Predictive nomogram for ischemic stroke risk in clear cell renal cell carcinoma patients Authors: Jie Wen, Yi Rong, Yinbo Kang, Dingyang Lv, Fan Cui, Huiyu Zhou, Mohan Jia, Qiwei Wang, Weibing Shuang Published: 2024 ABSTRACT Clear cell renal cell carcinoma (ccRCC) and ischemic stroke are critical global health challenges with a notable association. This study explores the correlation between tumor-related factors and ischemic stroke risk, aiming to construct a predictive nomogram model for ischemic stroke in ccRCC patients. We retrospectively analyzed data from ccRCC patients who underwent nephrectomy at the First Hospital of Shanxi Medical University between January 1, 2013, and May 31, 2022. The data were randomly divided into a training cohort (70%) and a validation cohort (30%). Predictive factors were identified using univariate logistic regression, least absolute shrinkage and selection operator regression, and multivariate logistic regression. A nomogram and a Shiny local calculator were developed using these predictors. We identified six predictors for the nomogram: WHO/ISUP grade, diabetes, hypertension, LDL-C, age, and D-dimer. The nomogram showed good discrimination, with an area under the ROC curve of 0.816 in the training cohort and 0.775 in the validation cohort. The optimal cutoff value was 53.7%. The model demonstrated excellent calibration and clinical applicability. WHO/ISUP grade correlates with ischemic stroke risk, offering insights into cancer-related ischemic stroke mechanisms. This nomogram aids in identifying high-risk individuals among ccRCC patients, facilitating early management and improved outcomes.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-024-82072-9. Introduction Renal Cell Carcinoma (RCC) stands as the most prevalent malignant tumor affecting the kidney, accounting for approximately 85% of all renal tumors1,2. Primarily arising in the tubular epithelial cells of the kidney, RCC encompasses three main histological subtypes: clear cell carcinoma (ccRCC; 70%), papillary renal cell carcinoma (pRCC; 10–15%), and chromophobe renal cell carcinoma (5%)3. Over the past decade, the incidence of RCC has gradually risen due to increased utilization of imaging examinations and improved detection rates for asymptomatic renal masses4–6. In 2020 alone, there were approximately 431,288 new cases of kidney cancer worldwide, constituting 2.2% of all newly diagnosed cancer cases. And kidney cancer claimed the lives of approximately 179,318 individuals, accounting for 1.8% of all cancer-related deaths7. This substantial burden underscores the global impact of kidney cancer8. Numerous studies have established a link between cancer and ischemic stroke9–13. Notably, around 14.6% of cancer patients experience ischemic stroke14. Case reports have also documented instances where RCC directly contributes to ischemic stroke15–18. Subah’s investigation further substantiated this association, revealing a significant correlation between RCC and ischemic stroke19. Ischemic stroke, also known as cerebral infarction, typically results from cerebral artery obstruction or interruption of blood flow, leading to brain tissue hypoxia and subsequent neurological damage. Globally, ischemic stroke constitutes the most prevalent form of stroke, with far-reaching consequences for long-term neurological and psychiatric health. Patients affected by ischemic stroke face diminished quality of life, placing substantial burdens on both families and society20–22. Despite these observations, the precise mechanism linking RCC to stroke remains elusive. Consequently, early clinical prevention and management of ischemic stroke in RCC patients pose significant challenges, profoundly impacting their health and well-being. The occurrence of ischemic stroke in RCC patients is influenced by various factors, including