A nomogram based on systemic inflammatory markers: development and validation for predicting postoperative overall survival in gastric cancer Authors: Liuyi Yang, Xuexian Cao, Yulan Gu, Jinghua Ye Published: 2026 ABSTRACT BackgroundGastric cancer (GC) remains a leading cause of cancer-related mortality worldwide. Emerging evidence links dysregulated lipid metabolism and chronic systemic inflammation to GC progression, and a bidirectional regulatory loop between lipid metabolism reprogramming and tumor inflammation; however, most prognostic studies have evaluated these markers in isolation. We hypothesized that integrating lipid derivatives and systemic inflammatory markers would provide complementary prognostic information. This study aimed to develop and validate a nomogram combining these markers to predict postoperative overall survival (OS) in patients with resectable GC.MethodsThis retrospective study included GC patients who underwent radical gastrectomy at Affiliated Changshu Hospital of Nantong University, between January 2017 and December 2022. Inclusion criteria included radical surgery without preoperative anticancer treatment, and expected postoperative survival ≥3 months. Exclusion criteria included concurrent or prior malignancies, preoperative infection, or use of lipid-altering medications within 6 months. Fasting venous blood samples were collected preoperatively for routine blood, lipid profile, and coagulation tests using standardized automated analyzers. Fourteen composite inflammatory and lipid indices were calculated. Univariate and multivariate Cox regression analyses identified the independent predictors of overall survival (OS). A nomogram was constructed and validated using Harrell’s concordance index (C-index) and calibration curves.ResultsA total of 295 patients with GC were included in the study. The patients were randomized into a training group (n=197) and a validation group (n=98) (2:1 ratio). Baseline characteristics were well balanced (all P>0.05), with only alcohol consumption rate differing significantly between groups (P=0.01). Univariate and multivariate Cox regression revealed that the pathological tumor-node-metastasis (pTNM) stage [hazard ratio (HR) =2.05; 95% confidence interval (CI): 1.09–3.84; P=0.03], systemic immune-inflammation index (SII) (HR =1.99; 95% CI: 1.08–3.68; P=0.03), systemic inflammatory response index (SIRI) (HR =1.88; 95% CI: 1.02–3.45; P=0.04), and fibrinogen-to-platelet ratio (FPR) (HR =2.14; 95% CI: 1.25–3.67; P=0.006) were independent predictors of postoperative OS, positively correlating with postoperative mortality. The nomogram integrating these four variables showed satisfactory discriminative ability, with C-indices of 0.699 in the training set and 0.719 in the validation set. Calibration curves demonstrated good agreement between the predicted and observed 1-, 2-, and 3-year OS rates.ConclusionsThe nomogram based on the pTNM stage, SII, SIRI, and FPR exhibits moderate prognostic performance for postoperative OS in GC patients. It provides a comprehensive tool for clinicians to assess prognosis, guiding adjuvant therapy and individualized follow-up strategies. Introduction Gastric cancer (GC), including gastroesophageal junction cancer, is one of the top five causes of cancer-related death worldwide (1,2). In 2021, approximately 1,226,056 individuals were diagnosed with GC in China, and 445,013 GC-related deaths were recorded (3). Although substantial progress has been made in the diagnosis and treatment of GC in recent years, the 5-year overall survival (OS) rate remains suboptimal due to local recurrence or metastasis following primary GC resection (4). Therefore, identifying independent and effective predictive factors is crucial for evaluating the prognosis of GC patients. Currently, the tumor-node-metastasis (TNM) staging system by the American Joint Committee on Cancer (AJCC) remains the most widely used prognostic assessment tool in clinical practice. It stages cancer based on primary tumor invasion depth, regional lymph node metastasis status, and distant metastasis status. However, the TNM system does not comprehensively encompass tumor-related biological information, fails to provide quantitative predictive outcomes, and exhibits limited performance. A study focusing on stage II/III GC patients revealed that the concordance index (C-index) for predicting OS based on the AJCC 8th edition TNM system was only 0.620 (5). Further, due to heterogeneity