A Study on Analyzing Risk Factors for Carotid Plaque Instability Using Machine Learning Models

A Study on Analyzing Risk Factors for Carotid Plaque Instability Using Machine Learning Models | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article A Study on Analyzing Risk Factors for Carotid Plaque Instability Using Machine Learning Models Tianwei Zhang, Huibo Ma, Xiaozhi Sun, Guofeng Li, Yongxin Li, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6966994/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 13 You are reading this latest preprint version Abstract Objective: The objective of this study was to develop and evaluate the effectiveness of machine learning (ML) models in predicting carotid plaque instability. Methods: A retrospective analysis was conducted on data from 449 patients with carotid plaques treated at The Affiliated Hospital of Qingdao University between July 2022 and November 2024. Patients were randomly divided into a training set and a testing set at a 7:3 ratio. Five ML algorithms were utilized to establish prediction models. The predictive performance of each model was assessed using the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity on the testing set. Results: Among the five ML algorithms, the light gradient boosting machine (LightGBM) demonstrated the best performance with an AUC of 0.921 (95% confidence interval [CI]: 0.887–0.956) and an accuracy of 0.837 in the training set, and an AUC of 0.864 (95% CI: 0.784–0.944) and an accuracy of 0.811 in the testing set. The feature importance results showed that C-reactive protein (CRP), leukocytes, high-density lipoprotein (HDL), and platelets were the four most significant contributors to carotid plaque instability. Conclusion: We developed and validated the ML models for predicting carotid plaque instability, and identified that LightGBM model with 9 relevant factors had the best performance and high levels of clinical applicability. Carotid plaque instability Machine learning Relevant factors Predictive model Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Background Atherosclerotic carotid plaques constitute a critical pathological substrate for ischemic cerebrovascular events 1 . The 2019 Global Burden of Disease Study revealed that stroke accounted for approximately 6.65 million global deaths, ranking as the second-leading cause of mortality and disability-adjusted life years, with ischemic stroke representing over 80% of total cases 2 . In China, stroke-related hospitalization costs reached US $ 12.2 billion in 2016, imposing substantial socioeconomic and healthcare burdens 3 . Evidence indicates that nearly 30% of ischemic strokes are directly attributable to thromboembolism originating from ruptured carotid atherosclerotic plaques 4 . While current clinical guidelines emphasize luminal stenosis severity for stroke risk stratification, emerging evidence demonstrates that plaque composition and biological activity fundamentally determine vulnerability rather than mere anatomical narrowing 5 – 7 . Clinically, unstable plaques characterized by thin fibrous caps and lipid-rich necrotic cores exhibit markedly higher rupture propensity compared to stable counterparts, frequently resulting in thromboembolic complications and disabling cerebrovascular outcomes 8 . Consequently, early detection of high-risk plaques and targeted mitigation of rupture risk represent urgent clinical priorities to reduce stroke incidence and improve population-level neurological prognosis. The current clinical evaluation of carotid plaque stability mainly depends on advanced imaging techniques like computed tomography angiography (CTA) and contrast-enhanced ultrasonography (CEUS) 9 , 10 . These methods offer detailed insights into plaque characteristics, such as lipid core size, fibrous cap integrity, and intraplaque hemorrhage (IPH). However, their clinical use is limited by several factors. CTA exposes patients to radiation and nephrotoxic contrast agents, while CEUS is operator-dependent and suffers from poor reproducibility 11 . In contrast, routine blood tests, commonly obtained during health screenings, present a promising noninvasive alternative for plaque stability assessment. They are cost-effective, scalable, and allow for ongoing monitoring. However, traditional statistical techniques struggle to account for the complex, nonlinear relationships between various biomarkers, which hampers prediction accuracy. Recently, machine learning algorithms have shown great potential in overcoming these challenges. By integrating high-dimensional data and identifying complex patterns, these algorithms improve cardiovascular risk prediction 12 , 13 . This study introduces a machine learning-based framework that uses accessible blood biomarkers to construct a predictive model for carotid plaque stability. This approach could facilitate early risk identification and personalized preventive measures. Methods Data Source and Study Design We conducted a retrospective study analyzing the clinical data of patients who had carotid plaque between July 2022 and November 2024 at the affiliated hospital of Qingdao University. Unstable plaques were defined according to criteria including irregular surface, ulcerations, low or mixed echogenicity on ultrasound, or recent ischemic events (e.g., TIA or stroke). We included adults aged 18 years and older with carotid atherosclerosis disease, who underwent CTA of the carotid arteries during the study period. Both carotid arteries were considered if plaques were present on both sides, otherwise, only the affected side was included. Patients were divided into training (70%) and testing (30%) sets by stratified random sampling using the Stratified Shuffle-Split function in Python. Exclusion criteria included serious systemic inflammatory diseases, autoimmune disorders, malignancies, recent infections, severe liver or kidney dysfunction, recent cardiovascular events, incomplete data, pregnant or breastfeeding women, and inability to comply with the study protocol. The patient selection flow diagram is shown in Fig. 1 . This retrospective study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Affiliated Hospital of Qingdao University (Protocol ID: QYFY WZLL 27010). Due to the retrospective nature of this study, the requirement for informed consent was waived by the Affiliated Hospital of Qingdao University Institutional Review Board. Clinical trial number: not applicable. Data collection The clinical data on the patients included: age, gender, leukocyte, neutrophil, lymphocyte, monocyte, platelet, hemoglobin, erythrocyte, neutrophil to lymphocyte ratio (NLR), platelet to lymphocyte ratio (PLR), lymphocyte to monocyte ratio (LMR), C-Reactive Protein (CRP), D-dimer, albumin, prealbumin, fibrinogen, alanine aminotransferase (ALT), aspartate aminotransferase (AST), triglyceride, cholesterol, high density lipoprotein, low density lipoprotein, smoke, drink, and hypertension. Feature Selection Firstly, we conducted the statistical Mann–Whitney test and chi-square test for clinical characteristics to determine whether there are differences between the groups. Our analysis considered the feature significant when its two-tailed p-value was p < 0.05. Secondly, spearman correlation analysis was performed to reduce collinearity among features. To reduce the risk of overfitting, the least absolute shrinkage and selection operator (LASSO) algorithm was applied to select features with non-zero coefficient values. Model Building To develop a predictive model for unstable carotid plaque, five machine learning (ML) algorithms were applied: k-nearest neighbor (KNN), random forest (RF), Extra-Trees classifier, eXtreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM). All patients were randomly divided into a training set (70%) and a testing set (30%). The training set was used to build the prediction models, while the testing set was employed to evaluate their performance using the area under the receiver operating characteristic curve (AUC) and corresponding 95% confidence intervals (95% CI). The model with the highest AUC was considered the optimal model. To evaluate the contribution of each variable to the model's prediction, we calculated feature importance scores based on the optimal model. A bar plot was used to visualize the top-ranked features, with higher values indicating greater contribution to the model's predictive performance. Additionally, decision curve analysis (DCA) was conducted to illustrate the net clinical benefit across a range of threshold probabilities. Statistical analyses The Stratified Shuffle-Split function was utilized in Python (version 3.7) to partition the dataset. Continuous variables with an abnormal distribution were described as medians and interquartile ranges, and categorical variables were described as frequencies and proportions. Continuous variables with an abnormal distribution, and categorical variables were analyzed using Mann-Whitney U-test, and Chi-squared test, respectively. Correlation analysis and feature selection using the least absolute shrinkage and selection operator (LASSO) algorithm, were conducted in Python (version 3.7) using the “scipy,” “numpy,” and “sklearn” packages. A bilateral P-value < 0.05 was considered as a measure of statistical significance. Results Patient Characteristics A total of 505 patients with carotid plaque were potentially eligible from the affiliated hospital of Qingdao University between July 2022 and November 2024. After the selecting process, 449 patients were finally enrolled in our study, and 149 of which had unstable carotid plaque. The training set included 215 patients with stable carotid plaque and 99 patients with unstable carotid plaque, while the testing set included 85 patients with stable carotid plaque and 50 patients with unstable carotid plaque (Fig. 1 ). The baseline data of the included patient are shown in Table 1 , which indicated that leukocyte, neutrophil, monocyte, platelet, LMR, CRP, fibrinogen, triglyceride, low density lipoprotein, and high density lipoprotein were significantly different between patients with stable carotid plaque and patients with unstable carotid plaque. The comparison of baseline characteristics between the training and testing sets with corresponding P values was shown in Table 2 . Notably, the incidence of unstable carotid plaque did not differ significantly between the training and testing sets (p = 0.841). Furthermore, all baseline characteristics were statistically insignificant between the two sets, indicating that the training and testing sets were well-balanced. Table 1 Baseline characteristics of the patients. Characteristics Stable carotid plaque Unstable carotid plaque P value Age 67.00 (62.00, 73.00) 67.00 (60.00, 71.00) 0.760 Leukocyte 5.85 (5.13, 6.93) 7.39 (6.21, 8.42) < 0.001 Neutrophil 3.50 (2.84, 4.45) 4.35 (3.44, 5.23) < 0.001 lymphocyte 1.83 (1.42, 2.26) 1.80 (1.45, 2.30) 0.971 Monocyte 0.43 (0.37, 0.60) 0.50 (0.41, 0.61) 0.004 Erythrocyte 4.52 (4.17, 4.83) 4.54 (4.21, 4.93) 0.214 Hemoglobin 138.00 (125.00, 147.00) 139.00 (129.00, 151.00) 0.066 Platelet 219.00 (184.00, 251.00) 234.00 (206.00, 273.00) 0.002 NLR 2.12 (1.67, 2.88) 2.27 (1.71, 3.27) 0.154 PLR 117.67 (93.24, 162.89) 128.74 (105.68, 168.89) 0.071 LMR 3.89 (2.56, 5.32) 3.61 (2.71, 4.49) 0.156 CRP 1.46 (0.50, 3.20) 5.47 (1.05, 8.43) < 0.001 D-dimer 350 (270, 440) 350 (280, 410) 0.964 Fibrinogen 3.03 (2.60, 3.52) 3.47 (2.92, 3.96) 