age, hypertension, diabetes, atrial fibrillation, lipid abnormalities, and coagulation disorders23. While prediction models for ischemic stroke have been developed24–27, none specifically address cancer patients. The accuracy and applicability of existing prediction models for RCC-related ischemic stroke remain uncertain and warrant further investigation. Therefore, it becomes imperative to explore the correlation between ischemic stroke and tumor-related factors such as tumor size, grade, and stage. Integrating statistically significant indicators into prediction models can enhance the ability to predict RCC-related ischemic stroke risk. Among grading systems for renal cell carcinoma, the WHO/ISUP classification is widely recognized28 and holds significant importance in the clinical management, risk stratification, and prognosis of patients with ccRCC29. In light of this, our retrospective study focuses exclusively on ccRCC patients, aiming to elucidate the relationship between tumor-related factors and ischemic stroke. Additionally, we construct a nomogram that combines tumor-related variables with traditional risk factors to predict ischemic stroke risk in ccRCC patients. This approach facilitates early identification of high-risk groups and enables targeted management strategies to improve outcomes. Materials and methods Study population We conducted this study using a cohort of patients with unilateral Renal Cell Carcinoma (RCC) who were hospitalized in the Department of Urology at the First Hospital of Shanxi Medical University between January 1, 2013, and May 31, 2022. The inclusion criteria were as follows: (1) Age ≥ 18 years; (2) Patients underwent radical nephrectomy (RN) or partial nephrectomy (PN); (3) Postoperative pathology confirmed as ccRCC; (4) No history of ischemic stroke before the first diagnosis of renal mass; (5) Complete clinical data; Exclusion criteria: (1) Missing follow-up data; (2) Combined with other malignant tumors; Ultimately, a total of 915 patients met the inclusion criteria and were included in this study (Fig. 1). This study was approved by the Ethics Committee of the First Hospital of Shanxi Medical University. Requirement for informed consent was waived by the approval body since the study was anonymous, involved very minimal risk to the subjects, did not include intentional deception, and did not involve sensitive populations or topics; this waiver does not adversely affect the rights and welfare of the subjects. All the methods were performed in accordance with the relevant guidelines and regulations. Fig. 1The process of patient selection for the retrospective study. Abbreviations: ccRCC Clear cell renal cell carcinoma. The process of patient selection for the retrospective study. Abbreviations: ccRCC Clear cell renal cell carcinoma. Data collection We collected clinical characteristics by reviewing patient records.The variables included gender, body mass index (BMI), age, smoking history, cardiovascular disease, hypertension, diabetes, tumor size, tumor laterality, WHO/ISUP grade, T stage, extent of resection (RN and PN), type of surgery (open surgery, laparoscopic surgery, robot-assisted surgery), and preoperative laboratory indicators (D-dimer, fibrinogen, Total Cholesterol [TC], Triglyceride [TG], High-Density Lipoprotein Cholesterol [HDL-C], and Low-Density Lipoprotein Cholesterol [LDL-C]). Patients who had smoked more than 100 cigarettes in their lifetime were considered to have a smoking history30. Cardiovascular disease encompassed atrial fibrillation, coronary atherosclerotic heart disease, and myocardial infarction. T stage classification followed guidelines established by the American Joint Committee on Cancer31. The WHO/ISUP grade, established by the International Society of Urological Pathology, defines grades 2 and 3 with relatively similar criteria. Consequently, this study combined