among patients, individuals with the same TNM stage may experience significantly different prognoses (6,7). Consequently, an increasing number of prognostic models integrate radiological features, treatment modalities, and biological markers with TNM stage to enhance model performance. A nomogram, based on multiple parameters for predicting postoperative OS in GC patients, demonstrated significantly superior discriminative ability compared to TNM alone (8). This finding shows the potential of incorporating multidimensional parameters for precise prognostic prediction in GC patients. Compared to other clinical parameters, biological markers possess significant advantages including high detection sensitivity, quantifiability, and low cost. Substantial research indicates that lipid metabolism participates in regulating tumor invasion and distant metastasis (9). Lipid metabolism indicators such as low-density lipoprotein cholesterol (LDL-C) and high-density lipoprotein cholesterol (HDL-C) have demonstrated significant correlations with GC prognosis and early recurrence (10). Further studies suggest that lipid derivatives such as the atherosclerotic index (AI), total cholesterol-to-high-density lipoprotein cholesterol ratio (THR), and low-density lipoprotein cholesterol-to-high-density lipoprotein cholesterol ratio (LHR) exhibit significant value in predicting postoperative mortality in GC (11). Additionally, the inflammatory response plays a critical role in carcinogenesis, tumor progression, and treatment resistance (12,13). The preoperative neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) show that elevated preoperative NLR and PLR are significantly associated with reduced postoperative OS, cancer-specific survival, and disease-free survival in GC patients. Incorporating these markers alongside TNM staging into a nomogram model significantly improved the C-index (14). In a study predicting postoperative GC mortality using the NLR and PLR, six models achieved area under the curve (AUC) values ranging from 0.753 to 0.943, indicating substantially improved predictive accuracy (15). Numerous previously published GC prognostic nomograms use either inflammatory or lipid markers alone, while few studies integrate both categories of biomarkers simultaneously (16). Mechanistically, systemic inflammatory markers, such as systemic immune-inflammation index (SII) and systemic inflammatory response index (SIRI), reflect systemic tumor-triggered chronic inflammatory microenvironment (17,18), while fibrinogen-to-platelet ratio (FPR) mirrors abnormal lipid metabolism and platelet activation induced by malignant lesions (19,20); these two categories of indicators reflect distinct tumor-promoting pathological pathways independently and can compensate for information defects of a single index when combined, thereby synergistically elevating predictive efficacy. Although abundant published studies have verified the prognostic value of individual inflammatory or lipid biomarkers in GC, few investigations combine SII, SIRI and FPR simultaneously to build a preoperative prognostic nomogram. Incorporating lipid metabolism indicators and inflammatory markers into prognostic models could potentially improve the accuracy of postoperative prognostic assessments in GC patients. Against this background, this study aimed to construct a nomogram model based on preoperative lipid derivatives and systemic inflammatory markers to predict OS in patients who had undergone radical gastrectomy for GC. Existing high-performance prediction models mostly focus on imaging or single-type laboratory indicators; our newly established model innovatively integrates dual-pathway biomarkers of inflammation and lipid metabolism, making up for the deficiency of current prediction systems and filling the research gap of combined application of SII, SIRI and FPR in gastric cancer prognostic evaluation. The findings may provide clinicians with a more comprehensive prognostic assessment tool that can be used to guide postoperative adjuvant therapy decisions and formulate individualized follow-up strategies. