RT Journal Article
SR Electronic
T1 Automated digital enumeration of plasma cells in bone marrow trephine biopsies of multiple myeloma
JF Journal of Clinical Pathology
JO J Clin Pathol
FD BMJ Publishing Group Ltd and Association of Clinical Pathologists
SP 50
OP 57
DO 10.1136/jclinpath-2020-207066
VO 75
IS 1
A1 Jacques A J Malherbe
A1 Kathryn A Fuller
A1 Bob Mirzai
A1 Bradley M Augustson
A1 Wendy N Erber
YR 2022
UL http://jcp.bmj.com/content/75/1/50.abstract
AB Aims Determination of the number of plasma cells in bone marrow biopsies is required for the diagnosis and ongoing evaluation of plasma cell neoplasms. We developed an automated digital enumeration platform to assess plasma cells identified by antigen expression in whole bone marrow sections in multiple myeloma, and compared it with manual assessments.Methods Bone marrow trephine biopsy specimens from 91 patients with multiple myeloma at diagnosis, remission and relapse were stained for CD138 and multiple myeloma oncogene 1 (MUM1). Manual assessment and digital quantification were performed for plasma cells in the entire trephine section. Concordance rates between manual and digital methods were evaluated for each antigen by intraclass correlation analyses (ICC) with associated Spearman’s correlations.Results The digital platform counted 16 484–1 118 868 cells and the per cent CD138 and MUM1-positive plasma cells ranged from 0.05% to 93.5%. Overall concordance between digital and manual methods was 0.63 for CD138 and 0.89 for MUM1. Concordance was highest with diffuse plasma cell infiltrates (MUM1: ICC=0.90) and lowest when in microaggregates (CD138: ICC=0.13). Manual counts exceeded digital quantifications for both antigens (CD138: mean=26.4%; MUM1: mean=9.7%). Diagnostic or relapse threshold counts, as determined by CD138 manual assessments, were not reached with digital counting for 16 cases (18%).Conclusions Automated digital enumeration of the entire, immunohistochemically stained bone marrow biopsy section can accurately determine plasma cell burden, irrespective of pattern and extent of disease (as low as 0.05%). This increases precision over manual visual assessments which tend to overestimate plasma burden, especially for CD138, and when plasma cells are in clusters.All data relevant to the study are included in the article or uploaded as online supplemental information.