grades 2 and 3 into a single group to improve model fit.For some patients, pathology reports included only the Fuhrman grade, without the WHO/ISUP grade. To address this limitation, two associate senior pathologists jointly re-evaluated these patients’ pathological results to assign the WHO/ISUP grade. As per the 2023 Chinese Blood Lipid Management Guidelines, the recommended cutoff values for TC, TG, HDL-C and LDL-C are 5.2 mmol/L, 1.7 mmol/L, 1.0 mmol/L, and 3.4 mmol/L, respectively32. Follow-up We obtained follow-up data by reviewing inpatient and outpatient medical records and conducting telephone interviews with patients and their relatives. The follow-up assessment included whether the patient had been diagnosed with ischemic stroke. Diagnostic criteria for ischemic stroke were based on cerebral infarction confirmed by computed tomography (CT) or magnetic resonance imaging (MRI) within 1 year of the initial renal mass diagnosis33.The outcome of this study was the occurrence of ischemic stroke within one year following the initial diagnosis of a renal mass. Only binary data on the presence or absence of ischemic stroke were recorded, without analysis of the exact timing of stroke events. Diagnosis of ischemic stroke was based solely on imaging, thereby including asymptomatic strokes as positive outcomes. Statistical analysis We conducted statistical analyses using R (version 4.3.2). Continuous variables were expressed as mean ± standard deviation (SD), and group differences were assessed using Student’s t-test. Categorical variables were presented as numbers (percentages), and differences between groups were evaluated using the chi-square test. For continuous variables with extreme distributions, we applied a logarithmic transformation (log10N). To develop and validate the nomogram, we randomly divided the entire dataset into a training cohort (70% of the data) and a validation cohort (30% of the data). Initially, univariate logistic regression was employed to screen variables, followed by the least absolute shrinkage and selection operator (LASSO) regression model to identify the best predictor variables in the training cohort. Subsequently, we incorporated the best predictor variables into a multivariate logistic regression model using the backward selection method. The statistically significant predictor variables identified from the multivariate logistic regression results were then utilized to construct the nomogram.To facilitate clinical use, we also developed a Shiny local calculator for rapid and accurate assessment of ischemic stroke risk in ccRCC patients. We quantified the discriminative ability of the nomogram using the area under the receiver operating characteristic (ROC) curve (AUC) and the C-index. The optimal cutoff value for clinical application was determined using the highest Youden index. Calibration curves and the Hosmer-Lemeshow test assessed the consistency between actual results and predicted probabilities. Additionally, clinical DCA evaluated the clinical efficacy of the nomogram. We considered statistical significance at P < 0.05. Results Patient characteristics Between January 1, 2013, and May 31, 2022, a total of 991 patients were enrolled in this study. After excluding patients with other tumors (n = 34) and those with follow-up failure (n = 42), 915 patients with ccRCC were included for statistical analysis. The cohort was divided into a training group (640 patients) and a validation group (275 patients) (Fig. 1). Due to the extreme distribution of D-dimer values and substantial data variance, we logarithmically transformed D-dimer (lgD-dimer) for subsequent statistical analyses. No significant differences were observed in various indicators between the two cohorts (P > 0.05) (Table 1). Table 1Characteristics of patients in the training cohort and validation cohorts.VariablesAll