as follows: Age ≥ 18 years; pathologically diagnosed with gastric adenocarcinoma; clinical stage T3-4N+M0; and underwent NICT and radical surgery. The exclusion criteria were as follows: Insufficient data for analysis; combined radiotherapy or targeted therapy; presence of distant metastasis on NICT; history of other malignant diseases; and death in hospital after surgery. The detailed screening process is illustrated in Figure 1. Screening identified 104 patients who fulfilled the established criteria and exhibited detectable plasma D-dimer and fibrinogen concentrations. The study was approved by the ethics committee (approval No. 2020-0447) and adhered to the Declaration of Helsinki guidelines.

Figure 1 Flow diagram for screening participants. LAGC: Locally advanced gastric cancer.

Treatment and follow-up

Before radical surgery, all enrolled patients underwent NICT every 21 days. The chemotherapy regimens were SOX (i.e., S-1 + oxaliplatin) and XELOX (i.e., oxaliplatin and capecitabine). Immunotherapy regimens included sintilimab, toripalimab, nivolumab and tislelizumab[26-29]. Details are summarized in Supplementary Table 1.

After 2 cycles of NICT, patients underwent re-evaluation of condition and treatment effects. Treatment effects were evaluated according to the Response Evaluation Criteria in Solid Tumors (i.e., “RECIST”) guideline, version 1.1. Patients with ineffective treatment discontinued NICT and withdrew from the study, whereas those with effective treatment completed the remaining treatment cycle. For patients with R0 resection possibility, radical resection should be performed. After conclusion of treatment, patients were followed-up periodically. The last follow-up for all the patients was in January 2024.

Coagulation factors

Plasma concentrations of D-dimer [assessed via latex agglutination (upper limit of normal, 0.5 mg/L)] and fibrinogen (upper limit of normal, 4.0 g/L) were determined 1 week before NICT and again one week before surgery. A non-hypercoagulation state was defined when both D-dimer and fibrinogen concentrations were less than their respective upper limits of normal, whereas a hypercoagulation state was defined when either of these two concentrations exceeded the upper limits of normal.

Statistical analysis

Differences in basic characteristics, clinical stages, and pathological characteristics between the 2 groups were analyzed using student’s t-test (numerical variables), χ² test, or Fisher’s exact test (categorical variables). The Kaplan-Meier method was used to plot survival curves and assess both survival rates and differences between the curves. Additionally, Cox regression analyses of hazard ratio (HR) with corresponding 95% confidence interval (CI) were performed to derive prognostic variables for OS and disease-free survival (DFS). Statistical analysis was performed using SPSS version 21.0 (IBM Corp., Armonk, NY, United States) and R software (Copenhagen: The Cochrane Collaboration, 2020). Differences with P < 0.05 were considered to be statistically significant.

RESULTS

Patient characteristics

Characteristics of the 104 eligible patients (23 females, 81 males; median age, 58 years; mean ± SD body mass index, 22.82 ± 2.81 kg/m²) are summarized in Table 1. The most common tumor location was the upper stomach (55/104, 52.9%). Clinical T stages were cT3 (n = 60) and cT4 (n = 44), whereas the clinical N stages were cN1 (n = 66), cN2 (n = 30), and cN3 (n = 8). Signet ring cell carcinoma accounted for 22.1% (23/104) of cases. Immunotherapy regimens included sintilimab (n = 13), toripalimab (n = 8), nivolumab (n = 6), and tislelizumab (n = 77). The median plasma D-dimer and fibrinogen concentrations before NICT were 0.32 mg/L (range 0.10-2.43 mg/L) and 3.30 mg/L (range 1.83-5.66 mg/L), respectively. The median plasma D-dimer and fibrinogen concentrations before surgery were 0.45 g/L (range 0.22-3.96 g/L) and 2.74 g/L (range 1.78-6.21 g/L), respectively. The median follow-up duration was 21 months.

Table 1 Patients’ characteristics.

