Diffuse Large B-Cell Lymphoma, Not Otherwise Specified: Discordance between Histology and Flow Cytometry in Bone Marrow Examination at Diagnosis

Background: Bone marrow (BM) examination is essential in the staging of diffuse large B-cell lymphoma (DLBCL). The assessment of BM involvement includes both histology (gold-standard) and flow cytometry (FC), but few studies have compared BM biopsy (BMB) histologic findings with results of FC analysis of BM aspirate. Discordance between both techniques generates debate about the staging and the prognostic significance in these cases. Methods: We performed a retrospective single-center analysis of patients with DLBCL, not otherwise specified (NOS) diagnosed during a 4-year period (2014-2017). Patients were divided in three groups according to BM findings of BMB and FC at diagnosis. Standard FC was performed by 4-color flow panel until 2016 and by 8-color FC since then. We described main characteristics of each group at diagnosis and analyzed survival outcomes. We applied means of descriptive statistics and Pearson's chi-squared test, and analyzed survival outcomes according to Kaplan-Meier, using Cox regression for comparisons. Results: We analyzed 59 cases, which were divided in three groups: 40 cases (67.8%) presented both negative histology and FC (BMB-/FC-), 10 (16.9%) showed BM involvement using both histology and FC (BMB+/FC+) and 9 cases (15.3%) presented discordant results, all of them with negative histology and positive FC (BMB-/FC+). Clinical and biological characteristics of each group at diagnosis are presented in Table 1. Median infiltration by FC analysis of the BMB-/FC+ group was 0.8% (0.1-2.9) and 3/9 patients presented discordant immunophenotype of lymphoma cells between BM and node biopsy. If we considered BM infiltration as positive in all BMB-/FC+ cases, 4/9 (6.8% of all patients) would be upstaged. First-line treatment was homogeneous in all patients. With a median observation time of 18 months, progression-free survival (PFS) after 2 years was 67%, 22% and 22% with BMB-/FC-, BMB-/FC+ and BMB+/FC+, respectively (Figure 1A), with a multivariate hazard ratio (HR) of 1.9 (95% CI 1.2-2.9, p=0.004) and an univariate HR for FC+ (BMB-/FC+ and BMB+/FC+) vs FC- (BMB-/FC-) of 3.3 (95% CI 1.5-7.3, p=0.003). Two-year overall survival (OS) was 68%, 41% and 33% with BMB-/FC-, BMB-/FC+ and BMB+/FC+, respectively (Figure 1B); multivariate HR was 1.6 (95% CI 1.1-2.6, p=0.042) and univariate HR for FC+ vs FC- was 2.5 (95% CI 1.1-5.9, p=0.035). We found no significant difference between BMB-/FC+ and BMB+/FC+ in survival outcomes. Conclusions: In our seri