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0916/rc). Methods Study design and population This study retrospectively collected the medical record data of GC patients who attended the Affiliated Changshu Hospital of Nantong University, from January 2017 to December 2022, and followed them up. All the GC patients were diagnosed by histopathology and staged according to the AJCC TNM staging system (8th edition, 2017). Follow-up was every 3 months for the first 2 years after surgery, every 6 months for the next 3 years, and annually thereafter until death or September 2024. The date of death was obtained by a telephone call from a physician or at a follow-up appointment. The inclusion criteria were as follows: (I) undergoing radical surgery; (II) no preoperative anti-cancer treatment, such as chemotherapy or radiotherapy; and (III) an expected postoperative survival ≥3 months. The exclusion criteria were as follows: (I) a history of previous or current concurrent other malignant tumors; (II) preoperative infection; and/or (III) use of any medications known to affect lipid levels, such as lipid-lowering drugs, glucocorticoids, and metformin, within the 6 months prior to blood specimen collection. This study was performed in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Affiliated Changshu Hospital of Nantong University (No. 2023-KY-ZX43), and all patients provided written informed consent. Sample size calculation At the study design stage, the exact number of predictors for the final multivariate Cox model could not be predefined. We planned to screen 35 systemic inflammatory, clinicopathological indicators via univariate analysis and conservatively assumed a maximum of 12 predictors to enter the final model. Based on the Events per Variable (EPV) ≥10 rule for Cox regression, no fewer than 120 endpoint deaths were required. Using a 40% postoperative mortality rate for gastric cancer in China (21) and an anticipated 10% loss to follow-up or incomplete data, the preliminary required total sample size was calculated as 503 patients. Restricted by the available eligible historical cases in our hospital, we finally enrolled 295 patients, who were randomly split at a 2:1 ratio into a training set (n=197) and a validation set (n=98). We performed 1,000-bootstrap resampling within the training set for internal validation, and the validation set was used to assess model discrimination and calibration. Data extraction All the clinical information, laboratory indicators, and pathological data of the patients were entered into a standardized electronic case system and integrated to create a study-specific database. The clinical information included gender, age, Eastern Cooperative Oncology Group performance status (ECOG-PS) score, smoking status, drinking status, and history of chronic diseases (e.g., hypertension and diabetes). The data were independently extracted and cross-checked by two researchers to ensure consistency. Laboratory indicators included blood lipid indicators, inflammatory indicators, and tumor markers. All blood samples were obtained from preoperative fasting venous blood. Blood lipid indicators included albumin, HDL-C, LDL-C, total cholesterol, prealbumin, alkaline phosphatase, fibrinogen, and lactate dehydrogenase (LDH). Inflammatory indicators included LDH, and counts of white blood cells, neutrophils, monocytes, lymphocytes, and platelets. Tumor markers included carcinoembryonic antigen (CEA), Ki-67, carbohydrate antigen 19-9 (CA19-9), and carbohydrate antigen 125 (CA125). The data were directly linked to the original test results in the hospital laboratory information management system (LIS) to avoid secondary input errors. Pathological data included the pathological tumor-node-metastasis (pTNM) stage (AJCC 8th edition), clinical grade, pT stage, pN stage, pM stage, histological subtype, degree of differentiation, and human epidermal growth factor receptor 2 (HER2) expression. The data were confirmed by the pathologist according to the original postoperative pathology report. The database was stored in an encrypted format and was accessible only to authorized researchers. Data cleansing and logical checks were conducted periodically by the investigators to ensure data integrity and accuracy. The preoperative lipid derivatives and systemic inflammatory markers used in this study were calculated from the raw data extracted from this database. The calculation formulas are presented in Table 1. Table 1Calculation formulas for the lipid derivatives and systemic inflammatory markersLaboratory DataFormulaTHRTC/HDL-CLHRLDL-C/HDL-CLMRLymphocyte count/mononuclear cell countNLRNeutrophil count/lymphocyte countdNLRNeutrophil count/(WBC count − neutrophil count)PLRPlatelet count/lymphocyte countNWRNeutrophil count/WBC countPNRPlatelet count/neutrophil countAPRAlkaline