Prognostic significance of preoperative C-reactive protein-triglyceride-glucose index in long-term outcomes after radical gastrectomy for gastric cancer

Abstract

BACKGROUND

Gastric cancer, a globally prevalent malignant tumor, continues to exhibit high incidence and mortality rates. Although radical gastrectomy remains the primary treatment for this disease, postoperative complications frequently arise, negatively impacting short-term recovery and significantly reducing patients’ quality of life. In this context, accurately predicting the risk of postoperative recurrence and metastasis, coupled with targeted interventions, could substantially improve patient outcomes. The C-reactive protein-triglyceride-glucose index (CTI), a composite biomarker that integrates metabolic disturbances and systemic inflammation, has garnered increasing attention in oncology. The prognostic nutritional index (PNI), a composite measure based on serum albumin and peripheral blood lymphocyte count, is used to evaluate both the nutritional status and systemic immune function of patients. In recent years, both the CTI and PNI have demonstrated significant prognostic value in predicting tumor outcomes, assessing treatment responses, and formulating personalized treatment strategies.

AIM

To investigate whether the combined inflammation and insulin resistance marker, the CTI, can serve as a prognostic indicator for patients undergoing radical gastrectomy for gastric cancer. Additionally, it seeks to develop a predictive model by incorporating the PNI alongside CTI.

METHODS

This retrospective study included a total of 300 patients who underwent radical gastrectomy. The patients were classified into high and low CTI groups based on their CTI index. Cox proportional hazards regression analysis was performed to identify independent prognostic factors influencing overall survival (OS) and disease-free survival (DFS), and two nomogram models were developed.

RESULTS

Of the included patients, 131 had a high CTI and 169 had a low CTI. The DFS period of the low CTI group was significantly longer than that of the high CTI group. The number of postoperative adjuvant treatments, T stage, N stage, CTI, and PNI were identified as independent prognostic factors for DFS. The hazard ratio for CTI was 2.07 (95% confidence interval: 1.36-3.17, P < 0.001). In terms of OS, the OS period of the low CTI group was significantly longer than that of the high CTI group. Whether adjuvant treatment was administered, T stage, CTI, and PNI were independent prognostic factors for OS. The hazard ratio for CTI was 2.47 (95% confidence interval: 1.44-4.23, P = 0.001). The nomogram models for OS and DFS further emphasized the importance of CTI as a key predictor of patient prognosis.

CONCLUSION

CTI is a long-term prognostic indicator for the outcome of radical gastrectomy for gastric cancer. Patients with lower CTI values have a better prognosis. The prediction models constructed by combining CTI with PNI has good predictive ability for DFS and OS after radical gastrectomy.

Key Words: C-reactive protein triglyceride glucose index; Prognostic nutritional index; Gastric cancer; Prognostic model; Radical gastrectomy

Core Tip: Gastric cancer is the third leading cause of cancer-related mortality worldwide. Radical surgery remains the primary treatment for early-stage gastric cancer, but the overall survival rate post-surgery remains relatively low. Identifying key prognostic factors is essential to optimize treatment strategies. The C-reactive protein-triglyceride-glucose (CTI) index, which integrates both inflammatory and metabolic markers, has garnered increasing attention in cancer research, particularly in gastric cancer prognosis. However, few studies have explored the impact of CTI on cancer patient outcomes. Therefore, we conducted a retrospective clinical study to examine the influence of CTI on the prognosis of patients undergoing radical gastrectomy for gastric cancer and developed a corresponding predictive model. Additionally, we incorporated the prognostic nutritional index coefficient to improve the predictive accuracy of the model.

INTRODUCTION

Gastric cancer (GC) is one of the most prevalent malignant tumors of the digestive system, ranking as the fifth most common cancer globally. It has a mortality rate that ranks third among all cancers worldwide, with more than 1 million new cases diagnosed each year[1,2]. In Asia, particularly in China and Japan, the incidence rate is notably higher than in other regions. In China, both the incidence and mortality rates of GC rank fifth, leading to substantial economic and social losses[3,4]. The primary treatment modalities for GC include surgical resection, radiotherapy, and chemotherapy. Radical surgery remains the cornerstone of treatment for early-stage GC (stages 1-3). The standard surgical approaches for radical gastrectomy include distal subtotal gastrectomy, proximal subtotal gastrectomy, and total gastrectomy. Postoperative digestive tract reconstruction techniques consist of Billroth I reconstruction, Billroth II reconstruction, Billroth II with Braun reconstruction, and Roux-en-Y reconstruction[5,6]. The selection of the surgical method is primarily determined by the tumor’s location and size[7].

