Progression Prediction with Machine Learning Methods among the Anterior Choroidal Artery Infarction

Progression Prediction with Machine Learning Methods among the Anterior Choroidal Artery Infarction | 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 Progression Prediction with Machine Learning Methods among the Anterior Choroidal Artery Infarction Mengying Chen, Jiaxin Fan, Shuyin Ma, Mengyuan Zhang, Xiaodong Zhang, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7910284/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Background Patients with anterior choroidal artery (AChA) infarction often experience fluctuating symptoms or neurological progression despite aggressive treatment. Machine learning, known for its high accuracy, is increasingly applied in medicine. This study explored the use of machine learning techniques to predict neurological progression in AChA infarction patients. Methods A total of 369 AChA infarction patients were retrospectively enrolled. Demographic and clinical characteristics were collected and analyzed. Clinical features were compared between patients with and without neurological progression. Six machine learning models—logistic regression (LR), random forest (RF), decision tree (DT), eXtreme Gradient Boosting (XGB), support vector machine (SVM), and K-Nearest Neighbour (KNN)—were employed to predict progression. Their performance was evaluated using the area under the receiver operating characteristic curve (AUC ROC), accuracy (ACC), and F1 score (F1). Results Neurological progression occurred in one-third of AChA infarction patients. Hemiparesis was the most common manifestation (87.8%), occurring more frequently in the progression group than in the no-progression group (95.1% vs. 84.1%, P = 0.002). Although more men were affected (75.3%), they had a lower likelihood of progression (66.7% vs. 79.7%, P = 0.006). No significant differences were observed in stroke location or intravenous rt-PA therapy between the groups. Among the models, RF demonstrated the best predictive performance, achieving the highest AUC ROC (0.851), accuracy (70.3%), and F1 score (52.2%). The top three predictors identified by RF were NIHSS score at peak, NIHSS score at admission, and age. Conclusion Motor deficits are the most frequent and characteristic symptoms in AChA infarcts. RF emerged as the most effective model for predicting clinical progression in these patients, offering a simple and useful tool for risk assessment. Special attention should be given to NIHSS scores at peak and admission, as well as patient age. Anterior choroidal artery Machine learning Clinical progression Prediction Figures Figure 1 Figure 2 Figure 3 1. Introduction Anterior choroidal artery (AChA) infarction represents a distinct clinical entity of cerebral infarction, often resulting in a wide spectrum of neurological deficits. Despite active management, including antiplatelet therapy and even thrombolysis, patients frequently experience symptom fluctuation or neurological progression, drawing increasing attention from clinicians. However, the underlying causes, pathogenesis, and mechanisms of disease progression in AChA infarction remain unclear. Several studies have attempted to predict progression in stroke patients using clinical observations, radiological findings, electrophysiological results, functional neuroimaging, diffusion tensor imaging (DTI), and even transcranial magnetic stimulation (TMS) [ 1 , 2 ] . Nevertheless, the application of machine learning methods for predicting motor outcomes in AChA infarction has not yet been reported. In this study, we retrospectively analyzed patients with newly diagnosed cerebral infarction confined to the AChA territory, admitted to the Second Affiliated Hospital of Xi'an Jiaotong University, and compared those with and without clinical progression. Six machine learning models—logistic regression (LR), random forest (RF), decision tree (DT), eXtreme Gradient Boosting (XGB), support vector machine (SVM), and K-Nearest Neighbour (KNN)—were employed to predict clinical progression. Their predictive performance was evaluated using the area under the receiver operating characteristic curve (AUC ROC), accuracy (ACC), and F1 score (F1). This study aims to provide a theoretical foundation for improving clinical management and reducing adverse outcomes in AChA infarction. 2. Methods 2.1. Data Source A total of 4,150 patients with acute ischemic stroke admitted to our hospital between January 2020 and January 2025 were initially screened. Among them, 369 patients were confirmed to have AChA infarction based on CT, MRI, and DWI findings. This study contributes a substantial cohort of patients with AChA-restricted infarcts diagnosed via DWI–MRI. Patients with liver or kidney dysfunction, cardiopulmonary failure, or new infarcts outside the AChA territory were excluded. The National Institutes of Health Stroke Scale (NIHSS) and modified Rankin Scale (mRS) were assessed at admission. The study was approved by the hospital’s Ethics Committee, and informed consent was obtained from all patients or their families. 2.2. Predictors and Data Preprocessing A total of 27 variables were selected initially as predictors: demographic characteristics (sex, age, and comorbidities including diabetes, hypertension, and heart disease), lifestyle factors (smoking and drinking), and clinical variables (clinical manifestations, intravenous thrombolysis therapy (IVT), blood glucose (GLU), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), ferritin, uric acid (UA), homocysteine (HCY), and glycosylated hemoglobin (GHb). AChA infarction primarily involves the posterior limb of the internal capsule (PLIC) and/or the posterior portion of the corona radiata (CR) [ 3 ] . Vascular assessments were performed using DSA, CTA, or MRA to evaluate stenosis or occlusion in the internal carotid artery (ICA), middle cerebral artery (MCA), and posterior cerebral artery (PCA) on the affected side. All imaging results were independently reviewed by two neuroradiologists and two neurologists. Final assessments of lesion location and vascular stenosis degree were determined by consensus. All predictors are detailed in Table 1 and Table S1 (Supplementary Materials). 2.3. Outcome Definition Patients were classified into a progressive group (NIHSS score increase > 2 points within one week) and a non-progressive group (NIHSS score increase ≤ 2 points within one week), in accordance with established criteria [ 4 ] . 2.4. Models Comparison Six machine learning algorithms—logistic regression (LR), random forest (RF), decision tree (DT), eXtreme Gradient Boosting (XGB), support vector machine (SVM), and K-Nearest Neighbour (KNN)—were applied to predict the binary outcome of clinical progression. These models were selected due to their widespread use and recognized performance in clinical data classification [ 5 ] . Among the initially selected 27 variables, ferritin, uric acid and glycation were deleted since their missing values exceed 10%. Meanwhile, we combined the analysis of motor and sensory disorder, dysarthria and visual impairment. Therefore, ultimately 22 variables were included in the machine learning analysis. Model performance was evaluated using accuracy, F1 score, specificity, precision, recall, and the area under the receiver operating characteristic curve (AUCROC). Among these, AUCROC was considered the primary performance indicator as it provides a comprehensive measure of model discriminative ability. Accuracy and F1 score served as secondary metrics. The best-performing model was further analyzed using a confusion matrix. 2.5. Statistical Analyses Statistical analyses were conducted using SPSS version 22. Categorical and continuous variables were compared using the chi-square test and Student’s t-test (or rank-sum test for non-parametric data), respectively. A two-tailed p-value < 0.05 was considered statistically significant. Machine learning models were implemented in Python version 3.12.3 using the Scikit-Learn library version 1.2.2. 3. Results 3.1. Patient Demographics, Risk Factors and Clinical Progression A total of 369 patients (8.89% of the screened cohort) were included with isolated AChA infarcts. The majority were male (75.3%). Hemiparesis (87.8%) was the most frequent clinical manifestation, followed by aphasia (55.3%), hemianesthesia (27.1%), and other syndromes. The most commonly involved infarct sites were the posterior limb of the internal capsule and the paraventricular corona radiata (58.5%). Ipsilateral carotid plaque or stenosis was observed in 108 patients (29.3%), and 24 patients (6.5%) received intravenous rt‑PA thrombolysis according to standard protocols. Clinical progression occurred in approximately one‑third of the patients (33.3%). No significant differences were observed in age, infarct location, or rt‑PA administration between the progression and non‑progression groups. Although the overall proportion of males was higher, progression was less frequent in men compared to women (66.7% vs. 79.7%, p = 0.006). Similarly, nicotine abuse was associated with a lower risk of progression (39.8% vs. 55.3%, p = 0.005). Hemiparesis was more common in the progression group than in the non‑progression group (95.1% vs. 84.1%, p = 0.002). The proportion of CA plaque or stenosis was significantly higher in the progression group (52.8% vs. 17.5%, p < 0.001), as was the proportion of MCA or PCA plaque or stenosis. Moreover, patients in the progression group exhibited more severe neurological deficits at both admission and peak assessment ( p < 0.001). Serum ferritin levels were significantly lower in the progression group (163.5 ± 89.8 vs. 203.1 ± 135.9, p = 0.015). Detailed results are summarized in Table 