phosphatase/prealbuminFLRFibrinogen/lymphocyte countFPRFibrinogen/platelet countPNI10 × serum albumin level (g/dL) + 0.005 × peripheral blood lymphocyte count (mm3)SIRI(Neutrophil × mononuclear cell count)/lymphocyte countSII(Platelet × neutrophil count)/lymphocyte countAPR, alkaline phosphatase-to-prealbumin ratio; dNLR, derived NLR; FLR, fibrinogen-to-lymphocyte ratio; FPR, fibrinogen-to-platelet ratio; HDL-C, high-density lipoprotein cholesterol; LHR, low-density lipoprotein cholesterol-to-high-density lipoprotein cholesterol ratio; LDL-C, low-density lipoprotein cholesterol; LMR, lymphocyte-to-monocyte ratio; NLR, neutrophil-to-lymphocyte ratio; NWR, neutrophil-to-white blood cell ratio; PLR, platelet-to-lymphocyte ratio; PNI, prognostic nutritional index; PNR, platelet-to-neutrophil ratio; SII, systemic immune-inflammation index; SIRI, systemic inflammatory response index; TC, total cholesterol; THR, total cholesterol-to-high-density lipoprotein cholesterol ratio; WBC, white blood cell. Calculation formulas for the lipid derivatives and systemic inflammatory markers APR, alkaline phosphatase-to-prealbumin ratio; dNLR, derived NLR; FLR, fibrinogen-to-lymphocyte ratio; FPR, fibrinogen-to-platelet ratio; HDL-C, high-density lipoprotein cholesterol; LHR, low-density lipoprotein cholesterol-to-high-density lipoprotein cholesterol ratio; LDL-C, low-density lipoprotein cholesterol; LMR, lymphocyte-to-monocyte ratio; NLR, neutrophil-to-lymphocyte ratio; NWR, neutrophil-to-white blood cell ratio; PLR, platelet-to-lymphocyte ratio; PNI, prognostic nutritional index; PNR, platelet-to-neutrophil ratio; SII, systemic immune-inflammation index; SIRI, systemic inflammatory response index; TC, total cholesterol; THR, total cholesterol-to-high-density lipoprotein cholesterol ratio; WBC, white blood cell. Outcome The primary endpoint of this study was OS in GC patients. OS was defined as the interval from the date of surgery to death from any cause or to the date of the final follow-up. Laboratory testing procedures For routine blood tests, 2 mL venous blood was harvested into ethylenediaminetetraacetic acid (EDTA) anticoagulant tubes and measured by the SYSMEX XT‑1800i hematology analyzer. Another 2 mL blood was drawn into plain serum tubes for liver function detection via a multichannel automatic biochemical analyzer. An additional 2 mL venous blood was collected in vacuum tubes prefilled with 0.2 mL 10% sodium citrate; after blending and centrifugation for plasma separation, coagulation parameters were assayed with the SYSMEX CA‑7000 coagulation analyzer. All laboratory examinations were implemented following standard laboratory internal quality control protocols. Statistical analysis Categorical variables were expressed as frequencies and percentages (%), and the Chi-squared test was used for inter-group comparisons. Optimal cutoff values for all continuous biomarkers were defined based on the median of the study cohort. Candidate prognostic factors were screened by univariate Cox regression, and variables with P<0.05 were further included into multivariate Cox regression to identify independent predictors of OS. The results were reported as the hazard ratio (HR) and its 95% confidence interval (CI). Statistically significant variables were used for nomogram modeling. The performance of the model was evaluated using calibration curve analysis and the C-index. For patients with missing core variable information, relevant cases were directly excluded from the final analysis; thus, no imputation was performed. All analyses were performed using R (version 4.4.3). All statistical tests were two-sided, with P<0.05 defined as statistically significant. The full mathematical expression of the Cox proportional hazards model and relevant baseline survival parameters are provided as follows and summarized in Table S1: where h(t,X) refers to instantaneous hazard at time t for specified covariates; h0(t) stands for baseline hazard when all predictors equal zero; X1−X7 indicate seven final independent predictors, and β7−β7 are corresponding raw regression coefficients. Results Clinical characteristics of patients A total of 295 patients with GC were enrolled in this study, of whom 202 (68.47%) were male and 93 (31.53%) were female. Among the patients, 80.34% had adenocarcinoma and 19.66% had non-adenocarcinoma. Patients with early stage (I, II) and intermediate-late stage (III, IV) disease accounted for 33.22% and 66.78%, respectively. The proportion of