The development and progression of GC are influenced by multiple factors, including family history, inflammation, metabolic disturbances, and infections. Chronic inflammation plays a significant role in the pathogenesis of GC. Factors such as Helicobacter pylori infection and chronic gastritis can induce persistent inflammation of the gastric mucosa, thereby facilitating the progression of GC[8]. During the inflammatory process, inflammatory cells release a variety of angiogenic factors, including vascular endothelial growth factor, which promotes the formation of tumor-associated blood vessels, providing essential nutrients to support tumor growth and enhancing tumor cell survival[9]. C-reactive protein (CRP), a classic acute-phase response protein, is closely linked to the activation of liver cells and macrophages during inflammation[10]. Therefore, CRP has become a critical marker for assessing the degree of systemic inflammation. In healthy individuals, CRP levels are typically low, but upon inflammatory stimulation, its concentration can increase rapidly, potentially rising by more than tenfold within 48-72 hours[11]. Numerous studies have demonstrated that elevated CRP levels are a significant indicator of the inflammatory status in patients with malignancies. Compared to healthy individuals, patients with various cancers exhibit markedly higher CRP concentrations[12]. Research by Zheng et al[13] and Mao et al[14] has shown that in GC patients, CRP not only reflects disease activity but also serves as an independent prognostic factor. CRP levels are significantly correlated with overall survival (OS) after radical surgery. Furthermore, CRP can be combined with other indicators, such as the CRP-albumin ratio, lymphocyte-CRP ratio, and the CRP-albumin-lymphocyte ratio (CALLY index), to collectively predict the prognosis of cancer patients[15,16]. In terms of metabolic factors, insulin resistance has been extensively studied as a key factor influencing the onset, progression, and prognosis of GC[17]. The high insulin-euglycemic clamp is internationally recognized as the gold standard for diagnosing insulin resistance. The triglyceride-glucose (TyG) index, a non-invasive marker for insulin resistance, is calculated based on fasting triglyceride and glucose levels and has been linked to the occurrence of various tumors, including gastric and colorectal cancers[18,19]. Furthermore, a higher TyG index is associated with tumor progression, recurrence, and shorter survival[20]. The CRP-TyG index (CTI), which integrates both inflammatory and metabolic factors, has been widely studied as a predictive marker in cardiovascular and endocrine disorders[21,22]. Zhao’s research highlighted that a higher CTI is positively correlated with cancer mortality risk in the general population, and this study partially validated its prognostic value[23].

The prognostic nutritional index (PNI) is calculated by combining serum albumin levels and peripheral blood lymphocyte counts. It not only reflects the nutritional status of the body but also provides an assessment of immune function. PNI has demonstrated a strong ability to evaluate both the nutritional and immune status of surgical patients. Impairment of the immune system and nutritional imbalance are both associated with poor prognosis. Previous studies have confirmed that PNI is also effective in predicting the prognosis of patients with various malignancies, such as lung cancer and esophageal cancer[24,25]. Therefore, this study aims to investigate whether the CTI and PNI serve as independent prognostic factors in patients undergoing radical gastrectomy for GC and to develop nomograms for their practical clinical application.

MATERIALS AND METHODS

Study design and patients

A retrospective analysis was conducted on the clinical data of patients who underwent radical gastrectomy at the Affiliated Hospital of Xuzhou Medical University from June 2018 to July 2021. Patients meeting the following criteria were included in the study: (1) Patients who underwent radical gastrectomy for GC; (2) Postoperative pathology was primary GC; (3) No preoperative adjuvant therapy; (4) No distant metastasis (M0); and (5) Complete follow-up data. Patients with the following characteristics were excluded from the study: (1) Complicated with malignant tumors in other organs or systems; (2) Received preoperative adjuvant therapy; (3) Had distant metastasis (M1); (4) Had clear pathogen infection [the presence of bacterial, viral, and fungal infections in patients can be determined through a combination of clinical symptoms (such as fever, cough, and dermatological manifestations), laboratory tests (including respiratory pathogen detection, urine analysis, and stool examination), and imaging studies (such as chest computed tomography and abdominal computed tomography scans)]; and (5) Incomplete follow-up data. The study process is shown in Figure 1. This retrospective study was approved by the Ethics Review Committee of the Affiliated Hospital of Xuzhou Medical University (No. XYFY2023-KL227-01) and was conducted in accordance with the principles outlined in the Declaration of Helsinki. Given the retrospective design of the study, written informed consent was not required.

Figure 1 Flow chart illustrates the screening situation of patients with gastric cancer who have undergone radical gastrectomy. CTI: C-reactive protein-triglyceride-glucose index; PNI: Prognostic nutritional index.

The routine clinical data extracted from the electronic medical record system included baseline characteristics (such as age, gender, and body mass index), laboratory parameters [such as lymphocyte count, fasting blood glucose, triglycerides, serum albumin, carcinoembryonic antigen, and carbohydrate antigen 199 (CA199)], and tumor characteristics. Quality assessment and risk of bias evaluation were independently performed by two researchers who were blinded to the other data.

