Researchers developed and validated a machine learning model using XGBoost to predict hyperglycemia risk in patients with:
This retrospective study
“Currently, there are no models that predict hyperglycemia risk in patients with psoriasis,” the study's researchers wrote. “Therefore, we propose to use clinical data of psoriasis patients and machine learning algorithms to build a predictive model of hyperglycemia risk in psoriasis patients, in order to screen high-risk groups and implement early intervention, and provide guidance for individualized treatment of psoriasis patients.”
Psoriasis is a condition associated with systemic inflammation, which impairs glucose metabolism and can lead to insulin resistance and hyperglycemia.2 Elevated stress-induced hyperglycemia in patients with psoriasis is also associated with poor outcomes, including increased all-cause mortality.
In this study, we collected clinical data from 575 psoriasis patients treated at the dermatology department of Guilin Medical University Hospital, China, and randomly divided the patients into a training set (70%) and an internal test set (30%).1 The external validation set included 135 patients from the 2003–2004 and 2011–2012 NHANES cohorts. Eleven machine learning algorithms were systematically trained and compared, and model performance was evaluated to assess both predictive accuracy and clinical utility.
The XGBoost model was chosen as the final prediction algorithm due to its excellent performance across multiple metrics. The model achieved an area under the curve (AUC) of 0.821 (95% CI, 0.775-0.866) on the training set, 0.820 (95% CI, 0.751-0.888) on the internal test set, and 0.788 (95% CI, 0.695-0.881) on the external NHANES test set, demonstrating consistent accuracy across. Dataset.
Additionally, the calibration curve showed strong agreement between predicted and observed hyperglycemic risk, while clinical decision curve analysis confirmed its potential usefulness in guiding treatment decisions. Additionally, a web-based calculator was developed to make the model easily accessible to clinicians managing psoriasis patients at risk for hyperglycemia.
However, the researchers noted some limitations. First, the data were obtained from a single hospital in China, which may have introduced bias due to population differences. Additionally, the retrospective design lacked key clinical indicators, limiting direct therapeutic use. Prospective multicenter studies are needed to validate the model.
Despite these limitations, researchers believe that the XGBoost model effectively predicts hyperglycemia risk in psoriasis patients and supports individualized management strategies. With further validation in diverse populations, this tool may help clinicians identify high-risk individuals and guide targeted interventions.
“Because hyperglycemia and psoriasis are associated, we used laboratory indicators and general data to develop a predictive model for the incidence of hyperglycemia in patients with psoriasis,” the researchers wrote. “We also investigated the relationship between the two in a preliminary manner, providing guidance for further research. Five of these indicators – age, BUN; [blood urea nitrogen],ALT [alanine aminotransferase]HDL-C [high-density lipoprotein cholesterol]T.G. [triglycerides]—There may be a pretty strong correlation with blood sugar levels. Patients with psoriasis who are hyperglycemic or at risk for hyperglycemia require a customized treatment plan to coordinate the management of psoriatic inflammation and glycemic progression. ”
References
1. Hu M, Chen D, Yu J. Predictive modeling of hyperglycemia risk in psoriasis patients using machine learning: a multicenter retrospective study. Clin Cosmet Investig Dermatol. 2025;18:3667-3680. doi:10.2147/CCID.S552796
2. Steinzor P. Stress hyperglycemia is associated with all-cause mortality in patients with psoriasis. AJMC®. November 7, 2024. Accessed January 5, 2026.