patients with a history of alcohol consumption was lower in the training set than in the validation set (93.91% vs. 100.00%; P=0.01); however, no significant differences were observed in the remaining baseline characteristics (Table 2). Table 2The clinical characteristics of GC patientsVariablesOverall (n=295), n (%)Training set (n=197), n (%)Validation set (n=98), n (%)PAge (years)0.42 <6059 (20.00)42 (21.32)17 (17.35) ≥60236 (80.00)155 (78.68)81 (82.65)Sex0.44 Male202 (68.47)132 (67.01)70 (71.43) Female93 (31.53)65 (32.99)28 (28.57)ECOG-PS score0.86 016 (5.42)11 (5.58)5 (5.10) 1–2279 (94.58)186 (94.42)93 (94.90)Smoking0.31 No277 (93.90)183 (92.89)94 (95.92) Yes18 (6.10)14 (7.11)4 (4.08)Drinking0.01 No283 (95.93)185 (93.91)98 (100.00) Yes12 (4.07)12 (6.09)0 (0.00)Hypertension0.83 No114 (38.64)77 (39.09)37 (37.76) Yes181 (61.36)120 (60.91)61 (62.24)Diabetes0.41 No251 (85.08)170 (86.29)81 (82.65) Yes44 (14.92)27 (13.71)17 (17.35)Cardiopathy0.53 No283 (95.93)190 (96.45)93 (94.90) Yes12 (4.07)7 (3.55)5 (5.10)Hepatitis B0.11 No285 (96.61)188 (95.43)97 (98.98) Yes10 (3.39)9 (4.57)1 (1.02)THR0.97 ≤3.519147 (49.83)98 (49.75)49 (50.00) >3.519148 (50.17)99 (50.25)49 (50.00)LHR0.39 ≤2.044143 (48.47)99 (50.25)44 (44.90) >2.044152 (51.53)98 (49.75)54 (55.10)LDH (IU/L)0.36 Normal279 (94.58)188 (95.43)91 (92.86) Abnormal16 (5.42)9 (4.57)7 (7.14)SII (1×109/L)0.77 ≤718.976145 (49.15)98 (49.75)47 (47.96) >718.976150 (50.85)99 (50.25)51 (52.04)SIRI (1×109/L)0.71 ≤1.233146 (49.49)99 (50.25)47 (47.96) >1.233149 (50.51)98 (49.75)51 (52.04)LMR0.34 ≤3.419154 (52.20)99 (50.25)55 (56.12) >3.419141 (47.80)98 (49.75)43 (43.88)NLR0.65 ≤3.471144 (48.81)98 (49.75)46 (46.94) >3.471151 (51.19)99 (50.25)52 (53.06)dNLR0.77 ≤2.263145 (49.15)98 (49.75)47 (47.96) >2.263150 (50.85)99 (50.25)51 (52.04)PLR0.77 ≤174.468145 (49.15)98 (49.75)47 (47.96) >174.468150 (50.85)99 (50.25)51 (52.04)NWR0.84 ≤0.694147 (49.83)99 (50.25)48 (48.98) >0.694148 (50.17)98 (49.75)50 (51.02)PNR0.30 ≤53.741153 (51.86)98 (49.75)55 (56.12) >53.741142 (48.14)99 (50.25)43 (43.88)PNI0.84 ≤45.800151 (51.19)100 (50.76)51 (52.04) >45.800144 (48.81)97 (49.24)47 (47.96)APR (IU/g)0.84 ≤0.522148 (50.17)98 (49.75)50 (51.02) >0.522147 (49.83)99 (50.25)48 (48.98)FLR (mg/109)0.34 ≤2.857141 (47.80)98 (49.75)43 (43.88) >2.857154 (52.20)99 (50.25)55 (56.12)FPR (mg/109)0.30 ≤1.538142 (48.14)99 (50.25)43 (43.88) >1.538153 (51.86)98 (49.75)55 (56.12)CEA (ng/mL)0.71 ≤5164 (55.59)108 (54.82)56 (57.14) >5131 (44.41)89 (45.18)42 (42.86)CA19-9 (U/mL)0.53 ≤35191 (64.75)130 (65.99)61 (62.24) >35104 (35.25)67 (34.01)37 (37.76)CA125 (U/mL)0.48 ≤35212 (71.86)139 (70.56)73 (74.49) >3583 (28.14)58 (29.44)25 (25.51)Ki-670.42 ≤45124 (42.03)86 (43.65)38 (38.78) >45171 (57.97)111 (56.35)60 (61.22)Histological type0.93 Non adenocarcinoma58 (19.66)39 (19.80)19 (19.39) Adenocarcinoma237 (80.34)158 (80.20)79 (80.61)Clinical grade0.19 High/moderate224 (75.93)145 (73.60)79 (80.61) Low71 (24.07)52 (26.40)19 (19.39)pTNM stage0.91 I/II98 (33.22)65 (32.99)33 (33.67) III/IV197 (66.78)132 (67.01)65 (66.33)pT stage0.39 1/250 (16.95)36 (18.27)14 (14.29) 3/4245 (83.05)161 (81.73)84 (85.71)pN stage0.49 074 (25.08)47 (23.86)27 (27.55) 1/2/3221 (74.92)150 (76.14)71 (72.45)pM stage0.71 0253 (85.76)170 (86.29)83 (84.69) 142 (14.24)27 (13.71)15 (15.31)HER2 expression0.57 Negative266 (90.17)179 (90.86)87 (88.78) Positive29 (9.83)18 (9.14)11 (11.22)APR, alkaline phosphatase-to-prealbumin ratio; CA125, carbohydrate antigen 125; CA19-9, carbohydrate antigen 19-9; CEA, carcinoembryonic antigen; dNLR, derived NLR; ECOG-PS, Eastern Cooperative Oncology Group performance status; FLR, fibrinogen-to-lymphocyte ratio; FPR, fibrinogen-to-platelet ratio; GC, gastric cancer; HER2, human epidermal growth factor receptor 2; LDH, lactate dehydrogenase; LHR, low-density lipoprotein cholesterol-to-high-density lipoprotein cholesterol ratio; LMR, lymphocyte-to-monocyte ratio; NLR, neutrophil-to-lymphocyte ratio; NWR, neutrophil-to-white blood cell ratio; PLR, platelet-to-lymphocyte ratio; PNI, prognostic nutritional index; PNR, platelet-to-neutrophil ratio; pTNM Stage, pathological tumor-node-metastasis stage; SII, systemic immune-inflammation index; SIRI, systemic inflammatory response index; THR, total cholesterol-to-high-density lipoprotein cholesterol ratio. The clinical characteristics of GC patients APR, alkaline phosphatase-to-prealbumin ratio; CA125, carbohydrate antigen 125; CA19-9, carbohydrate antigen 19-9; CEA, carcinoembryonic antigen; dNLR, derived