Definition

Use the following formula to calculate CTI: CTI = 0.412 × ln [CRP (mg/L)] + ln [fasting triglycerides (mg/dL) × fasting blood glucose (mg/dL)/2], and the calculation formula for PNI is: PNI = [serum albumin (g/L) + 5] × lymphocyte count (10⁹/L).

After an overnight fast, fasting glucose, triglyceride, and serum albumin levels were measured using the enzymatic assay method on an automatic biochemical analyzer (Olympus, AU 2700). Lymphocyte count and CRP levels were determined using an automatic blood cell analyzer (Mindray, BC-5390CRP), following standard laboratory protocols to ensure accuracy and consistency across patients. In this study, all hematological measurements were performed by certified laboratory personnel, and standardized operating procedures were followed for each measurement batch. Additionally, laboratory equipment was regularly calibrated, and any abnormal laboratory results were retested to ensure consistency.

The preoperative CTI values of patients undergoing radical gastrectomy for GC were utilized to construct the receiver operating characteristic curve. Diagnostic performance was assessed by calculating the area under the curve (AUC). The optimal cut-off value for the CTI index was determined by identifying the point that maximized the product of specificity and sensitivity. Based on this cut-off value, patients were stratified into high and low CTI index groups. A total of 300 patients were categorized into a high CTI group (≥ 9.07) and a low CTI group (< 9.07).

Evaluation

The tumor treatment status was evaluated using the Response Evaluation Criteria in Solid Tumors version 1.1. Preoperative and postoperative computed tomography examinations were conducted to determine the baseline values. Subsequently, abdominal computed tomography assessments were performed every 3 months to monitor the disease condition. Postoperative follow-up was conducted through inpatient stays, outpatient visits, or telephone follow-ups to understand the patient’s postoperative condition and survival time. All patients were followed until death or until June 30, 2025, to determine OS and disease-free survival (DFS). To maintain consistency, two independent radiologists with over 5 years of experience in tumor imaging independently assessed all imaging data. In cases of disagreement, a third radiologist reviewed the images to achieve a consensus.

Statistical analysis

Statistical analysis was performed using R software (version 4.5.0). Prior to hypothesis testing, the Kolmogorov-Smirnov test was applied to assess the normality of continuous data distributions. Continuous variables are presented as mean ± SD, while categorical variables are expressed as counts and corresponding percentages. Fisher’s exact test, χ² test, and the Kaplan-Meier method were employed to estimate the median OS and DFS. Univariate and multivariate analyses were conducted using appropriate statistical methods. Multivariate analysis was performed using Cox proportional hazards regression, with a focus on variables exhibiting P < 0.05 to identify independent prognostic factors for OS and DFS. Based on these independent prognostic factors, two nomogram prediction models were developed to estimate patient outcomes at 1, 3, and 5 years. Variables were selected based on both statistical significance and clinical relevance, and each was weighted according to its hazard ratio (HR). Internal validation of the models was performed using a bootstrap method with 1000 resamples to assess predictive accuracy and minimize overfitting. All statistical tests were two-sided, with a P value < 0.05 considered statistically significant.

RESULTS

Patient characteristics

This retrospective study analyzed 300 GC patients (Figure 1) who underwent radical gastrectomy between June 2018 and July 2021. Among these patients, 169 were classified into the low CTI (< 9.07) group, and 131 into the high CTI (≥ 9.07) group. Table 1 presents the baseline characteristics of the low and high CTI groups, including clinical and pathological features such as age, gender, body mass index, receipt of postoperative adjuvant therapy, tumor size, T stage, N stage, lymph node metastasis, vascular invasion, and neural invasion. For patients who experience progression or recurrence after the surgery, only the number of times of adjuvant therapy before the occurrence of progression or recurrence should be recorded. Additionally, various laboratory parameters, including carcinoembryonic antigen and CA199, were recorded. All blood parameters were measured following a standardized clinical protocol after an overnight fast (≥ 8 hours). Significant differences were observed between the two groups in body mass index and tumor size (P < 0.05).

Table 1 Baseline characteristics of C-reactive protein-triglyceride-glucose index < 9.07 and C-reactive protein-triglyceride-glucose index ≥ 9.07, mean ± SD/n (%).

