Identification of Neoplastic Infiltration of the Cerebrospinal Fluid (CSF) in Patients with Aggressive B-Cell Non-Hodgkin’s Lymphoma (B-NHL) without Clinical Evidence of Leptomeningeal Disease: A Comparative Analysis of the Utility of Flow Cytometry (FCM) Versus Conventional Cytology (CC).

Introduction: CC analysis of the CSF is considered as the reference method to diagnose meningeal disease in patients with B-NHL. However, recent studies suggest that in patients with B-NHL who are at high risk of CNS relapse, FCM could be more sensitive than CC for detecting meningeal disease in CSF.

Objective: To evaluate the sensitivity and specificity of a standardized FCM immunophenotypic approach vs CC for detecting the presence of neoplastic cells in CSF, in patients with aggressive B-NHL who are at high risk of CNS relapse.

Patients and methods: A total of 29 CSF samples were analysed (total volume: 0.8 to 4ml; median: 2.4) in newly diagnosed patients with aggressive B-NHL, from a total of 14 different hospitals (diffuse large B cell lymphoma-DLBCL: 17; Burkitt’s lymphoma-BL: 9; follicular lymphoma transformed to DLBCL -tFL: 2; and T-cell-rich B-NHL: 1). Of the 29 patients studied, 15 were men (52%) and 14 women (48%) with a mean age of 55 ± 19 years (range: 16–86). In all cases, the CSF samples were analysed simultaneously by CC at the institution of origin and FCM, centrally one institution. For the FCM analysis of the CSF, stabilised samples (Transfix, CYTOMARK) were systematically stained with the following combination of monoclonal antibodies: CD8-sIgl/CD56-sIgk/CD4-CD19/CD3/CD20/CD45 (FITC/PE/PERCPCY5.5/PECY7/APC/APCCY7). If the FCM test showed infiltration, an additional 6-color antibody panel was used for full phenotypic characterisation of the disease.

Results: Haematopoietic cells were detected in all cases (mean: 2.2 ± 4 cells/μl; range: 0.2–14). Systematically, these cells included T cells (mean: 0.6±0.8/μl; range: 0.1–4) and monocytes (mean: 1±2/μl; range: 0.1–9). Furthermore, in 31%, 3%, 3% and 14% of cases the following cell populations were also detected: polyclonal B-lymphocytes (mean: 0.06±0.07/μl; range: 0.01–0,2), plasma cells (0.09/μl), natural killer cells (0.03/μl) and neutrophils (mean: 0.6±1/μl; range: 0.03–2), respectively. Of the 29 cases studied, 5 (17%) showed infiltration by neoplastic B-cells by FCM, while CC only showed infiltration in two of these patients (7%). Cases with CNS infiltration by both methods included one tFL (82%; 14 neoplastic cells/μl by FCM) and one BL (68%; 8 neoplastic cells/μl using FCM). The three patients with FCM+/CC- presented a lower percentage of pathological B-cells (2% in one DLBCL and 1% and 0.1% in two BL, respectively). One of these three patients presented neurological symptoms (meningism).

Conclusion: Although preliminary, these results suggest that FCM is more sensitive that CC in detecting CSF infiltration by neoplastic B-cells in aggressive B-NHL, especially when these cells are present in a relatively low numbers.