Development and validation of a predictive model for white matter lesions in young- and middle-aged people.

Background: White matter lesion (WML) is an age-related disorder associated with stroke and cognitive impairment. This study aimed to investigate the risk factors and build a predictive model of WML in young- and middle-aged people. Methods: We performed a second analysis of the data from the Dryad Digital Repository. We selected those people who are <60 years old and randomly divided them into the training group and the validation group. We investigated the risk factors of WML in the training group with logistic regression analysis and built a prediction nomogram based on multivariate logistic regression analysis; finally, the performance of the prediction nomogram was evaluated for discrimination, accuracy, and clinical utility. Results: There were 308 people in the training group and 723 people in the validation group. Multivariate regression analysis showed that the age (OR = 1.49, 95% CI: 1.31-1.70), diastolic blood pressure (OR = 1.02, 95% CI: 1.00-1.03), carotid plaque score (OR = 1.31, 95% CI: 1.14-1.50), female gender (OR = 2.27, 95% CI: 1.56-3.30), and metabolic syndrome (OR = 2.12, 95% CI: 1.22-3.70) were significantly associated with white matter lesions. The area under the curve value (AUC) of the receiver operating curve (ROC) was 0.734 for the training group and 0.642 for the validation group. The calibration curve and clinical impact curve showed that the prediction nomogram has good accuracy and clinical application value. Conclusion: Age, diastolic blood pressure, carotid plaque score, female gender, and metabolic syndrome were risk factors in young- and middle-aged people <60 years old with WML, and the nomogram based on these risk factors showed good discrimination, accuracy, and clinical utility.