No abstract

Bone marrow (BM) cellularity assessment is a crucial step in the evaluation of BM trephine biopsies for hematological and non-hematological disorders. Clinical assessment is based on a labor-intensive, semi-quantitative visual estimation of the hematopoietic and adipocytic components by hematopathologists, which does not provide quantitative information on other stromal compartments. In this study, we developed and validated MarrowQuant 2.0, an efficient user-friendly digital hematopathology workflow integrated within QuPath software which serves as BM quantifier for five mutually-exclusive compartments (bone, hematopoietic, adipocytic, interstitial/microvasculature areas, and “Other”) to derive cellularity of human BM trephine biopsies. Instance segmentation of individual adipocytes is realized via adaptation of the machine-learning-based algorithm StarDist. We calculated BM compartments and adipocyte size distributions of haematoxylin and eosin (H&E) images obtained from a total of 250 bone specimens, from control and acute myeloid leukemia or myelodysplastic patients at diagnosis or follow-up, then measured the agreement of cellularity estimates by MarrowQuant 2.0 against visual scores from four hematopathologists. The algorithm was capable of robust BM compartment segmentation with average mask accuracy of 86%, maximal for bone (99%), hematopoietic (92%) and adipocyte (98%) areas. MarrowQuant 2.0 cellularity score and hematopathologist estimations were highly correlated (R2 = 0.92–0.98, Intraclass-Correlation-Coefficient ICC = 0.98; inter-observer ICC 0.96). BM compartment segmentation quantitatively confirmed reciprocity of the hematopoietic and adipocytic compartments. MarrowQuant 2.0 performance was additionally tested for cellularity assessment of specimens prospectively collected from clinical routine diagnosis. After special consideration for the choice of the cellularity equation in specimens with expanded stroma, performance was similar in this setting (R2 = 0.86, n = 42). We thus conclude that MarrowQuant 2.0 can be applied in a clinical setting. We expect this workflow will contribute to improving the speed and ease of diagnosis in hematopathology and serve as a clinical research tool to explore novel biomarkers related to BM stromal components.