patientsTraining cohortValidation cohortP-value(N = 915)(N = 640)(N = 275)GenderFemale313 (34.2%)212 (33.1%)101 (36.7%)0.292Male602 (65.8%)428 (66.9%)174 (63.3%)LateralityLeft443 (48.4%)305 (47.7%)138 (50.2%)0.483Right472 (51.6%)335 (52.3%)137 (49.8%)T stage1796 (87.0%)555 (86.7%)241 (87.6%)0.370287 (9.5%)59 (9.2%)28 (10.2%)327 (3.0%)21 (3.3%)6 (2.2%)45 (0.5%)5 (0.8%)0 (0%)WHO/ISUP1154 (16.8%)110 (17.2%)44 (16.0%)0.3582 & 3730 (79.8%)505(78.9%)225 (81.8%)431 (3.4%)25 (3.9%)6 (2.2%)Extent of resectionPN300 (32.8%)203 (31.7%)97 (35.3%)0.294RN615 (67.2%)437 (68.3%)178 (64.7%)Type of surgeryOpen105 (11.5%)72 (11.3%)33 (12.0%)0.105Laparoscopy736 (80.4%)524 (81.9%)212 (77.1%)Robotic-assisted74 (8.1%)44 (6.9%)30 (10.9%)Smoking historyNo656 (71.7%)460 (71.9%)196 (71.3%)0.853Yes259 (28.3%)180 (28.1%)79 (28.7%)DiabetesNo777 (84.9%)542 (84.7%)235 (85.5%)0.766Yes138 (15.1%)98 (15.3%)40 (14.5%)Cardiovascular diseasesNo836 (91.4%)587 (91.7%)249 (90.5%)0.562Yes79 (8.6%)53 (8.3%)26 (9.5%)HypertensionNo535 (58.5%)371 (58.0%)164 (59.6%)0.639Yes380 (41.5%)269 (42.0%)111 (40.4%)TC<5.2mmol/L713 (77.9%)496 (77.5%)217 (78.9%)0.638≥ 5.2mmol/L202 (22.1%)144 (22.5%)58 (21.1%)TG<1.7mmol/L563 (61.5%)397 (62.0%)166 (60.4%)0.635≥ 1.7mmol/L352 (38.5%)243 (38.0%)109 (39.6%)HDL-C≥ 1.0mmol/L548 (59.9%)385 (60.2%)163 (59.3%)0.803<1.0mmol/L367 (40.1%)255 (39.8%)112 (40.7%)LDL-C<3.4mmol/L726 (79.3%)510 (79.7%)216 (78.5%)0.696≥ 3.4mmol/L189 (20.7%)130 (20.3%)59 (21.5%)Ischemic strokeNo829 (90.6%)579 (90.5%)250 (90.9%)0.834Yes86 (9.4%)61 (9.5%)25 (9.1%)Age57.8 ± 10.757.8 ± 10.758.0 ± 10.80.785BMI25.0 ± 3.1624.8 ± 3.1425.2 ± 3.210.128D-dimer208 ± 334204 ± 316215 ± 3750.680lgD-dimer2.12 ± 0.3872.11 ± 0.3902.13 ± 0.3800.470Fibrinogen3.27 ± 0.9973.24 ± 0.9903.33 ± 1.010.193Tumor size (cm)4.42 ± 2.314.45 ± 2.284.35 ± 2.360.576Continuous variables were analyzed by the Student’s t-test, while categorical variables were analyzed by the Chi-squared test. RN Radical nephrectomy; PN Partial nephrectomy; TC Total cholesterol; TG Triglyceride; HDL-C High-density lipoprotein cholesterol; LDL-C Low-Density Lipoprotein Cholesterol; BMI Body mass index; Table 1Characteristics of patients in the training cohort and validation cohorts.VariablesAll patientsTraining cohortValidation cohortP-value(N = 915)(N = 640)(N = 275)GenderFemale313 (34.2%)212 (33.1%)101 (36.7%)0.292Male602 (65.8%)428 (66.9%)174 (63.3%)LateralityLeft443 (48.4%)305 (47.7%)138 (50.2%)0.483Right472 (51.6%)335 (52.3%)137 (49.8%)T stage1796 (87.0%)555 (86.7%)241 (87.6%)0.370287 (9.5%)59 (9.2%)28 (10.2%)327 (3.0%)21 (3.3%)6 (2.2%)45 (0.5%)5 (0.8%)0 (0%)WHO/ISUP1154 (16.8%)110 (17.2%)44 (16.0%)0.3582 & 3730 (79.8%)505(78.9%)225 (81.8%)431 (3.4%)25 (3.9%)6 (2.2%)Extent of resectionPN300 (32.8%)203 (31.7%)97 (35.3%)0.294RN615 (67.2%)437 (68.3%)178 (64.7%)Type of surgeryOpen105 (11.5%)72 (11.3%)33 (12.0%)0.105Laparoscopy736 (80.4%)524 (81.9%)212 (77.1%)Robotic-assisted74 (8.1%)44 (6.9%)30 (10.9%)Smoking historyNo656 (71.7%)460 (71.9%)196 (71.3%)0.853Yes259 (28.3%)180 (28.1%)79 (28.7%)DiabetesNo777 (84.9%)542 (84.7%)235 (85.5%)0.766Yes138 (15.1%)98 (15.3%)40 (14.5%)Cardiovascular diseasesNo836 (91.4%)587 (91.7%)249 (90.5%)0.562Yes79 (8.6%)53 (8.3%)26 (9.5%)HypertensionNo535 (58.5%)371 (58.0%)164 (59.6%)0.639Yes380 (41.5%)269 (42.0%)111 (40.4%)TC<5.2mmol/L713 (77.9%)496 (77.5%)217 (78.9%)0.638≥ 5.2mmol/L202 (22.1%)144 (22.5%)58 (21.1%)TG<1.7mmol/L563 (61.5%)397 (62.0%)166 (60.4%)0.635≥ 1.7mmol/L352 (38.5%)243 (38.0%)109 (39.6%)HDL-C≥ 1.0mmol/L548 (59.9%)385 (60.2%)163 (59.3%)0.803<1.0mmol/L367 (40.1%)255 (39.8%)112 (40.7%)LDL-C<3.4mmol/L726 (79.3%)510 (79.7%)216 (78.5%)0.696≥ 3.4mmol/L189 (20.7%)130 (20.3%)59 (21.5%)Ischemic strokeNo829 (90.6%)579 (90.5%)250 (90.9%)0.834Yes86 (9.4%)61 (9.5%)25 (9.1%)Age57.8 ± 10.757.8 ± 10.758.0 ± 10.80.785BMI25.0 ± 3.1624.8 ± 3.1425.2 ± 3.210.128D-dimer208 ± 334204 ± 316215 ± 