AI-Based Hematological Malignancy Prediction From Peripheral Blood Smears in a Large Diagnostic Laboratory Cohort

MCML Authors

Ario Sadafi

→ Group Carsten Marr
 Artificial Intelligence in Cell Therapy and Hematology
→ Co-Group Nassir Navab

Carsten Marr

Prof. Dr.

Principal Investigator

Artificial Intelligence in Cell Therapy and Hematology

Abstract

Hematological malignancies represent a wide range of disease entities, most of which arise from dysfunctional proliferation and differentiation of hematopoietic stem and progenitor cells in the bone marrow [1]. Diagnosis requires integration of cytomorphology, molecular genetics, and immunophenotyping from blood or bone marrow. Unlike bone marrow aspiration, assessing cytomorphology in a blood smear is fast, minimally invasive, and provides information on differential cell counts and morphological abnormalities that guide follow-up diagnostic pathways. However, conventional peripheral blood smear analysis involves labor-intensive manual examination of hundreds of cells, which is subject to inter-observer variability. Previous work explored machine-learning for single-cell classification [2,3], and disease detection [4,5,6,7,8,9] on curated cohorts. Systematic evaluation across multiple malignancies at their natural clinical distribution remains unexplored.

article DKL+26

---

Leukemia

Mar. 2026.

---

Authors

M. F. Dasdelen • I. Kukuljan • P. Lienemann • F. Ozlugedik • A. Sadafi • M. Hehr • K. Spiekermann • C. Pohlkamp • C. Marr

---

Links

DOI

---

Research Area

 C1 | Medicine