Hierarchical model improves bone marrow cell classification and maturity ordering
HemaHier: Chain-Conditioned Ordinal Hierarchies for Lineage-Aware Bone-Marrow Cytology
Computer Vision and Pattern Recognition
Summary
Classifying bone marrow cells is tricky because cells belong to families and mature in order. The authors designed a model called HemaHier that respects these family trees and maturity stages, unlike common models that treat all mistakes equally. HemaHier reduces serious errors and adds maturity ordering inside cell families, helping better identify healthy and abnormal cells. They tested it on three bone marrow datasets and shared the code for others to use.
What this means in practice
- •For clinical diagnostic teams: Improve bone marrow cell classification systems to reduce critical misclassifications and provide maturity-based insights for better diagnosis.
- •For medical imaging software developers: Incorporate hierarchy- and maturity-aware models into software that analyzes bone marrow cytology images for enhanced cell typing.
Authors
Afshin Bozorgpour, Peter Schüffler, Edgar Jost, Dorit Merhof
Abstract
Bone-marrow cytology is inherently structured: each cell belongs to a hematopoietic lineage, and many cell types lie on ordered maturation trajectories. Standard flat classifiers ignore this structure, treating a mild same-lineage confusion the same as a severe cross-lineage mistake and predicting only discrete labels. We propose HemaHier, an ordinal-hierarchical prediction head for a frozen or lightly adapted cytology foundation model. Its central component is a chain-conditioned maturity score that reads a single maturity value under a per-chain query, supervised only on biologically valid healthy chains, while dysplastic and off-chain cell types remain classes but are excluded from maturity supervision. Fine and lineage predictions are coupled through a shared posterior that guarantees hierarchical consistency, and a staged objective first stabilizes recognition, then adds lineage and maturity supervision. On three bone-marrow datasets under a shared ontology, HemaHier achieves competitive recognition while reducing biologically severe errors and adding a within-lineage maturity ordering that flat classifiers lack. Code is available at https://github.com/xmindflow/HemaHier.