Predicting the efficacy of Recombinant Human Thrombopoietin in Treating Cancer Therapy-Related Thrombocytopenia:based on stacking ensemble methods

The project aimed to develop a data-driven approach for predicting platelet recovery in cancer treatment-induced thrombocytopenia (CTIT) patients receiving recombinant human thrombopoietin (Rh-TPO). By integrating key clinical indicators into a predictive modeling framework, the study sought to enhance understanding of individual treatment responses and facilitate timely clinical decision-making. A retrospective two-stage modeling analysis was conducted on 400 hospitalized CTIT patients who received Rh-TPO therapy in 2023, with data randomly split into training and testing sets. Following rigorous feature selection, multiple machine learning regression models were trained, and a stacking ensemble model was developed to leverage their combined strengths. Model performance was evaluated on the test set, and an explainable AI analysis was subsequently performed to identify the most influential clinical features driving platelet recovery. Eight variables--baseline platelet count before Rh-TPO initiation, duration of Rh-TPO therapy, pre-anticancer treatment platelet count, timing of follow-up blood tests, hemoglobin level, age, height, and ethnicity--were consistently identified as key predictors. The stacking ensemble model achieved the highest predictive accuracy, demonstrating superior agreement between predicted and observed platelet recovery trajectories compared to individual models. Improved prediction stability was associated with more comprehensive pretreatment assessments and well-defined treatment timelines. These results underscore the value of integrating heterogeneous clinical data through ensemble learning and suggest that such predictive models could support earlier identification of suboptimal platelet recovery, thereby improving CTIT management.