Risk Stratification in ALS Using XGBoost and LSTM with Biomarkers and Vital Signs

Amyotrophic Lateral Sclerosis (ALS) is a rapidly progressing neurodegenerative disease with no known cure and limited treatment options, making early intervention and targeted therapeutic strategies essential. This study proposes a hybrid machine learning approach for risk stratification by leveraging structured biomarkers and time-series vital signs. Biomarker data, including C-Reactive Protein, albumin, and creatinine, are processed using XGBoost to generate a biomarker risk score, while time-dependent vital signs-such as systolic and diastolic blood pressure, pulse, and respiratory rate-are analyzed with a Long Short-Term Memory (LSTM) model to compute a vital signs risk score. These risk scores are concatenated and passed through a final predictive model to generate an ALS Risk Progression Score. The findings contribute to advancing ALS research, offering better survival predictions and paving the way for more effective patient management,