0.038 Albumin 39.98 (37.40, 43.33) 40.50 (37.60, 43.30) 0.624 Prealbumin 277.05 (220.00, 309.00) 267.00 (220.00, 327.30) 0.787 ALT 19.00 (13.16, 29.80) 21.00 (15.70, 30.00) 0.067 AST 19.00 (16.00, 24.50) 19.20 (16.00, 26.00) 0.614 Triglyceride 1.01 (0.77, 1.39) 1.23 (0.86, 1.66) 0.002 Cholesterol 3.87 (3.04, 4.58) 3.88 (3.00, 4.58) 0.728 High density lipoprotein 1.19 (0.97, 1.42) 1.02 (0.81, 1.22) < 0.001 Low density lipoprotein 1.85 (1.49, 2.47) 2.25 (1.56, 3.15) 0.045 Gender, n (%) 0.284 Female 226 (75.3%) 119 (79.9%) Male 74 (24.7%) 30 (20.1%) Smoke, n (%) 0.315 No 178 (59.3%) 81 (54.4%) Yes 122 (40.7%) 68 (45.6%) Drink, n (%) 0.191 No 200 (66.7%) 90 (60.4%) Yes 100 (33.3%) 59 (39.6%) Hypertension, n (%) 0.372 No 50 (16.7%) 20 (13.4%) Yes 250 (83.3%) 129 (86.6%) NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; LMR, lymphocyte -to- monocyte ratio; CRP, C-Reactive Protein; ALT, Alanine Aminotransferase; AST, Aspartate Aminotransferase. Table 2 Comparison of demographic data between the training and test sets. Characteristics Training set Testing set P value Age 67.00 (62.00, 72.00) 66.00 (60.00, 71.00) 0.181 Leukocyte 6.27 (5.40, 7.63) 6.45 (5.41, 7.82) 0.555 Neutrophil 3.76 (3.03, 4.73) 3.68 (3.11, 5.08) 0.778 lymphocyte 1.81 (1.46, 2.27) 1.83 (1.42, 2.29) 0.941 Monocyte 0.46 (0.37, 0.61) 0.47 (0.38, 0.58) 0.846 Erythrocyte 4.54 (4.20, 4.85) 4.53 (4.17, 4.84) 0.899 Hemoglobin 139.00 (126.00, 148.00) 138.00 (128.00, 147.00) 0.822 Platelet 223.00 (186.00, 261.00) 227.00 (198.50, 259.50) 0.939 NLR 2.16 (1.68, 3.08) 2.18 (1.65, 2.89) 0.850 PLR 121.13 (94.29, 162.89) 119.05 (99.14, 164.25) 0.656 LMR 3.77 (2.69, 5.22) 3.88 (2.69, 4.85) 0.820 CRP 2.31 (0.50, 5.02) 1.56 (0.42, 4.46) 0.145 D-dimer 350.00 (280.00, 440.00) 340.00 (255.00, 430.00) 0.498 Fibrinogen 3.08 (2.63, 3.66) 2.99 (2.66, 3.41) 0.456 Albumin 40.30 (37.41, 43.65) 40.20 (37.54, 43.23) 0.649 Prealbumin 263.90 (218.00, 309.00) 286.00 (241.70, 315.20) 0.066 ALT 19.00 (14.00, 28.00) 23.00 (14.00, 31.60) 0.051 AST 19.00 (16.00, 25.00) 19 (16.00, 23.40) 0.985 Triglyceride 1.07 (0.76, 1.48) 1.06 (0.86, 1.53) 0.622 Cholesterol 3.78 (3.01, 4.58) 3.94 (3.11, 4.64) 0.286 High density lipoprotein 1.10 (0.92, 1.36) 1.13 (0.95, 1.41) 0.651 Low density lipoprotein 2.16 (1.61, 2.80) 2.21 (1.71, 2.87) 0.408 Gender, n (%) 0.245 Female 236 (75.16%) 109 (80.74%) Male 78 (24.84%) 26 (19.26%) Smoke, n (%) 0.420 No 185 (58.92%) 74 (54.81%) Yes 129 (41.08%) 61 (45.19%) Drink, n (%) 0.367 No 207 (65.92%) 83 (61.48%) Yes 107 (34.08%) 52 (38.52%) Hypertension, n (%) 0.402 No 46 (14.65%) 24 (17.78%) Yes 268 (85.35%) 111 (82.22%) NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; LMR, lymphocyte -to- monocyte ratio; CRP, C-Reactive Protein; ALT, Alanine Aminotransferase; AST, Aspartate Aminotransferase. Table 3 Comparison of the performance of machine learning models in the training and testing set. Set Model name Accuracy AUC 95% CI Sensitivity Specificity Training set LR 0.689 0.738 0.672–0.806 0.727 0.655 NaiveBayes 0.675 0.720 0.651–0.789 0.545 0.791 KNN 0.732 0.841 0.791–0.891 0.465 0.973 XGBoost 0.990 0.999 0.998–1.000 0.98 1.000 LightGBM 0.837 0.921 0.887–0.956 0.859 0.818 Testing set LR 0.778 0.839 0.757–0.923 0.840 0.700 NaiveBayes 0.756 0.772 0.673–0.872 0.840 0.650 KNN 0.656 0.746 0.644–0.849 0.600 0.725 XGBoost 0.778 0.849 0.770–0.928 0.720 0.850 LightGBM 0.811 0.864 0.784–0.944 0.800 0.825 AUC, area under the curve; 95%CI, 95% confidence intervals; LR, Logistic Regression; KNN, K-Nearest Neighbor; XGBoost, eXtreme gradient boosting; LightGBM, light gradient boosting machine. Feature Selection Using spearman correlation analysis and the lasso algorithm with tenfold cross-validation (Fig. 2A, B), the 26 variables were ultimately reduced to 9 potential predictors for the unstable carotid plaque, which were incorporated into the construction of the predictive model in our study. Model Building Five machine learning algorithms were employed to construct prediction models using the nine selected factors as input variables. Model performance was assessed on the testing set and evaluated in terms of AUC, accuracy, sensitivity, and specificity. The predictive performance of each model in both the training and testing sets is summarized in Table 4. Additionally, the receiver operating characteristic (ROC) curves and corresponding area under the curve (AUC) values for each model in the testing set are illustrated in Fig. 3a. The LightGBM model performed the best prediction ability with an AUC of 0.921(95%CI: 0.887–0.956) and accuracy of 0.837 in the training set, and an AUC of 0.864 (95%CI: 0.784–0.944) and accuracy of 0.811 in the testing set (Fig. 3b). Importance of features of the best model Based on the LightGBM model, we calculated and ranked the importance of factors associated with carotid plaque instability (Fig. 4 ). C-reactive protein (CRP) was identified as the most important predictive variable, showing the highest level of importance in the model. This was followed by leukocyte and high-density lipoprotein (HDL). Other influential factors included platelet, neutrophil, lymphocyte-to-monocyte ratio (LMR), fibrinogen, low-density lipoprotein (LDL), and monocyte. Furthermore, decision curve analysis (DCA) of the LightGBM model in the testing set demonstrated a favorable net clinical benefit across a range of threshold probabilities, suggesting that the model may offer meaningful utility in clinical decision-making (Fig. 5 ). Discussion The rupture of unstable carotid plaques and subsequent thromboembolic events constitute pivotal pathophysiological mechanisms driving ischemic stroke pathogenesis 14 . Precise stratification of plaque stability enables optimization of secondary prevention protocols while establishing therapeutic windows for targeted interventions in high-risk cohorts 9 . Our machine learning framework identified CRP, lymphocyte count, HDL, and LMR as robust predictors of plaque vulnerability. Elevated CRP levels reflect macrophage-mediated matrix metalloproteinase activation within unstable plaques, whereas reduced HDL and aberrant LMR collectively indicate dysregulated lipid metabolism and immune homeostasis that potentiate necrotic core expansion. Crucially, these hematological biomarkers facilitate noninvasive detection of molecular vulnerability signatures, outperforming conventional stenosis-based anatomical assessments. This paradigm shift toward functional biomarker-driven evaluation equips clinicians with scalable tools for preemptive risk mitigation across diverse care settings. Chronic vascular inflammation, a key factor in atherosclerosis progression, destabilizes plaques through complex mechanisms 15 , 16 . Our study reveals strong positive correlations between elevated CRP, neutrophils, monocytes, and plaque instability. Additionally, a reduced LMR indicates a Th1/Th2 immune imbalance, increasing plaque vulnerability. Neutrophils contribute to oxidative stress by releasing myeloperoxidase (MPO) and forming neutrophil extracellular traps (NETs), which accelerate endothelial apoptosis and necrotic core expansion 17 , 18 . Monocyte-derived macrophages produce proinflammatory cytokines, promoting foam cell formation and sustaining an "inflammation-oxidation" cycle 19 . CRP further degrades plaques by activating the complement system and inducing matrix metalloproteinases (MMPs), while a low LMR reflects weakened immunoregulatory function 20 , 21 . These results highlight the role of systemic inflammation and localized immune dysregulation in plaque destabilization, supporting the use of integrated biomarker panels to identify lesions at risk of rupture. Lipid metabolic disturbances play dual roles in regulating plaque stability. Accumulation of oxidized low-density lipoprotein (oxLDL) in subendothelial spaces fosters foam cell formation and promotes NF-κB-mediated inflammation 20 . In contrast, HDL mitigates atherogenesis by facilitating reverse cholesterol transport (RCT) and preserving endothelial nitric oxide (NO) 24 , 25 . Feature importance analysis supports that HDL is also an important predictor and contributes significantly to the model. However, targeting LDL reduction or HDL elevation alone produces limited benefits. Therefore, combinatorial strategies are needed to effectively modulate the lipid-inflammatory interaction. Thrombogenic propensity following plaque rupture is a key factor in ischemic stroke development 23 . Elevated fibrinogen levels promote thrombus formation by increasing blood viscosity, enhancing fibrin cross-linking, and stimulating intraplaque neovascularization through vascular endothelial growth factor (VEGF) 24 , 25 . Platelets display functional complexity: while CD40L and P-selectin release contribute to thrombosis, platelet-derived growth factor (PDGF) may aid in fibrous cap repair 26 , 27 . As shown by the feature importance analysis in our study, platelets and fibrinogen are also important influencing factors. This suggests that prothrombotic effects dominate during vulnerable phases. Therefore, combined anticoagulant-antiplatelet treatments, guided by monitoring fibrinogen and platelet activity, are crucial for preventing thromboembolic events. Conventional plaque stability assessment relies heavily on structural imaging modalities, while emerging single-cell omics reveal immune infiltrates in unstable plaques 28 – 30 . However, no prior studies have leveraged routine blood biomarkers for predictive modeling. Compared to imaging-dependent approaches, hematological profiling offers distinct advantages: noninvasiveness, cost-effectiveness, and dynamic monitoring feasibility—particularly valuable for renal-impaired or contrast-allergic populations. Blood biomarkers (CRP, HDL, LMR) uniquely capture pathophysiological processes (inflammation, lipid dysregulation, immune imbalance) underlying vulnerability, while machine learning deciphers complex nonlinear interactions among multidimensional variables, transcending stenosis-centric anatomical evaluations. Our machine learning models outperform traditional linear methods by autonomously identifying nonlinear associations and mitigating overfitting through LASSO regularization. The model leverages data-driven feature importance analysis to highlight the most influential predictors, thereby enhancing interpretability and supporting clinical relevance. To our knowledge, this study represents the first integration of blood-derived inflammatory biomarkers into a machine learning framework for predicting carotid plaque stability. Among the five machine learning models, LightGBM achieves the best performance due to its ability to capture complex nonlinear interactions, its efficient leaf-wise tree growth strategy, and its built-in feature selection mechanism, making it particularly suitable for structured clinical data with moderate sample size. Our study also has certain limitations. First, due to the retrospective nature of the study, the possibility of inaccurate data selection and the influence of unmeasured confounding variables could not be fully excluded. Second, only internal validation of the machine learning models was performed, and the LightGBM model was initially identified as the best-performing algorithm. Finally, the analysis did not incorporate multimodal data such as imaging, histopathology, or genetic information, thereby restricting the comprehensiveness of plaque vulnerability assessment from an integrative, systems-level perspective. In the future, we intend to conduct external validation using larger and more diverse cohorts to enhance the model’s generalizability and robustness. Conclusions We developed and validated the ML models to predict carotid plaque instability, and identified that the LightGBM model with 9 variables had the best performance. The LightGBM model demonstrated high accuracy and clinical applicability based on internal validation. Abbreviations ML: machine learning; AUC: area under the curve; ROC: receiver operating characteristics; 95%CI: 95% confidence intervals; CRP: C-Reactive Protein; HDL: high-density lipoprotein; LDL: low-density lipoprotein; CTA: computed tomography angiography; CEUS: contrast enhanced ultrasonography; IPH: intraplaque hemorrhage; NLR: neutrophil-to-lymphocyte ratio; PLR: platelet-to-lymphocyte ratio; LMR: lymphocyte-to-monocyte ratio; ALT: alanine aminotransferase; AST: aspartate aminotransferase; KNN: K-Nearest Neighbor; RF: Random Forest; XGBoost: eXtreme Gradient Boosting; LightGBM: Light Gradient Boosting Machine; LASSO: Least Absolute Shrinkage and Selection Operator; DCA: Decision Curve Analysis. Declarations Ethics approval and consent to participate This study was approved by the Ethics Committee of the Affiliated Hospital of Qingdao University (Protocol ID: QYFY WZLL 27010) and was carried out following the Declaration of Helsinki of the World Medical Association. Consent for publication Not applicable. Data availability The datasets used and analyzed during the current study are available from the corresponding author on reasonable request. Competing interests The authors declare no competing interests. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Author contributions TWZ contributed to data collection, processing, and analysis. HBM performed data analysis and manuscript writing. XZS participated in data collection and manuscript preparation. GFL assisted in data collection. YXL was responsible for manuscript writing, revision, and critical review. MJG contributed to manuscript revision, data verification, and final approval. All authors have read and agreed to the published version of the manuscript. Acknowledgements None. Clinical trial number: not applicable. References Tan J, Liang Y, Yang Z, He Q, Tong J, Deng Y, et al. Single-Cell Transcriptomics Reveals Crucial Cell Subsets and Functional Heterogeneity Associated With Carotid Atherosclerosis and Cerebrovascular Events. Arterioscler Thromb Vasc Biol. 2023 Dec;43(12):2312–32. Feigin VL, Brainin M, Norrving B, Martins S, Sacco RL, Hacke W, et al. World Stroke Organization (WSO): Global Stroke Fact Sheet 2022. Int J Stroke. 2022 Jan;17(1):18–29. Wang YJ, Li ZX, Gu HQ, Zhai Y, Zhou Q, Jiang Y, et al. China Stroke Statistics: an update on the 2019 report from the National Center for Healthcare Quality Management in Neurological Diseases, China National Clinical Research Center for Neurological Diseases, the Chinese Stroke Association, National Center for Chronic and Non-communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention and Institute for Global Neuroscience and Stroke Collaborations. Stroke Vasc Neurol. 2022 Oct;7(5):415–50. Kopczak A, Schindler A, Sepp D, Bayer-Karpinska A, Malik R, Koch ML, et al. Complicated Carotid Artery Plaques and Risk of Recurrent Ischemic Stroke or TIA. J Am Coll Cardiol. 2022 Jun 7;79(22):2189–99. van Dam-Nolen DHK, Truijman MTB, van der Kolk AG, Liem MI, Schreuder FHBM, Boersma E, et al. Carotid Plaque Characteristics Predict Recurrent Ischemic Stroke and TIA: The PARISK (Plaque At RISK) Study. JACC Cardiovasc Imaging. 2022 Oct;15(10):1715–26. Howard DP, van Lammeren GW, Rothwell PM, Redgrave JN, Moll FL, de Vries JPP, et al. Symptomatic carotid atherosclerotic disease: correlations between plaque composition and ipsilateral stroke risk. Stroke. 2015 Jan;46(1):182–9. van Dam-Nolen DHK, van Egmond NCM, Dilba K, Nies K, van der Kolk AG, Liem MI, et al. Sex Differences in Plaque Composition and Morphology Among Symptomatic Patients With Mild-to-Moderate Carotid Artery Stenosis. Stroke. 2022 Feb;53(2):370–8. Miceli G, Basso MG, Pintus C, Pennacchio AR, Cocciola E, Cuffaro M, et al. Molecular Pathways of Vulnerable Carotid Plaques at Risk of Ischemic Stroke: A Narrative Review. Int J Mol Sci. 2024 Apr 15;25(8):4351. Karlöf E, Buckler A, Liljeqvist ML, Lengquist M, Kronqvist M, Toonsi MA, et al. Carotid Plaque Phenotyping by Correlating Plaque Morphology from Computed Tomography Angiography with Transcriptional Profiling. Eur J Vasc Endovasc Surg. 2021 Nov;62(5):716–26. Zamani M, Skagen K, Scott H, Lindberg B, Russell D, Skjelland M. Carotid Plaque Neovascularization Detected With Superb Microvascular Imaging Ultrasound Without Using Contrast Media. Stroke. 2019 Nov;50(11):3121–7. Rudnick MR, Leonberg-Yoo AK, Litt HI, Cohen RM, Hilton S, Reese PP. The Controversy of Contrast-Induced Nephropathy With Intravenous Contrast: What Is the Risk? Am J Kidney Dis. 2020 Jan;75(1):105–13. Lundberg SM, Nair B, Vavilala MS, Horibe M, Eisses MJ, Adams T, et al. Explainable machine-learning predictions for the prevention of hypoxaemia during surgery. Nat Biomed Eng. 2018 Oct;2(10):749–60. Lundberg SM, Erion G, Chen H, DeGrave A, Prutkin JM, Nair B, et al. From Local Explanations to Global Understanding with Explainable AI for Trees. Nat Mach Intell. 2020 Jan;2(1):56–67. Noubiap JJ, Thomas G, Kamtchum-Tatuene J, Middeldorp ME, Sanders P. High-risk carotid plaques and incident ischemic stroke in patients with atrial fibrillation in the Cardiovascular Health Study. Eur J Neurol. 2023 Jul;30(7):2042–50. Ajoolabady A, Pratico D, Lin L, Mantzoros CS, Bahijri S, Tuomilehto J, et al. Inflammation in atherosclerosis: pathophysiology and mechanisms. Cell Death Dis. 2024 Nov 11;15(11):817. Weber C, Habenicht AJR, von Hundelshausen P. Novel mechanisms and therapeutic targets in atherosclerosis: inflammation and beyond. Eur Heart J. 2023 Aug 1;44(29):2672–81. Natorska J, Ząbczyk M, Undas A. Neutrophil extracellular traps (NETs) in cardiovascular diseases: From molecular mechanisms to therapeutic interventions. Kardiol Pol. 2023;81(12):1205–16. Shamsuzzaman S, Deaton RA, Salamon A, Doviak H, Serbulea V, Milosek VM, et al. Novel Mouse Model of Myocardial Infarction, Plaque Rupture, and Stroke Shows Improved Survival With Myeloperoxidase Inhibition. Circulation. 2024 Aug 27;150(9):687–705. Wang B, Tang X, Yao L, Wang Y, Chen Z, Li M, et al. Disruption of USP9X in macrophages promotes foam cell formation and atherosclerosis. J Clin Invest. 2022 May 16;132(10):e154217. Kuppa A, Tripathi H, Al-Darraji A, Tarhuni WM, Abdel-Latif A. C-Reactive Protein Levels and Risk of Cardiovascular Diseases: A Two-Sample Bidirectional Mendelian Randomization Study. Int J Mol Sci. 2023 May 23;24(11):9129. Xu L, Liu JT, Li K, Wang SY, Xu S. Genistein inhibits Ang II-induced CRP and MMP-9 generations via the ER-p38/ERK1/2-PPARγ-NF-κB signaling pathway in rat vascular smooth muscle cells. Life Sci. 2019 Jan 1;216:140–6. Huang Z, Shen S, Han X, Li W, Luo W, Lin L, et al. Macrophage DCLK1 promotes atherosclerosis via binding to IKKβ and inducing inflammatory responses. EMBO Mol Med. 2023 May 8;15(5):e17198. Kang H, Song J, Cheng Y. HDL regulates the risk of cardiometabolic and inflammatory-related diseases: Focusing on cholesterol efflux capacity. Int Immunopharmacol. 2024 Sep 10;138:112622. Hu S, Zhu Y, Zhao X, Li R, Shao G, Gong D, et al. Hepatocytic lipocalin-2 controls HDL metabolism and atherosclerosis via Nedd4-1-SR-BI axis in mice. Dev Cell. 2023 Nov 6;58(21):2326-2337.e5. Miceli G, Basso MG, Pintus C, Pennacchio AR, Cocciola E, Cuffaro M, et al. Molecular Pathways of Vulnerable Carotid Plaques at Risk of Ischemic Stroke: A Narrative Review. Int J Mol Sci. 2024 Apr 15;25(8):4351. Wolberg AS. Fibrinogen and fibrin: synthesis, structure, and function in health and disease. J Thromb Haemost. 2023 Nov;21(11):3005–15. Jiang S, Ai Y, Ni L, Wu L, Huang X, Chen S. Platelet-derived TGF-β1 is related to portal vein thrombosis in cirrhosis by promoting hypercoagulability and endothelial dysfunction. Front Cardiovasc Med. 2022;9:938397. L S, V N, R C, A G, H K, Js S, et al. Carotid Artery Plaque Calcifications: Lessons From Histopathology to Diagnostic Imaging. Stroke [Internet]. 2022 Jan [cited 2025 Apr 13];53(1). Available from: https://pubmed.ncbi.nlm.nih.gov/34753301/ Zhang S, Jiang S, Wang C, Han C. Comparison of ultrasonic shear wave elastography, AngioPLUS planewave ultrasensitive imaging, and optimized high-resolution magnetic resonance imaging in evaluating carotid plaque stability. PeerJ. 2023;11:e16150. Narayanan S, Vuckovic S, Bergman O, Wirka R, Verdezoto Mosquera J, Chen QS, et al. Atheroma transcriptomics identifies ARNTL as a smooth muscle cell regulator and with clinical and genetic data improves risk stratification. Eur Heart J. 2025 Jan 16;46(3):308–22. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 10 Sep, 2025 Reviews received at journal 25 Aug, 2025 Reviewers agreed at journal 25 Aug, 2025 Reviews received at journal 24 Aug, 2025 Reviewers agreed at journal 23 Aug, 2025 Reviewers agreed at journal 23 Aug, 2025 Reviews received at journal 16 Aug, 2025 Reviewers agreed at journal 07 Aug, 2025 Reviewers invited by journal 05 Aug, 2025 Editor invited by journal 14 Jul, 2025 Editor assigned by journal 03 Jul, 2025 Submission checks completed at journal 03 Jul, 2025 First submitted to journal 24 Jun, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6966994","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":496249994,"identity":"6225e219-2e39-46cb-8469-7a9313900ab2","order_by":0,"name":"Tianwei Zhang","email":"","orcid":"","institution":"Affiliated Hospital of Qingdao University","correspondingAuthor":false,"prefix":"","firstName":"Tianwei","middleName":"","lastName":"Zhang","suffix":""},{"id":496249995,"identity":"249cbc5b-f93b-4483-966f-9481b481956d","order_by":1,"name":"Huibo Ma","email":"","orcid":"","institution":"Peking Union Medical College Hospital, Chines Academy of Medical Sciences","correspondingAuthor":false,"prefix":"","firstName":"Huibo","middleName":"","lastName":"Ma","suffix":""},{"id":496249996,"identity":"4bc7153b-30ff-410d-8b4c-bef5c8f2dec9","order_by":2,"name":"Xiaozhi Sun","email":"","orcid":"","institution":"Affiliated Hospital of Qingdao University","correspondingAuthor":false,"prefix":"","firstName":"Xiaozhi","middleName":"","lastName":"Sun","suffix":""},{"id":496249997,"identity":"c0859eb3-71b7-495d-9373-fe762ac24df9","order_by":3,"name":"Guofeng Li","email":"","orcid":"","institution":"Juxian People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Guofeng","middleName":"","lastName":"Li","suffix":""},{"id":496249998,"identity":"6b5ebfeb-814f-403d-862c-9bfe2a4b16a2","order_by":4,"name":"Yongxin Li","email":"","orcid":"","institution":"Affiliated Hospital of Qingdao University","correspondingAuthor":false,"prefix":"","firstName":"Yongxin","middleName":"","lastName":"Li","suffix":""},{"id":496249999,"identity":"a28469b2-8695-4fc2-ab96-649b2028f5ce","order_by":5,"name":"Mingjin Guo","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAw0lEQVRIiWNgGAWjYFCCBAZmMM3e2PjwA2laeA43G0uQpkUivU2AhxgNfMdzDD8XttnZ88982MYgwWAnp9tAQIvkmTfG0jPbkhNn3E5se1DAkGxsdoCAFoMbOQbSPGeYEwykE9sNJBgOJG4jQovxb54z9fYGkgfbJHiI1GImzVNxmHGDBCORWiTPPCuz5qk4njjjTCIwkA2I8Avf8eTNt3kMqu35248/fPihwk6OoBYGVAUGhJRjahkFo2AUjIJRgAUAAGmzQWVVEYINAAAAAElFTkSuQmCC","orcid":"","institution":"Affiliated Hospital of Qingdao University","correspondingAuthor":true,"prefix":"","firstName":"Mingjin","middleName":"","lastName":"Guo","suffix":""}],"badges":[],"createdAt":"2025-06-24 14:53:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6966994/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6966994/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":88774755,"identity":"65f5e392-00e6-428a-a5a5-d3ed12db308e","added_by":"auto","created_at":"2025-08-11 10:03:13","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":68101,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFlow chart of the process of patients’ selection.