Neutrophil to lymphocyte ratio, platelet to lymphocyte ratio, and systemic immune-inflammation index as predictors of venous thromboembolism in patients with diffuse large B-cell lymphoma Authors: Waad Almalki, Atheer Alenezi, Rawan Almutairi, Anfal Alsadhan Published: 2026 ABSTRACT BackgroundNeutrophil-lymphocyte ratio (NLR), platelet-lymphocyte ratio (PLR), and systemic immune-inflammation index (SII) have been linked to venous thromboembolism (VTE) across various cancer types; however, their relationship with diffuse large B-cell lymphoma (DLBCL) remains largely unexplored. This study aimed to investigate the role of these inflammatory markers in predicting the risk of VTE in patients with DLBCL.MethodA retrospective cohort study was conducted to analyze NLR, PLR, and SII in 224 DLBCL patients prior to chemotherapy. Of them, 31 (13.8%) patients were diagnosed with VTE, while the remaining 193 patients were considered as the control group.ResultsThe median age was 55 (30) years, with men representing 57% of the patients. The majority of VTE events occurred during active treatment (58%), however; 42% of VTE cases were diagnosed prior to the initiation of treatment. NLR, PLR, and SII were significantly higher in DLBCL patients who developed VTE compared with the control (P = 0.003, P = 0.015, and P = 0.027, respectively). The optimal cut-off values were > 2.41 for NLR (P = 0.003, with a sensitivity of 68% and specificity of 63%), > 185 for PLR (P = 0.016, with a sensitivity of 71% and specificity of 67%), and > 656 for SII (P = 0.027, with a sensitivity of 64% and specificity of 55%). Area under the curve values were 66% for NLR, 63% for PLR, and 62% for SII. The multivariate logistic regression model showed that NLR (ORa 3.750, P = 0.001) and PLR (ORa 5.211, P < 0.001) were independent predictors of VTE in DLBCL patients.ConclusionsNLR and PLR could be beneficial predictors of potential VTE in DLBCL patients without additional costs to the healthcare systems. Introduction Diffuse large B-cell lymphoma (DLBCL) is the most prevalent form of non-Hodgkin lymphoma (NHL), representing 25–40% of cases [1, 2]. Venous thromboembolism (VTE), which includes deep vein thrombosis (DVT) and pulmonary embolism (PE), is a frequent complication among cancer patients and is regarded as one of the primary causes of mortality [3, 4]. The incidence of VTE varies based on the histology of NHL, with the highest rates seen in aggressive lymphomas, such as DLBCL [5–7]. Multiple factors may contribute to the etiopathogenesis of VTE, including endothelial injury, inflammation, and hypercoagulability. These factors can activate platelets and the coagulation cascade, resulting in thrombus formation [8, 9]. Lymphoma is marked by a state of immune dysregulation that induces systemic inflammation, which interferes with various stages of hemostasis through the release of several pro-inflammatory cytokines [10, 11]. Therefore, thrombocytosis may be a general response to the inflammation induced by cancer. The association of VTE with inflammatory markers, including the neutrophil/lymphocyte ratio (NLR), platelet/lymphocyte ratio (PLR), and systemic immune-inflammatory index (SII), has gained more attention in recent years. Elevated levels of NLR and PLR have been linked with VTE in various cancers, including lymphoma [12–16]. Numerous studies have shown that higher NLR and PLR values correlate with the incidence of VTE, with NLR specifically suggested as a potential early predictor of acute VTE events [13]. In patients with NHL undergoing chemotherapy, both NLR and PLR have been identified as independent prognostic markers for the development of VTE [15]. However, other research findings have indicated that only NLR, not PLR, serves as an independent predictor of VTE risk in patients with Hodgkin and non-Hodgkin lymphoma [16]. These discrepancies in predictive outcomes may be attributed to differences in study designs (retrospective versus prospective), sample sizes, and the extensive biological variability of