1 . Table 1 Patient demographics, risk factors and clinical progression Variables Total (n = 369) Clinical progression (n = 123) No-clinical progression (n = 246) p -value Age (mean ± sd) 60.8 ± 11.2 61.6 ± 11.8 60.3 ± 10.9 0.305 Male, n (%) 278(75.3) 82(66.7) 196(79.7) 0.006 # Risk factors, n (%) Diabetes 110(29.8) 35(28.5) 75(30.5) 0.687 Hypertension 246(66.7) 80(65.0) 166(67.5) 0.639 Coronary heart disease 30(8.1) 11(8.9) 19(7.7) 0.686 Atrial fibrillation 7(1.9) 1(0.8) 6(2.4) 0.500 Nicotine abuse 185(50.1) 49(39.8) 136(55.3) 0.005 # Alcohol abuse 84(22.8) 30(24.4) 54(22.0) 0.598 Symptom, n (%) Hemiparesis 324(87.8) 117(95.1) 207(84.1) 0.002 # Hemianesthesia 100(27.1) 39(31.7) 61(24.8) 0.159 Aphasia 204(55.3) 67(54.5) 137(55.7) 0.824 Hemianopia 3(0.8) 0(0) 3(1.2) 0.539 Location, n (%) PLIC 78(21.1) 22(17.9) 56(22.8) 0.279 CR 75(20.3) 22(17.9) 53(21.5) 0.410 PLIC + CR 216(58.5) 79(64.2) 137(55.7) 0.117 CA plaque or stenosis 108(29.3) 65(52.8) 43(17.5) ๤0.001 # MCA or PCA plaque or stenosis 122(33.1) 57(46.3) 65(26.4) ๤0.001 # NIHSS score, median (IQR) At admission 3(2–4) 3(2-5.5) 2(2–3) ๤0.001 # At peak 3(2–5) 4(3–6) 2(2–3) ๤0.001 # mRS score, median (IQR) 2(1–3) 2(2–4) 2(1–2) ๤0.001 # blood test (mean ± sd) GLU, mmol/L 6.1 ± 2.4 6.1 ± 2.4 6.1 ± 2.4 0.901 GHb, % 6.5 ± 1.7 6.5 ± 1.8 6.4 ± 1.7 0.672 TG, mmol/L 1.7 ± 1.1 1.6 ± 0.8 1.8 ± 1.3 0.063 HDL-C, mmol/L 1.1 ± 0.3 1.1 ± 0.2 1.1 ± 0.3 0.563 LDL-C, mmol/L 2.9 ± 0.9 3.0 ± 0.9 2.9 ± 0.9 0.605 UA, µmol/L 321.8 ± 86.8 308.6 ± 86.0 328.4 ± 86.7 0.051 HCY, µmol/L 23.0 ± 16.4 23.9 ± 18.2 21.8 ± 14.7 0.663 Ferritin, µg/L 189.8 ± 122.3 163.5 ± 89.8 203.1 ± 135.9 0.015 # IV thrombolysis, n (%) 24(6.5) 12(9.8) 12(4.9) 0.073 IV, intravenous; GLU, blood glucose; TG, triglyceride; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; UA, uric acid; HCY, homocysteine; GHb, glycosylated hemoglobin; PLIC, the posterior limb of the internal capsule; CR, corona radiata; MCA, middle cerebral artery; PCA, posterior cerebral artery. 3.2. Model comparison for binary classification problem A comparison of the receiver operating characteristic curve for six models is shown in Fig. 1 and Table 2 . The differences between these curves were slight, but we could still clearly recognize each model. KNN showed the poorest predictive performance, with the lowest AUCROC of 0.713, an accuracy of 66.6%, and a F1 score of 46.3%. In comparison the other five models improved prediction: LR (AUCROC: 0.799, accuracy: 71.2% and F1 score: 46.7%); DT (0.755, 74.8%, and 65.8% respectively); XGB (0.804, 63.9%, and 0 respectively); SVM (0.759, 63.9%, and 0 respectively); and RF, which had the highest AUCROC of 0.851, an accuracy of 70.3%, and a F1 score of 52.2%. Detailed predictions of RF were presented in the form of a confusion matrix (Fig. 2 for more details). In the present context, RF was able to predict 87/123 progressive AChA infarction (sensitivity 70.3%) and 208/246 non- progressive AChA infarction (specificity 84.5%). The results of the tenfold cross-validation showed that RF had a better discriminative ability for progressive AChA infarction than the other five models (Table 3 ). Table 2 Comparison of the predictive performance for six models Model Accuracy(%) Specificity(%) Precision(%) Recall(%) F1 score(%) AUCROC LR 71.2 91.5 70.0 35.0 46.7 0.799 RF 70.3 84.5 62.1 45.0 52.2 0.851 DT 74.8 0.788732 64.3 67.5 65.8 0.755 XGB 63.9 100.0 0 0 0 0.804 SVM 63.9 100.0 0 0 0 0.759 KNN 66.6 81.7 55.2 40.0 46.3 0.713 LR, logistic regression; RF, random forest; DT, decision tree; XGB, eXtreme Gradient Boosting; SVM, support vector machine; KNN, K-Nearest Neighbour. Table 3 Comparison of the performance for six models (tenfold cross-validation) Model AUCROC Model AUCROC LR 0.693 XGB 0.639 RF 0.783 SVM 0.639 DT 0.684 KNN 0.720 LR, logistic regression; RF, random forest; DT, decision tree. Moreover, according to the information gain values of RF model, we further ranked those 22 features, as shown in Fig. 3 . The NIHSS score at peak contributed the most to the progressive AChA infarction outcome, followed by NIHSS score at admission, age, high density lipoprotein, homocysteine and so on. 4. Discussion Neurological deterioration is a well-recognized phenomenon in AChA infarction and is independently associated with poor outcomes [ 6 ] . The perforating arteries involved are terminal arterioles with limited collateral circulation, rendering them particularly vulnerable to hemodynamic fluctuations. Moreover, the high density of nerve fibers in the paraventricular and pontine regions means that even a small expansion of the infarct can lead to progressive neurological deficits [ 7 ] . Previous studies have reported clinical progression rates in AChA infarcts ranging from 7% to 43% [ 8 , 9 ] , which is notably higher than in hemispheric or deep infarcts [ 9 ] . In our cohort, clinical progression occurred in 33.3% of patients, predominantly among those with higher baseline NIHSS scores. No AChA infarction-related fatalities were observed, reaffirming its generally non-lethal nature. We also identified a significantly higher prevalence of ipsilateral severe carotid stenosis or occlusion in patients with clinical progression, suggesting that carotid atherosclerosis may contribute substantially to disease progression. While small vessel disease remains a key etiological factor, the presence of large artery disease combined with multiple risk factors appears to be associated with clinical worsening in AChA infarcts [ 8 , 9 ] . Intravenous thrombolysis did not significantly prevent clinical progression in our study, possibly due to the small caliber of the affected artery and thrombus size [ 8 ] . Nevertheless, thrombolysis has been linked to favorable functional outcomes and a low risk of hemorrhagic transformation [ 10 ] . In contrast, Wu et al. reported that thrombolysis reduced the risk of stroke evolution in larger AChA infarcts [ 11 ] , highlighting a potential size-dependent treatment effect. Previous studies have explored machine learning for predicting stroke outcomes [ 12 , 13 ] , and various biomarkers—such as neutrophil-related inflammatory indices, ADC values, and infarct dimensions—have been investigated for forecasting AChA infarction progression [ 14 , 15 ] . In this study, we demonstrated that machine learning models can effectively predict clinical progression in acute AChA infarction. Using 22 input variables and alongside extensive hyperparameter tuning, we compared six machine learning algorithms at their optimized performance. Among them, Random Forest (RF) achieved the highest predictive performance, with an AUCROC of 0.851, accuracy of 70.3%, and an F1 score of 52.2% on the hold-out test set. This result aligns with prior studies applying machine learning to clinical event prediction [ 16 , 17 ] . As an ensemble method, RF aggregates multiple decision trees to enhance predictive accuracy and robustness [ 18 ] . The tenfold cross-validation further confirmed its superior discriminative capability. Based on 5 clinical data, Fanhai Bu et al. developed prediction models for acute ischemic stroke following intravenous thrombolysis using eight machine learning algorithms, with their optimal logistic regression model achieving an AUC of 0.792 [ 19 ] . In contrast, our RF model demonstrated superior performance, attaining an AUC of 0.851—a 6% improvement in predictive accuracy that holds significant clinical utility. This enhancement may be attributed to our comprehensive approach: 1) addressing missing data through multiple imputation methods; 2) incorporating more extensive predictive variables including radiomic features from neuroimaging modalities; and 3) implementing meticulous hyperparameter optimization to refine model architecture. Notably, NIHSS scores were among the most important predictors identified by the RF model, underscoring that patients with severe initial neurological impairment are at higher risk of progression. This finding should be considered during clinical communication with patients and families. Several limitations of this study should be acknowledged. Firstly, the sample size for analyzing AChA infarction progression using machine learning was relatively small, and future studies should seek to expand the cohort. Secondly, stratifying AChA infarcts by size (e.g., small vs. large) in subsequent analyses could yield more nuanced insights. Thirdly, although we used multiple imputation for missing data, imputation model specifications may introduce uncertainty affecting model validity. Additionally, our prediction model was validated only internally, external validation in larger, multicenter, prospective cohorts is needed to confirm generalizability. 5. Conclusions Motor deficits represent the most frequent and characteristic clinical manifestation in AChA infarcts. This study identifies Random Forest as the most effective machine learning model for predicting clinical progression in AChA infarction, demonstrating its utility as a practical and accessible tool for risk stratification. Particular attention should be given to NIHSS scores at admission and peak, along with patient age, as these factors were identified as key predictors of neurological deterioration. Declarations All machine learning methods were carried out in accordance with the “Guidelines for Developing and Reporting Machine Learning Predictive Models in Biomedical Research”. And the protocol conformed to the Declaration of Helsinki. Ethics approval and consent to participate The ethics committee at The Second Affiliated Hospital of Xi’an Jiaotong University had approved this study (No.2024-047). Written informed consent was obtained from individual orguardian participants. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Author details 1 Department of Neurology, The Second Affiliated Hospital of Xi’an Jiaotong University, No. 157 West Five Road, Xi’an 710004, Shaanxi, China. 2 Department of Neurology, Norinco General Hospital, No. 12 Zhangba East Road, Xi'an, 710077, Shaanxi, China. Funding This study is supported by the National Natural Science Foundation of China (No. 81070999); the Foundation of Shaanxi social development and technology research project (No. 2016SF-020); the Foundation of Xi’an Science and technology plan project (No. 2019114613YX001SF039(2)); the new medical technology of the Second Affiliated Hospital of Xi’an Jiaotong University (No. 2019-32, 2018-16, 2010-22); the Fundamental Research Funds for the Central Universities (Xi’an Jiaotong University, No. xjj2014153, 2009-95); the Foundation of Second Affiliated Hospital of Xi’an Jiaotong University (No. RC(GG)201109). Author Contribution MYC and SQZ conceived and designed the study. MYC and SYM analyzed the data, wrote, and revised the manuscript. JXF developed the models and guided the Python code. MYZ, QLY, and XDZ collected the data. SD, HYQ, and YXC conducted the data. HY obtained ethical approval. SQZ monitored the entire planning, revision, and provided funding. All authors read and approved the final manuscript. Acknowledgements Not applicable. Data Availability The datasets generated and/or analysed during the current study are not publicly available due privacy but are available from the corresponding author on reasonable request. References Kwon YH, Son SM, Lee J, et al. Combined study of transcranial magnetic stimulation and diffusion tensor tractography for prediction of motor outcome in patients with corona radiata infarct. J Rehabil Med. 2011;43(5):430–4. Nelles M, Gieseke J, Flacke S, et al. Diffusion tensor pyramidal tractography in patients with anterior choroidal artery infarcts. AJNR Am J Neuroradiol. 2008;29(3):488–93. Sohn H, Kang DW, Kwon SU, et al. Anterior choroidal artery territory infarction: lesions confined to versus beyond the internal capsule. Cerebrovasc Dis. 2013;35(3):228–34. Poh KW, Er CK, Hoh WH, et al. Neurological deterioration and its risk score in total anterior circulation infarct. Clin Neurol Neurosurg. 2020;191:105684. Kop R, Hoogendoorn M, Teije AT, et al. Predictive modeling of colorectal cancer using a dedicated pre-processing pipeline on routine electronic medical records. Comput Biol Med. 2016;76:30–8. Chausson N, Joux J, Saint-Vil M, et al. Infarction in the anterior choroidal artery territory: clinical progression and prognosis factors. J Stroke Cerebrovasc Dis. 2014;23(8):2012–7. Yamamoto Y, Nagakane Y, Tomii Y, et al. The relationship between progressive motor deficits and lesion location in patients with single infarction in the lenticulostriate artery territory. J Neurol. 2017;264(7):1381–7. Cheng Z, Duan H, Meng F, et al. Acute Anterior Choroidal Artery Territory Infarction: A Retrospective Study. Clin Neurol Neurosurg. 2020;195:105826. Ois A, Cuadrado-Godia E, Solano A, et al. Acute ischemic stroke in anterior choroidal artery territory. J Neurol Sci. 2009;281(1–2):80–4. Trouillas P, Derex L, Nighoghossian N, et al. rtPA intravenous thrombolysis in anterior choroidal artery territory stroke. Neurology. 2000;54(3):666–73. Wu MC, Tsai LK, Wu CC, et al. Thrombolytic therapy is an only determinant factor for stroke evolution in large anterior choroidal artery infarcts. J Stroke Cerebrovasc Dis. 2014;23(5):1089–93. Wu Y, Fang Y. Stroke Prediction with Machine Learning Methods among Older Chinese. Int J Environ Res Public Health, 2020, 17 (6). Heo J, Yoon JG, Park H, et al. Machine Learning-Based Model for Prediction of Outcomes in Acute Stroke. Stroke. 2019;50(5):1263–5. Zhao N, Li J, Zhang QX, et al. Elevated neutrophil-related immune-inflammatory biomarkers in acute anterior choroidal artery territory infarction with early progression. Clin Neurol Neurosurg. 2023;229:107720. Yu W, Yang J, Liu L, et al. The value of diffusion weighted imaging in predicting the clinical progression of perforator artery cerebral infarction. Neuroimage Clin. 2022;35:103117. Lu Y, Ning Y, Li Y, et al. Risk factor mining and prediction of urine protein progression in chronic kidney disease: a machine learning- based study. BMC Med Inf Decis Mak. 2023;23(1):173. Shi Y, Zhang G, Ma C et al. Machine learning algorithms to predict intraoperative hemorrhage in surgical patients: a modeling study of real-world data in Shanghai, China. BMC Med Inform Decis Mak. 2023, 23 (1): 156. Breiman L. Random Forests. Machine Learning, 2001, 45 (1): 5–32. Bu F, Cai R, Zhang W, et al. Explainable machine learning model for predicting the outcome of acute ischemic stroke after intravenous thrombolysis. Front Neurol. 2025;16:1668816. Additional Declarations No competing interests reported. Supplementary Files Supplementarymaterial.xls Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 24 Nov, 2025 Reviewers agreed at journal 19 Nov, 2025 Reviewers invited by journal 19 Nov, 2025 Editor invited by journal 24 Oct, 2025 Editor assigned by journal 23 Oct, 2025 Submission checks completed at journal 23 Oct, 2025 First submitted to journal 20 Oct, 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-7910284","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":551336948,"identity":"e8fedfb3-31f3-4baf-a671-0b404dc4a0ca","order_by":0,"name":"Mengying Chen","email":"","orcid":"","institution":"the Second Affiliated Hospital of Xi’an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Mengying","middleName":"","lastName":"Chen","suffix":""},{"id":551336949,"identity":"27ccc2ea-57f0-428c-8880-1205ac5e3e53","order_by":1,"name":"Jiaxin Fan","email":"","orcid":"","institution":"the Second Affiliated Hospital of Xi’an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Jiaxin","middleName":"","lastName":"Fan","suffix":""},{"id":551336950,"identity":"d6cec477-ba3c-4b1b-bb0a-5af4ec4d227f","order_by":2,"name":"Shuyin Ma","email":"","orcid":"","institution":"the Second Affiliated Hospital of Xi’an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Shuyin","middleName":"","lastName":"Ma","suffix":""},{"id":551336951,"identity":"6ea95969-e921-4cfa-a85c-4f2a002f286c","order_by":3,"name":"Mengyuan Zhang","email":"","orcid":"","institution":"Norinco General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Mengyuan","middleName":"","lastName":"Zhang","suffix":""},{"id":551336953,"identity":"241d1136-3d53-4997-b364-7bae5d510fdd","order_by":4,"name":"Xiaodong Zhang","email":"","orcid":"","institution":"the Second Affiliated Hospital of Xi’an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Xiaodong","middleName":"","lastName":"Zhang","suffix":""},{"id":551336958,"identity":"35efcdff-f534-4617-82c9-386b013f9efb","order_by":5,"name":"Minyu Duan","email":"","orcid":"","institution":"the Second Affiliated Hospital of Xi’an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Minyu","middleName":"","lastName":"Duan","suffix":""},{"id":551336959,"identity":"6c5ba2b1-1a07-456a-8a60-b11167a5dcae","order_by":6,"name":"Qingling Yao","email":"","orcid":"","institution":"the Second Affiliated Hospital of Xi’an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Qingling","middleName":"","lastName":"Yao","suffix":""},{"id":551336960,"identity":"a47c4acd-65a1-47ca-a798-dff9c3d7cfa3","order_by":7,"name":"Han Yang","email":"","orcid":"","institution":"the Second Affiliated Hospital of Xi’an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Han","middleName":"","lastName":"Yang","suffix":""},{"id":551336961,"identity":"85cd3a7f-615d-479b-b3cf-cddd9e2e2d8d","order_by":8,"name":"Shuang Du","email":"","orcid":"","institution":"the Second Affiliated Hospital of Xi’an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Shuang","middleName":"","lastName":"Du","suffix":""},{"id":551336962,"identity":"5c72983d-6731-48c6-96d8-3626d22afb8d","order_by":9,"name":"Yuxuan Cheng","email":"","orcid":"","institution":"the Second Affiliated Hospital of Xi’an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Yuxuan","middleName":"","lastName":"Cheng","suffix":""},{"id":551336963,"identity":"15608ad7-15f0-4744-8cdf-a3bdaf934d68","order_by":10,"name":"Huiyang Qu","email":"","orcid":"","institution":"the Second Affiliated Hospital of Xi’an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Huiyang","middleName":"","lastName":"Qu","suffix":""},{"id":551336964,"identity":"611329c6-6e38-4827-92f8-4cada9362c83","order_by":11,"name":"Hong Sun","email":"","orcid":"","institution":"the Second Affiliated Hospital of Xi’an Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Hong","middleName":"","lastName":"Sun","suffix":""},{"id":551336965,"identity":"35169b57-4506-4ae4-98ff-e2e1a3274bbe","order_by":12,"name":"Shuqin Zhan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+klEQVRIiWNgGAWjYHACNiC2YWBgZm5gYOABixgQoyUNqIURpMWAaC2HgRikhYEILfLt6c8e/NxxPpq/nbGBmUfmT2IDe/M2CYaaOzi1GJx5kG7Ye+Z27ozDjA2MM3gMEht4jpVJMBx7hluLRMIxCd6227kNQC0MH0BaJHLMJBgbDuN22IzENsm/bedy54O0JIC0yL/Br4XhRjKbNG/bgdwNCFt48GsxOPOMTVq2LTl3I1DLwRk8xsZtPGnFFgnH8DgMGGKSb9vscuedP3zwMW+PnGw/++GNNz7U4HEYQwKCeYCxBxJNyIL4tTAw/MCrdBSMglEwCkYoAAA261PYvMQNvAAAAABJRU5ErkJggg==","orcid":"","institution":"the Second Affiliated Hospital of Xi’an Jiaotong University","correspondingAuthor":true,"prefix":"","firstName":"Shuqin","middleName":"","lastName":"Zhan","suffix":""}],"badges":[],"createdAt":"2025-10-21 04:08:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7910284/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7910284/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97095930,"identity":"6e9ee034-891b-4691-9efc-3eb367fb43bf","added_by":"auto","created_at":"2025-11-30 