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6966994/v1/38267af5070ddf080d0475ab.jpg"},{"id":88774754,"identity":"630076d1-4f96-44ec-be58-fddc432011dc","added_by":"auto","created_at":"2025-08-11 10:03:13","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":42007,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLASSO regression analysis for feature selection. \u003c/strong\u003e(A) The process of feature selection. We used the LASSO regression model with regularization parameter (λ) tuning conducted by tenfold cross validation according to the minimum mean squared error (MSE) criteria. Based on the minimum MSE criterion, the vertical dotted line is plotted at the optimal value λ = 0.0339. (B) At this optimal λ, 9 features retain nonzero coefficients. LASSO, least absolute shrinkage and selection operator.\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6966994/v1/b0f510969cef36db0d3acd15.jpg"},{"id":88777902,"identity":"537e44bd-246e-4380-85ca-c29894d30fca","added_by":"auto","created_at":"2025-08-11 10:19:13","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":59750,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eReceiver operating characteristic curves of the five models.\u003c/strong\u003e (A) Performance for machine learning models in the testing set based on the AUC of the ROC curve. (B) AUC and the ROC curve of LightGBM model in the training set and the testing set. AUC, area under the curve; ROC, receiver operating characteristics; LR, Logistic Regression; KNN, K-Nearest Neighbor; XGBoost, eXtreme gradient boosting; LightGBM, light gradient boosting machine.\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6966994/v1/6506c85e7484f95de67c68e6.jpg"},{"id":88777158,"identity":"d1934ddc-c593-4557-b5bf-4ee5090357df","added_by":"auto","created_at":"2025-08-11 10:11:13","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":46337,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe importance rankings of the different predictive factors in LightGBM model\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6966994/v1/24b454d6669faf52635f05d1.jpg"},{"id":88774756,"identity":"5205a946-9e2a-49d9-9bc5-ad0c6d77d3e8","added_by":"auto","created_at":"2025-08-11 10:03:13","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":41969,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDecision curve analysis of LightGBM model in testing set\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6966994/v1/8b90c577c57f9a1e7e9df50a.jpg"},{"id":88779271,"identity":"26f19ae3-784c-4268-86e2-237bacd55905","added_by":"auto","created_at":"2025-08-11 10:35:13","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1222870,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6966994/v1/bd359650-fb49-4254-8e5b-eb415d8799c2.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"A Study on Analyzing Risk Factors for Carotid Plaque Instability Using Machine Learning Models","fulltext":[{"header":"Background","content":"\u003cp\u003eAtherosclerotic carotid plaques constitute a critical pathological substrate for ischemic cerebrovascular events\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. The 2019 Global Burden of Disease Study revealed that stroke accounted for approximately 6.65\u0026nbsp;million global deaths, ranking as the second-leading cause of mortality and disability-adjusted life years, with ischemic stroke representing over 80% of total cases\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. In China, stroke-related hospitalization costs reached US\u003cspan\u003e$\u003c/span\u003e12.2\u0026nbsp;billion in 2016, imposing substantial socioeconomic and healthcare burdens\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. Evidence indicates that nearly 30% of ischemic strokes are directly attributable to thromboembolism originating from ruptured carotid atherosclerotic plaques\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. While current clinical guidelines emphasize luminal stenosis severity for stroke risk stratification, emerging evidence demonstrates that plaque composition and biological activity fundamentally determine vulnerability rather than mere anatomical narrowing\u003csup\u003e\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e–\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Clinically, unstable plaques characterized by thin fibrous caps and lipid-rich necrotic cores exhibit markedly higher rupture propensity compared to stable counterparts, frequently resulting in thromboembolic complications and disabling cerebrovascular outcomes\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. Consequently, early detection of high-risk plaques and targeted mitigation of rupture risk represent urgent clinical priorities to reduce stroke incidence and improve population-level neurological prognosis.\u003c/p\u003e\u003cp\u003eThe current clinical evaluation of carotid plaque stability mainly depends on advanced imaging techniques like computed tomography angiography (CTA) and contrast-enhanced ultrasonography (CEUS) \u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. These methods offer detailed insights into plaque characteristics, such as lipid core size, fibrous cap integrity, and intraplaque hemorrhage (IPH). However, their clinical use is limited by several factors. CTA exposes patients to radiation and nephrotoxic contrast agents, while CEUS is operator-dependent and suffers from poor reproducibility\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. In contrast, routine blood tests, commonly obtained during health screenings, present a promising noninvasive alternative for plaque stability assessment. They are cost-effective, scalable, and allow for ongoing monitoring. However, traditional statistical techniques struggle to account for the complex, nonlinear relationships between various biomarkers, which hampers prediction accuracy. Recently, machine learning algorithms have shown great potential in overcoming these challenges. By integrating high-dimensional data and identifying complex patterns, these algorithms improve cardiovascular risk prediction\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. This study introduces a machine learning-based framework that uses accessible blood biomarkers to construct a predictive model for carotid plaque stability. This approach could facilitate early risk identification and personalized preventive measures.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cb\u003eData Source and Study Design\u003c/b\u003e\u003c/p\u003e\u003cp\u003eWe conducted a retrospective study analyzing the clinical data of patients who had carotid plaque between July 2022 and November 2024 at the affiliated hospital of Qingdao University. Unstable plaques were defined according to criteria including irregular surface, ulcerations, low or mixed echogenicity on ultrasound, or recent ischemic events (e.g., TIA or stroke). We included adults aged 18 years and older with carotid atherosclerosis disease, who underwent CTA of the carotid arteries during the study period. Both carotid arteries were considered if plaques were present on both sides, otherwise, only the affected side was included. Patients were divided into training (70%) and testing (30%) sets by stratified random sampling using the Stratified Shuffle-Split function in Python. Exclusion criteria included serious systemic inflammatory diseases, autoimmune disorders, malignancies, recent infections, severe liver or kidney dysfunction, recent cardiovascular events, incomplete data, pregnant or breastfeeding women, and inability to comply with the study protocol. The patient selection flow diagram is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. This retrospective study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Affiliated Hospital of Qingdao University (Protocol ID: QYFY WZLL 27010). Due to the retrospective nature of this study, the requirement for informed consent was waived by the Affiliated Hospital of Qingdao University Institutional Review Board.\u003c/p\u003e\u003cp\u003eClinical trial number: not applicable.\u003c/p\u003e\u003cp\u003e\u003cb\u003eData collection\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe clinical data on the patients included: age, gender, leukocyte, neutrophil, lymphocyte, monocyte, platelet, hemoglobin, erythrocyte, neutrophil to lymphocyte ratio (NLR), platelet to lymphocyte ratio (PLR), lymphocyte to monocyte ratio (LMR), C-Reactive Protein (CRP), D-dimer, albumin, prealbumin, fibrinogen, alanine aminotransferase (ALT), aspartate aminotransferase (AST), triglyceride, cholesterol, high density lipoprotein, low density lipoprotein, smoke, drink, and hypertension.\u003c/p\u003e\u003cp\u003e\u003cb\u003eFeature Selection\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFirstly, we conducted the statistical Mann–Whitney test and chi-square test for clinical characteristics to determine whether there are differences between the groups. Our analysis considered the feature significant when its two-tailed p-value was p \u0026lt; 0.05. Secondly, spearman correlation analysis was performed to reduce collinearity among features. To reduce the risk of overfitting, the least absolute shrinkage and selection operator (LASSO) algorithm was applied to select features with non-zero coefficient values.\u003c/p\u003e\u003cp\u003e\u003cb\u003eModel Building\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo develop a predictive model for unstable carotid plaque, five machine learning (ML) algorithms were applied: k-nearest neighbor (KNN), random forest (RF), Extra-Trees classifier, eXtreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM). All patients were randomly divided into a training set (70%) and a testing set (30%). The training set was used to build the prediction models, while the testing set was employed to evaluate their performance using the area under the receiver operating characteristic curve (AUC) and corresponding 95% confidence intervals (95% CI). The model with the highest AUC was considered the optimal model. To evaluate the contribution of each variable to the model's prediction, we calculated feature importance scores based on the optimal model. A bar plot was used to visualize the top-ranked features, with higher values indicating greater contribution to the model's predictive performance. Additionally, decision curve analysis (DCA) was conducted to illustrate the net clinical benefit across a range of threshold probabilities.\u003c/p\u003e\u003cp\u003e\u003cb\u003eStatistical analyses\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe Stratified Shuffle-Split function was utilized in Python (version 3.7) to partition the dataset. Continuous variables with an abnormal distribution were described as medians and interquartile ranges, and categorical variables were described as frequencies and proportions. Continuous variables with an abnormal distribution, and categorical variables were analyzed using Mann-Whitney U-test, and Chi-squared test, respectively. Correlation analysis and feature selection using the least absolute shrinkage and selection operator (LASSO) algorithm, were conducted in Python (version 3.7) using the “scipy,” “numpy,” and “sklearn” packages. A bilateral P-value \u0026lt; 0.05 was considered as a measure of statistical significance.