University of Göttingen Researchers Utilize Machine Learning to Predict Clinical Risks During Stem Cell Mobilization for Multiple Myeloma Patients

University of Göttingen researchers use AI to predict side effects during stem cell therapy, identifying which myeloma patients can safely receive outpatient care.

By: AXL Media

Published: Apr 1, 2026, 8:53 AM EDT

Source: Information for this report was sourced from University of Göttingen

Modernizing the Autologous Stem Cell Transplantation Process

Multiple myeloma remains a challenging cancer of the plasma cells with no current cure, though autologous stem cell transplantation offers a vital pathway for disease stabilization. Historically, the "mobilization phase"—where chemotherapy is used to flush stem cells from the bone marrow into the blood for collection—has required patients to spend two to three weeks in a hospital setting. This prolonged stay is designed to provide immediate intervention for life-threatening side effects, yet many patients experience these complications late in the process or not at all. Researchers from the University of Göttingen and clinical partners have now challenged this universal inpatient requirement by using predictive algorithms to identify low-risk time windows for individual patients.

Data-Driven Risk Stratification in Hematology

The research team analyzed treatment data from 109 patients who underwent stem cell mobilization at the University Medical Center Göttingen (UMG). By applying machine learning methods to this clinical dataset, the investigators were able to develop a "therapy roadmap" that predicts the onset of adverse events with high accuracy. This model allows oncologists to distinguish between patients who strictly require inpatient monitoring and those who can safely remain in their home environment while receiving close outpatient care. Dr. Enver Aydilek noted that the primary objective was to determine if the traditional long-term hospital stay was medically necessary for every individual, or if a more nuanced approach could be adopted.

Enhancing Patient Quality of Life Through Outpatient Care

The shift toward outpatient treatment pathways is not merely a logistical adjustment but a significant improvement in the patient experience. Simulations conducted during the study indicate that recovering in a familiar home environment reduces the psychological and physical stress associated with a three-week hospital stay. Friedrich Schwarz, the study’s first author, emphasized that a data-based roadmap allows for a more "patient-friendly and modern" version of cancer care. By tailoring the intensity of monitoring to the actual risk profile of the patient, the medical system can respect the individual’s quality of life without compromising their clinical safety.