NLR; ECOG-PS, Eastern Cooperative Oncology Group performance status; FLR, fibrinogen-to-lymphocyte ratio; FPR, fibrinogen-to-platelet ratio; GC, gastric cancer; HER2, human epidermal growth factor receptor 2; LDH, lactate dehydrogenase; LHR, low-density lipoprotein cholesterol-to-high-density lipoprotein cholesterol ratio; LMR, lymphocyte-to-monocyte ratio; NLR, neutrophil-to-lymphocyte ratio; NWR, neutrophil-to-white blood cell ratio; PLR, platelet-to-lymphocyte ratio; PNI, prognostic nutritional index; PNR, platelet-to-neutrophil ratio; pTNM Stage, pathological tumor-node-metastasis stage; SII, systemic immune-inflammation index; SIRI, systemic inflammatory response index; THR, total cholesterol-to-high-density lipoprotein cholesterol ratio. Selection of predictive factors Univariate Cox regression models were used to analyze the associations between 35 variables from baseline and OS. The results indicated that hypertension (HR =1.76; 95% CI: 1.01–3.09; P=0.047), the systemic immune-inflammation index (SII) (HR =2.51; 95% CI: 1.46–4.32; P=0.001), systemic inflammatory response index (SIRI) (HR =2.85; 95% CI: 1.64–4.95; P=0.001), the alkaline phosphatase-to-prealbumin ratio (APR) (HR =1.76; 95% CI: 1.05–2.96; P=0.03), the fibrinogen-to-platelet ratio (FPR) (HR =1.78; 95% CI: 1.06–2.98; P=0.03), the pTNM stage (HR =2.26; 95% CI: 1.22–4.18; P=0.009), and the pM stage (HR =2.29; 95% CI: 1.26–4.17; P=0.006) were all risk factors for postoperative mortality in GC patients (Table 3). Further, the multivariate Cox regression analysis of those seven variables revealed that the SII (HR =1.99; 95% CI: 1.08–3.68; P=0.03), SIRI (HR =1.88, 95% CI: 1.02–3.45; P=0.04), FPR (HR =2.14; 95% CI: 1.25–3.67; P=0.006), and pTNM stage (HR =2.05; 95% CI: 1.09–3.84; P=0.03) were independent predictors of postoperative OS in GC patients (Table 3). Table 3Univariate and multivariate analyses of prognostic factors in Cox regression modelCharacteristicsUnivariate CoxMultivariate CoxHR (95% CI)PHR (95% CI)PAge1.48 (0.75–2.92)0.26––Sex0.76 (0.44–1.34)0.35––ECOG-PS score3.47 (0.48–25.12)0.22––Smoking1.67 (0.76–3.66)0.20––Drinking0.92 (0.33–2.54)0.87––Hypertension1.76 (1.01–3.09)0.0471.54 (0.87–2.75)0.14Diabetes0.93 (0.44–1.95)0.80––Cardiopathy0.93 (0.23–3.79)0.91––Hepatitis B0.28 (0.04–2.03)0.21––THR0.74 (0.45–1.23)0.25––LHR0.71 (0.43–1.18)0.19––LDH0.64 (0.16–2.63)0.54––SII2.51 (1.46–4.32)0.0011.99 (1.08–3.68)0.03SIRI2.85 (1.64–4.95)0.0011.88 (1.02–3.45)0.042LMR0.74 (0.44–1.23)0.24––NLR1.08 (0.65–1.78)0.77––dNLR0.85 (0.52–1.41)0.54––PLR0.87 (0.53–1.45)0.60––NWR0.87 (0.53–1.44)0.58––PNR0.69 (0.41–1.14)0.15––PNI0.89 (0.54–1.47)0.64––APR1.76 (1.05–2.96)0.031.52 (0.9–2.57)0.12FLR0.97 (0.59–1.61)0.92––FPR1.78 (1.06–2.98)0.032.14 (1.25–3.67)0.006CEA 0.75 (0.45–1.25)0.27––CA19-9 0.81 (0.47–1.38)0.44––CA1250.93 (0.54–1.61)0.80––Ki-670.73 (0.44–1.21)0.23––Histological type1.36 (0.67–2.76)0.39––Clinical grade0.81 (0.45–1.45)0.48––pTNM stage2.26 (1.22–4.18)0.0092.05 (1.09–3.84)0.03pT stage1.26 (0.64–2.48)0.50––pN stage1.43 (0.76–2.7)0.27––pM stage2.29 (1.26–4.17)0.0061.27 (0.68–2.38)0.45HER2 expression1.21 (0.52–2.82)0.66––APR, alkaline phosphatase-to-prealbumin ratio; CA125, carbohydrate antigen 125; CA19-9, carbohydrate antigen 19-9; CEA, carcinoembryonic antigen; CI, confidence interval; dNLR, derived NLR; ECOG-PS, Eastern Cooperative Oncology Group performance status; FLR, fibrinogen-to-lymphocyte ratio; FPR, fibrinogen-to-platelet ratio; HER2, human epidermal growth factor receptor 2; HR, hazard ratio; LDH, lactate dehydrogenase; LHR, low-density lipoprotein cholesterol-to-high-density lipoprotein cholesterol ratio; LMR, lymphocyte-to-monocyte ratio; NLR, neutrophil-to-lymphocyte ratio; NWR, neutrophil-to-white blood cell ratio; PLR, platelet-to-lymphocyte ratio; PNI, prognostic nutritional index; PNR, platelet-to-neutrophil ratio; pTNM stage, pathological tumor-node-metastasis stage; SII, systemic immune-inflammation index; SIRI, systemic inflammatory response index; THR, total cholesterol-to-high-density lipoprotein cholesterol ratio. Univariate and multivariate analyses of prognostic factors in Cox regression model APR, alkaline phosphatase-to-prealbumin ratio; CA125, carbohydrate antigen 125; CA19-9, carbohydrate antigen 19-9; CEA, carcinoembryonic antigen; CI, confidence interval; dNLR, derived NLR; ECOG-PS, Eastern Cooperative Oncology Group performance status; FLR, fibrinogen-to-lymphocyte ratio; FPR, fibrinogen-to-platelet ratio; HER2, human epidermal