2026 ASH®
Annual Meeting & Exposition December 12-15, 2026 ▪ New Orleans, Louisiana

Paper Details

Abstracts, authors, affiliations, and disclosures.

SESSION

201. Granulocytes, Monocytes, and Macrophages: Poster III

Program: Poster Abstract Session
Monday, December 14, 2026
06:00 p.m. - 08:00 p.m.
MCCNO - Hall F
Technology and Procedures, Assays, Bioinformatics, Treatment Considerations, Combination Therapy, Supportive Care, Biological Therapies, Antibody Therapy, (CAR)-T Cell Therapies, Immunotherapy, Monoclonal Antibody Therapy, Non-Biological Therapies, Chemotherapy, Women’s Health, Adverse Events, Artificial Intelligence (AI), Machine Learning, Biological Processes, Pathogenesis, Genomics, Hematopoiesis, Immune Mechanism, Clinical Practice (Health Services and Quality), Clinical Guidelines, Health Economics, Diseases, Bone Marrow Failure Syndromes, Immunology, Autoimmune Disorders, Immunodeficiency, Neutropenia, White Blood Cell Disorders, Infectious Diseases, Viral, Bacterial, Fungal, Lymphoid Malignancies, Lymphoid Leukemias, Lymphomas, Myeloid Malignancies, Acute Myeloid Malignancies, Chronic Myeloid Malignancies, Measurable Residual Disease, ALL, B Cell Lymphoma, T Cell Lymphoma, Research, Fundamental Science, AML, Translational Research, Clinical Research, Clinical Trials, Epidemiology, Health Outcomes Research, Real-World Evidence, Study Population, Human, Adult, Pediatric, Pregnant
PAPER

4735 The HLH-risk-calculator: External validation and real-world applicability of a machine learning-based tool predicting clinical course and mortality of secondary hemophagocytic lymphohistiocytosis

Michael Ruzicka, Josia Fauser, MD2, Manuel Trebo3, Aiman Gamal Abdelrahim4, Maximilian Al-Bazaz, MD5, Ricardo Ahrens6, Klaus-Georg Kreitmeier7, Bruno Loureiro Cadilha8, Sabine Janina Ehrlich9, Hans-Joachim Stemmler10, Stephanie Susanne Stecher11, Andreas Pircher12, Michael Von Bergwelt, Prof. Dr. med.13, Marion Subklewe I, MD14, Karsten Spiekermann, MD15
1 German Cancer Consortium (DKTK), Partner Site Munich, Germany, Ludwig Maximilian University Hospital, Munich, Germany
2 Medical University of Innsbruck, Innsbruck, Austria
3 Medical University of Innsbruck, Innsbruck, Austria
4 University Medical Center Hamburg-Eppendorf, Hamburg, Germany
5 University Medical Center Hamburg-Eppendorf, Hamburg, Germany
6 University Medical Center Hamburg-Eppendorf, Hamburg, Germany
7 University Hospital Regensburg, Regensburg, Germany
8 University Hospital Regensburg, Regensburg, Germany
9 Ludwig Maximilian University Hospital, Munich, Germany
10 Ludwig Maximilian University Hospital, Munich, Germany
11 LMU University Hospital, Munich, Bavaria, Germany
12 Medical University of Innsbruck, Innsbruck, Austria
13 Bavarian Cancer Research Center (BZKF), Munich, Germany, Comprehensive Cancer Center Munich (CCC), Munich, Germany, German Cancer Consortium (DKTK), Partner Site Munich, Germany, Ludwig Maximilian University Hospital, Munich, Germany
14 Ludwig Maximilian University Hospital, Munich, Germany
15 German Cancer Consortium (DKTK), Partner Site Munich, Germany, Ludwig Maximilian University Hospital, Munich, Germany

Background:

Hemophagocytic lymphohistiocytosis (HLH) is a highly lethal hyperinflammatory syndrome that may arise in the context of various underlying conditions. Due to its orphan disease status and heterogeneous nature, understanding of disease biology and determinants of outcome remains limited. We recently introduced the HLH-Risk-Calculator, a machine learning-based tool providing individualized estimates of timepoint-specific mortality and initial disease severity (IDS; defined as admission to intensive care units (ICU) or early death without ICU admission) in adult patients with secondary HLH (sHLH). Predictions are based on sets of eight routine laboratory and clinical parameters, with serum levels of the soluble interleukin-2 receptor (sIL-2R), platelet counts, and serum albumin among the strongest predictors. Internal validation in held-out test cohorts demonstrated promising predictive performance. The HLH-Risk-Calculator may thereby provide a framework for individualized risk assessment and, following independent validation, support risk-based stratification in prospective clinical trials.

Methods:

We initiated an international, multicenter validation effort across Germany, Austria, Switzerland, and additional participating centers. Predictive performance is being evaluated in patients treated at centers contributing to the original model development as well as independent external cohorts of adults with sHLH. Model predictions of IDS and mortality risks are compared with prospectively or independently documented clinical outcomes to assess discrimination, calibration, and generalizability across heterogeneous real-world sHLH populations.

Results:

Ongoing validation demonstrates the feasibility of implementing the HLH-Risk-Calculator across international study sites and applying its underlying models to real-world sHLH cohorts. Preliminary analyses indicate preserved risk discrimination across heterogeneous populations composed of sHLH patients with various underlying disease triggers. Comprehensive analyses of predictive performance, including discrimination and calibration of risk estimates, are ongoing and will incorporate additional independent cohorts as data acquisition progresses.

A relevant practical limitation identified during ongoing validation efforts is the heterogeneous availability and reporting of sIL-2R, which emerged as a key prognostic marker in all of our models. In particular, measurement of sIL-2R is not universally established, and may be reported only relative to pre-defined cutoffs, limiting its use for detailed predictions due to the lack of quantitative values.

Conclusions:

External validation in international real-world patient cohorts represents a critical step toward potential clinical implementation of the HLH-Risk-Calculator. Beyond validating an individualized prognostic tool for adult sHLH, our ongoing study underscores the importance of comprehensive, quantitative measurement of sIL-2R. Based on its prominent role as both a diagnostic and prognostic marker, we encourage broad implementation and high-resolution reporting of sIL-2R across centers treating patients with sHLH. Upon successful external validation, the HLH-Risk-Calculator holds potential to support clinical decision-making and provide a basis for risk-adapted stratification in future prospective therapeutic trials.

Michael Ruzicka: No relevant conflicts of interest to declare.