A retrospective analysis was conducted on medical records of lung cancer patients with or without VTE admitted to the Second Xiangya Hospital of Central South University between January 2019 and December 2022. Lung cancer patients were searched in the electronic medical record system according to the tenth revision of the International Classification of Diseases (ICD-10) codes. The diagnosis of lung cancer was pathologically confirmed. The diagnosis of APE was confirmed by computed tomography pulmonary angiography (CTPA) or pulmonary angiography and adhered to the 2019 European Society of Cardiology (ESC) guideline (19). APE was classified as symptomatic if clinical signs (e.g., acute dyspnea, chest pain, syncope) prompted diagnostic imaging, or incidental if discovered on imaging performed for other reasons (e.g., cancer staging). The diagnosis of DVT adhered to the National Institute for Health and Care Excellence (NICE) guideline (20). Control measurements were generated by analysing lung cancer patients with DVT and lung cancer patients without VTE by using Propensity Score Matching (PSM) based on sex and age. Ethical approval was obtained from the Second Xiangya Hospital of Central South University (2022-107).

The inclusion criteria comprised: 1) Confirmation of APE through computed tomography pulmonary angiography (CTPA) or pulmonary angiography; 2) hospitalisation time longer than 24 h.

Exclusion criteria encompassed: 1) Comorbidity of APE and DVT at the current presentation; 2) A history of previous DVT or APE (to minimize confounding effects from prior thrombotic events and their treatments); 3) Having different hospital admission numbers, but actually the same patient; 4) combined with diseases that can cause fluctuations in blood counts such as infections, blood disorders, chronic renal insufficiency, leukaemia, and others. The flowchart was illustrated as Figure 1.

We identified lung cancer cases from the electronic medical record system of the Second Xiangya Hospital of Central South University with ICD-10 codes. Demographic, clinical, and laboratory data were extracted, including age, gender, body mass index (BMI), medical history, vital signs, comorbidities, blood routine, coagulation function, and biochemical indices. Laboratory values, including complete blood count (CBC), coagulation parameters, and biochemical indices (including uric acid), were measured from blood samples collected at the time of hospital admission, prior to the initiation of specific therapeutic interventions for the current event (e.g., anticoagulation for APE). NLR was defined as the absolute neutrophil count divided by the absolute lymphocyte count. PLR was defined as the absolute platelet count divided by the absolute lymphocyte count.

Characteristics were expressed as mean ± standard deviation (SD) or median (P25, P75) for continuous variables and the number of participants (percentage) for categorical variables. The statistical analysis of the differences in groups was performed by t-test or Mann–Whitney U test for continuous data and χ² test for categorical data. The stepwise forward logistic regression model was used for identifying the risk factors of APE. The Cox proportional risk model was used for multivariate and univariate analyses to determine prognostic factors related to the disease, and hazard ratios were reported as relative risks with corresponding 95% confidence intervals. The receiver operating characteristic (ROC) curve analysis was used to determine the optimal cut-off value of index. The Kaplan–Meier method was used to estimate the proportion whether patients is admitted to the ICU unit. In this study, “poor prognosis” was primarily defined by the composite outcome of disease severity leading to ICU admission. All comparisons were employed with a two-sided test through SPSS (Statistical Package for Social Science, Chicago, IL) version 26.0, and a P value less than 0.05 was considered significant. All figures were concluded using the Medcalc version 20.1 software.

The demographical data of the enrolled patients are summarized in Table 1. The APE group patients were an average of 58.50 (54.75, 67.00) years old and composed of a relative even split of males (56.7%) and females (43.3%). Age, sex, BMI, vital signs, and previous history (exclude admitted to ICU, p < 0.001) were similar between three groups. However, the APE group exhibited a higher proportion of shortness of breath (46.7% vs. 13.3%, p = 0.010), chest distress (20% vs. 0, p = 0.024) and lower proportion of legs pain (0 vs. 13.3% and 0, p = 0.032) levels when compared to the Non-VTE group. When the APE group compare to DVT group, it has higher proportion of surgical treatment (46.7% vs. 13.3%, p = 0.005) and lower proportion of chemotherapy (13.3% vs. 40%, p = 0.020).