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cb\u003ePatient Characteristics\u003c/b\u003e\u003c/p\u003e\u003cp\u003eA total of 505 patients with carotid plaque were potentially eligible from the affiliated hospital of Qingdao University between July 2022 and November 2024. After the selecting process, 449 patients were finally enrolled in our study, and 149 of which had unstable carotid plaque. The training set included 215 patients with stable carotid plaque and 99 patients with unstable carotid plaque, while the testing set included 85 patients with stable carotid plaque and 50 patients with unstable carotid plaque (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The baseline data of the included patient are shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, which indicated that leukocyte, neutrophil, monocyte, platelet, LMR, CRP, fibrinogen, triglyceride, low density lipoprotein, and high density lipoprotein were significantly different between patients with stable carotid plaque and patients with unstable carotid plaque. The comparison of baseline characteristics between the training and testing sets with corresponding P values was shown in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. Notably, the incidence of unstable carotid plaque did not differ significantly between the training and testing sets (p\u0026thinsp;=\u0026thinsp;0.841). Furthermore, all baseline characteristics were statistically insignificant between the two sets, indicating that the training and testing sets were well-balanced.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eBaseline characteristics of the patients.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCharacteristics\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eStable carotid plaque\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eUnstable carotid plaque\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eP value\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAge\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e67.00 (62.00, 73.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e67.00 (60.00, 71.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.760\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLeukocyte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e5.85 (5.13, 6.93)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e7.39 (6.21, 8.42)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNeutrophil\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e3.50 (2.84, 4.45)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e4.35 (3.44, 5.23)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003elymphocyte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1.83 (1.42, 2.26)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.80 (1.45, 2.30)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.971\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMonocyte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.43 (0.37, 0.60)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.50 (0.41, 0.61)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.004\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eErythrocyte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e4.52 (4.17, 4.83)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e4.54 (4.21, 4.93)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.214\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHemoglobin\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e138.00 (125.00, 147.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e139.00 (129.00, 151.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.066\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePlatelet\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e219.00 (184.00, 251.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e234.00 (206.00, 273.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.002\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNLR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e2.12 (1.67, 2.88)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2.27 (1.71, 3.27)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.154\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePLR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e117.67 (93.24, 162.89)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e128.74 (105.68, 168.89)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.071\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLMR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e3.89 (2.56, 5.32)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3.61 (2.71, 4.49)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.156\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCRP\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1.46 (0.50, 3.20)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5.47 (1.05, 8.43)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eD-dimer\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e350 (270, 440)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e350 (280, 410)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.964\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFibrinogen\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e3.03 (2.60, 3.52)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3.47 (2.92, 3.96)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.038\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAlbumin\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e39.98 (37.40, 43.33)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e40.50 (37.60, 43.30)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.624\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePrealbumin\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e277.05 (220.00, 309.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e267.00 (220.00, 327.30)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.787\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eALT\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e19.00 (13.16, 29.80)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e21.00 (15.70, 30.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.067\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAST\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e19.00 (16.00, 24.50)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e19.20 (16.00, 26.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.614\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTriglyceride\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1.01 (0.77, 1.39)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.23 (0.86, 1.66)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.002\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCholesterol\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e3.87 (3.04, 4.58)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3.88 (3.00, 4.58)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.728\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHigh density lipoprotein\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1.19 (0.97, 1.42)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.02 (0.81, 1.22)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLow density lipoprotein\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1.85 (1.49, 2.47)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2.25 (1.56, 3.15)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.045\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGender, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.284\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFemale\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e226 (75.3%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e119 (79.9%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMale\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e74 (24.7%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e30 (20.1%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSmoke, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.315\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e178 (59.3%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e81 (54.4%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e122 (40.7%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e68 (45.6%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDrink, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.191\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e200 (66.7%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e90 (60.4%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e100 (33.3%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e59 (39.6%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHypertension, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.372\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e50 (16.7%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e20 (13.4%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e250 (83.3%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e129 (86.6%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"4\"\u003eNLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; LMR, lymphocyte -to- monocyte ratio; CRP, C-Reactive Protein; ALT, Alanine Aminotransferase; AST, Aspartate Aminotransferase.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eComparison of demographic data between the training and test sets.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eCharacteristics\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eTraining set\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eTesting set\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eP value\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eAge\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e67.00 (62.00, 72.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e66.00 (60.00, 71.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.181\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eLeukocyte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6.27 (5.40, 7.63)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e6.45 (5.41, 7.82)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.555\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eNeutrophil\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3.76 (3.03, 4.73)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3.68 (3.11, 5.08)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.778\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003elymphocyte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.81 (1.46, 2.27)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.83 (1.42, 2.29)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.941\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eMonocyte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.46 (0.37, 0.61)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.47 (0.38, 0.58)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.846\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eErythrocyte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e4.54 (4.20, 4.85)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e4.53 (4.17, 4.84)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.899\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eHemoglobin\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e139.00 (126.00, 148.