growth factor receptor 2; HR, hazard ratio; LDH, lactate dehydrogenase; LHR, low-density lipoprotein cholesterol-to-high-density lipoprotein cholesterol ratio; LMR, lymphocyte-to-monocyte ratio; NLR, neutrophil-to-lymphocyte ratio; NWR, neutrophil-to-white blood cell ratio; PLR, platelet-to-lymphocyte ratio; PNI, prognostic nutritional index; PNR, platelet-to-neutrophil ratio; pTNM stage, pathological tumor-node-metastasis stage; SII, systemic immune-inflammation index; SIRI, systemic inflammatory response index; THR, total cholesterol-to-high-density lipoprotein cholesterol ratio. Performance of the nomogram Based on the Cox regression model, a nomogram incorporating the pTNM stage, SII, SIRI, and FPR was constructed. The nomogram was used to predict 1-, 2-, and 3-year OS rates after radical gastrectomy in patients with GC (Figure 1). A nomogram to predict OS rates at 1, 2, and 3 years after radical resection in GC patients. AJCC: pTNM stage, 0 represents stage I/II and 1 represents stage III/IV; SII: 0 for SII ≤718.976×109/L, 1 for SII >718.976×109/L; SIRI: 0 for SIRI ≤1.233×109/L, 1 for SIRI >1.233×109/L; FPR: 0 for FPR ≤1.538 mg/109, 1 for FPR >1.538 mg/109. AJCC, American Joint Committee on Cancer; FPR, fibrinogen-to-platelet ratio; GC, gastric cancer; OS, overall survival; pTNM, pathological tumor-node-metastasis; SII, systemic immune-inflammation index; SIRI, systemic inflammatory response index. In the training set, the C-index of the nomogram was 0.699, estimated via bootstrap resampling. In the validation set, the C-index reached 0.719, indicating satisfactory discriminative ability. Further, the calibration curve analysis demonstrated excellent agreement between the predicted and observed survival probabilities (Figure 2). Calibration curve analysis of nomogram prediction probabilities. (A-C) Calibration curves for 1-, 2-, and 3-year OS in the training set. (D-F) Calibration curves for 1-, 2-, and 3-year OS in the validation set. OS, overall survival. Discussion The present study developed a nomogram integrating the pTNM stage and three preoperative biomarkers (the SII, SIRI, and FPR) to predict postoperative OS rates in GC patients who underwent radical gastrectomy. This model demonstrated satisfactory discriminative ability, with C-indices of 0.699 in the training set and 0.719 in the validation set. The calibration curves confirmed good agreement between the predicted and observed survival probabilities. These findings highlight the potential of systemic inflammatory markers in GC prognosis. Previous studies have produced various prognostic models for GC patients. For example, Tonello et al. developed a nomogram incorporating age, tumor size/location, TNM stage, and complications to predict OS in GC patients (22). In a large Surveillance, Epidemiology, and End Results (SEER)-based analysis, Hu et al. built grade-specific OS models using demographic and tumor variables. The 5-year OS AUC reached 0.828 in the training set and 0.758 in the external validation set, exceeding that of TNM staging (23). These models rely on standard clinical and pathological predictors and underscore that integrating variables beyond the TNM stage can improve discrimination. Several recent prognostic models have incorporated genetic molecular signatures, modestly improving risk stratification compared with the TNM staging system (24-26). However, their routine use is limited by assay costs and laboratory availability. In addition, advanced imaging techniques have yielded high-performance models for GC (27). A study showed that a higher radiomics score was highly associated with poorer survival outcomes after radical resection in patients with GC. The nomogram based on quantitative computed tomography texture features achieved C-indices of 0.782 in predicting OS and 0.791 in predicting recurrence-free survival (28). Although these models show excellent discrimination, advanced image acquisition, high-throughput feature extraction, and computational analysis are needed. Therefore, radiomics may not be widely available in all practice settings. In this study, the SII, SIRI, and FPR supplement TNM staging by enabling assessment of the tumor microenvironment. The findings align with previous reports indicating that the SII, SIRI, the fibrinogen-SII score, and the fibrinogen-SIRI score serve as valuable predictors of OS in patients with GC (29-31). A previous model based on the fibrinogen-SII score for postoperative OS prediction in GC