Variables	Total (n = 300)	CTI < 9.07 (n = 169)	CTI ≥ 9.07 (n = 131)	Statistic	P value
Size (cm)	4.85 ± 3.05	4.52 ± 3.03	5.28 ± 3.04	t = -2.15	0.032
BMI	23.15 ± 3.79	22.71 ± 3.83	23.72 ± 3.68	t = -2.29	0.023
Gender				χ² = 0.23	0.635
Female	89 (29.67)	52 (30.77)	37 (28.24)
Male	211 (70.33)	117 (69.23)	94 (71.76)
Age				χ² = 0.71	0.401
< 50	118 (39.33)	70 (41.42)	48 (36.64)
≥ 50	182 (60.67)	99 (58.58)	83 (63.36)
Adjuvant therapy				χ² = 0.94	0.332
No	72 (24.00)	37 (21.89)	35 (26.72)
Yes	228 (76.00)	132 (78.11)	96 (73.28)
T stage				χ² = 2.96	0.061
T1	50 (16.67)	37 (21.89)	13 (9.92)
T2	49 (16.33)	29 (17.16)	20 (15.27)
T3	129 (43.00)	71 (42.01)	58 (44.27)
T4	72 (24.00)	32 (18.93)	40 (30.53)
N stage				χ² = 1.20	0.273
N0	100 (33.33)	65 (38.46)	35 (26.72)
N1	54 (18.00)	34 (20.12)	20 (15.27)
N2	50 (16.67)	28 (16.57)	22 (16.79)
N3	96 (32.00)	42 (24.85)	54 (41.22)
LN meta				χ² = 3.67	0.055
No	105 (35.00)	67 (39.64)	38 (29.01)
Yes	195 (65.00)	102 (60.36)	93 (70.99)
Vein inv				χ² = 1.40	0.236
No	126 (42.00)	76 (44.97)	50 (38.17)
Yes	174 (58.00)	93 (55.03)	81 (61.83)
Nerve inv				χ² = 0.78	0.376
No	130 (43.33)	77 (45.56)	53 (40.46)
Yes	170 (56.67)	92 (54.44)	78 (59.54)
CEA				χ² = 0.03	0.857
< 3	170 (56.67)	95 (56.21)	75 (57.25)
≥ 3	130 (43.33)	74 (43.79)	56 (42.75)
CA199				χ² = 2.48	0.115
< 37	256 (85.33)	149 (88.17)	107 (81.68)
≥ 37	44 (14.67)	20 (11.83)	24 (18.32)

CTI: C-reactive protein-triglyceride-glucose index; T: Tumor; N: Node; BMI: Body mass index; CEA: Carcinoembryonic antigen; CA199: Carbohydrate antigen 199; LN meta: Lymph node metastasis; Vein inv: Vein invasion; Nerve inv: Nerve invasion.

DFS and OS

Due to the fact that only 84 GC patients in this study had died postoperatively, which is less than half of the total cohort, insufficient data were available to calculate the median survival time. Kaplan-Meier analysis revealed that, compared to the high CTI group, the DFS of the low CTI group was significantly prolonged [P < 0.001, HR = 2.743, 95% confidence interval (CI): 1.938-3.882, Figure 2A]. The OS of the low CTI group was also significantly higher than that of the high CTI group (P < 0.001, HR = 3.494, 95%CI: 2.199-5.553, Figure 2B).

Figure 2 Effects of different C-reactive protein-triglyceride-glucose index on the long-term prognosis of gastric cancer patients. A: Kaplan-Meier plot of disease-free survival in the C-reactive protein-triglyceride-glucose index (CTI) < 9.07 and CTI ≥ 9.07 groups; B: Kaplan-Meier plot of overall survival in the CTI < 9.07 and CTI ≥ 9.07 groups. CTI: C-reactive protein triglyceride-glucose index; HR: Hazard ratio; CI: Confidence interval; DFS: Disease-free survival; OS: Overall survival.

Univariate and multifactorial analyses of DFS and OS

Univariate analysis indicated that the number of postoperative adjuvant treatments, T stage, N stage, presence or absence of lymph node metastasis, CTI, and PNI were significantly associated with DFS outcomes (P < 0.05) (Table 2). In terms of OS, whether adjuvant treatment was administered, T stage, N stage, presence or absence of lymph node metastasis, CTI, PNI, and CA199 were statistically significant (P < 0.05) (Table 3). Multivariate analysis revealed that the number of postoperative adjuvant treatments, T stage, N stage, CTI, and PNI were independent prognostic factors for DFS. For OS, the independent prognostic factors included whether adjuvant treatment was administered, T stage, CTI, and PNI. Based on these identified independent prognostic factors, nomogram prediction models were developed for 1-year, 3-year, and 5-year survival predictions for DFS and OS (Figure 3). For DFS, the concordance index (C-index) was 0.768 for the training dataset (95%CI: 0.719-0.818) and 0.778 for the validation dataset (95%CI: 0.713-0.844), indicating good predictive accuracy. For OS, the C-index of the training dataset was 0.810 (95%CI: 0.750-0.869), and the C-index of the validation dataset was 0.791 (95%CI: 0.691-0.890), also indicating good predictive accuracy.

Figure 3 Graph depicting the prognostic model for predicting 1-, 3-, and 5-year. A: Disease-free survival; B: Overall survival. AT: Adjuvant therapy; T: Tumor; N: Node; CTI: C-reactive protein-triglyceride-glucose index; PNI: Prognostic nutritional index; DFS: Disease-free survival; OS: Overall survival.