The clinical indices of the enrolled patients are summarized in Table 2. Statistically significant differences were observed in white blood cell (WBC, p = 0.028), red blood cell (RBC, p = 0.025), neutrophil (N, p = 0.011), NLR (p < 0.037), PLR (p < 0.001), uric acid (UA, p = 0.020), and estimated glomerular filtration rate (eGFR, p = 0.025) between the APE and DVT groups.

Additionally, statistically significant differences were found in WBC (p < 0.001), RBC (p < 0.001), N (p < 0.001), L (p = 0.026), NLR (p < 0.001), monocyte (M, p < 0.001) count, PLR (p < 0.001), D-Dimer (p < 0.001), prothrombin time (PT, p = 0.003), fibrinogen (FIB, p < 0.001), international standardized ratio (INR, p = 0.003), aspartic transaminase (AST, p = 0.002), and albumin (ALB, p < 0.008) between the DVT and non-VTE groups.

Moreover, statistically significant differences were noted in WBC (p < 0.001), RBC (p = 0.005), platelet (PLT, p = 0.022), N, L (p < 0.001), NLR (p < 0.001), E (p < 0.001), M (p < 0.001), PLR (p < 0.001), D-Dimer (p < 0.001), activated partial thromboplastin time (APTT, p = 0.031), PT (p = 0.031), FIB (p = 0.016), ALT (p = 0.009), AST (p < 0.001), ALB (p < 0.001), and UA (p = 0.001) between the APE and non-VTE groups.

Clinical indices including WBC (p < 0.001), RBC (p < 0.001), N (p < 0.001), L (p < 0.001), NLR (p < 0.001), E, M (p < 0.001), PLR (p < 0.001), D-Dimer (p < 0.001), APTT (p = 0.040), PT (p = 0.009), FIB (p = 0.009), INR (p = 0.012), ALT (p = 0.028), AST (p < 0.001), ALB (p < 0.001), UA (p = 0.005), and eGFR (p = 0.048) showed statistically significant differences among the three groups.

WBC, RBC, N, NLR, and PLR can be found statistically significant in all of two groups comparisons and three groups comparison. CRP levels were not routinely measured in all patients at admission in this retrospective cohort and were therefore not included in the analysis.

Based on the above data and inter-group comparison results, we conducted logistic regression analysis for the three groups. In the logistic regression analysis between APE and DVT groups, the adjusted stepwise forward logistic regression model indicated that the expressed level of neutrophil [odds ratio (OR) = 0.853, 95% CI: 0.741–0.982, p = 0.027], and red blood cell (OR = 0.418, 95% CI: 1.178–0.979, p = 0.045) were independent determinants, even after adjustment for demographics, therapy methods, and laboratory indicators. In the logistic regression analysis between DVT and non-VTE groups, the adjusted stepwise forward logistic regression model indicated that the expressed level of neutrophil (OR = 1.959, 95% CI: 1.102–3.484, p = 0.022), and D-Dimer (OR = 1.301, 95% CI: 1.017–1.664, p = 0.036) were independent determinants, even after adjustment for demographics, therapy methods, and laboratory indicators. In the logistic regression analysis between APE and non-VTE groups, the adjusted stepwise forward logistic regression model indicated that the expressed level of neutrophil (OR = 3.068, 95% CI: 1.472–6.394, p = 0.003) and aspartic transaminase (OR = 1.268, 95% CI: 1.012–1.589, p = 0.039) were independent determinants, even after adjustment for demographics, therapy methods, and laboratory indicators, as shown in Table 3.

The receiver operating characteristic (ROC) curves was constructed for neutrophil count to predict VTE in lung cancer patients, as depicted in Figure 2. Receiver operating characteristic (ROC) curves analyzing the predictive ability of neutrophil count for differentiating between lung cancer patient groups. The area under curve (AUC) values were 0.691 (APE group vs. DVT group), 0.822 (DVT group vs. non-VTE group) and 0.952 (APE group vs. non-VTE group), respectively. The sensitivity and specificity values for APE and DVT groups are 80% and 56.7%. The sensitivity and specificity values for APE and non-VTE groups are 90% and 90%. The sensitivity and specificity values for DVT and non-VTE groups are 80% and 76.7%.