23:27:11","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":174792,"visible":true,"origin":"","legend":"","description":"","filename":"CMYProgressionPredictionwithMachineLearningMethodsamongtheAnteriorChoroidalArteryInfarction.docx","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/a792034fd5f20a30de20651c.docx"},{"id":97140251,"identity":"dcdcb268-843b-4a9d-ae0b-fb3bc402db4e","added_by":"auto","created_at":"2025-12-01 10:04:20","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13105,"visible":true,"origin":"","legend":"","description":"","filename":"433aba9622f64faf9b3f65bb087ef502.json","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/34d25adb49783079a6bb25fe.json"},{"id":97140998,"identity":"debef530-b881-4961-b29b-4b77fc335e21","added_by":"auto","created_at":"2025-12-01 10:06:05","extension":"xls","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":111616,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial.xls","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/debce6b23605b082d0e7f5a2.xls"},{"id":97141033,"identity":"eec0f1fb-d6d4-41cf-825b-5b610b74f2f1","added_by":"auto","created_at":"2025-12-01 10:06:09","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":77037,"visible":true,"origin":"","legend":"","description":"","filename":"433aba9622f64faf9b3f65bb087ef5021enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/5d813cecc72e188320d94e2c.xml"},{"id":97141002,"identity":"aee87dfc-24b5-4a00-91ca-75921e7853be","added_by":"auto","created_at":"2025-12-01 10:06:06","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11828,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/2ab6523ed24c9d8beb61893d.png"},{"id":97095934,"identity":"9529b03b-4d75-4e2c-8fb9-167b22225fdc","added_by":"auto","created_at":"2025-11-30 23:27:12","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6892,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/8953aa5a8b07cf3d16a23327.png"},{"id":97140303,"identity":"6a2ee44e-6fe5-420e-a9e4-705705c0fe6b","added_by":"auto","created_at":"2025-12-01 10:04:32","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14003,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/d505180cda425de1bcf584dd.png"},{"id":97139912,"identity":"300109de-632f-43ec-ac76-3b69fb062ed0","added_by":"auto","created_at":"2025-12-01 10:03:16","extension":"xml","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":75052,"visible":true,"origin":"","legend":"","description":"","filename":"433aba9622f64faf9b3f65bb087ef5021structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/8ff84c5ef713558d536a7f81.xml"},{"id":97095936,"identity":"8c489b96-bdf9-41e4-81d7-663554341c57","added_by":"auto","created_at":"2025-11-30 23:27:12","extension":"html","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":82239,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/4642716762c7ecccdcce343e.html"},{"id":97095926,"identity":"3899df7c-a538-4116-b6da-ad2462d3ef2a","added_by":"auto","created_at":"2025-11-30 23:27:11","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":41905,"visible":true,"origin":"","legend":"\u003cp\u003ePerformance characteristic curves for six models. LR, logistic regression;RF, random forest; DT, decision tree; XGB, eXtreme Gradient Boosting; SVM, support vector machine; KNN, K-Nearest Neighbour.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/6f2404244b412fb83f9f82e4.png"},{"id":97095925,"identity":"af8a3391-5c45-49a3-a043-86f7487a9910","added_by":"auto","created_at":"2025-11-30 23:27:11","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":22510,"visible":true,"origin":"","legend":"\u003cp\u003eDetailed predictions of RF were presented in the form of a confusion matrix\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/fca22c114aef660f7b7f271c.png"},{"id":97095928,"identity":"fb5282e0-53ad-45a0-a0d7-1871c9c00721","added_by":"auto","created_at":"2025-11-30 23:27:11","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":49259,"visible":true,"origin":"","legend":"\u003cp\u003eStandard feature importance bar chart shows the importance of each predictor in the RF model. RF, random forest. GLU, blood glucose; TG, triglyceride; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; HCY, homocysteine; NIHSS, National Institutes of Health Stroke Scale; MPPS, MCA or PCA plaque or stenosis; mRS, modified Rankin Scale; CAPS, CA plaque or stenosis; IL, infarction location; DSI, dysarthria and visual impairment; HBP, Hypertension; MSD, motor sensory disorder; DM, Diabetes Mellitus; CVD, cerebrovascular disease; AF, atrial fibrillation.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/71233fd52f544b3d9cb5cf3f.png"},{"id":97145185,"identity":"d2ee6a34-0800-4fcc-9ca8-9c116e958175","added_by":"auto","created_at":"2025-12-01 10:13:14","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":932275,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/ea662ff8-36c4-4029-b070-e9ecd3fcf8e4.pdf"},{"id":97141539,"identity":"dce4b690-6a71-4522-bc5b-43d35426a3be","added_by":"auto","created_at":"2025-12-01 10:06:47","extension":"xls","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":111616,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial.xls","url":"https://assets-eu.researchsquare.com/files/rs-7910284/v1/1807f191ae3353114f736eb1.xls"}],"financialInterests":"No competing interests reported.","formattedTitle":"Progression Prediction with Machine Learning Methods among the Anterior Choroidal Artery Infarction","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eAnterior choroidal artery (AChA) infarction represents a distinct clinical entity of cerebral infarction, often resulting in a wide spectrum of neurological deficits. Despite active management, including antiplatelet therapy and even thrombolysis, patients frequently experience symptom fluctuation or neurological progression, drawing increasing attention from clinicians. However, the underlying causes, pathogenesis, and mechanisms of disease progression in AChA infarction remain unclear.\u003c/p\u003e\u003cp\u003eSeveral studies have attempted to predict progression in stroke patients using clinical observations, radiological findings, electrophysiological results, functional neuroimaging, diffusion tensor imaging (DTI), and even transcranial magnetic stimulation (TMS) \u003csup\u003e[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]\u003c/sup\u003e. Nevertheless, the application of machine learning methods for predicting motor outcomes in AChA infarction has not yet been reported.\u003c/p\u003e\u003cp\u003eIn this study, we retrospectively analyzed patients with newly diagnosed cerebral infarction confined to the AChA territory, admitted to the Second Affiliated Hospital of Xi'an Jiaotong University, and compared those with and without clinical progression. Six machine learning models\u0026mdash;logistic regression (LR), random forest (RF), decision tree (DT), eXtreme Gradient Boosting (XGB), support vector machine (SVM), and K-Nearest Neighbour (KNN)\u0026mdash;were employed to predict clinical progression. Their predictive performance was evaluated using the area under the receiver operating characteristic curve (AUC ROC), accuracy (ACC), and F1 score (F1). This study aims to provide a theoretical foundation for improving clinical management and reducing adverse outcomes in AChA infarction.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1. Data Source\u003c/h2\u003e\u003cp\u003eA total of 4,150 patients with acute ischemic stroke admitted to our hospital between January 2020 and January 2025 were initially screened. Among them, 369 patients were confirmed to have AChA infarction based on CT, MRI, and DWI findings. This study contributes a substantial cohort of patients with AChA-restricted infarcts diagnosed via DWI\u0026ndash;MRI. Patients with liver or kidney dysfunction, cardiopulmonary failure, or new infarcts outside the AChA territory were excluded. The National Institutes of Health Stroke Scale (NIHSS) and modified Rankin Scale (mRS) were assessed at admission. The study was approved by the hospital\u0026rsquo;s Ethics Committee, and informed consent was obtained from all patients or their families.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2. Predictors and Data Preprocessing\u003c/h2\u003e\u003cp\u003eA total of 27 variables were selected initially as predictors: demographic characteristics (sex, age, and comorbidities including diabetes, hypertension, and heart disease), lifestyle factors (smoking and drinking), and clinical variables (clinical manifestations, intravenous thrombolysis therapy (IVT), blood glucose (GLU), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), ferritin, uric acid (UA), homocysteine (HCY), and glycosylated hemoglobin (GHb). AChA infarction primarily involves the posterior limb of the internal capsule (PLIC) and/or the posterior portion of the corona radiata (CR) \u003csup\u003e[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]\u003c/sup\u003e. Vascular assessments were performed using DSA, CTA, or MRA to evaluate stenosis or occlusion in the internal carotid artery (ICA), middle cerebral artery (MCA), and posterior cerebral artery (PCA) on the affected side. All imaging results were independently reviewed by two neuroradiologists and two neurologists. Final assessments of lesion location and vascular stenosis degree were determined by consensus. All predictors are detailed in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e (Supplementary Materials).