Development and validation of a prognostic nomogram based on the hub genes of patients with hypertensive intracerebral hemorrhage.

Objective: Hypertensive intracerebral hemorrhage (HICH) is frequently associated with high disability, high mortality, and poor prognosis. The present study aimed to identify genes associated with HICH to construct prognostic models to improve accuracy in predicting HICH prognosis.

Methods: Hub genes were identified by screening out differentially expressed genes from data in the Gene Expression Omnibus database and conducting weighted gene co-expression network analysis. 68 patients with HICH were recruited and categorized based on prognosis. The univariate logistic, least absolute shrinkage and selection operator, and multivariate logistic regression models were then established based on clinical data and the identified hub genes. A prognostic model was constructed based on the nomogram score. The model was validated using receiver operating characteristic curve, C-index, calibration plots, and decision curve analysis. It was also compared to a prognostic model constructed based on clinical data alone. The prognostic value of the nomogram score was assessed in different subgroups.

Results: Three hub genes: pro-platelet basic protein (PPBP), PDZ and LIM domain protein 1 (PDLIM1), and metalloproteinase 1 (TIMP1) were identified as significantly correlated to adverse outcomes in HICH. These hub genes, in combination with the clinical data, were used to construct a nomogram score system, which exhibited strong predictive power, excellent consistency between actual outcomes and predictions, and a higher net clinical benefit. HICH patients with high scores presented significantly worse outcome. Importantly, the developed nomogram score system was superior to the use of clinicopathological features in predicting HICH prognosis. The nomogram score system exhibited adequate predictive performance in different subgroups as well.

Conclusion: The nomogram score system based on PPBP, PDLIM1, and TIMP1 genes, along with clinical data, exhibited superior performance in predicting adverse outcome in HICH patients. This system could, therefore, be useful for guiding clinical decisions and providing valuable insight for designing individualized treatments for HICH patients.