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e138.00 (128.00, 147.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.822\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003ePlatelet\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e223.00 (186.00, 261.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e227.00 (198.50, 259.50)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.939\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eNLR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2.16 (1.68, 3.08)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2.18 (1.65, 2.89)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.850\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003ePLR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e121.13 (94.29, 162.89)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e119.05 (99.14, 164.25)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.656\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eLMR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3.77 (2.69, 5.22)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3.88 (2.69, 4.85)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.820\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eCRP\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2.31 (0.50, 5.02)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.56 (0.42, 4.46)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.145\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eD-dimer\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e350.00 (280.00, 440.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e340.00 (255.00, 430.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.498\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eFibrinogen\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3.08 (2.63, 3.66)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2.99 (2.66, 3.41)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.456\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eAlbumin\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e40.30 (37.41, 43.65)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e40.20 (37.54, 43.23)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.649\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003ePrealbumin\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e263.90 (218.00, 309.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e286.00 (241.70, 315.20)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.066\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eALT\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e19.00 (14.00, 28.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e23.00 (14.00, 31.60)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.051\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eAST\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e19.00 (16.00, 25.00)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e19 (16.00, 23.40)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.985\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eTriglyceride\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.07 (0.76, 1.48)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.06 (0.86, 1.53)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.622\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eCholesterol\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3.78 (3.01, 4.58)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3.94 (3.11, 4.64)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.286\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eHigh density lipoprotein\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.10 (0.92, 1.36)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.13 (0.95, 1.41)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.651\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eLow density lipoprotein\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2.16 (1.61, 2.80)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2.21 (1.71, 2.87)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.408\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eGender, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.245\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eFemale\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e236 (75.16%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e109 (80.74%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eMale\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e78 (24.84%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e26 (19.26%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eSmoke, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.420\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e185 (58.92%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e74 (54.81%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e129 (41.08%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e61 (45.19%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eDrink, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.367\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e207 (65.92%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e83 (61.48%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e107 (34.08%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e52 (38.52%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eHypertension, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.402\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e46 (14.65%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e24 (17.78%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003e268 (85.35%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e111 (82.22%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"5\"\u003eNLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; LMR, lymphocyte -to- monocyte ratio; CRP, C-Reactive Protein; ALT, Alanine Aminotransferase; AST, Aspartate Aminotransferase.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eComparison of the performance of machine learning models in the training and testing set.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSet\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eModel name\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAccuracy\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eAUC\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003e95% CI\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eSensitivity\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eSpecificity\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTraining set\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.689\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.738\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.672\u0026ndash;0.806\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.727\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.655\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNaiveBayes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.675\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.720\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.651\u0026ndash;0.789\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.545\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.791\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eKNN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.732\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.841\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.791\u0026ndash;0.891\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.465\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.973\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eXGBoost\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.990\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.999\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.998\u0026ndash;1.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.98\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e1.000\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLightGBM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.837\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.921\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.887\u0026ndash;0.956\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.859\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.818\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTesting set\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.778\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.839\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.757\u0026ndash;0.923\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.840\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.700\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNaiveBayes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.756\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.772\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.673\u0026ndash;0.872\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.840\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.650\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eKNN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.656\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.746\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.644\u0026ndash;0.849\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.600\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.725\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eXGBoost\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.778\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.849\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.770\u0026ndash;0.928\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.720\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.850\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLightGBM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.811\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.864\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.784\u0026ndash;0.944\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.800\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.825\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"7\"\u003eAUC, area under the curve; 95%CI, 95% confidence intervals; LR, Logistic Regression; KNN, K-Nearest Neighbor; XGBoost, eXtreme gradient boosting; LightGBM, light gradient boosting machine.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eFeature Selection\u003c/b\u003e\u003c/p\u003e\u003cp\u003eUsing spearman correlation analysis and the lasso algorithm with tenfold cross-validation (Fig.\u0026nbsp;2A, B), the 26 variables were ultimately reduced to 9 potential predictors for the unstable carotid plaque, which were incorporated into the construction of the predictive model in our study.