patients achieved a C-index of 0.72, a performance comparable to that of the present model (32). Although the predictive accuracy of the present model does not surpass that of genomic or radiomics models, it offers a clear advantage in terms of clinical feasibility. These lipid derivatives and systemic inflammatory markers are obtained from routine preoperative blood work; thus, they impose no additional financial burden on patients. Moreover, laboratory assays provide rapid turnaround, enhancing their utility across diverse healthcare settings and enabling real-time monitoring during follow-up. In our cohort, the SII, SIRI, and FPR were all associated with poorer postoperative OS. Biologically, the SII and SIRI, which integrate neutrophil, lymphocyte, monocyte, and platelet counts, reflect the complex crosstalk between systemic inflammation and immune suppression (33). Studies have shown that neutrophilia and monocytosis promote tumor progression through the secretion of pro-angiogenic and pro-metastatic mediators, creating an immunosuppressive microenvironment (34,35). Lymphopenia signals impairs cellular anti-tumor responses (36). Platelets facilitate tumor cell survival in the circulation and support metastatic seeding (37). Notably, this study highlights the prognostic significance of the FPR, a less explored marker. An elevated FPR likely reflects a pro-thrombotic state and an intensified inflammatory response associated with tumor progression and metastasis. This finding resonates with research linking coagulation pathways and platelet activation to cancer biology, suggesting that the FPR may capture a dimension of host-tumor interaction not fully assessed by other inflammatory indices (38). The independent prognostic value of lipid derivatives observed in earlier studies, such as the ratios involving triglycerides, LDL-C, and HDL-C, was partially reflected in our univariate analysis (10,11). However, the FPR emerged as the most robust lipid-related predictor in our multivariate model, suggesting its potential superiority or complementary value within this specific biomarker panel. In conclusion, by integrating readily available preoperative blood-based markers with standard pathological staging, this tool provides a practical and relatively inexpensive method for refining postoperative risk stratification. Clinicians can use it to identify high-risk patients who may benefit from intensified adjuvant therapy, closer surveillance, or enrollment in clinical trials evaluating novel therapeutic strategies. Further, these blood biomarkers may serve as inexpensive signals of concurrent comorbid inflammatory or coagulation disorders that warrant clinical attention and might themselves influence perioperative management and long-term outcomes. Together, these features make the SII, SIRI, and FPR attractive translational biomarkers for integration into everyday clinical workflows. Limitations Several limitations of this study should be acknowledged. First, this was a retrospective, single-center study, potentially introducing selection bias and limiting the generalizability of our findings. External validation in independent, multi-center cohorts with diverse demographics is essential to confirm the robustness and applicability of model. Second, the exclusion of patients receiving neoadjuvant therapy, while methodologically sound for isolating the prognostic value of preoperative markers, means the model is not directly applicable to the increasingly common cohort of patients treated with perioperative chemotherapy. Third, the total number of observed death events was 50 against seven final predictors, yielding an EPV value of approximately 7.14, slightly below the conventional threshold of 10. Such relatively limited endpoint events may raise a modest risk of model overfitting. Finally, certain variables potentially affecting OS, such as postoperative complications or changes in lifestyle factors, were not included in the analysis, which may have residual confounding effects. Conclusions This study developed and validated a nomogram that integrates the pTNM stage with the SII, SIRI, and FPR, to predict postoperative OS in GC patients. This model offers a practical tool for enhancing individualized risk assessment. The findings underscore the importance of the systemic inflammatory response pathways in GC progression. Supplementary The article’s supplementary files as Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0916/prf