Association of immunophenotype with cerebrospinal fluid involvement in childhood B-lineage acute lymphoblastic leukemia

The records of 71 pediatric patients who had B-lineage acute lymphoblastic leukemia (ALL) were studied retrospectively to document the correlation between antigenic phenotype of leukemic blasts in bone marrow (BM) and cerebrospinal fluid (CSF) involvement. Immunofluorescence assay for terminal deoxynucleotidyl transferase (TdT-IF) was used to examine specimens of CSF for the presence (CSF+) or absence (CSF-) of leukemic blasts at diagnosis and at various times after the induction of chemotherapy. The results of immunophenotyping of the leukemic blasts in the BM of the CSF+ and CSF- patient groups were then compared, and the relative risk ratio and positive predictive value were calculated for each marker. In addition other possible prognostic factors, such as gender, race and ethnicity, age, and initial leukocyte count were compared between the two groups. Thirty-eight percent of all patients, including those who had cytologically confirmed cases of CNS leukemia, had TdT(+) cells in the CSF at diagnosis or during the course of the disease. No differences were observed between the CSF+ and CSF- groups with respect to clinical and demographic factors. However, the cases of CSF+ leukemia had the almost exclusive expression of cytoplasmic immunoglobulin heavy chain (c mu) or surface CD22 (Leu-14), CD23 (B6), and/or IgM markers, which normally characterize the most developmentally mature members of the B-cell series in BM. The expression of any of these antigenic markers at diagnosis appears to identify at least 88% of cases of B-lineage ALL that have or ultimately may have TdT(+) cells in CSF. The results of this study therefore may have both basic and clinical implications: The former concerns the mechanisms by which these phenotypically defined forms of ALL invade and persist in the CNS; the latter concerns the utility of routine immunophenotyping of BM blasts at diagnosis to assess the biologic predisposition to CNS involvement in individual cases of ALL.