3750.680lgD-dimer2.12 ± 0.3872.11 ± 0.3902.13 ± 0.3800.470Fibrinogen3.27 ± 0.9973.24 ± 0.9903.33 ± 1.010.193Tumor size (cm)4.42 ± 2.314.45 ± 2.284.35 ± 2.360.576Continuous variables were analyzed by the Student’s t-test, while categorical variables were analyzed by the Chi-squared test. RN Radical nephrectomy; PN Partial nephrectomy; TC Total cholesterol; TG Triglyceride; HDL-C High-density lipoprotein cholesterol; LDL-C Low-Density Lipoprotein Cholesterol; BMI Body mass index; Characteristics of patients in the training cohort and validation cohorts. Continuous variables were analyzed by the Student’s t-test, while categorical variables were analyzed by the Chi-squared test. RN Radical nephrectomy; PN Partial nephrectomy; TC Total cholesterol; TG Triglyceride; HDL-C High-density lipoprotein cholesterol; LDL-C Low-Density Lipoprotein Cholesterol; BMI Body mass index; Univariate logistic regression and LASSO regression analysis In the training cohort, univariate logistic regression analysis revealed significant associations between the risk of ischemic stroke and 7 variables: WHO/ISUP grade, diabetes, hypertension, TC, LDL-C, age, and lgD-dimer (P < 0.05) (Table 2). Among the ccRCC patients in the training cohort, 61 experienced ischemic stroke. Considering sample size requirements, up to 6 variables were allowed to enter multivariate logistic regression34. Therefore, we employed LASSO regression model to further identify the best predictor variables. TC was excluded based on the minimum Lambda value criterion (Fig. 2). Table 2The result of univariate logistic regression.VariablesORCI P Gender Male vs. Female0.7410.431–1.2760.279Laterality Right vs. Left1.2520.735–2.1340.409T stage 2 vs. 10.6750.236–1.9330.464T stage 3 vs. 11.5460.441–5.4210.496T stage 4 vs. 1<0.0010-Inf0.984WHO/ISUP 2 & 3 vs. 14.0941.256–13.3490.019WHO/ISUP 4 vs. 111.2632.59-48.9860.001Extent of resection RN vs. PN1.1220.629-2.0000.697Type of surgery Laparoscopy vs. Open2.5340.771–8.3270.126Type of surgery Robotic-assisted vs. Open3.6320.86-15.3370.079Smoking history Yes vs. No0.9000.495–1.6360.729Diabetes Yes vs. No3.4231.916–6.114<0.001Cardiovascular diseases Yes vs. No2.1040.972–4.5550.059Hypertension Yes vs. No5.9813.170-11.287<0.001TC ≥ 5.2mmol/L vs. <5.2mmol/L1.7901.012–3.1660.045TG ≥ 1.7mmol/L vs. <1.7mmol/L0.7790.445–1.3640.382HDL <1.0mmol/L vs. ≥1.0mmol/L0.9050.526–1.5580.720LDL ≥ 3.4mmol/L vs. <3.4mmol/L2.0801.171–3.6930.012Age1.0761.045–1.108<0.001BMI1.0160.936–1.1030.708lgD-dimer3.5771.852–6.911<0.001Fibrinogen1.2090.953–1.5320.116Tumor size (cm)0.9710.862–1.0950.634RN Radical nephrectomy; PN Partial nephrectomy; TC Total cholesterol; TG Triglyceride; HDL-C High-density lipoprotein cholesterol; LDL-C Low-Density Lipoprotein Cholesterol; BMI Body mass index. The result of univariate logistic regression. RN Radical nephrectomy; PN Partial nephrectomy; TC Total cholesterol; TG Triglyceride; HDL-C High-density lipoprotein cholesterol; LDL-C Low-Density Lipoprotein Cholesterol; BMI Body mass index. Fig. 2Demographic and clinical feature selection through the LASSO regression model. (A) According to the logarithmic (lambda) sequence, a coefficient profile was generated, and non-zero coefficients were produced by the optimal lambda. (B) The optimal parameter (lambda) in the LASSO model was selected via minimum criterion (left vertical line). Demographic and clinical feature selection through the LASSO regression model. (A) According to the logarithmic (lambda) sequence, a coefficient profile was generated, and non-zero coefficients were produced by the optimal lambda. (B) The optimal parameter (lambda) in the LASSO model was selected via minimum criterion (left vertical line). Multivariate logistic regression analysis The meaningful variables identified above were included in multivariate logistic regression analysis using the backward method. The results