Association of composite biomarkers with imaging burden in cerebral small vessel disease

BACKGROUND: Chronic inflammation and dysregulated lipid metabolism may contribute to the pathogenesis of cerebral small vessel disease (CSVD). This study aimed to establish composite serum inflammation/metabolism biomarkers and evaluate their association with total neuroimaging burden in acute ischemic CSVD. METHODS: This study enrolled a cohort of 328 patients with acute ischemic CSVD who met the predefined selection criteria, were hospitalized in the neurology department between March 2023 and October 2024, and underwent standardized assessments. The total CSVD burden was quantified using the modified Rotterdam criteria. Participants were stratified into the low burden (0–1 point; n = 153) and high burden (2–4 points; n = 175) groups. Composite biomarkers, including the neutrophil-to-high-density lipoprotein (HDL) (NHR), monocyte-to-HDL (MHR), lymphocyte-to-HDL (LHR), platelet-to-HDL (PHR) ratios, as well as systemic immune inflammation (SII) and systemic inflammation response (SIRI) indices, were compared. A predictive model was developed using least absolute shrinkage and selection operator and multivariable logistic regression analyses, and its discriminatory performance was validated by the receiver operating characteristic (ROC) curve analysis and bootstrap resampling with 1,000 repetitions. Subgroup analyses (based on age, sex, etc.) were conducted to evaluate the associations between the biomarkers and disease burden. RESULTS: The high burden group demonstrated significantly higher values for age, hypertension prevalence, and levels of several composite biomarkers (NHR, LHR, PHR, SIRI, and SII) than the low burden group. Multivariate logistic regression revealed that NHR, PHR, SIRI, and SII were independent risk factors for CSVD burden. ROC analysis showed superior predictive performance for NHR. The combined biomarker model demonstrated a significantly superior predictive value compared with any single biomarker, with an initial area under the ROC curve (AUC) value of 0.816 and a corrected AUC value of 0.803 after internal validation. In the analysis that excluded individuals under 60 years of age, the model maintained robust predictive performance (AUC = 0.828, corrected AUC = 0.808). Subgroup analyses further confirmed that NHR and SII were significantly associated with CSVD severity across all subgroups. CONCLUSION: The findings indicate an association between composite biomarkers and CSVD burden, supporting the likely implication of chronic inflammation and metabolic dysfunction in disease progression. The combined biomarker panel demonstrated superior performance in identifying acute ischemic CSVD neuroimaging burden, suggesting its potential as a clinical tool for early risk stratification. This approach could facilitate earlier identification of and intervention in high-risk patients, thereby contributing to strategies aimed at reducing the long-term burden of stroke and cognitive impairment.