Biomathematical Modeling in Clinical Intelligent Diagnosis and Treatment

Abstract

Hemorrhagic stroke, a severe cerebrovascular disorder caused by the rupture of brain blood vessels, is characterized by high acute mortality rates and enduring neurological impairments. To provide more targeted clinical recommendations, this study, based on the clinical data of patients with hemorrhagic stroke, extensively explores the relationships between edema changes, treatment conditions, and modified Rankin Scale (MRS) scores. The HemExPred, EdemaVolReg, and PrognosisPred models are meticulously designed to address these three critical aspects. In this study, we skilfully employed machine learning techniques to identify characteristics associated with hematoma expansion events, with the effectiveness of the ElasticNet technique being particularly notable. Additionally, polynomial regression demonstrated exceptional fitting capabilities in deciphering the complexities of edema volume changes. Utilizing the ExtraTreesRegressor model, with an R^2 of 94.6% and 90.6%, we successfully predicted the future trajectories of patients, affirming that hematoma volume, edema volume, and age are key determinants of prognosis. Through in-depth analysis, this research provides valuable insights for clinical decision-making in hemorrhagic stroke, aiding physicians in devising more precise and effective treatment strategies.