indicated that diabetes, hypertension, LDL-C, age, and lgD-dimer independently predicted ischemic stroke in ccRCC patients (Fig. 3). Patients with WHO/ISUP grade 4 had a significantly higher risk of ischemic stroke compared to those with WHO/ISUP grade 1 (P = 0.008), making it a valuable independent predictor. Although the risk of ischemic stroke was not significantly increased in patients with WHO/ISUP grade 2&3 compared to grade I (P = 0.091 > 0.05), we retained it in the nomogram model due to its excellent C-index and log-likelihood ratio. Nomogram construction and validation We constructed a nomogram (Fig. 4A) based on the 6 independent predictors mentioned above and developed a Shiny local calculator (Fig. 4B) for clinical use. The nomogram demonstrated good discriminative ability, with an AUC of 0.816 (95% CI, 0.762–0.871) in the training cohort (Fig. 5A) and 0.775 (95% CI, 0.686–0.864) in the validation cohort (Fig. 5B). Using the highest Youden index, the calculated value from logistic regression was 0.149 (Fig. 5A). The optimal cutoff value for clinical application, based on the logistic regression formula, was 53.7%. Individuals with predicted probabilities greater than 53.7% were considered at high risk. The calibration curve of the prediction nomogram demonstrated good accuracy in both the training cohort (Fig. 6A) (Hosmer-Lemeshow test P = 0.561) and the validation cohort (Fig. 6B) (Hosmer-Lemeshow test P = 0.144). Furthermore, the decision curve analysis (DCA) results confirmed the significant clinical utility of the nomogram in identifying ischemic stroke risk in ccRCC patients (Fig. 7). Overall, the nomogram model holds substantial clinical significance for predicting ischemic stroke risk in ccRCC patients. Fig. 3Forest plot of independent predictors of ischemic stroke. Forest plot of independent predictors of ischemic stroke. Fig. 4Nomogram (A) and shiny local calculator (B) to predict ischemic stroke risk in ccRCC patients. Nomogram (A) and shiny local calculator (B) to predict ischemic stroke risk in ccRCC patients. Fig. 5ROC curves for predicting ischemic stroke among ccRCC patients in the training cohort (A) and the validation cohort (B). ROC curves for predicting ischemic stroke among ccRCC patients in the training cohort (A) and the validation cohort (B). Fig. 6Calibration curves of the nomogram prediction in the training cohort (A) and the validation cohort (B). Calibration curves of the nomogram prediction in the training cohort (A) and the validation cohort (B). Fig. 7DCA of the nomogram prediction in the training cohort (A) and the validation cohort (B). DCA of the nomogram prediction in the training cohort (A) and the validation cohort (B). Discussion To date, there remains a dearth of satisfactory predictive tools for identifying individuals at high risk of ischemic stroke among patients with ccRCC. In response, we developed a nomogram model and a Shiny local calculator utilizing clinical data and follow-up information collected by the First Hospital of Shanxi Medical University. This model effectively estimates the probability of ischemic stroke within this patient population. The nomogram comprises six predictors: WHO/ISUP grade, diabetes, hypertension, LDL-C levels, age, and lgD-dimer. Notably, the model demonstrated robust discriminative ability in both the training and validation cohorts, achieving AUC values of 0.816 and 0.764, respectively. Furthermore, calibration curves and the Hosmer-Lemeshow test confirmed the model’s accuracy across both cohorts. Additionally, DCA underscored the model’s clinical applicability in both groups. In our study, we observed a positive correlation between WHO/ISUP grade and the risk of ischemic stroke in ccRCC patients. However, tumor size and tumor stage did not exhibit statistically significant associations. These findings suggest that cancer-related