Parameters	Number
Age [median (range)]	58 (31-75)
Gender, male/female (n)	81/23
BMI (kg/m2) (mean ± SD)	22.82 ± 2.81
Smoking status, never smoker/ever smoker (n)	76/28
Drinking status, never drinker/ever drinker (n)	87/17
Plasma D-dimer (mg/L) [median (range)]
Before neoadjuvant treatment	0.32 (0.10-2.43)
Before surgery	0.45 (0.22-3.96)
Plasma fibrinogen (g/L) [median (range)]
Before neoadjuvant treatment	3.30 (1.83-5.66)
Before surgery	2.74 (1.78-6.21)
Tumor location (n)
Upper stomach	55
Middle stomach	14
Lower stomach	35
Clinical T stage (n)
cT3	60
cT4	44
Clinical N stage (n)
cN1	66
cN2	30
cN3	8
Pathological type (n)
Non- signet ring cell carcinoma	81
Signet ring cell carcinoma	23
Chemotherapy regimen, sox/xelox (n)	101/3
Immunotherapy regimen (n)
Sintilimab	13
Toripalimab	8
Nivolumab	6
Tislelizumab	77
Median follow-up (months)	21

BMI: Body mass index.

Comparison of clinical parameters

Fifty (48.1%) patients developed preoperative hypercoagulation, among whom 37 (35.6%) exhibited high D-dimer concentrations, 4 (3.8%) had high fibrinogen concentrations, and 9 (8.7%) had high concentrations of both. Comparisons of clinical parameters between the hypercoagulation and non-hypercoagulation groups are reported in Supplementary Table 2. Patients in the hypercoagulation group were significantly older than those in the other group (61.30 ± 7.88 years vs 54.93 ± 9.46 years, P < 0.001). However, differences in other factors were not statistically significant. R software was then used to determine the optimal age cut-off value for age (66 years); patients were then and divided into 2 age groups accordingly: < 66 years and ≥ 66 years (Supplementary Figure 1).

Prognostic role of coagulation state before and after NICT

According to results of analysis, 14 of the 104 patients died: 7 of those who were in a hypercoagulable state before NICT; and 11 of those who were in a hypercoagulable state after NICT. High pre-NICT D-dimer and fibrinogen concentrations were not significantly associated with OS (D-dimer, P = 0.506; fibrinogen, P = 0.090) (Figure 2A). There was no significant difference in OS between the hypercoagulation and non-hypercoagulation groups (P = 0.443) (Figure 2B). However, patients with D-dimer concentration below the upper limit of normal after NICT experienced significantly better OS than those with D-dimer concentrations above the upper limit of normal, as did those with fibrinogen concentrations (D-dimer, P = 0.028; fibrinogen, P = 0.003) (Figure 3A). The non-hypercoagulation group also experienced a longer OS than the hypercoagulation group (P = 0.019) (Figure 3B).

Figure 2 Comparison of overall survival according to pretherapeutic coagulation factors. A: Overall survival (OS) associated with pretherapeutic plasma D-dimer and fibrinogen concentrations; B: Comparison of OS between pretherapeutic hypercoagulation and non-hypercoagulation; C: Comparison among pretherapeutic non-hypercoagulation, high concentrations of either (either elevated), and high concentrations of both (both elevated) after neoadjuvant immunochemotherapy.

Figure 3 Comparison of overall survival according to preoperative coagulation factors. A: Overall survival (OS) associated with preoperative plasma D-dimer and fibrinogen concentrations; B: Comparison of OS between preoperative hypercoagulation and non-hypercoagulation; C: Comparison among preoperative non-hypercoagulation, high concentrations of either (either elevated), and high concentrations of both (both elevated) after neoadjuvant immunochemotherapy.

Subsequently, patients in the hypercoagulation group were categorized into 2 distinct groups: Those with elevated D-dimer or fibrinogen concentrations, and those with increased concentrations of both D-dimer and fibrinogen. The 2 groups were then compared with the non-hypercoagulation group (Figure 2C; Figure 3C). There was a significant difference between the groups only after NICT (P = 0.001).

