Flow cytometry to detect bone marrow involvement by follicular lymphoma.

Abstract

High-quality data on bone marrow involvement (BMI) assessed by flow cytometry (FC) in follicular lymphoma (FL) is lacking. We set up a prospective protocol with a 10-color tube and acquisition of 500.000 leukocytes on a Nav flow cytometer for evaluation of BMI in FL by FC. FC was compared with a combination of histopathology and IGH gene rearrangement, which were considered the gold standard. We also compared BMI by FC with PET. Fifty-two patients were included (median 67 years, 54% female). BMI by FC was seen in 35 (67%), with a median involvement of 1.2% (interquartile range: 0.3%-7%) of leukocytes. Comparison with the gold standard revealed two false negatives and two false positives (potentially true involvement undetected by the gold standard). BMI by PET was seen in 14/46 (30%). Immunophenotype of FL in the bone marrow was highly heterogeneous. The most common phenotypic abnormality was dim expression of CD19 (>0.5 log loss in 30% of patients). CD10 was negative in 13 (37%) and incompletely positive (overlap with the negative population) in a further 8 (28%) while entirely positive only in 14 (48%). Other abnormalities (loss of CD20, gain or loss of CD79b, expression of CD43, and substantial loss of CD45) were rare. Computational analysis by means of FlowSOM confirmed the heterogeneous phenotype, with FL from different patients clustering in unrelated metaclusters. BMI by FL was frequent and immunophenotype was heterogeneous. However, this protocol enabled detection of FL in bone marrow in the vast majority of patients with bone marrow involvement by the gold standard.