\u003c/p\u003e\u003cp\u003e\u003cb\u003eModel Building\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFive machine learning algorithms were employed to construct prediction models using the nine selected factors as input variables. Model performance was assessed on the testing set and evaluated in terms of AUC, accuracy, sensitivity, and specificity. The predictive performance of each model in both the training and testing sets is summarized in Table\u0026nbsp;4. Additionally, the receiver operating characteristic (ROC) curves and corresponding area under the curve (AUC) values for each model in the testing set are illustrated in Fig.\u0026nbsp;3a. The LightGBM model performed the best prediction ability with an AUC of 0.921(95%CI: 0.887\u0026ndash;0.956) and accuracy of 0.837 in the training set, and an AUC of 0.864 (95%CI: 0.784\u0026ndash;0.944) and accuracy of 0.811 in the testing set (Fig.\u0026nbsp;3b).\u003c/p\u003e\u003cp\u003e\u003cb\u003eImportance of features of the best model\u003c/b\u003e\u003c/p\u003e\u003cp\u003eBased on the LightGBM model, we calculated and ranked the importance of factors associated with carotid plaque instability (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e4\u003c/span\u003e). C-reactive protein (CRP) was identified as the most important predictive variable, showing the highest level of importance in the model. This was followed by leukocyte and high-density lipoprotein (HDL). Other influential factors included platelet, neutrophil, lymphocyte-to-monocyte ratio (LMR), fibrinogen, low-density lipoprotein (LDL), and monocyte. Furthermore, decision curve analysis (DCA) of the LightGBM model in the testing set demonstrated a favorable net clinical benefit across a range of threshold probabilities, suggesting that the model may offer meaningful utility in clinical decision-making (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe rupture of unstable carotid plaques and subsequent thromboembolic events constitute pivotal pathophysiological mechanisms driving ischemic stroke pathogenesis\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. Precise stratification of plaque stability enables optimization of secondary prevention protocols while establishing therapeutic windows for targeted interventions in high-risk cohorts\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. Our machine learning framework identified CRP, lymphocyte count, HDL, and LMR as robust predictors of plaque vulnerability. Elevated CRP levels reflect macrophage-mediated matrix metalloproteinase activation within unstable plaques, whereas reduced HDL and aberrant LMR collectively indicate dysregulated lipid metabolism and immune homeostasis that potentiate necrotic core expansion. Crucially, these hematological biomarkers facilitate noninvasive detection of molecular vulnerability signatures, outperforming conventional stenosis-based anatomical assessments. This paradigm shift toward functional biomarker-driven evaluation equips clinicians with scalable tools for preemptive risk mitigation across diverse care settings.\u003c/p\u003e\u003cp\u003eChronic vascular inflammation, a key factor in atherosclerosis progression, destabilizes plaques through complex mechanisms\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e,\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. Our study reveals strong positive correlations between elevated CRP, neutrophils, monocytes, and plaque instability. Additionally, a reduced LMR indicates a Th1/Th2 immune imbalance, increasing plaque vulnerability. Neutrophils contribute to oxidative stress by releasing myeloperoxidase (MPO) and forming neutrophil extracellular traps (NETs), which accelerate endothelial apoptosis and necrotic core expansion\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e,\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. Monocyte-derived macrophages produce proinflammatory cytokines, promoting foam cell formation and sustaining an \"inflammation-oxidation\" cycle\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. CRP further degrades plaques by activating the complement system and inducing matrix metalloproteinases (MMPs), while a low LMR reflects weakened immunoregulatory function\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. These results highlight the role of systemic inflammation and localized immune dysregulation in plaque destabilization, supporting the use of integrated biomarker panels to identify lesions at risk of rupture.\u003c/p\u003e\u003cp\u003eLipid metabolic disturbances play dual roles in regulating plaque stability. Accumulation of oxidized low-density lipoprotein (oxLDL) in subendothelial spaces fosters foam cell formation and promotes NF-κB-mediated inflammation\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. In contrast, HDL mitigates atherogenesis by facilitating reverse cholesterol transport (RCT) and preserving endothelial nitric oxide (NO)\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. Feature importance analysis supports that HDL is also an important predictor and contributes significantly to the model. However, targeting LDL reduction or HDL elevation alone produces limited benefits. Therefore, combinatorial strategies are needed to effectively modulate the lipid-inflammatory interaction.\u003c/p\u003e\u003cp\u003eThrombogenic propensity following plaque rupture is a key factor in ischemic stroke development\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. Elevated fibrinogen levels promote thrombus formation by increasing blood viscosity, enhancing fibrin cross-linking, and stimulating intraplaque neovascularization through vascular endothelial growth factor (VEGF)\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. Platelets display functional complexity: while CD40L and P-selectin release contribute to thrombosis, platelet-derived growth factor (PDGF) may aid in fibrous cap repair\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. As shown by the feature importance analysis in our study, platelets and fibrinogen are also important influencing factors. This suggests that prothrombotic effects dominate during vulnerable phases. Therefore, combined anticoagulant-antiplatelet treatments, guided by monitoring fibrinogen and platelet activity, are crucial for preventing thromboembolic events.\u003c/p\u003e\u003cp\u003eConventional plaque stability assessment relies heavily on structural imaging modalities, while emerging single-cell omics reveal immune infiltrates in unstable plaques\u003csup\u003e\u003cspan additionalcitationids=\"CR29\" citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. However, no prior studies have leveraged routine blood biomarkers for predictive modeling. Compared to imaging-dependent approaches, hematological profiling offers distinct advantages: noninvasiveness, cost-effectiveness, and dynamic monitoring feasibility\u0026mdash;particularly valuable for renal-impaired or contrast-allergic populations. Blood biomarkers (CRP, HDL, LMR) uniquely capture pathophysiological processes (inflammation, lipid dysregulation, immune imbalance) underlying vulnerability, while machine learning deciphers complex nonlinear interactions among multidimensional variables, transcending stenosis-centric anatomical evaluations.\u003c/p\u003e\u003cp\u003eOur machine learning models outperform traditional linear methods by autonomously identifying nonlinear associations and mitigating overfitting through LASSO regularization. The model leverages data-driven feature importance analysis to highlight the most influential predictors, thereby enhancing interpretability and supporting clinical relevance. To our knowledge, this study represents the first integration of blood-derived inflammatory biomarkers into a machine learning framework for predicting carotid plaque stability. Among the five machine learning models, LightGBM achieves the best performance due to its ability to capture complex nonlinear interactions, its efficient leaf-wise tree growth strategy, and its built-in feature selection mechanism, making it particularly suitable for structured clinical data with moderate sample size.\u003c/p\u003e\u003cp\u003eOur study also has certain limitations. First, due to the retrospective nature of the study, the possibility of inaccurate data selection and the influence of unmeasured confounding variables could not be fully excluded. Second, only internal validation of the machine learning models was performed, and the LightGBM model was initially identified as the best-performing algorithm. Finally, the analysis did not incorporate multimodal data such as imaging, histopathology, or genetic information, thereby restricting the comprehensiveness of plaque vulnerability assessment from an integrative, systems-level perspective. In the future, we intend to conduct external validation using larger and more diverse cohorts to enhance the model\u0026rsquo;s generalizability and robustness.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eWe developed and validated the ML models to predict carotid plaque instability, and identified that the LightGBM model with 9 variables had the best performance. The LightGBM model demonstrated high accuracy and clinical applicability based on internal validation.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eML: machine learning; AUC: area under the curve; ROC: receiver operating characteristics; 95%CI: 95% confidence intervals; CRP: C-Reactive Protein; HDL: high-density lipoprotein; LDL: low-density lipoprotein; CTA: computed tomography angiography; CEUS: contrast enhanced ultrasonography; IPH: intraplaque hemorrhage; NLR: neutrophil-to-lymphocyte ratio; PLR: platelet-to-lymphocyte ratio; LMR: lymphocyte-to-monocyte ratio; ALT: alanine aminotransferase; AST: aspartate aminotransferase; KNN: K-Nearest Neighbor; RF: Random Forest; XGBoost: eXtreme Gradient Boosting; LightGBM: Light Gradient Boosting Machine; LASSO: Least Absolute Shrinkage and Selection Operator; DCA: Decision Curve Analysis.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Ethics Committee of the Affiliated Hospital of Qingdao University (Protocol ID: QYFY WZLL 27010) and was carried out following the Declaration of Helsinki of the World Medical Association.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and analyzed during the current study are available from the corresponding author on reasonable request.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTWZ contributed to data collection, processing, and analysis. HBM performed data analysis and manuscript writing. XZS participated in data collection and manuscript preparation. GFL assisted in data collection. YXL was responsible for manuscript writing, revision, and critical review. MJG contributed to manuscript revision, data verification, and final approval. All authors have read and agreed to the published version of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNone.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical trial number:\u0026nbsp;\u003c/strong\u003enot applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eTan J, Liang Y, Yang Z, He Q, Tong J, Deng Y, et al. Single-Cell Transcriptomics Reveals Crucial Cell Subsets and Functional Heterogeneity Associated With Carotid Atherosclerosis and Cerebrovascular Events. Arterioscler Thromb Vasc Biol. 2023 Dec;43(12):2312\u0026ndash;32. \u003c/li\u003e\n\u003cli\u003eFeigin VL, Brainin M, Norrving B, Martins S, Sacco RL, Hacke W, et al. World Stroke Organization (WSO): Global Stroke Fact Sheet 2022. Int J Stroke. 