DFS after NICT

Patients with low D-dimer concentrations after NICT experienced better DFS (P = 0.025) (Figure 4A), and the same result was observed for fibrinogen concentrations (P = 0.018) (Figure 4A). DFS of the non-hypercoagulation group was also significantly longer than that of the hypercoagulation group (P = 0.027) (Figure 4B). The overall recurrence rate was 22.1%. The recurrence rate in the hypercoagulation group was significantly higher than that in the non-hypercoagulation group (13% vs 32%, P = 0.019) (Figure 4C).

Figure 4 Comparison of recurrence according to levels of preoperative coagulation factors. A: Disease-free survival (DFS) associated with preoperative D-dimer and fibrinogen concentrations; B: Comparison of DFS between hypercoagulation and non-hypercoagulation; C: Frequency of the recurrences in each coagulation group.

Predictors of DFS and OS according to Cox regression analysis

D-dimer and fibrinogen concentrations before NICT had no significant effect on OS. Similarly, based on the concentrations of these 2 markers before NICT, there was no significant difference in OS between the hypercoagulation and non-hypercoagulation groups. Therefore, marker concentrations after NICT were chosen for analysis. Prognostic factors related to survival identified using Cox regression analyses based on various clinical characteristics are summarized in Tables 2 and 3. The data revealed that neurological invasion, signet-ring cell carcinoma, and hypercoagulation were associated with prognosis. Multivariate Cox regression analysis indicated that hypercoagulation was an independent predictor of OS (HR 4.436, 95%CI: 1.227-16.045; P = 0.023) and DFS (HR 2.551, 95%CI: 1.047-6.217; P = 0.039).

Table 2 Preoperatively determined prognostic factors for overall survival in locally advanced gastric cancer.

Variable	Univariate			Multivariate
	Hazard ratio	95%CI	P value	Hazard ratio	95%CI	P value
Gender			0.587
Male	Reference
Female	0.660	0.147-2.956
Age (years)			0.138
< 66	Reference
≥ 66	2.235	0.773-6.464
Smoking history			0.682
Never smoker	Reference
Ever smoker	0.766	0.213-2.753
Drinking history			0.922
Never drinker	Reference
Ever drinker	1.078	0.240-4.844
BMI (kg/m2)			0.670
< 22	Reference
≥ 22	0.796	0.278-2.279
Tumor location			0.593
Upper stomach	Reference
Middle stomach	1.556	0.410-5.904	0.516
Lower stomach	0.676	0.178-2.567	0.565
Vascular invasion			0.079
Yes	Reference
No	2.662	0.893-7.934
Neurological invasion			0.015			0.055
Yes	Reference			Reference
No	4.324	1.328-14.079		3.329	0.975-11.366
Clinical T stage			0.109
cT3	Reference
cT4	2.494	0.815-7.625
Clinical N stage			0.258
cN1	Reference
cN2-3	1.851	0.637-5.381
Pathological type			0.024			0.082
Non- signet ring cell carcinoma	Reference			Reference
Signet ring cell carcinoma	3.428	1.177-9.981		2.694	0.882-8.231
Hypercoagulation			0.031			0.023
No	Reference			Reference
Yes	4.078	1.136-14.643		4.436	1.227-16.045

CI: Confidence interval; BMI: Body mass index.

Table 3 Preoperatively determined prognostic factors for disease-free survival in locally advanced gastric cancer.