ischemic stroke is intricately linked to the degree of tumor cell differentiation but lacks a substantial correlation with tumor size or invasion range. It is plausible that certain substances secreted by poorly differentiated tumor cells contribute to the occurrence of ischemic stroke, opening new avenues for investigating the underlying mechanisms of cancer-related ischemic stroke. The pathogenesis of cancer-related ischemic stroke remains incompletely understood. Currently recognized mechanisms include tumor embolism, nonbacterial thrombotic endocarditis (NBTE), and hypercoagulable states. Tumor embolism refers to the direct embolization of tumor fragments into blood vessels, and numerous case reports document ischemic stroke resulting from tumor embolism35.36. NBTE, often considered the predominant cause of ischemic stroke in malignancy patients, manifests as vegetations on heart valves. These vegetations consist of fibrin thrombi devoid of bacterial involvement and frequently co-occur with malignancies37–39. The fragile nature of these vegetations renders them prone to detachment, leading to vascular embolism and subsequent ischemic stroke40,41. Cancer-related hypercoagulability contributes to both venous and arterial thromboembolic events42, which can readily precipitate ischemic stroke. Mechanistically, tumor cells stimulate endothelial cells to secrete procoagulant factors, thereby promoting fibrin coagulation. Additionally, pro-inflammatory factors (such as IL-6 and TNF-α) mediate this process43–46. Furthermore, the inflammatory response induced by tumor cells may facilitate thrombosis by mediating the formation of neutrophil extracellular traps47. Elevated levels of plasminogen activator inhibitors in cancer patients have also been reported, inhibiting plasminogen activation and consequently impairing fibrinolysis48,49. In 1865, Trousseau initially defined thrombophlebitis in the context of cancer as Trousseau syndrome50. Subsequently, in 1977, Sack et al. expanded this definition to encompass various venous and arterial thromboembolic events associated with malignancy51. These events include venous thromboembolism, arterial thrombosis, nonbacterial thromboendocarditis, and acute or chronic disseminated intravascular coagulation. Over time, the concept of Trousseau syndrome has evolved into broader definitions: “hypercoagulable disorder in patients with malignant tumors” or “any type of coagulopathy occurring in any type of malignancy”52,53. Trousseau syndrome is believed to represent the coagulation and fibrinolytic systems’ response to the presence of a tumor. An appropriate level of response can mitigate or even prevent tumor metastasis, while an excessive response may contribute to tumor cell proliferation, metastasis, angiogenesis within tumor tissue, and various thromboembolic events54. Currently, Trousseau syndrome ranks as the second most common cause of death in cancer patients, surpassed only by the cancer itself55. Consequently, early detection and effective management of cancer patients are critical. Compared with other cancers, ischemic stroke associated with RCC also presents some distinct mechanisms. Numerous studies have confirmed that RCC can negatively affect renal function56,57, leading to the accumulation of toxins, which in turn triggers adverse effects such as neuroinflammation, oxidative stress, microvascular alterations, and cerebrovascular ischemia58. Research has also shown that impaired renal function increases the risk of cerebral small vessel disease59–61, suggesting that kidney function-related factors may play a mediating role in RCC-related ischemic stroke. Vitamin D, a fat-soluble vitamin primarily responsible for promoting calcium and phosphorus absorption, has been linked to RCC risk. Wu et al. found an inverse correlation between serum vitamin D levels and vitamin D intake