Table 2 Univariate and multivariate analyses of prognostic factors for disease-free survival.

Variables	Univariate		Multivariate
	P value	HR (95%CI)	P value	HR (95%CI)
Gender
Female		1.00 (Ref.)
Male	0.956	1.01 (0.64-1.59)
Age
< 50		1.00 (Ref.)
≥ 50	0.592	0.89 (0.59-1.35)
AT times
0		1.00 (Ref.)		1.00 (Ref.)
1-3	0.011	2.06 (1.18-3.60)	0.779	0.91 (0.48-1.72)
4	0.022	0.56 (0.34-0.92)	< 0.001	0.17 (0.09-0.31)
T stage
T1		1.00 (Ref.)		1.00 (Ref.)
T2	0.057	2.59 (0.97-6.91)	0.110	2.28 (0.83-6.28)
T3	< 0.001	4.24 (1.81-9.89)	0.001	4.78 (1.88-12.20)
T4	< 0.001	5.16 (2.13-12.52)	0.004	4.62 (1.65-12.94)
N stage
N0		1.00 (Ref.)		1.00 (Ref.)
N1	0.356	1.36 (0.71-2.60)	0.112	1.79 (0.87-3.66)
N2	0.002	2.59 (1.42-4.72)	< 0.001	3.10 (1.62-5.93)
N3	< 0.001	2.76 (1.61-4.73)	< 0.001	3.49 (1.76-6.92)
LN meta
No		1.00 (Ref.)
Yes	0.002	2.06 (1.30-3.27)
CTI
< 9.07		1.00 (Ref.)		1.00 (Ref.)
≥ 9.07	< 0.001	2.46 (1.63-3.71)	< 0.001	2.07 (1.36-3.17)
PNI	< 0.001	0.93 (0.90-0.97)	0.044	0.96 (0.92-0.99)
CEA	0.937	1.00 (0.99-1.01)
CA199	0.040	1.01 (1.01-1.01)

HR: Hazards ratio; CI: Confidence interval; AT: Adjuvant therapy; T: Tumor; N: Node; CTI: C-reactive protein-triglyceride-glucose index; PNI: Prognostic nutritional index; CEA: Carcinoembryonic antigen; CA199: Carbohydrate antigen 199; LN meta: Lymph node metastasis.

Table 3 Univariate and multivariate analyses of prognostic factors for overall survival.

Variables	Univariate		Multivariate
	P value	HR (95%CI)	P value	HR (95%CI)
Gender
Female		1.00 (Ref.)
Male	0.667	1.14 (0.63-2.04)
Age
< 50		1.00 (Ref.)
≥ 50	0.346	1.31 (0.75-2.27)
AT
No		1.00 (Ref.)		1.00 (Ref.)
Yes	0.015	0.51 (0.29-0.88)	< 0.001	0.15 (0.08-0.30)
T stage
T1		1.00 (Ref.)
T2	0.016	4.84 (1.35-17.36)
T3	0.017	4.24 (1.29-13.96)
T4	0.006	5.53 (1.62-18.88)
N stage
N0		1.00 (Ref.)		1.00 (Ref.)
N1	0.028	2.89 (1.12-7.46)	0.001	5.87 (2.02-17.06)
N2	0.003	4.09 (1.63-10.26)	< 0.001	6.53 (2.48-17.18)
N3	< 0.001	5.83 (2.55-13.36)	< 0.001	12.18 (4.65-31.91)
LN meta
No		1.00 (Ref.)
Yes	< 0.001	3.20 (1.62-6.32)
CTI
< 9.07		1.00 (Ref.)		1.00 (Ref.)
≥ 9.07	< 0.001	2.77 (1.63-4.73)	0.001	2.47 (1.44-4.23)
PNI	< 0.001	0.90 (0.86-0.94)	< 0.001	0.91 (0.87-0.95)
CEA	0.884	1.00 (0.99-1.01)
CA199	0.004	1.01 (1.01-1.01)

HR: Hazards ratio; CI: Confidence interval; AT: Adjuvant therapy; T: Tumor; N: Node; LN meta: Lymph node metastasis; CTI: C-reactive protein-triglyceride-glucose index; PNI: Prognostic nutritional index; CEA: Carcinoembryonic antigen; CA199: Carbohydrate antigen 199.

