Using deep learning for quantification of cellularity and cell lineages in bone marrow biopsies and comparison to normal age-related variation

Leander Van Eekelen,Hans Pinckaers,Michiel Van Den Brand,Konnie M Hebeda,Geert Litjens

doi:10.1016/j.pathol.2021.07.011

Abstract

Cellularity estimation forms an important aspect of the visual examination of bone marrow biopsies. In clinical practice, cellularity is estimated by eye under a microscope, which is rapid, but subjective and subject to inter- and intraobserver variability. In addition, there is little consensus in the literature on the normal variation of cellularity with age. Digital image analysis may be used for more objective quantification of cellularity. As such, we developed a deep neural network for the segmentation of six major cell and tissue types in digitized bone marrow trephine biopsies. Using this segmentation, we calculated the overall bone marrow cellularity in a series of biopsies from 130 patients across a wide age range. Using intraclass correlation coefficients (ICC), we measured the agreement between the quantification by the neural network and visual estimation by two pathologists and compared it to baseline human performance. We also examined the age-related changes of cellularity and cell lineages in bone marrow and compared our results to those found in the literature. The network was capable of accurate segmentation (average accuracy and dice score of 0.95 and 0.76, respectively). There was good neural network-pathologist agreement on cellularity measurements (ICC=0.78, 95% CI 0.58-0.85). We found a statistically significant downward trend for cellularity, myelopoiesis and megakaryocytes with age in our cohort. The mean cellularity began at approximately 50% in the third decade of life and then decreased ±2% per decade to 40% in the seventh and eighth decade, but the normal range was very wide (30-70%).

Highlights

Bone marrow trephine biopsies are routinely used in haematology to investigate myeloid diseases and for staging lymphomas
We evaluate the agreement between the neural network and two pathologists and compare this to human-level performance
We developed a neural network for the segmentation of six major cell and tissue types in bone marrow biopsies: erythropoiesis, myelopoiesis, megakaryocytes, lipocytes, trabecular bone, and erythrocytes

Summary

Introduction

Bone marrow trephine biopsies are routinely used in haematology to investigate myeloid diseases and for staging lymphomas. During the visual examination of a biopsy, various aspects of the tissue are evaluated, such as the marrow architecture and the cellularity of haematopoietic and stromal components.[1] The estimation of cellularity allows for a biopsy to be roughly categorised as hypo-, normo- or hypercellular, depending on the age and the clinical circumstances, which can give an indication of the activity of haematopoiesis. Many bone marrow diseases are characterised by either hypercellularity (e.g., myeloproliferative neoplasms) or hypocellularity (e.g., aplastic anaemia). Categorising biopsies as hypo-, normo- or hypercellular relies on the subjective internal reference frame of the pathologist and knowledge of the normal variations of bone marrow cellularity with age. Several studies published in the last 60 years found results that disagree on the rate of decrease and at which ages this decrease takes place.[2,3,4,5]

Methods

Results

Discussion

Conclusion