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3. Outcome Definition\u003c/h2\u003e\u003cp\u003ePatients were classified into a progressive group (NIHSS score increase\u0026thinsp;\u0026gt;\u0026thinsp;2 points within one week) and a non-progressive group (NIHSS score increase\u0026thinsp;\u0026le;\u0026thinsp;2 points within one week), in accordance with established criteria \u003csup\u003e[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.4. Models Comparison\u003c/h2\u003e\u003cp\u003eSix machine learning algorithms\u0026mdash;logistic regression (LR), random forest (RF), decision tree (DT), eXtreme Gradient Boosting (XGB), support vector machine (SVM), and K-Nearest Neighbour (KNN)\u0026mdash;were applied to predict the binary outcome of clinical progression. These models were selected due to their widespread use and recognized performance in clinical data classification \u003csup\u003e[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eAmong the initially selected 27 variables, ferritin, uric acid and glycation were deleted since their missing values exceed 10%. Meanwhile, we combined the analysis of motor and sensory disorder, dysarthria and visual impairment. Therefore, ultimately 22 variables were included in the machine learning analysis.\u003c/p\u003e\u003cp\u003eModel performance was evaluated using accuracy, F1 score, specificity, precision, recall, and the area under the receiver operating characteristic curve (AUCROC). Among these, AUCROC was considered the primary performance indicator as it provides a comprehensive measure of model discriminative ability. Accuracy and F1 score served as secondary metrics. The best-performing model was further analyzed using a confusion matrix.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.5. Statistical Analyses\u003c/h2\u003e\u003cp\u003eStatistical analyses were conducted using SPSS version 22. Categorical and continuous variables were compared using the chi-square test and Student\u0026rsquo;s t-test (or rank-sum test for non-parametric data), respectively. A two-tailed p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant. Machine learning models were implemented in Python version 3.12.3 using the Scikit-Learn library version 1.2.2.\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e3.1. Patient Demographics, Risk Factors and Clinical Progression\u003c/h2\u003e\u003cp\u003eA total of 369 patients (8.89% of the screened cohort) were included with isolated AChA infarcts. The majority were male (75.3%). Hemiparesis (87.8%) was the most frequent clinical manifestation, followed by aphasia (55.3%), hemianesthesia (27.1%), and other syndromes. The most commonly involved infarct sites were the posterior limb of the internal capsule and the paraventricular corona radiata (58.5%). Ipsilateral carotid plaque or stenosis was observed in 108 patients (29.3%), and 24 patients (6.5%) received intravenous rt‑PA thrombolysis according to standard protocols.\u003c/p\u003e\u003cp\u003eClinical progression occurred in approximately one‑third of the patients (33.3%). No significant differences were observed in age, infarct location, or rt‑PA administration between the progression and non‑progression groups. Although the overall proportion of males was higher, progression was less frequent in men compared to women (66.7% vs. 79.7%, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.006). Similarly, nicotine abuse was associated with a lower risk of progression (39.8% vs. 55.3%, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.005). Hemiparesis was more common in the progression group than in the non‑progression group (95.1% vs. 84.1%, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.002).\u003c/p\u003e\u003cp\u003eThe proportion of CA plaque or stenosis was significantly higher in the progression group (52.8% vs. 17.5%, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), as was the proportion of MCA or PCA plaque or stenosis. Moreover, patients in the progression group exhibited more severe neurological deficits at both admission and peak assessment (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Serum ferritin levels were significantly lower in the progression group (163.5\u0026thinsp;\u0026plusmn;\u0026thinsp;89.8 vs. 203.1\u0026thinsp;\u0026plusmn;\u0026thinsp;135.9, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.015). Detailed results are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\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\u003ePatient demographics, risk factors and clinical progression\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"6\"\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\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eVariables\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eTotal\u003c/p\u003e\u003cp\u003e(n\u0026thinsp;=\u0026thinsp;369)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eClinical progression\u003c/p\u003e\u003cp\u003e(n\u0026thinsp;=\u0026thinsp;123)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eNo-clinical progression\u003c/p\u003e\u003cp\u003e(n\u0026thinsp;=\u0026thinsp;246)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u003cem\u003ep\u003c/em\u003e-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 (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;sd)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e60.8\u0026thinsp;\u0026plusmn;\u0026thinsp;11.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e61.6\u0026thinsp;\u0026plusmn;\u0026thinsp;11.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e60.3\u0026thinsp;\u0026plusmn;\u0026thinsp;10.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.305\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eMale, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e278(75.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e82(66.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e196(79.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.006\u003csup\u003e#\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"5\" rowspan=\"6\"\u003e\u003cp\u003eRisk factors, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDiabetes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e110(29.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e35(28.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e75(30.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.687\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHypertension\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e246(66.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e80(65.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e166(67.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.639\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCoronary heart disease\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e30(8.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e11(8.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e19(7.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.686\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAtrial fibrillation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e7(1.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1(0.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e6(2.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.500\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNicotine abuse\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e185(50.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e49(39.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e136(55.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.005\u003csup\u003e#\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAlcohol abuse\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e84(22.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e30(24.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e54(22.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.598\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003eSymptom, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHemiparesis\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e324(87.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e117(95.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e207(84.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.002\u003csup\u003e#\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHemianesthesia\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e100(27.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e39(31.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e61(24.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.159\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAphasia\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e204(55.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e67(54.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e137(55.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.824\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHemianopia\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3(0.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0(0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e3(1.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.539\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e\u003cp\u003eLocation, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePLIC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e78(21.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e22(17.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e56(22.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.279\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e75(20.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e22(17.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e53(21.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.410\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePLIC\u0026thinsp;+\u0026thinsp;CR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e216(58.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e79(64.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e137(55.