In January 2023, a study led by Professor Xiao Jun Huang from Peking University People's Hospital was published in the international academic journal ——Blood Science . The title of the study is "Machine learning algorithm as a prognostic tool for Epstein-Barr virus reactivation after haploidentical hematopoietic stem cell transplantation". This study elucidates the machine learning algorithm-based model for predicting Epstein-Barr Virus (EBV) reactivation after haploidentical Hematopoietic Stem Cell Transplantation (HSCT) demonstrates not only promising results but also a translational potential for personalized post-transplant care.
Haploidentical hematopoietic stem cell transplantation (HSCT) has emerged as a pivotal therapeutic intervention for acute leukemia patients. However, post-transplant complications, particularly Epstein-Barr virus (EBV) reactivation, pose significant risks to patient outcomes. This study presents a groundbreaking machine learning algorithm designed to serve as a prognostic tool for identifying patients at high risk of EBV reactivation following haploidentical HSCT.
The study enrolled 470 consecutive acute leukemia patients undergoing haploidentical HSCT with anti-thymocyte globulin (ATG) for graft-versus-host disease (GVHD) prophylaxis. The cohort was split into a training cohort (n = 282) and a validation cohort (n = 188). A logistic regression model with L2 regularization was employed, incorporating 13 variables encompassing patient demographics, disease characteristics, and transplant details.
The machine learning algorithm-based model successfully predicted the risk of EBV reactivation, differentiating between low- and high-risk groups. Patients identified as high-risk exhibited a significantly higher incidence of EBV reactivation compared to their low-risk counterparts. The secondary outcomes, including non-relapse mortality, overall survival, and leukemia-free survival, also displayed noteworthy differences between the two risk groups.
In conclusion, the machine learning algorithm-based model for predicting EBV reactivation after haploidentical HSCT demonstrates not only promising results but also a transformative potential for personalized post-transplant care. The robust differentiation between low- and high-risk groups, coupled with the significant impact on secondary outcomes, emphasizes the clinical relevance of this prognostic tool.
Reference：Fan S, et al. Machine learning algorithm as a prognostic tool for Epstein-Barr virus reactivation after haploidentical hematopoietic stem cell transplantation. Blood Science. 5(1):51-59, January 2023.
