A novel predictive model for the recurrence of pediatric alopecia areata by bioinformatics analysis and a single-center prospective study

Alopecia areata (AA) is a disease featured by recurrent, non-scarring hair loss with a variety of clinical manifestations. The outcome of AA patients varies greatly. When they progress to the subtypes of alopecia totalis (AT) or alopecia universalis (AU), the outcome is unfavorable. Therefore, identifying clinically available biomarkers that predict the risk of AA recurrence could improve the prognosis for AA patients.

In the present study, we construct a novel model based on serum levels of BMP2, CD8A, PRF1, and XCL1, which served as a potential non-invasive prognostic biomarker for forecasting the recurrence of AA patients with high accuracy.

In this study, we conducted weighted gene co-expression network analysis (WGCNA) and functional annotation analysis to identify key genes that correlated to the severity of AA. Then, 80 AA children were enrolled at the Department of Dermatology, Wuhan Children's Hospital between January 2020 to December 2020. Clinical information and serum samples were collected before and after treatment. And the serum level of proteins coded by key genes were quantitatively detected by ELISA. Moreover, 40 serum samples of healthy children from the Department of Health Care, Wuhan Children's Hospital were used for healthy control.

We identified four key genes that significantly increased (CD8A, PRF1, and XCL1) or decreased (BMP2) in AA tissues, especially in the subtypes of AT and AU. Then, the serum levels of these markers in different groups of AA patients were detected to validate the results of bioinformatics analysis. Similarly, the serum levels of these markers were found remarkedly correlated with the Severity of Alopecia Tool (SALT) score. Finally, a prediction model that combined multiple markers was established by conducting a logistic regression analysis.