Response: Genes associated with response to therapy in childhood acute lymphoblastic leukemia

We thank Dr Catchpoole and colleagues for examining the set of 14 genes that we recently reported.¹  These genes were among those significantly associated with minimal residual disease (MRD) during remission induction therapy in children with acute lymphoblastic leukemia (ALL); they were also independent predictors of relapse in a separate patient cohort enrolled in our Total XIII study. The analysis of Catchpoole et al confirmed that expression of the 14-gene set can distinguish 2 subsets of childhood ALL. However, this classification lacked prognostic significance in their series.

We, of course, agree that the value of gene expression profiles as prognostic indicators depends on the treatment regimen. In fact, this caveat applies to any prognostic factor. Nevertheless, we would like to point to differences in clinical outcome and in follow-up observation time between the 2 studies that should be considered when interpreting the data. All patients in the cohort studied by Catchpoole et al relapsed within 3 years of therapy whereas most of the relapses in our series occurred after that time. This suggests significant differences in the antileukemic effect of the treatment administered in the Total XIII study and that given in the Australian Berlin-Frankfurt-Munster (BFM) 95/Australian and New Zealand Children's Haematology and Oncology Group (ANZCHOG) VIII protocol. It seems that patients with a “high-proliferative” profile did particularly well in Total XIII (and subsequent trials), whereas they did not show a superior response in the Australian study. We think that it is premature to conclude, as Catchpoole et al do, that BFM95/ANZCHOG VIII regimen might be more effective for patients with “low proliferative” profile. The follow-up time in the Australian cohort appears to be significantly shorter than ours and many relapses among patients with this profile in our cohort occurred late.

Cario et al²  reported that expression of MADL2, one of our 14 independent predictors of relapse, was associated with MRD on week 12 in the BFM2000 study. It is not clear whether genes strongly associated with MRD and/or relapse were identified by Catchpoole et al. If this is the case, it would be interesting to know if any also appeared in the list of 40 genes associated with relapse in our study (Table S1 in the article). Searching for genes related to treatment response across different regimens might provide clues about general mechanisms that regulate drug sensitivity in leukemic cells.