lymphomas, which range from indolent to highly aggressive forms [17]. A disease-specific evaluation may yield more precise insights into the predictive value of these inflammatory markers. Furthermore, the systemic immune-inflammation index (SII) is a novel biomarker that has recently gained the attention of researchers, as it reflects immune thrombosis and inflammatory status in the body by combining neutrophil, platelet, and lymphocyte counts [18]. A recent study suggested that SII could serve as an auxiliary diagnostic tool to help identify patients with DVT [19]. Higher SII values have been observed in patients with DVT and VTE, supporting its role as an independent predictor of thromboembolic events [20, 21]. Additionally, SII has been incorporated into predictive models for VTE risk in gastrointestinal and lung cancers [22, 23]. Although NLR, PLR, and SII have been proposed as predictors of VTE development in various cancers, data specific to patients with DLBCL remain limited. Identifying DLBCL patients with a greater susceptibility to VTE using simple clinical parameters could facilitate early intervention and management strategies. Therefore, this study aimed to evaluate the predictive value of NLR, PLR, and SII for VTE risk in DLBCL patients. Methods Study design A retrospective cohort study was conducted between the years 2015 and 2023 at King Abdulaziz Medical City (KAMC) in Saudi Arabia. The study received approval from the Institutional Review Board of the King Abdullah International Medical Research Centre (KAIMRC). All patients over the age of 18 who have been newly diagnosed with diffuse large B-cell lymphoma (DLBCL) were included in the study. Patients were excluded if they had incomplete medical records, prior history of VTE, renal or hepatic disease, cardiac conditions, hematological disorders, or other cancers besides DLBCL, as these factors may influence the differential white blood cell count. A total of 224 patients who met the inclusion and exclusion criteria were enrolled. Expert radiologists and physicians documented VTE in patients’ medical records based on radiographic studies, clinical examinations, and laboratory evaluations. Data and variables Laboratory data, including complete blood count (CBC) and derived parameters (NLR, PLR, and SII), were obtained at the time of diagnosis, before the initiation of any treatments or the occurrence of VTE events. Additionally, clinical and epidemiological data, including age, sex, body mass index (BMI), Ann Arbor tumor Stage, and Eastern Oncology Cooperative Group performance status (EOCG PS) were recorded. NLR and PLR were calculated as the ratios of absolute neutrophil count to absolute lymphocyte count and absolute platelet count to absolute lymphocyte count, respectively. SII was derived by multiplying NLR by the absolute platelet count. Statistical analysis Categorical parameters were reported as numbers and percentages, while continuous parameters were expressed as medians with interquartile ranges (IQR). The normality of the data was evaluated using the Shapiro–Wilk test. Differences between DLBCL patients with VTE complications and those without were analyzed using the Mann-Whitney test for continuous variables and the Fisher’s exact test for categorical variables. Receiver operating characteristic (ROC) curve analysis was employed to determine the optimal cut-off values for NLR, PLR, and SII by measuring the Youden index. Both simple and multivariable logistic regressions were conducted to assess the association between NLR, PLR, and SII with the development of VTE. Results were presented as odds ratios (OR) and 95% confidence intervals (CI). A P-value of less than 0.05 was deemed statistically significant. Data were analyzed using GraphPad Prism version 10 (GraphPad Software, San Diego, CA, USA). Results Patients’ characteristics The median age of all the patients (n = 224) was 55 (30) years, with men comprising 57% of the patient population. During the study period, 31 (13.8%) of the included patients developed