2022 Jan;17(1):18\u0026ndash;29. \u003c/li\u003e\n\u003cli\u003eWang YJ, Li ZX, Gu HQ, Zhai Y, Zhou Q, Jiang Y, et al. China Stroke Statistics: an update on the 2019 report from the National Center for Healthcare Quality Management in Neurological Diseases, China National Clinical Research Center for Neurological Diseases, the Chinese Stroke Association, National Center for Chronic and Non-communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention and Institute for Global Neuroscience and Stroke Collaborations. Stroke Vasc Neurol. 2022 Oct;7(5):415\u0026ndash;50. \u003c/li\u003e\n\u003cli\u003eKopczak A, Schindler A, Sepp D, Bayer-Karpinska A, Malik R, Koch ML, et al. Complicated Carotid Artery Plaques and Risk of Recurrent Ischemic Stroke or TIA. J Am Coll Cardiol. 2022 Jun 7;79(22):2189\u0026ndash;99. \u003c/li\u003e\n\u003cli\u003evan Dam-Nolen DHK, Truijman MTB, van der Kolk AG, Liem MI, Schreuder FHBM, Boersma E, et al. Carotid Plaque Characteristics Predict Recurrent Ischemic Stroke and TIA: The PARISK (Plaque At RISK) Study. JACC Cardiovasc Imaging. 2022 Oct;15(10):1715\u0026ndash;26. \u003c/li\u003e\n\u003cli\u003eHoward DP, van Lammeren GW, Rothwell PM, Redgrave JN, Moll FL, de Vries JPP, et al. Symptomatic carotid atherosclerotic disease: correlations between plaque composition and ipsilateral stroke risk. Stroke. 2015 Jan;46(1):182\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003evan Dam-Nolen DHK, van Egmond NCM, Dilba K, Nies K, van der Kolk AG, Liem MI, et al. Sex Differences in Plaque Composition and Morphology Among Symptomatic Patients With Mild-to-Moderate Carotid Artery Stenosis. Stroke. 2022 Feb;53(2):370\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eMiceli G, Basso MG, Pintus C, Pennacchio AR, Cocciola E, Cuffaro M, et al. Molecular Pathways of Vulnerable Carotid Plaques at Risk of Ischemic Stroke: A Narrative Review. Int J Mol Sci. 2024 Apr 15;25(8):4351. \u003c/li\u003e\n\u003cli\u003eKarl\u0026ouml;f E, Buckler A, Liljeqvist ML, Lengquist M, Kronqvist M, Toonsi MA, et al. Carotid Plaque Phenotyping by Correlating Plaque Morphology from Computed Tomography Angiography with Transcriptional Profiling. Eur J Vasc Endovasc Surg. 2021 Nov;62(5):716\u0026ndash;26. \u003c/li\u003e\n\u003cli\u003eZamani M, Skagen K, Scott H, Lindberg B, Russell D, Skjelland M. Carotid Plaque Neovascularization Detected With Superb Microvascular Imaging Ultrasound Without Using Contrast Media. Stroke. 2019 Nov;50(11):3121\u0026ndash;7. \u003c/li\u003e\n\u003cli\u003eRudnick MR, Leonberg-Yoo AK, Litt HI, Cohen RM, Hilton S, Reese PP. The Controversy of Contrast-Induced Nephropathy With Intravenous Contrast: What Is the Risk? Am J Kidney Dis. 2020 Jan;75(1):105\u0026ndash;13. \u003c/li\u003e\n\u003cli\u003eLundberg SM, Nair B, Vavilala MS, Horibe M, Eisses MJ, Adams T, et al. Explainable machine-learning predictions for the prevention of hypoxaemia during surgery. Nat Biomed Eng. 2018 Oct;2(10):749\u0026ndash;60. \u003c/li\u003e\n\u003cli\u003eLundberg SM, Erion G, Chen H, DeGrave A, Prutkin JM, Nair B, et al. From Local Explanations to Global Understanding with Explainable AI for Trees. Nat Mach Intell. 2020 Jan;2(1):56\u0026ndash;67. \u003c/li\u003e\n\u003cli\u003eNoubiap JJ, Thomas G, Kamtchum-Tatuene J, Middeldorp ME, Sanders P. High-risk carotid plaques and incident ischemic stroke in patients with atrial fibrillation in the Cardiovascular Health Study. Eur J Neurol. 2023 Jul;30(7):2042\u0026ndash;50. \u003c/li\u003e\n\u003cli\u003eAjoolabady A, Pratico D, Lin L, Mantzoros CS, Bahijri S, Tuomilehto J, et al. Inflammation in atherosclerosis: pathophysiology and mechanisms. Cell Death Dis. 2024 Nov 11;15(11):817. \u003c/li\u003e\n\u003cli\u003eWeber C, Habenicht AJR, von Hundelshausen P. Novel mechanisms and therapeutic targets in atherosclerosis: inflammation and beyond. Eur Heart J. 2023 Aug 1;44(29):2672\u0026ndash;81. \u003c/li\u003e\n\u003cli\u003eNatorska J, Ząbczyk M, Undas A. Neutrophil extracellular traps (NETs) in cardiovascular diseases: From molecular mechanisms to therapeutic interventions. Kardiol Pol. 2023;81(12):1205\u0026ndash;16. \u003c/li\u003e\n\u003cli\u003eShamsuzzaman S, Deaton RA, Salamon A, Doviak H, Serbulea V, Milosek VM, et al. Novel Mouse Model of Myocardial Infarction, Plaque Rupture, and Stroke Shows Improved Survival With Myeloperoxidase Inhibition. Circulation. 2024 Aug 27;150(9):687\u0026ndash;705. \u003c/li\u003e\n\u003cli\u003eWang B, Tang X, Yao L, Wang Y, Chen Z, Li M, et al. Disruption of USP9X in macrophages promotes foam cell formation and atherosclerosis. J Clin Invest. 2022 May 16;132(10):e154217. \u003c/li\u003e\n\u003cli\u003eKuppa A, Tripathi H, Al-Darraji A, Tarhuni WM, Abdel-Latif A. C-Reactive Protein Levels and Risk of Cardiovascular Diseases: A Two-Sample Bidirectional Mendelian Randomization Study. Int J Mol Sci. 2023 May 23;24(11):9129. \u003c/li\u003e\n\u003cli\u003eXu L, Liu JT, Li K, Wang SY, Xu S. Genistein inhibits Ang II-induced CRP and MMP-9 generations via the ER-p38/ERK1/2-PPAR\u0026gamma;-NF-\u0026kappa;B signaling pathway in rat vascular smooth muscle cells. Life Sci. 2019 Jan 1;216:140\u0026ndash;6. \u003c/li\u003e\n\u003cli\u003eHuang Z, Shen S, Han X, Li W, Luo W, Lin L, et al. Macrophage DCLK1 promotes atherosclerosis via binding to IKK\u0026beta; and inducing inflammatory responses. EMBO Mol Med. 2023 May 8;15(5):e17198. \u003c/li\u003e\n\u003cli\u003eKang H, Song J, Cheng Y. HDL regulates the risk of cardiometabolic and inflammatory-related diseases: Focusing on cholesterol efflux capacity. Int Immunopharmacol. 2024 Sep 10;138:112622. \u003c/li\u003e\n\u003cli\u003eHu S, Zhu Y, Zhao X, Li R, Shao G, Gong D, et al. Hepatocytic lipocalin-2 controls HDL metabolism and atherosclerosis via Nedd4-1-SR-BI axis in mice. Dev Cell. 2023 Nov 6;58(21):2326-2337.e5. \u003c/li\u003e\n\u003cli\u003eMiceli G, Basso MG, Pintus C, Pennacchio AR, Cocciola E, Cuffaro M, et al. Molecular Pathways of Vulnerable Carotid Plaques at Risk of Ischemic Stroke: A Narrative Review. Int J Mol Sci. 2024 Apr 15;25(8):4351. \u003c/li\u003e\n\u003cli\u003eWolberg AS. Fibrinogen and fibrin: synthesis, structure, and function in health and disease. J Thromb Haemost. 2023 Nov;21(11):3005\u0026ndash;15. \u003c/li\u003e\n\u003cli\u003eJiang S, Ai Y, Ni L, Wu L, Huang X, Chen S. Platelet-derived TGF-\u0026beta;1 is related to portal vein thrombosis in cirrhosis by promoting hypercoagulability and endothelial dysfunction. Front Cardiovasc Med. 2022;9:938397. \u003c/li\u003e\n\u003cli\u003eL S, V N, R C, A G, H K, Js S, et al. Carotid Artery Plaque Calcifications: Lessons From Histopathology to Diagnostic Imaging. Stroke [Internet]. 2022 Jan [cited 2025 Apr 13];53(1). Available from: https://pubmed.ncbi.nlm.nih.gov/34753301/\u003c/li\u003e\n\u003cli\u003eZhang S, Jiang S, Wang C, Han C. Comparison of ultrasonic shear wave elastography, AngioPLUS planewave ultrasensitive imaging, and optimized high-resolution magnetic resonance imaging in evaluating carotid plaque stability. PeerJ. 2023;11:e16150. \u003c/li\u003e\n\u003cli\u003eNarayanan S, Vuckovic S, Bergman O, Wirka R, Verdezoto Mosquera J, Chen QS, et al. Atheroma transcriptomics identifies ARNTL as a smooth muscle cell regulator and with clinical and genetic data improves risk stratification. Eur Heart J. 2025 Jan 16;46(3):308\u0026ndash;22. \u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-cardiovascular-disorders","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bcar","sideBox":"Learn more about [BMC Cardiovascular Disorders](http://bmccardiovascdisord.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bcar/default.aspx","title":"BMC Cardiovascular Disorders","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Carotid plaque instability, Machine learning, Relevant factors, Predictive model","lastPublishedDoi":"10.21203/rs.3.rs-6966994/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6966994/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eObjective: \u003c/strong\u003eThe objective of this study was to develop and evaluate the effectiveness of machine learning (ML) models in predicting carotid plaque instability.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003eA retrospective analysis was conducted on data from 449 patients with carotid plaques treated at The Affiliated Hospital of Qingdao University between July 2022 and November 2024. Patients were randomly divided into a training set and a testing set at a 7:3 ratio. Five ML algorithms were utilized to establish prediction models. The predictive performance of each model was assessed using the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity on the testing set.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003eAmong the five ML algorithms, the light gradient boosting machine (LightGBM) demonstrated the best performance with an AUC of 0.921 (95% confidence interval [CI]: 0.887–0.956) and an accuracy of 0.837 in the training set, and an AUC of 0.864 (95% CI: 0.784–0.944) and an accuracy of 0.811 in the testing set. The feature importance results showed that C-reactive protein (CRP), leukocytes, high-density lipoprotein (HDL), and platelets were the four most significant contributors to carotid plaque instability.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion:\u003c/strong\u003eWe developed and validated the ML models for predicting carotid plaque instability, and identified that LightGBM model with 9 relevant factors had the best performance and high levels of clinical applicability.\u003c/p\u003e","manuscriptTitle":"A Study on Analyzing Risk Factors for Carotid Plaque Instability Using Machine Learning Models","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-11 10:03:08","doi":"10.21203/rs.3.rs-6966994/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-09-10T11:39:43+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-25T14:16:21+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"76840995776167125228827206743668185965","date":"2025-08-25T08:42:42+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-24T09:17:42+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"298246404414036809261381601259630143125","date":"2025-08-24T00:17:05+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"123811750672211740087462136262284385980","date":"2025-08-23T20:33:58+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-16T07:23:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"142765057726893595113956952013115607560","date":"2025-08-08T03:44:04+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-08-05T19:24:38+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-07-14T14:39:53+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-07-04T03:55:17+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-07-04T03:54:51+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Cardiovascular Disorders","date":"2025-06-24T14:48:19+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-cardiovascular-disorders","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bcar","sideBox":"Learn more about [BMC Cardiovascular Disorders](http://bmccardiovascdisord.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bcar/default.aspx","title":"BMC Cardiovascular Disorders","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0d38eae7-6ecc-4186-b1e5-b59b7f249dca","owner":[],"postedDate":"August 11th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2025-09-25T18:08:20+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-11 10:03:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6966994","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6966994","identity":"rs-6966994","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}