Dr Hernández-Boluda on Machine Learning for Identifying Transplant Risk in Myelofibrosis

"The technology [of this machine learning model] is completely different and more accurate than the available models. It’s also far more comprehensive because it is based on data from [5183] patients transplanted within EBMT centers and includes [data like] graft source and methods of graft-vs-host disease."

Juan Carlos Hernández-Boluda, MD, PhD, an associated physician at the Hematology and Medical Oncology Service of the Hospital Clínico of Valencia, discussed the development of a unique EBMT machine learning model to predict survival outcomes following allogeneic hematopoietic cell transplant (allo-HCT) in patients with myelofibrosis and expanded on how this model potentially improves upon existing transplant risk assessment tools.

Hernández-Boluda and colleagues at the EBMT developed a random survival forest (RSF) machine learning model using data from 5183 adult patients with primary or secondary myelofibrosis who first underwent allo-HCT between 2005 and 2020 across EBMT centers. The RSF model demonstrated superior predictive performance compared with other machine learning approaches, yielding higher concordance indices across subgroups of patients with both primary and secondary myelofibrosis. The model also stratified 25% of patients as high risk for non-relapse mortality. The open-access tool was officially launched on March 27, 2025, and is now freely available online.

What distinguishes this EBMT model from prior transplant risk assessment tools is its integration of contemporary, large-scale data and use of RSF algorithms, which can handle complex, nonlinear interactions without assumptions about variable distributions, Hernández-Boluda explained. The model’s accuracy reflects the inclusion of detailed clinical parameters—such as graft source and graft-vs-host disease prophylaxis strategies—collected from a total of 288 EBMT centers, which currently represent standards of care in allo-HCT for patients with myelofibrosis, he detailed. Unlike older models that relied on smaller cohorts or less granular data, this RSF tool provides more individualized risk prediction and is aligned with modern transplant practices, Hernández-Boluda concluded.