VTE. DLBCL patients with VTE had a higher prevalence of advanced Ann Arbor Stage (III/IV) compared to the non-VTE group (51.5% vs. 43.5%, respectively). The clinical characteristics of the included patients are shown in Table 1. The median age was 55 (33) for DLBCL patients without VTE complications and 56 (24) for those who developed VTE. There were no statistically significant differences in age, sex, BMI, tumor stage or ECOG PS between the two groups. Regarding laboratory data, there were no statistical differences between the groups in terms of absolute lymphocyte, neutrophil, and platelet counts. However, the calculated NLR, PLR, and SII parameters were significantly higher in DLBCL patients who developed VTE compared to the controls (P = 0.003, P = 0.015, and P = 0.027, respectively (Table 1). Table 1Baseline characteristics of DLBCL patients with and without VTECharacteristicsAll Subjectn= 224Presence of VTEP-valueNon = 193 (86%)Yesn = 31(14%)Age [Median (IQR)]55 (30)55 (33)56 (24)0.720Sex [n (%)]MaleFemale128 (57%)96 (43%)115 (60%)78 (40%)13 (42%)18 (58%)0.079BMI [Median (IQR)]27.5 (8.2)27.5 (7.65)24.2 (12.8)0.388Ann Arbor Stage [n (%)]I-IIIII-IV124 (55.4%)100 (44.6%)109 (56.5%)84 (43.5%)15 (48.5%)16 (51.5%)0.439ECOG PS< 2≥ 2186 (83%)38 (17%)163 (84.5%)30 (15.5%)23 (74.2%)8 (25.8%)0.195 Laboratory Parameters [Median (IQR)] Lymphocyte (109/L)1.52 (1.08)1.53 (1.11)1.47 (1.09)0.080Neutrophil3.83 (2.47)3.25 (2.57)3.80 (1.91)0.222Platelet (109/L)273.5 (168.2)274 (171.5)265 (150)0.726NLR2.28 (1.78)2.25 (1.53)3.3 (3.67)0.003PLR176.5 (119.6)175 (103.1)219.7 (261.4)0.015SII595.8 (506.1)578.8 (478.2)754.1 (1165)0.027Abbreviations: DLBCL, Diffuse large B-cell lymphoma; IQR, interquartile ranges, n, number; NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; SII, systemic immune-inflammation index, VTE; ECOG PS, Eastern Oncology Cooperative Group Performance Status Table 1Baseline characteristics of DLBCL patients with and without VTECharacteristicsAll Subjectn= 224Presence of VTEP-valueNon = 193 (86%)Yesn = 31(14%)Age [Median (IQR)]55 (30)55 (33)56 (24)0.720Sex [n (%)]MaleFemale128 (57%)96 (43%)115 (60%)78 (40%)13 (42%)18 (58%)0.079BMI [Median (IQR)]27.5 (8.2)27.5 (7.65)24.2 (12.8)0.388Ann Arbor Stage [n (%)]I-IIIII-IV124 (55.4%)100 (44.6%)109 (56.5%)84 (43.5%)15 (48.5%)16 (51.5%)0.439ECOG PS< 2≥ 2186 (83%)38 (17%)163 (84.5%)30 (15.5%)23 (74.2%)8 (25.8%)0.195 Laboratory Parameters [Median (IQR)] Lymphocyte (109/L)1.52 (1.08)1.53 (1.11)1.47 (1.09)0.080Neutrophil3.83 (2.47)3.25 (2.57)3.80 (1.91)0.222Platelet (109/L)273.5 (168.2)274 (171.5)265 (150)0.726NLR2.28 (1.78)2.25 (1.53)3.3 (3.67)0.003PLR176.5 (119.6)175 (103.1)219.7 (261.4)0.015SII595.8 (506.1)578.8 (478.2)754.1 (1165)0.027Abbreviations: DLBCL, Diffuse large B-cell lymphoma; IQR, interquartile ranges, n, number; NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; SII, systemic immune-inflammation index, VTE; ECOG PS, Eastern Oncology Cooperative Group Performance Status Baseline characteristics of DLBCL patients with and without VTE Sex [n (%)] Male Female 128 (57%) 96 (43%) 115 (60%) 78 (40%) 13 (42%) 18 (58%) Ann Arbor Stage [n (%)] I-II III-IV 124 (55.4%) 100 (44.6%) 109 (56.5%) 84 (43.5%) 15 (48.5%) 16 (51.5%) ECOG PS < 2 ≥ 2 186 (83%) 38 (17%) 163 (84.5%) 30 (15.5%) 23 (74.2%) 8 (25.8%) Abbreviations: DLBCL, Diffuse large B-cell lymphoma; IQR, interquartile ranges, n, number; NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; SII, systemic immune-inflammation index, VTE; ECOG PS, Eastern Oncology Cooperative Group Performance Status First-line therapy was predominantly anthracycline-based, with 93.4% of the cohort receiving R-CHOP (Rituximab, Cyclophosphamide, Doxorubicin, Vincristine, and Prednisone). The remaining 6.6% of patients received alternative treatment regimens, which included DA-EPOCH, non-anthracycline protocols, or localised radiotherapy. For hospitalised patients, primary prophylaxis with low-molecular-weight heparin (LMWH) was the standard of care, unless patients showed a high risk of bleeding