The reproducibility of acute lymphoblastic leukemia phenotype determinations: evaluation of monoclonal antibody and conventional hematopoietic markers

Peripheral blood and/or bone marrow lymphoblasts from 25 patients with acute lymphoblastic leukemia (ALL) were tested with a panel of monoclonal antibodies (MoAbs) and conventional hematopoietic markers by three different laboratories. The results were analysed to evaluate the reproducibility of ALL phenotype determinations. Specimens were transported between laboratories by 24-h courier service and were classified on the basis of indirect immunofluorescence MoAb reactivities as follows: B-lineage ALL (BA-1⁺T^-MCS-2^-); T-lineage ALL (T⁺BA-1^-MCS-2^-); myeloid antigen ALL (MCS-2⁺BA-1^-CALLA^-T^-) and unclassified ALL (BA-1^-MCS-2^-CALLA^-T^-). Conventional marker studies for surface immunoglobulin (sIg), cytoplasmic immunoglobulin (cIg), sheep erythrocyte rosette formation (E) and nuclear terminal deoxynucleotidyl transferase (TdT) were also performed. In the cases with sufficient marker data to permit classification, 90% (18/20) were identically classified by different laboratories and this concordance was statistically significant (p < 0.05). The agreement between laboratories for individual MoAb and conventional marker analyses was statistically significant (p < 0.05) for all markers with the exception of BA-2, cIg and TdT determinations. Six of 7 discordant BA-2 cases represented BA-2⁺ evaluations which had subsequent BA-2^- results following specimen transportation. These findings suggest instability of the BA-2 antigen to transport conditions. A similar pattern of positive to negative evaluations following transportation was observed in 5/5 discordant results involving other MoAbs. Disagreement between laboratories for cIg and TdT determinations implies that the detection of cytoplasmic or nuclear antigens may be more prone to subjective interpretation than cell surface antigen marker analyses. Our findings suggest that immunofluorescence marker studies employing MoAbs to cell surface antigens are in general highly reproducible. Our results also indicate that specimen storage conditions and the cellular location of target antigens are important variables which may affect the reproducibility of ALL phenotype determinations.