Validation of the prognostic model

The patients were divided into two groups in a 7:3 ratio, with 210 patients assigned to the training dataset and 90 patients assigned to the internal validation set. Table 4 presents the baseline characteristics of both datasets. As depicted in Figure 4A, the AUC for 1-year DFS was 0.828 for the training set and 0.837 for the validation set. At 3 years, the AUC for DFS was 0.816 in the training set and 0.857 in the validation set (Figure 4B). At 5 years, the corresponding AUC values were 0.811 and 0.832, respectively (Figure 4C). The calibration curves for DFS at 1, 3, and 5 years for both the training and validation sets are presented in Figure 5A-F. For OS, the 1-year AUC was 0.946 for the training set and 0.846 for the validation set (Figure 4D). At 3 years, the AUC for OS was 0.841 for the training set and 0.845 for the validation set (Figure 4E), while at 5 years, the AUC values were 0.838 and 0.795, respectively (Figure 4F). The calibration curves for OS at 1, 3, and 5 years for both the training and validation sets are shown in Figure 5G-L. These results demonstrate that the predicted risk values from the model closely correspond to the actual observed outcomes.

Figure 4 Graph depicting the operating characteristic evaluation plot for prognostic models. A: Graph showing the training set and validation receiver operating characteristic (ROC) evaluation plots for 1-year disease-free survival (DFS) prognostic prediction model; B: Graph showing the training set and validation set ROC evaluation plots for 3-year DFS prognostic prediction model; C: Graph showing the training set and validation set ROC evaluation plots for 5-year DFS prognostic prediction model; D: Graph showing the training set and validation set ROC evaluation plots for 1-year overall survival (OS) prognostic prediction model; E: Graph showing the training set and validation set ROC evaluation plots for 3-year OS prognostic prediction model; F: Graph showing the training set and validation set ROC evaluation plots for 5-year OS prognostic prediction model. AUC: Area under the curve; CI: Confidence interval.

Figure 5 Graph illustrating the calibration plots for a prognostic model. A: Calibration plots for the training set 1-year disease-free survival (DFS); B: Calibration plots for the validation set 1-year DFS; C: Calibration plots for the training set 3-year DFS; D: Calibration plots for the validation set 3-year DFS; E: Calibration plots for the training set 5-year DFS; F: Calibration plots for the validation set 5-year DFS; G: Calibration plots for the training set 1-year overall survival (OS); H: Calibration plots for the validation set 1-year OS; I: Calibration plots for the training set 3-year OS; J: Calibration plots for the validation set 3-year OS; K: Calibration plots for the training set 5-year OS; L: Calibration plots for the validation set 5-year OS.

Table 4 Comparison of features between the training and validation sets, n (%).

Variables	Total (n = 300)	Validation (n = 90)	Training (n = 210)	Statistic	P value
Gender				χ² = 0.22	0.639
Female	89 (29.67)	25 (27.78)	64 (30.48)
Male	211 (70.33)	65 (72.22)	146 (69.52)
Age				χ² = 0.77	0.381
< 50	118 (39.33)	32 (35.56)	86 (40.95)
≥ 50	182 (60.67)	58 (64.44)	124 (59.05)
BMI				χ² = 0.26	0.614
< 25	216 (72.00)	63 (70.00)	153 (72.86)
≥ 25	84 (28.00)	27 (30.00)	57 (27.14)
Adjuvant therapy				χ² = 1.13	0.288
No	72 (24.00)	18 (20.00)	54 (25.71)
Yes	228 (76.00)	72 (80.00)	156 (74.29)
T stage				χ² = 1.52	0.678
T1	50 (16.67)	14 (15.56)	36 (17.14)
T2	49 (16.33)	16 (17.78)	33 (15.71)
T3	129 (43.00)	42 (46.67)	87 (41.43)
T4	72 (24.00)	18 (20.00)	54 (25.71)
N stage				χ² = 2.46	0.483
N0	100 (33.33)	28 (31.11)	72 (34.29)
N1	54 (18.00)	19 (21.11)	35 (16.67)
N2	50 (16.67)	18 (20.00)	32 (15.24)
N3	96 (32.00)	25 (27.78)	71 (33.81)
LN meta				χ² = 0.85	0.355
No	105 (35.00)	28 (31.11)	77 (36.67)
Yes	195 (65.00)	62 (68.89)	133 (63.33)
Vein inv				χ² = 0.21	0.646
No	126 (42.00)	36 (40.00)	90 (42.86)
Yes	174 (58.00)	54 (60.00)	120 (57.14)
Nerve inv				χ² = 1.03	0.309
No	130 (43.33)	43 (47.78)	87 (41.43)
Yes	170 (56.67)	47 (52.22)	123 (58.57)
CTI				χ² = 0.11	0.741
< 9.07	169 (56.33)	52 (57.78)	117 (55.71)
≥ 9.07	131 (43.67)	38 (42.22)	93 (44.29)
PNI				χ² = 0.25	0.616
< 45	86 (28.67)	24 (26.67)	62 (29.52)
≥ 45	214 (71.33)	66 (73.33)	148 (70.48)
CEA				χ² = 0.58	0.446
< 3	170 (56.67)	54 (60.00)	116 (55.24)
≥ 3	130 (43.33)	36 (40.00)	94 (44.76)
CA199				χ² = 2.24	0.135
< 37	256 (85.33)	81 (90.00)	175 (83.33)
≥ 37	44 (14.67)	9 (10.00)	35 (16.67)

CTI: C-reactive protein-triglyceride-glucose index; T: Tumor; N: Node; BMI: Body mass index; PNI: Prognostic nutritional index; CEA: Carcinoembryonic antigen; CA199: Carbohydrate antigen 199; LN meta: Lymph node metastasis; Vein inv: Vein invasion; Nerve inv: Nerve invasion.