with the risk of RCC62. However, findings on the relationship between vitamin D and ischemic stroke have been inconsistent. For instance, Navale et al., using data from the UK Biobank, suggested that vitamin D deficiency increases the risk of ischemic stroke, though subsequent Mendelian randomization studies failed to achieve statistical significance63. Notably, no clinical trials have demonstrated that vitamin D supplementation can prevent ischemic stroke in the general population. In a study involving 35,517 samples, a 10 nmol/L decrease in vitamin D levels was associated with an odds ratio of 0.98 for ischemic stroke (95% CI, 0.86–1.13), but this result was not statistically significant64. Conversely, other studies have suggested a positive correlation between elevated vitamin D levels and ischemic stroke risk65. Vitamin D can raise serum calcium and phosphorus levels, and current evidence indicates that these minerals may influence ischemic stroke risk66,67. Moreover, serum calcium and phosphorus are regulated by various factors, including parathyroid hormone and calcitonin, complicating the relationship between vitamin D and ischemic stroke. This complexity may explain the significant variability in research outcomes on this topic. Additionally, RCC can induce renal hypertension in some patients68. Tumor mass can compress or invade the renal artery, reducing renal blood flow. In response, juxtaglomerular cells activate the renin-angiotensin system, potentially leading to renal hypertension. Since hypertension is a major risk factor for ischemic stroke69, RCC may contribute to stroke occurrence through its effect on renal hypertension. Our findings align with existing evidence, identifying hypertension, diabetes, LDL-C levels, age, and D-dimer as significant risk factors for ischemic stroke69–73. However, contrary to prevailing studies, this investigation did not yield similar results regarding cardiovascular disease and smoking history74–77. Notably, the cardiovascular disease and smoking history data in our study relied on patients’ subjective descriptions during hospitalization history collection, introducing potential information bias. While some studies have linked fibrinogen, BMI, TC, TG, and HDL-C to ischemic stroke occurrence72,78–83, our study in patients with ccRCC does not yet support these conclusions. Further research, utilizing larger sample sizes, is warranted to validate these associations. This study innovatively explored the association between tumor-related factors and ischemic stroke and combined meaningful factors with traditional risk factors to develop a predictive nomogram. These research findings can aid clinicians in early identification of high-risk groups for ischemic stroke among ccRCC patients, facilitating timely intervention. However, we acknowledge several limitations: firstly, the retrospective nature of this study may introduce patient selection bias, and secondly, our investigation was conducted at a single center with a moderate sample size. Therefore, large-scale, multi-ethnic, and multi-center prospective studies are necessary to confirm the clinical significance of our findings. Conclusion Our investigation reveals a possible association between WHO/ISUP grade and the risk of ischemic stroke. This finding opens new avenues for exploring the underlying mechanisms of cancer-related ischemic stroke in future research. To facilitate clinical decision-making, we developed a nomogram specifically tailored to assess ischemic stroke risk in ccRCC patients. Clinicians can utilize this predictive model to identify high-risk individuals, enabling closer monitoring and more assertive treatment strategies. By doing so, we have the potential to reduce ischemic stroke incidence and significantly improve outcomes for patients with ccRCC. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1