Variable	Univariate			Multivariate
	Hazard ratio	95%CI	P value	Hazard ratio	95%CI	P value
Gender			0.747
Male	Reference
Female	1.177	0.437-3.172
Age (years)			0.108
< 66	Reference
≥ 66	1.988	0.859-4.598
Smoking history			0.874
Never smoker	Reference
Ever smoker	0.927	0.365-2.352
Drinking history			0.745
Never drinker	Reference
Ever drinker	1.196	0.406-3.520
BMI (kg/m2)			0.998
< 22	Reference
≥ 22	1.001	0.438-2.286
Tumor location			0.951
Upper stomach	Reference
Middle stomach	0.850	0.242-2.984	0.800
Lower stomach	0.890	0.355-2.232	0.804
Vascular invasion			0.088
Yes	Reference
No	2.074	0.897-4.798
Neurological invasion			0.038			0.044
Yes	Reference			Reference
No	2.432	1.051-5.627		2.372	1.024-5.492
Clinical T stage			0.066
cT3	Reference
cT4	0.459	0.200-1.054
Clinical N stage			0.744
cN1	Reference
cN2-3	1.154	0.489-2.722
Pathological type			0.439
Non- signet ring cell carcinoma	Reference
Signet ring cell carcinoma	1.445	0.569-3.669
Hypercoagulation			0.034			0.039
No	Reference			Reference
Yes	2.611	1.073-6.356		2.551	1.047-6.217

CI: Confidence interval; BMI: Body mass index.

DISCUSSION

The present study focused on patients with LAGC who underwent curative gastrectomy after NICT, and analyzed the associations between coagulation markers and patient prognosis to identify the most useful prognostic indicators. Results indicated that the coagulation markers D-dimer and fibrinogen after NICT were correlated with a favorable prognosis, and the predictive effect was better when they were used in combination. The three-year OS and DFS rates were significantly higher in the non-hypercoagulation group than those in the hypercoagulation group, and hypercoagulation was an independent prognostic variable.

Malignant tumors affect the coagulation system, and vice-versa[14]. The presence of tumors, release of tissue factors, and vascular endothelial damage caused by the destruction of tumor cells by chemotherapy can lead to abnormal coagulation and thrombosis. These events can lead to cancer progression, particularly hematogenous metastasis. An increase in blood viscosity and the impact of the coagulation system on immune function provide more opportunities for tumor cells to implant[14,17,30-33]. In addition, abnormalities in the coagulation system can lead to decreased postoperative treatment efficacy and disease recurrence. A hypercoagulable state can also induce other risks for mortality, such as myocardial infarction, cerebral infarction, and pulmonary embolism caused by thrombosis, which further affect patient prognosis[34,35].

D-dimer is derived from the degradation of cross-linked fibrin clots by plasmin and is the simplest fibrin-specific degradation product. An increased concentration indicates the presence of hypercoagulation and secondary hyperfibrinolysis in the body[36]. Fibrinogen is a large molecular protein synthesized by the liver that has coagulation functions. Thrombin is then transformed into fibrin, which directly participates in coagulation. An increased fibrinogen concentration indicates that the blood is in a hypercoagulable state with increased viscosity[15]. These 2 markers are easy to measure, repeatable, and change significantly, which makes them convenient for evaluating the risk status of patients. This study found that both markers were associated with patient prognosis, and that their combination had a stronger correlation with prognosis.

We first defined patients in whom D-dimer and fibrinogen concentrations were below the upper limit of normal as the non-hypercoagulation group and those in whom D-dimer or fibrinogen concentrations were above the upper limit of normal as the hypercoagulation group[24]. Results of analysis revealed a correlation with prognosis. Subsequently, we refined the grouping by separating patients in the hypercoagulation group with a single marker above the upper limit of normal from those with both markers above the upper limit of normal. This grouping highlights the effect of coagulation marker concentrations on patient prognosis. The difference in prognosis between the different groups was more significant; more specifically, the prognosis of patients with both markers above the upper limit of normal was the worst, and the prognosis of the non-hypercoagulation group was the best. This strongly suggests that normal D-dimer and fibrinogen concentrations are associated with a favorable prognosis and, moreover, that hypercoagulation may serve as an excellent new prognostic indicator. We also assessed DFS to determine the correlation between hypercoagulation and LAGC progression. Consistent with the OS results, when the two coagulation markers were combined, the correlation with DFS was more convincing. This may identify an abnormal coagulation system that induces tumor progression[14].