DISCUSSION

In this study, we conducted a comprehensive investigation into the roles of the CTI and the PNI in the prognosis of patients following radical gastrectomy for GC. Our findings indicate that both CTI and PNI are independent prognostic factors that influence DFS and OS in patients with GC. Notably, higher levels of CTI are significantly associated with shorter DFS and OS. Postoperative adjuvant therapy, T stage, and N stage were also identified as key factors impacting both DFS and OS. These results provide valuable insights for clinical prognosis assessment and further underscore the prognostic significance of CTI and PNI in GC. The independent prognostic factors for DFS include the number of postoperative adjuvant treatments, T stage, N stage, CTI, and PNI. The independent prognostic factors for OS include whether adjuvant treatment was administered, T stage, CTI, and PNI.

We observed that a higher CTI was significantly associated with shorter DFS and OS, which is consistent with previous studies. As a composite marker, CTI integrates the levels of inflammation and insulin resistance. Specifically, CTI combines three routine biomarkers: CRP, triglycerides, and glucose, which reflect the inflammatory response, metabolic disorders, and insulin resistance status of patients, respectively. A high CTI generally indicates the presence of a relatively severe inflammatory response and metabolic abnormalities, both of which have been shown to be closely related to the onset, progression, and prognosis of various tumors[23]. Furthermore, these factors play a crucial role in the progression and drug resistance of GC. Studies have demonstrated that CRP, as an acute-phase response protein, is involved in immune evasion and angiogenesis within the tumor microenvironment, thereby promoting tumor growth and metastasis[11]. A previous meta-analysis including 2597 patients with GC revealed that elevated preoperative CRP levels (≥ 10 mg/L) were significantly associated with poorer OS, with a HR of 1.77 (95%CI: 1.56-2.00)[26]. Chronic inflammation facilitates tumor initiation and progression through various mechanisms, including inducing DNA damage, promoting tumor cell proliferation, inhibiting apoptosis, stimulating angiogenesis, and enabling immune evasion. The tumor microenvironment contains various immune cells, such as tumor-associated macrophages, regulatory T cells, and myeloid-derived suppressor cells, which promote tumor growth, metastasis, and drug resistance by secreting cytokines and growth factors[27,28].

The role of insulin resistance in the onset, progression, and prognosis of tumors has become a key focus in recent tumor metabolism research[17]. Insulin resistance is not only a hallmark of type 2 diabetes and metabolic syndrome but also influences tumor cell proliferation, metabolism, metastasis, and drug resistance through multiple molecular pathways. A previous systematic review and meta-analysis revealed that cancer patients exhibited significantly reduced insulin sensitivity, with an average glucose disposal rate of 4.7 mg/kg/minute in cancer patients compared to 7.5 mg/kg/minute in healthy control groups, indicating the high prevalence of insulin resistance among cancer patients[29]. The TyG index, an indirect marker of insulin resistance, has garnered considerable attention in GC research in recent years. A cohort study on health examinations found that elevated TyG index was significantly associated with gastric precancerous lesions, such as atrophic gastritis and intestinal metaplasia, as well as the development of GC. Specifically, the higher the quartile of the TyG index, the greater the risk of GC, suggesting that the TyG index may serve as a predictive marker for GC occurrence[20,30]. In patients with advanced GC receiving combined immunotherapy and chemotherapy, those with higher TyG indices demonstrated better objective response rates and disease control rates, indicating that the TyG index may enhance the patient’s response to immunotherapy by improving metabolic and immune status[31].

The PNI is a comprehensive indicator that integrates both nutritional and immune status by combining serum albumin concentration and peripheral blood lymphocyte count. PNI has been shown to correlate with survival outcomes in patients with various malignancies, particularly in GC patients, where a lower PNI is significantly associated with poorer prognosis[25,32]. In this study, we found that patients with lower PNI values had a worse prognosis in both DFS and OS. A low PNI suggests that the patient may have poor nutritional status and compromised immune function, which can impair the anti-tumor immune response and hinder postoperative recovery[33]. Chronic inflammation and diminished immune function not only promote tumor cell growth, metastasis, and immune evasion but may also contribute to postoperative infections and complications, thus shortening the patient’s survival time.