or their platelet count was less than 60 × 10⁹/L. Routine prophylaxis for outpatients is not standard and follows risk assessment scores (Khorana or Caprini). The majority of VTE events occurred during active treatment (58%), with 6.4% of these cases belonging to the alternative regimen group. However, 42% of VTE cases were diagnosed prior to the initiation of treatment. One instance of suspected catheter-associated thrombosis was specifically identified in a patient undergoing the DA-EPOCH regimen. Following first-line therapy, complete remission (CR) rate was 56.2%, with an additional 25.3% achieving partial remission (PR). At the end of the study period, 66% (n = 148) of the participants were alive. The mortality rate was recorded at 15.6% (n = 35), while 18.3% (n = 41) were lost to follow-up. Optimal cut-off analysis The optimal cut-off values determined by ROC analysis were > 2.41 for NLR, > 185 for PLR, and > 656 for SII (Table 2). The area under the curve (AUC) for NLR, PLR, and SII were 66%, 63%, and 62%, respectively, as shown in Fig. 1. Accordingly, patients were classified into two groups (low and high) based on the cut-off value of each parameter. Table 2Receiver-operating characteristic (ROC) curve of NLR, PLR, and SIITest Result VariablesSensitivity %Specificity %95% CICut-offP-value NLR 68630.66 ± 0.060.540 to 0.785> 2.410.003 PLR 71670.63 ± 0.060.511 to 0.758> 1850.016 SII 64550.62 ± 0.060.506 to 0.740> 6560.027Abbreviations: CI, confidence interval; NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; SII, systemic immune-inflammation indexa ROC curve analysis (Wilson/Brown method)b Area (Mean± Stander Error) of venous thromboembolism in diffuse large B-cell lymphoma patients Table 2Receiver-operating characteristic (ROC) curve of NLR, PLR, and SIITest Result VariablesSensitivity %Specificity %95% CICut-offP-value NLR 68630.66 ± 0.060.540 to 0.785> 2.410.003 PLR 71670.63 ± 0.060.511 to 0.758> 1850.016 SII 64550.62 ± 0.060.506 to 0.740> 6560.027Abbreviations: CI, confidence interval; NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; SII, systemic immune-inflammation indexa ROC curve analysis (Wilson/Brown method)b Area (Mean± Stander Error) of venous thromboembolism in diffuse large B-cell lymphoma patients Receiver-operating characteristic (ROC) curve of NLR, PLR, and SII Abbreviations: CI, confidence interval; NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; SII, systemic immune-inflammation index a ROC curve analysis (Wilson/Brown method) b Area (Mean± Stander Error) of venous thromboembolism in diffuse large B-cell lymphoma patients Fig. 1Receiver-operating characteristic (ROC) curve of NLR, PLR, and SII. Abbreviations: NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; SII, systemic immune-inflammation index Receiver-operating characteristic (ROC) curve of NLR, PLR, and SII. Abbreviations: NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; SII, systemic immune-inflammation index Association of NLR, PLR, and SII with VTE To identify the factors that may affect VTE development, we calculated the crude odds ratio (OR) following a logistic regression analysis of NLR, utilizing the cut-off value derived from the ROC curve. The analyses were adjusted for age and tumor stage (III/IV) through a multivariable logistic regression model and were reported as the adjusted odds ratio (ORa) (Table 3). According to the multivariate logistic regression analysis, only NLR and PLR were significantly elevated in patients who developed VTE compared to controls. Table 3Univariate and multivariate logistic regression model for the association between VTE and NLR, PLR, and SII in DLBCL patientsIndicatorsCrude odds ratio (OR)95% CIP-valueAdjusted odds ratio (ORa)95% CIP-valueNLR3.6081.645 to 8.4080.0013.7501.698 to 8.8060.001PLR4.9272.209 to 11.86< 0.0005.2112.316 to 12.67< 0.001SII2.2151.023 to 5.0250.0482.0200.922 to 4.6250.079Abbreviations: CI, confidence interval; DLBCL, Diffuse large B-cell lymphoma; NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; SII, systemic