A clinical and head CT-based scoring system for predicting high cerebral microbleeds burden in hypertensive patients

This study aimed to establish and validate a scoring system utilizing clinical and head CT features to predict high cerebral microbleeds (CMBs) burden (> 10 CMBs) in hypertensive patients.

A retrospective review was conducted on 458 hypertensive patients from two centers, including 244 patients in the training cohort, 101 patients in the internal validation cohort, and 113 patients in the external validation cohort. Clinical data, as well as head CT and MRI findings, were collected. Head MRI results were used to classify patients into two groups: those with > 10 CMBs and those with ≤ 10 CMBs. Statistically significant clinical and CT features distinguishing the groups were identified through univariate and multivariate logistic regression analyses. These features were subsequently weighted and scored to establish a scoring system. The model's effectiveness was assessed through receiver operating characteristic (ROC) curves and decision curve analysis (DCA). To enhance clinical applicability, the scoring system was stratified into three score ranges.

Four features were ultimately incorporated into the scoring system: hypertension duration, lacunar infarcts count grade, lacunar infarcts location grade, and leukoaraiosis grade. The area under the curve (AUC) values for the training, internal validation, and external validation cohorts were 0.898 (95% confidence interval [CI]: 0.857-0.938), 0.840 (95% CI: 0.760-0.920), and 0.810 (95% CI: 0.725-0.877), respectively. At a cutoff value of 4.5 points, the sensitivity and specificity were 85.7% and 76.9%, respectively. The DCA further demonstrated the clinical utility of the model. The scoring system was categorized into three ranges: 0-1, 2-4, and 5-12. As the score increased, the incidence of high CMBs burden (> 10 CMBs) in the training, internal validation, and external validation cohorts progressively increased.

The scoring system, which incorporates clinical and head CT features, has proven to be a valuable tool for predicting high CMBs burden in hypertensive patients and provides significant support for clinical decision-making.