Development and internal validation of a spatiotemporal gait parameter-based diagnostic model for cerebral small vessel disease.

Introduction: Cerebral small vessel disease (CSVD) is a common cause of functional decline in the elderly, yet its diagnosis often relies on neuroimaging, which may be inaccessible in routine practice. Given that gait impairment is a core feature of CSVD, we aimed to develop and validate a clinically applicable diagnostic model by integrating quantitative spatiotemporal gait parameters with conventional clinical features. Methods: This case-control study included 417 healthy controls from a community-based cohort and 117 hospital-based CSVD patients. Conventional clinical characteristics and quantitative spatiotemporal gait parameters were collected from all participants. A two-stage modeling approach was used, in which least absolute shrinkage and selection operator (LASSO) regression was first applied for predictor screening, followed by multivariable logistic regression for constructing the final diagnostic model. Model performance was assessed by discrimination (area under the curve [AUC]), calibration, and clinical utility (decision curve analysis [DCA]). Results: Six key variables were included in the final diagnostic model: sex, hypertension, body mass index (BMI), stride length, step frequency, and step width. The model exhibited excellent discrimination, achieving an AUC of 0.914 (95% CI: 0.886-0.943), along with strong calibration. DCA further confirmed its clinical utility, showing a greater net benefit across a wide range of threshold probabilities compared to default screening strategies. Conclusion: The diagnostic model developed in this study effectively identifies individuals at high risk of CSVD by leveraging quantitative spatiotemporal gait parameters alongside conventional clinical features.