According to the patients whether admitted to ICU, the APE groups was divided into the APE-ICU and the APE-nonICU groups. The demographical data and clinical indices of the enrolled patients were summarized in Tables 4, 5. A statistically significant difference was found in the history of operation (p = 0.007). When comparing the clinical indicators between the two groups, statistically significant differences were observed in PLT (p = 0.015), M (p = 0.023), PLR (p = 0.008), PT (p = 0.032), UA (p < 0.001), and eGFR (p = 0.002). In the multivariate analysis between those two groups (Table 6), the adjusted stepwise forward Cox proportional hazards model revealed that higher expressed levels of UA [hazard ratio (HR): 1.006, 95% confidence interval (CI): 1.000–1.012, p = 0.035] were associated with an increased hazard of lung cancer patients with APE which was independent from age, BMI, and sex. And it (HR: 1.006; 95% CI: 1.000–1.012, p = 0.035) was still shown as an independent risk factor even after adjustment for age, BMI, sex, PLR, PLT, M, PT and eGFR. Furthermore, in the multivariate analysis between APE-ICU and APE-nonICU patients, the adjusted stepwise forward logistic regression model indicated that the expressed level of UA [odds ratio (OR) = 1.017, 95% CI: 1.002–1.033, p = 0.028] and PLR (OR = 0.990, 95% CI: 0.981–0.999, p = 0.037) were associated with the severity of APE, even after adjustment for demographics, PLT, M, PT, and eGFR.

The ROC curves were constructed for UA and PLR to predict severity in lung cancer patients with APE, as depicted in Figuress 3, 4. In Figure 3, ROC curve for Uric Acid (UA) predicting ICU admission (severity) in lung cancer patients with APE (AUC = 0.782, p = 0.001). In Figure 4, ROC curve for Platelet-to-Lymphocyte Ratio (PLR) predicting ICU admission (severity) in lung cancer patients with APE (AUC = 0.792, p = 0.001). The end-off time is defined as the duration from the time of VTE diagnosis to ICU admission. If the patient was diagnosed after entering the ICU, the end-off time was considered as the 0 day. The AUC value for UA was 0.782 and P value was 0.001. The AUC value for PLR was 0.792 and P value was 0.001. Meanwhile, the optimal predictive threshold for the expressed level of UA was determined to be 260.1 µmol/L according to the AUC, yielding a sensitivity of 61.11% and specificity of 83.33%. The optimal predictive threshold for the expressed level of PLR was determined to be 318.83 according to the AUC, yielding a sensitivity of 88.89% and specificity of 66.67%.

Kaplan–Meier analysis was conducted base on different level of PLR (p = 0.005) and UA (p = 0.004), as separately depicted in Figures 5a, b. Kaplan–Meier curves for ICU admission-free survival in lung cancer patients with APE, stratified by (a) Platelet-to-Lymphocyte Ratio (PLR) levels (cut-off 318.83, Log-rank p = 0.005), (b) Uric Acid (UA) levels (cut-off 260.1 µmol/L, Log-rank p = 0.004), and (c) the combination of high UA and low PLR (Log-rank p = 0.009). The graphs are of uniform size, and labels (a, b, c) are positioned above the respective panels. Subsequently, the lung cancer patients with APE of UA (associated criterion > 260.1 µmol/L) were compared to those with PLR (associated criterion ≤ 381.83). The difference in ICU admitted proportion was statistically significant (p = 0.009, Figure 5c).

This study has several limitations. First, the sample size was relatively small, which was attributable to the strict inclusion criteria requiring definitive diagnosis, detailed clinical examination results, and minimization of potential confounding factors in this single-center study. Furthermore, important variables such as detailed cancer stage, specific cancer treatment regimens, and the presence of concurrent DVT at diagnosis could not be fully accounted for in the analysis due to missing data or sample size constraints, which may limit the validity and generalizability of the findings. The lack of data on additional inflammatory markers like CRP is another limitation. In addition, the lack of long-term follow-up limited our ability to evaluate prognostic outcomes over time. Future prospective studies should prioritize larger sample sizes, standardized data collection on cancer characteristics, treatment details, and inflammatory biomarkers, as well as incorporate extended follow-up to improve the generalizability and clinical relevance of the findings.