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.117\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eCA plaque or stenosis\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e108(29.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e65(52.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e43(17.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e๤0.001\u003csup\u003e#\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eMCA or PCA plaque or stenosis\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e122(33.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e57(46.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e65(26.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e๤0.001\u003csup\u003e#\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eNIHSS score, median (IQR)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAt admission\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3(2\u0026ndash;4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3(2-5.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2(2\u0026ndash;3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e๤0.001\u003csup\u003e#\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAt peak\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3(2\u0026ndash;5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e4(3\u0026ndash;6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2(2\u0026ndash;3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e๤0.001\u003csup\u003e#\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003emRS score, median (IQR)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2(1\u0026ndash;3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2(2\u0026ndash;4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2(1\u0026ndash;2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e๤0.001\u003csup\u003e#\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"7\" rowspan=\"8\"\u003e\u003cp\u003eblood test (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;sd)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eGLU, mmol/L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6.1\u0026thinsp;\u0026plusmn;\u0026thinsp;2.4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e6.1\u0026thinsp;\u0026plusmn;\u0026thinsp;2.4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e6.1\u0026thinsp;\u0026plusmn;\u0026thinsp;2.4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.901\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eGHb, %\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6.5\u0026thinsp;\u0026plusmn;\u0026thinsp;1.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e6.5\u0026thinsp;\u0026plusmn;\u0026thinsp;1.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e6.4\u0026thinsp;\u0026plusmn;\u0026thinsp;1.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.672\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTG, mmol/L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.7\u0026thinsp;\u0026plusmn;\u0026thinsp;1.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.6\u0026thinsp;\u0026plusmn;\u0026thinsp;0.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1.8\u0026thinsp;\u0026plusmn;\u0026thinsp;1.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.063\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHDL-C, mmol/L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.1\u0026thinsp;\u0026plusmn;\u0026thinsp;0.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.1\u0026thinsp;\u0026plusmn;\u0026thinsp;0.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1.1\u0026thinsp;\u0026plusmn;\u0026thinsp;0.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.563\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLDL-C, mmol/L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2.9\u0026thinsp;\u0026plusmn;\u0026thinsp;0.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3.0\u0026thinsp;\u0026plusmn;\u0026thinsp;0.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2.9\u0026thinsp;\u0026plusmn;\u0026thinsp;0.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.605\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eUA, \u0026micro;mol/L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e321.8\u0026thinsp;\u0026plusmn;\u0026thinsp;86.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e308.6\u0026thinsp;\u0026plusmn;\u0026thinsp;86.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e328.4\u0026thinsp;\u0026plusmn;\u0026thinsp;86.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.051\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHCY, \u0026micro;mol/L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e23.0\u0026thinsp;\u0026plusmn;\u0026thinsp;16.4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e23.9\u0026thinsp;\u0026plusmn;\u0026thinsp;18.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e21.8\u0026thinsp;\u0026plusmn;\u0026thinsp;14.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.663\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eFerritin, \u0026micro;g/L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e189.8\u0026thinsp;\u0026plusmn;\u0026thinsp;122.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e163.5\u0026thinsp;\u0026plusmn;\u0026thinsp;89.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e203.1\u0026thinsp;\u0026plusmn;\u0026thinsp;135.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.015\u003csup\u003e#\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eIV thrombolysis, n (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e24(6.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e12(9.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e12(4.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.073\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eIV, intravenous; GLU, blood glucose; TG, triglyceride; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; UA, uric acid; HCY, homocysteine; GHb, glycosylated hemoglobin; PLIC, the posterior limb of the internal capsule; CR, corona radiata; MCA, middle cerebral artery; PCA, posterior cerebral artery.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e3.2. Model comparison for binary classification problem\u003c/h2\u003e\u003cp\u003eA comparison of the receiver operating characteristic curve for six models is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The differences between these curves were slight, but we could still clearly recognize each model. KNN showed the poorest predictive performance, with the lowest AUCROC of 0.713, an accuracy of 66.6%, and a F1 score of 46.3%. In comparison the other five models improved prediction: LR (AUCROC: 0.799, accuracy: 71.2% and F1 score: 46.7%); DT (0.755, 74.8%, and 65.8% respectively); XGB (0.804, 63.9%, and 0 respectively); SVM (0.759, 63.9%, and 0 respectively); and RF, which had the highest AUCROC of 0.851, an accuracy of 70.3%, and a F1 score of 52.2%. Detailed predictions of RF were presented in the form of a confusion matrix (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e for more details). In the present context, RF was able to predict 87/123 progressive AChA infarction (sensitivity 70.3%) and 208/246 non- progressive AChA infarction (specificity 84.5%). The results of the tenfold cross-validation showed that RF had a better discriminative ability for progressive AChA infarction than the other five models (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e\u003cp\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 the predictive performance for six models\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\u003eModel\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAccuracy(%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSpecificity(%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003ePrecision(%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eRecall(%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eF1 score(%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eAUCROC\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e71.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e91.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e70.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e35.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e46.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.799\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRF\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e70.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e84.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e62.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e45.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e52.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.851\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDT\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e74.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.788732\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e64.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e67.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e65.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.755\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eXGB\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e63.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e100.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.804\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSVM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e63.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e100.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.759\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eKNN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e66.6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e81.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e55.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e40.