Chemotherapy and immunotherapy have a profound impact on the coagulation system, including the components of blood cells and concentrations of various coagulation factors[30]. One study reported that platelet count decreased and activated partial thromboplastin time was significantly shortened after chemotherapy, especially among patients who later developed venous thromboembolism[37]. Studies have shown that neoadjuvant chemotherapy can activate the coagulation system. Specifically, chemotherapeutic agents, including cisplatin and fluorouracil, can induce thrombogenic effects through various mechanisms, including the secretion of immunomodulatory and pro-angiogenic cytokines by tumor cells, increased expression of tissue factors in vascular endothelial cells, direct vascular endothelial toxicity, and a reduction in protein C. In one study, upregulation of the procoagulant response was observed even after 1 week of combined chemoradiation therapy, associated with increased concentrations of FVIII: C, F1+2, and thrombin-antithrombin complex, as well as a corresponding decrease in protein C[38]. Immune checkpoint inhibitors (ICIs) activate the immune system by blocking inhibitory pathways, but they may also cause immune-related adverse events, including thrombotic events, such as deep vein thrombosis or pulmonary embolism[39]. Studies have shown that ICIs and BRAF/MEK inhibitors can induce a procoagulant state, which is characterized by increased expression and release of procoagulant factors, such as factor VIII and von Willebrand factor, along with reduced activity of coagulation-inhibiting factors such as protein S. This shift toward a procoagulant state can lead to hemostatic imbalances and an increased risk for thromboembolic events in patients undergoing these therapies[40].

Preoperative coagulation test results reflect patient coagulation status at the time of surgery and tolerance to invasive surgeries. In addition, the extent to which invasive surgery induces the release of inflammatory cytokines, suppresses tumor immunity, and other factors that have adverse effects on tumor progression may be more pronounced in patients with high coagulation risk before surgery.

Our study may provide insights into existing clinical issues. It may be possible to improve the prognosis of patients with LAGC and hypercoagulation after NICT by adjusting patient management and treatment. First, preventive treatments can be administered during the preoperative period to reduce the risk for coagulation. Thrombosis is caused by the complex interaction of various factors, such as tissue factors, coagulation abnormalities, activated platelet activation, activated adhesion activation, and endothelial cell dysfunction. During NICT, these factors work together to induce a hypercoagulable state. Hypercoagulation can be prevented by dietary adjustment, appropriate exercise, fluid therapy, and early anticoagulation, thereby reducing the negative effects of NICT on the coagulation system. Second, the poor prognosis of patients with high coagulation risk after surgery is noteworthy. For patients in a hypercoagulable state, management should be strengthened, monitoring should be performed, and measures to prevent postoperative complications should be implemented. In addition, due to the high recurrence rate in patients with hypercoagulation, a prediction model based on coagulation function assessment can be established to measure patient recurrence probability, thereby personalizing patient postoperative monitoring and timely assessment of tumor progression[41,42].

The present study had several limitations, the first of which were its retrospective, single-center design and limited sample size. Future studies with larger multicenter cohorts are warranted to enhance the generalizability of our findings. Second, the short follow-up period may have introduced some bias in the prognosis analysis. Nevertheless, even with a short follow-up, these data revealed the effects of NICT in patients with LAGC. Third, patients with hypercoagulation are usually treated with anticoagulation therapy, which may improve their prognosis. However, further research is required to verify these hypotheses.

CONCLUSION

In conclusion, this study identifies hypercoagulation after NICT as a strong, independent prognostic indicator in LAGC patients undergoing radical gastrectomy. It highlights the need for coagulation monitoring and suggests that early detection could guide personalized treatment strategies to improve outcomes. Further validation in larger studies is warranted.

ACKNOWLEDGEMENTS

The authors thank the patients who agreed to participate in this study and all researchers, pathologists, and clinical staff who supported this study.