Prognostic Value of Immunophenotype In Infant ALL – Results of the INTERFANT'99 Study

Infant acute lymphoblastic leukemia (ALL) is characterized by distinct biological and clinical features. In this study, we aimed at detailed evaluation of infant ALL immunophenotype based on the large cohort of patients treated homogeneously on INTERFANT'99 protocol.

We were able to obtain sufficient flow cytometry based immunophenotypic data from 286 (130 boys and 156 girls) of 483 infant ALL cases treated on INTERFANT'99 protocols. The distribution of patient and leukemia characteristics as well as clinical outcome were very similar to the entire study cohort (Pieters et al, Lancet 2007; 370: 240–250). The positivity for a certain antigen was defined at the cut-off value of >20% of cells expressing this antigen. In B-cell precursor ALL (BCP-ALL), evaluated antigens included CD10, CD20, CD34, myeloid markers (CD13, CD33, CD15 and/or CD65) and NG2 antigen assessed with 7.1 antibody. The strength of association between characteristics (patient's age and gender, initial white blood count [WBC], CNS involvement, presence of MLL gene rearrangements, prednisone response [PR]) and antigen expression was analyzed with Fisher exact test adjusted for multiple comparisons. Univariate analysis of outcome by immunophenotype was based on event free survival (EFS) and compared by log-rank test. Multivariate analysis was performed using Cox model in MLL-rearranged cases. P values < 0.05 were considered to be significant.

Altogether, 272 cases of infant BCP-ALL and 14 of infant T-ALL were analyzed. Most BCP-ALL cases were CD10 negative (66%), CD20 negative (88%), CD34 positive (60%), and NG2 positive (58%). Co-expression of myeloid antigens was less frequent and concerned CD13 in 8%, CD33 in 17%, CD15 in 31%, and CD65 in 28%. CD10 negativity and NG2 expression were significantly more frequent in younger infants, patients with initial high leukocytosis and with MLL rearrangements. CD34 expression was observed more frequently in younger infants, while CD20 positivity was significantly associated with lower initial leukocytosis and absence of MLL rearrangement. For myeloid antigen expression, the only significant association was more frequent CNS involvement in infants with CD13+ and/or CD33+ BCP-ALL. Analysis of outcome showed significantly better prognosis for CD10 positivity (4-year EFS of 56.3 as compared to 37.2 for CD10 negative BCP-ALL), absence of NG2 expression (4-year EFS of 63.9 as compared to 34.9 for NG2 positive BCP-ALL) and CD33 negativity (4-year EFS of 47.6 as compared to 27.7 for CD33 positive BCP-ALL). There was also a trend towards better outcome in CD20 positive infant BCP-ALL. While CD10 negativity and NG2 positivity were associated with well-known high-risk features, prognostic significance of CD33 expression is somewhat more difficult to explain. Of note, CD10 negativity was not associated to a worst outcome in MLL-rearranged patients. Moreover, in this subset, multivariate analysis showed that none of the markers analyzed had independent prognostic significance, after adjustment for sex, age at diagnosis, WBC and PR. In rare infant T-ALL cases, we observed equal distribution to different immunophenotypic subgroups as defined by EGIL classification.

Infant BCP-ALL has distinct immunophenotypic features. However, immunophenotype has no independent prognostic relevance when MLL, age, WBC, PR and other factors are included.

No relevant conflicts of interest to declare.