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e46.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.713\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eLR, logistic regression; RF, random forest; DT, decision tree; XGB, eXtreme Gradient Boosting; SVM, support vector machine; KNN, K-Nearest Neighbour.\u003c/p\u003e\u003cp\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 for six models (tenfold cross-validation)\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=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eModel\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAUCROC\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eModel\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eAUCROC\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.693\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eXGB\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.639\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRF\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.783\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSVM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.639\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDT\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.684\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eKNN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.720\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eLR, logistic regression; RF, random forest; DT, decision tree.\u003c/p\u003e\u003cp\u003eMoreover, according to the information gain values of RF model, we further ranked those 22 features, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. The NIHSS score at peak contributed the most to the progressive AChA infarction outcome, followed by NIHSS score at admission, age, high density lipoprotein, homocysteine and so on.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eNeurological deterioration is a well-recognized phenomenon in AChA infarction and is independently associated with poor outcomes \u003csup\u003e[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]\u003c/sup\u003e. The perforating arteries involved are terminal arterioles with limited collateral circulation, rendering them particularly vulnerable to hemodynamic fluctuations. Moreover, the high density of nerve fibers in the paraventricular and pontine regions means that even a small expansion of the infarct can lead to progressive neurological deficits \u003csup\u003e[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]\u003c/sup\u003e. Previous studies have reported clinical progression rates in AChA infarcts ranging from 7% to 43% \u003csup\u003e[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/sup\u003e, which is notably higher than in hemispheric or deep infarcts \u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/sup\u003e. In our cohort, clinical progression occurred in 33.3% of patients, predominantly among those with higher baseline NIHSS scores. No AChA infarction-related fatalities were observed, reaffirming its generally non-lethal nature.\u003c/p\u003e\u003cp\u003eWe also identified a significantly higher prevalence of ipsilateral severe carotid stenosis or occlusion in patients with clinical progression, suggesting that carotid atherosclerosis may contribute substantially to disease progression. While small vessel disease remains a key etiological factor, the presence of large artery disease combined with multiple risk factors appears to be associated with clinical worsening in AChA infarcts \u003csup\u003e[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/sup\u003e. Intravenous thrombolysis did not significantly prevent clinical progression in our study, possibly due to the small caliber of the affected artery and thrombus size \u003csup\u003e[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]\u003c/sup\u003e. Nevertheless, thrombolysis has been linked to favorable functional outcomes and a low risk of hemorrhagic transformation \u003csup\u003e[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]\u003c/sup\u003e. In contrast, Wu et al. reported that thrombolysis reduced the risk of stroke evolution in larger AChA infarcts \u003csup\u003e[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]\u003c/sup\u003e, highlighting a potential size-dependent treatment effect.\u003c/p\u003e\u003cp\u003ePrevious studies have explored machine learning for predicting stroke outcomes \u003csup\u003e[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]\u003c/sup\u003e, and various biomarkers\u0026mdash;such as neutrophil-related inflammatory indices, ADC values, and infarct dimensions\u0026mdash;have been investigated for forecasting AChA infarction progression \u003csup\u003e[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]\u003c/sup\u003e. In this study, we demonstrated that machine learning models can effectively predict clinical progression in acute AChA infarction. Using 22 input variables and alongside extensive hyperparameter tuning, we compared six machine learning algorithms at their optimized performance. Among them, Random Forest (RF) achieved the highest predictive performance, with an AUCROC of 0.851, accuracy of 70.3%, and an F1 score of 52.2% on the hold-out test set. This result aligns with prior studies applying machine learning to clinical event prediction \u003csup\u003e[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/sup\u003e. As an ensemble method, RF aggregates multiple decision trees to enhance predictive accuracy and robustness \u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/sup\u003e. The tenfold cross-validation further confirmed its superior discriminative capability. Based on 5 clinical data, Fanhai Bu et al. developed prediction models for acute ischemic stroke following intravenous thrombolysis using eight machine learning algorithms, with their optimal logistic regression model achieving an AUC of 0.792 \u003csup\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]\u003c/sup\u003e. In contrast, our RF model demonstrated superior performance, attaining an AUC of 0.851\u0026mdash;a 6% improvement in predictive accuracy that holds significant clinical utility. This enhancement may be attributed to our comprehensive approach: 1) addressing missing data through multiple imputation methods; 2) incorporating more extensive predictive variables including radiomic features from neuroimaging modalities; and 3) implementing meticulous hyperparameter optimization to refine model architecture. Notably, NIHSS scores were among the most important predictors identified by the RF model, underscoring that patients with severe initial neurological impairment are at higher risk of progression. This finding should be considered during clinical communication with patients and families.\u003c/p\u003e\u003cp\u003eSeveral limitations of this study should be acknowledged. Firstly, the sample size for analyzing AChA infarction progression using machine learning was relatively small, and future studies should seek to expand the cohort. Secondly, stratifying AChA infarcts by size (e.g., small vs. large) in subsequent analyses could yield more nuanced insights. Thirdly, although we used multiple imputation for missing data, imputation model specifications may introduce uncertainty affecting model validity. Additionally, our prediction model was validated only internally, external validation in larger, multicenter, prospective cohorts is needed to confirm generalizability.\u003c/p\u003e"},{"header":"5. Conclusions","content":"\u003cp\u003eMotor deficits represent the most frequent and characteristic clinical manifestation in AChA infarcts. This study identifies Random Forest as the most effective machine learning model for predicting clinical progression in AChA infarction, demonstrating its utility as a practical and accessible tool for risk stratification. Particular attention should be given to NIHSS scores at admission and peak, along with patient age, as these factors were identified as key predictors of neurological deterioration.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eAll machine learning methods were carried out in accordance with the\u003c/p\u003e\u003cp\u003e \u0026ldquo;Guidelines for Developing and Reporting Machine Learning Predictive Models in Biomedical Research\u0026rdquo;. And the protocol conformed to the Declaration of Helsinki.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003cp\u003e The ethics committee at The Second Affiliated Hospital of Xi\u0026rsquo;an Jiaotong University had approved this study (No.2024-047). Written informed consent was obtained from individual orguardian participants.\u003c/p\u003e\u003ch2\u003eConsent for publication\u003c/h2\u003e\u003cp\u003eNot applicable.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\u003ch2\u003eAuthor details\u003c/h2\u003e\u003cp\u003e1 Department of Neurology, The Second Affiliated Hospital of Xi\u0026rsquo;an Jiaotong University, No. 157 West Five Road, Xi\u0026rsquo;an 710004, Shaanxi, China.\u003c/p\u003e\u003cp\u003e2 Department of Neurology, Norinco General Hospital, No. 12 Zhangba East Road, Xi'an, 710077, Shaanxi, China.\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e\u003cp\u003eThis study is supported by the National Natural Science Foundation of China (No. 81070999); the Foundation of Shaanxi social development and technology research project (No. 2016SF-020); the Foundation of Xi\u0026rsquo;an Science and technology plan project (No. 2019114613YX001SF039(2)); the new medical technology of the Second Affiliated Hospital of Xi\u0026rsquo;an Jiaotong University (No. 2019-32, 2018-16, 2010-22); the Fundamental Research Funds for the Central Universities (Xi\u0026rsquo;an Jiaotong University, No. xjj2014153, 2009-95); the Foundation of Second Affiliated Hospital of Xi\u0026rsquo;an Jiaotong University (No. RC(GG)201109).