immune-inflammation index, VTE; Venous thromboembolism Table 3Univariate and multivariate logistic regression model for the association between VTE and NLR, PLR, and SII in DLBCL patientsIndicatorsCrude odds ratio (OR)95% CIP-valueAdjusted odds ratio (ORa)95% CIP-valueNLR3.6081.645 to 8.4080.0013.7501.698 to 8.8060.001PLR4.9272.209 to 11.86< 0.0005.2112.316 to 12.67< 0.001SII2.2151.023 to 5.0250.0482.0200.922 to 4.6250.079Abbreviations: CI, confidence interval; DLBCL, Diffuse large B-cell lymphoma; NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; SII, systemic immune-inflammation index, VTE; Venous thromboembolism Univariate and multivariate logistic regression model for the association between VTE and NLR, PLR, and SII in DLBCL patients Abbreviations: CI, confidence interval; DLBCL, Diffuse large B-cell lymphoma; NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; SII, systemic immune-inflammation index, VTE; Venous thromboembolism Discussion and conclusion Identifying risk factors for venous thromboembolism is crucial in managing individuals at risk. The primary objective of this study was to determine whether there is a predictive association between inflammatory markers, specifically NLR, PLR, and SII, and the future occurrence of VTE in patients with DLBCL. Our analysis demonstrated that NLR, PLR, and SII were significantly elevated in VTE patients compared to the control group, while no statistical differences were noted concerning clinical characteristics. Furthermore, the ROC curve analysis indicated acceptable specificity and sensitivity of these inflammatory markers in identifying DLBCL patients with VTE. Applying the derived cut-off values, the multivariate regression analysis revealed that only NLR and PLR were independently associated with VTE. Our findings conclude that NLR and PLR may serve as valuable predictors of potential VTE in patients with DLBCL. Simple inflammatory markers, including NLR, PLR, and SII, can be easily calculated from the routinely performed CBC test. Research indicates that patients with non-Hodgkin lymphoma have a significantly higher rate of VTE events compared to those with Hodgkin lymphoma, with estimates for DLBCL ranging from 11.1% to 14.5%. Our data reveal that 13.8% of DLBCL patients experienced VTE, supporting the notion that aggressive lymphoma subtypes are risk factors for thrombosis [5–7, 24–26]. Furthermore, the elevated levels of NLR, PLR, and SII in our VTE cohort are consistent with the wider literature, reinforcing the strong association between thrombosis and inflammatory status [8–11, 13–16, 19–23, 27]. There are variations in the literature regarding the absolute counts of lymphocytes, neutrophils, and platelets between groups with VTE and those without [13–16, 19–23, 27]. Song et al. [28] suggested that inflammatory markers are superior to absolute WBC count in predicting the clinical outcomes of patients with stroke. Moreover, Otasevic et al. [16] identified NLR, but not PLR, as an independent risk factor for VTE in lymphoma patients. Conversely, Artoni et al. [29] found no association between NLR and PLR and the risk of developing VTE in lymphoma patients. However, similar to our findings, El-Ashwah et al. [15] identified NLR and PLR as independent predictive factors of VTE. In their study population, they focused exclusively on NHL, where DLBCL accounts for 92% of the cohort. Aggressive types of lymphoma are known to produce elevated levels of pro-inflammatory cytokines, which exacerbate systemic inflammation by recruiting and activating neutrophils and platelets [10, 30]. Therefore, the predictive values of these inflammatory markers may differ based on lymphoma histology. Additionally, the interaction between neutrophils and platelets during inflammation is crucial for the formation of VTE. Neutrophils are responsible for initiating thrombus formation during the inflammatory process, while platelets contribute to the propagation of thrombosis [31, 32]. Briefly, platelets adhere to the endothelium