CTI reflects metabolic dysfunction and inflammation, which may impact cancer prognosis by promoting inflammation, altering immune responses, and influencing the tumor microenvironment. The TyG index within the CTI reflects insulin resistance and metabolic dysfunction, factors that can impair the body’s ability to fight cancer. PNI is commonly used to assess a patient’s nutritional status and immune function. A low PNI typically indicates malnutrition and impaired immune function, both of which are associated with poorer prognosis in cancer patients. Furthermore, nutritional deficiencies further compromise the immune system, exacerbating the cancer prognosis. Patients with a low PNI often experience malnutrition, which may increase the risk of infection, delay wound healing, and affect chemotherapy tolerance, leading to worse outcomes. CTI and PNI are critical biomarkers for evaluating immune response, metabolic status, and nutritional condition in cancer patients. These factors play a pivotal role in determining cancer prognosis.

Postoperative adjuvant therapy plays a crucial role in the prognosis of GC patients. By improving OS, reducing local recurrence rates, and enhancing quality of life, it has become an essential component of the comprehensive treatment plan for GC. In this study, the number of postoperative adjuvant treatments was found to be significantly correlated with DFS. Postoperative chemotherapy, radiotherapy, immunotherapy, and targeted therapy can effectively reduce the risk of recurrence and improve patient survival, particularly in high-risk patients, where the therapeutic effect is more pronounced[34,35]. Previous studies have demonstrated that chemoradiotherapy (CRT) can effectively reduce the local recurrence rate. CRT helps eliminate micro-metastases that may have been missed during surgery, thereby decreasing the risk of distant metastasis. For patients with a lower PNI, multiple postoperative adjuvant treatments may be especially important, as these patients often have compromised immune function and may require more aggressive treatment to prevent tumor recurrence[33]. We also observed that the presence or absence of adjuvant treatment is closely associated with OS, further highlighting the necessity of postoperative adjuvant therapy. Adjuvant treatment offers patients a better chance of survival by reducing the growth of microlesions or residual tumor cells[36,37].

Therefore, for patients with a high CTI in preoperative hematological tests, it is important to promptly control their lipid levels postoperatively and dynamically monitor their blood glucose levels. For patients with a low PNI, early postoperative nutritional support should be provided, including intravenous nutrition, enteral nutrition, or high-quality diet during the recovery phase. Furthermore, postoperative adjuvant therapy is essential. Patients should, if their physical condition permits, receive timely individualized postoperative adjuvant therapy to improve prognosis and extend survival.

To further improve the accuracy of prognostic assessment, we developed two prognostic models incorporating factors such as the CTI, PNI, T stage, N stage, and postoperative adjuvant therapy to predict the 1-year, 3-year, and 5-year DFS and OS in GC patients. To assess the performance of the nomogram prediction models, the C-index was calculated, and calibration curves were generated. In the internal validation set, the C-index values for DFS and OS were 0.778 and 0.791, respectively. These models provide quantitative and individualized tools for clinicians to accurately assess the postoperative prognosis of GC patients and formulate appropriate treatment plans. The application of these models can offer early intervention for high-risk patients and further enhance their survival outcomes. However, although the predictive performance of the nomogram was assessed through internal validation in this study, due to its nature as a single-center retrospective analysis, there are inherent data limitations, and external validation has not been conducted. In future studies, we plan to collect patient data from multiple centers or different national populations for external validation, in order to further evaluate the predictive efficacy of the model.

This study demonstrates that the CTI and the PNI are crucial for prognostic assessment following radical gastrectomy for GC. Higher CTI levels are significantly associated with poorer DFS and OS. Additionally, PNI, which reflects both nutritional and immune status, plays a significant role in the long-term prognosis of patients. By constructing a prognostic model, we have provided a valuable tool for clinical practice, aiding in the accurate prediction of postoperative recurrence and survival risk in GC patients. However, there are several limitations to this study. First, as a single-center retrospective study, although we minimized the impact of potential biases through appropriate statistical analysis, control of confounding factors, ensuring data integrity, and proper handling of missing data, factors such as selection bias, information bias, and recall bias remain inevitable. Therefore, these findings should be validated in future multicenter prospective studies. Secondly, the critical threshold for the CTI may differ depending on the population and disease characteristics, necessitating further investigation to determine whether this threshold is applicable to various tumor types. This could affect the broader applicability of our results. Finally, the clinical and pathological data collected in this study are relatively limited, and further exploration of additional potential prognostic factors is warranted to provide more personalized treatment strategies for GC patients.

CONCLUSION

Patients with lower preoperative CTI have a better prognosis, with longer DFS and OS. In this study, CTI became an independent prognostic factor for DFS and OS, indicating its significant role as a predictor of long-term prognosis for patients undergoing radical gastrectomy. Moreover, the nomogram model constructed by combining CTI and PNI showed high predictive accuracy in predicting DFS and OS, further confirming the importance of CTI. These findings support the use of CTI as a valuable indicator for guiding treatment and improving the prognosis assessment of patients undergoing radical gastrectomy for GC.