\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eMYC and SQZ conceived and designed the study. MYC and SYM analyzed the data, wrote, and revised the manuscript. JXF developed the models and guided the Python code. MYZ, QLY, and XDZ collected the data. SD, HYQ, and YXC conducted the data. HY obtained ethical approval. SQZ monitored the entire planning, revision, and provided funding. All authors read and approved the final manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgements\u003c/h2\u003e\u003cp\u003eNot applicable.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe datasets generated and/or analysed during the current study are not publicly available due privacy but are available from the corresponding author on reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eKwon YH, Son SM, Lee J, et al. Combined study of transcranial magnetic stimulation and diffusion tensor tractography for prediction of motor outcome in patients with corona radiata infarct. J Rehabil Med. 2011;43(5):430\u0026ndash;4.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNelles M, Gieseke J, Flacke S, et al. Diffusion tensor pyramidal tractography in patients with anterior choroidal artery infarcts. AJNR Am J Neuroradiol. 2008;29(3):488\u0026ndash;93.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSohn H, Kang DW, Kwon SU, et al. Anterior choroidal artery territory infarction: lesions confined to versus beyond the internal capsule. Cerebrovasc Dis. 2013;35(3):228\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePoh KW, Er CK, Hoh WH, et al. Neurological deterioration and its risk score in total anterior circulation infarct. Clin Neurol Neurosurg. 2020;191:105684.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKop R, Hoogendoorn M, Teije AT, et al. Predictive modeling of colorectal cancer using a dedicated pre-processing pipeline on routine electronic medical records. Comput Biol Med. 2016;76:30\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChausson N, Joux J, Saint-Vil M, et al. Infarction in the anterior choroidal artery territory: clinical progression and prognosis factors. J Stroke Cerebrovasc Dis. 2014;23(8):2012\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYamamoto Y, Nagakane Y, Tomii Y, et al. The relationship between progressive motor deficits and lesion location in patients with single infarction in the lenticulostriate artery territory. J Neurol. 2017;264(7):1381\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCheng Z, Duan H, Meng F, et al. Acute Anterior Choroidal Artery Territory Infarction: A Retrospective Study. Clin Neurol Neurosurg. 2020;195:105826.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOis A, Cuadrado-Godia E, Solano A, et al. Acute ischemic stroke in anterior choroidal artery territory. J Neurol Sci. 2009;281(1\u0026ndash;2):80\u0026ndash;4.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTrouillas P, Derex L, Nighoghossian N, et al. rtPA intravenous thrombolysis in anterior choroidal artery territory stroke. Neurology. 2000;54(3):666\u0026ndash;73.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWu MC, Tsai LK, Wu CC, et al. Thrombolytic therapy is an only determinant factor for stroke evolution in large anterior choroidal artery infarcts. J Stroke Cerebrovasc Dis. 2014;23(5):1089\u0026ndash;93.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWu Y, Fang Y. Stroke Prediction with Machine Learning Methods among Older Chinese. Int J Environ Res Public Health, 2020, 17 (6).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHeo J, Yoon JG, Park H, et al. Machine Learning-Based Model for Prediction of Outcomes in Acute Stroke. Stroke. 2019;50(5):1263\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhao N, Li J, Zhang QX, et al. Elevated neutrophil-related immune-inflammatory biomarkers in acute anterior choroidal artery territory infarction with early progression. Clin Neurol Neurosurg. 2023;229:107720.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYu W, Yang J, Liu L, et al. The value of diffusion weighted imaging in predicting the clinical progression of perforator artery cerebral infarction. Neuroimage Clin. 2022;35:103117.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLu Y, Ning Y, Li Y, et al. Risk factor mining and prediction of urine protein progression in chronic kidney disease: a machine learning- based study. BMC Med Inf Decis Mak. 2023;23(1):173.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShi Y, Zhang G, Ma C et al. Machine learning algorithms to predict intraoperative hemorrhage in surgical patients: a modeling study of real-world data in Shanghai, China.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBMC Med Inform Decis Mak. 2023, 23 (1): 156. Breiman L. Random Forests. Machine Learning, 2001, 45 (1): 5\u0026ndash;32.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBu F, Cai R, Zhang W, et al. Explainable machine learning model for predicting the outcome of acute ischemic stroke after intravenous thrombolysis. Front Neurol. 2025;16:1668816.\u003c/span\u003e\u003c/li\u003e\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-medical-informatics-and-decision-making","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"midm","sideBox":"Learn more about [BMC Medical Informatics and Decision Making](http://bmcmedinformdecismak.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/midm/default.aspx","title":"BMC Medical Informatics and Decision Making","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Anterior choroidal artery, Machine learning, Clinical progression, Prediction","lastPublishedDoi":"10.21203/rs.3.rs-7910284/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7910284/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003ePatients with anterior choroidal artery (AChA) infarction often experience fluctuating symptoms or neurological progression despite aggressive treatment. Machine learning, known for its high accuracy, is increasingly applied in medicine. This study explored the use of machine learning techniques to predict neurological progression in AChA infarction patients.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eA total of 369 AChA infarction patients were retrospectively enrolled. Demographic and clinical characteristics were collected and analyzed. Clinical features were compared between patients with and without neurological progression. Six machine learning models\u0026mdash;logistic regression (LR), random forest (RF), decision tree (DT), eXtreme Gradient Boosting (XGB), support vector machine (SVM), and K-Nearest Neighbour (KNN)\u0026mdash;were employed to predict progression. Their performance was evaluated using the area under the receiver operating characteristic curve (AUC ROC), accuracy (ACC), and F1 score (F1).\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eNeurological progression occurred in one-third of AChA infarction patients. Hemiparesis was the most common manifestation (87.8%), occurring more frequently in the progression group than in the no-progression group (95.1% vs. 84.1%, P\u0026thinsp;=\u0026thinsp;0.002). Although more men were affected (75.3%), they had a lower likelihood of progression (66.7% vs. 79.7%, P\u0026thinsp;=\u0026thinsp;0.006). No significant differences were observed in stroke location or intravenous rt-PA therapy between the groups. Among the models, RF demonstrated the best predictive performance, achieving the highest AUC ROC (0.851), accuracy (70.3%), and F1 score (52.2%). The top three predictors identified by RF were NIHSS score at peak, NIHSS score at admission, and age.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e\u003cp\u003eMotor deficits are the most frequent and characteristic symptoms in AChA infarcts. RF emerged as the most effective model for predicting clinical progression in these patients, offering a simple and useful tool for risk assessment. Special attention should be given to NIHSS scores at peak and admission, as well as patient age.\u003c/p\u003e","manuscriptTitle":"Progression Prediction with Machine Learning Methods among the Anterior Choroidal Artery Infarction","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-30 23:27:07","doi":"10.21203/rs.3.rs-7910284/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2025-11-24T16:11:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"309892930125967229067862686821743741726","date":"2025-11-19T18:29:26+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-11-19T16:33:46+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-10-24T11:40:47+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-10-23T09:13:19+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-10-23T09:12:37+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medical Informatics and Decision Making","date":"2025-10-21T03:58:28+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-informatics-and-decision-making","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"midm","sideBox":"Learn more about [BMC Medical Informatics and Decision Making](http://bmcmedinformdecismak.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/midm/default.aspx","title":"BMC Medical Informatics and Decision Making","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e67aaba4-aaf3-4be9-8476-874ae0e20795","owner":[],"postedDate":"November 30th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2025-11-30T23:27:07+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-30 23:27:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7910284","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7910284","identity":"rs-7910284","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}