and interact with neutrophils to initiate clot formation. This interaction is facilitated by the binding of P-selectin glycoprotein ligand-1 (PSGL-1) on neutrophils to the P-selectin receptor on platelets and endothelial cells, prompting platelet recruitment and enhancing thrombosis [33]. Furthermore, neutrophils extrude neutrophil extracellular traps (NETs), which trigger platelet aggregation and create a scaffold for clot formation [34]. The activated platelets release prothrombotic mediators, including ADP, PF4, and thromboxane A2, which further enhance platelet recruitment and promote clot formation [35]. Furthermore, a reduced lymphocyte count reflects a dysregulated inflammatory response that can promote thrombosis [31]. Higher values of the NLR and PLR may indicate the formation of VTE, as they represent the relationship between lymphocytes and neutrophils or platelets. In our study, 42% of VTE events occurred prior to the treatment initiation. This suggests that a high tumor burden and presence of inflammatory cytokines in newly diagnosed DLBCL may contribute to a hypercoagulable state that exists before the thrombogenic effects of chemotherapy manifest. A meta-analysis concluded that NLR could serve as a diagnostic factor for detecting potential VTE with moderate accuracy [27]. Likewise, elevated NLR and PLR have been linked to poor outcomes in various hematological malignancies, including DLBCL [14, 16, 36, 37]. SII represents a more comprehensive view of an individual’s thrombosis and inflammation state, and elevated levels have been associated with thromboembolic events [19–23]. In our study, SII was significantly higher in DLBCL patients with VTE compared to those without. Other studies reported similar findings, indicating that SII levels were higher in the thrombotic group compared to the control group [19–21]. Moreover, SII has been demonstrated to be a useful indicator for predicting clinical outcomes in VTE patients, with some studies suggesting it is a more specific inflammatory biomarker than the NLR and PLR [19]. Peng et al. [21] conducted a multivariate analysis showing that SII was an independent predictor of VTE risk following hip fracture in the elderly. In our study, univariate analysis revealed that SII was significantly higher in the VTE group; however, no significant association was observed in the multivariate analysis. Further studies need to confirm whether high SII values serve as risk factors for VTE, particularly in patients with DLBCL. Several limitations may affect our study, including its retrospective nature, single-center design, and the relatively small sample size due to the rarity of DLBCL. However, our cohort reflects specific treatment practices such as high adherence to R-CHOP and risk-stratified thromboprophylaxis guidelines that are consistent with international standards of care. Furthermore, the ethnicity of our population, primarily Arab, provides valuable regional data that is often underrepresented in large Western registries, though it may limit the direct application of these inflammatory thresholds to other ethnic groups. Additionally, numerous underlying conditions can influence systemic inflammation markers, potentially affecting the results and subsequent conclusions. Therefore, multicenter prospective studies are recommended to verify the clinical value of NLR, PLR, and SII in comparison, or in combination with other standard inflammatory markers to refine VTE risk stratification in aggressive lymphoma across diverse populations. In conclusion, systemic inflammation significantly contributes to the onset of venous thromboembolism in lymphoma settings. Simple and easily accessible inflammatory parameters such as NLR and PLR could aid the assessment of VTE risk in DLBCL patients and guide treatment strategies. Further prospective research is required to standardise the cut-offs for these inflammatory markers and validate their predictive value by utilising time-to-event and competing risk methodologies in larger cohort studies.