Recently we showed that in 20-30% of children diagnosed as BCR::ABL1-positive (Ph+) acute lymphoblastic leukemia (ALL) the BCR::ABL1 fusion is present in a wider clone, involving myeloid cells, non-ALL B-cells and T-cells. This leads to discordant minimal residual disease (MRD) levels assessed by the quantification of two available MRD targets - immunoglobulin (IG)/T-cell receptor (TR) genes rearrangements and BCR::ABL1 fusion. As the cases with multilineage BCR::ABL1 involvement resemble lymphoid blast crisis (LBC) of chronic myeloid leukemia (CML) we named these leukemias "CML-like", as distinct from "typical Ph+ ALL". To reveal biological and clinical differences between the two subtypes, we retrospectively analyzed prognostic relevance of MRD and other features in 147 children with BCR::ABL1-positive ALL treated according to EsPhALL (n = 133) or other (n = 14) protocols. MRD was assessed by DNA-based monitoring of BCR::ABL1 genomic breakpoint and of clonal IG/TR rearrangements. Although overall prognosis of CML-like (n = 48) and typical Ph+ ALL (n = 99) was similar (5-year EFS 60% and 49%; 5-year OS 75% and 73%, respectively), typical Ph+ ALL presented more relapses (12/48 vs. 42/99; p = 0.046) while CML-like patients more often died in the first remission (6/48 vs. 3/99; p = 0.059 for CR1 deaths in total and p = 0.01 for CR1 deaths out of the total deaths). Prognostic role of MRD measured at standard treatment timepoints (TP1 - end of induction IA, day 33; TP2 - end of consolidation IB, week 12) was highly significant in the typical Ph+ ALL (p = 0.0005 for EFS). In contrast, in CML-like patients prognostic impact of MRD was not significant and thus inapplicable for therapy adjustment. Similarly, although diagnostic white blood cell counts were similar in the two subgroups, impact of the hyperleukocytosis ≥ 50 x 109/l on outcome was highly significant in the typical Ph+ ALL (and remained significant in multivariate analyses, being apparent even in patients with fast MRD response), but no prognostic impact was found in CML-like patients. While in the typical Ph+ ALL the NCI risk tended to impact 5-year OS (89% vs. 67%; p = 0.03), there was no impact on prognosis in the CML-like patients. The BCR::ABL1 transcript type (minor/p190 vs. Major/p210), IKZF1 deletion, sex, or the age at diagnosis (< 10 vs. ≥ 10 years) did not show any prognostic relevance. Earlier start of continuous tyrosine kinase inhibitors (TKI) treatment (day 15 vs. day 33) was associated with lower MRD levels in the typical Ph+ ALL but not in CML-like patients. However, earlier TKI start tended to result in better 5-year OS in both subtypes (78% vs. 67%; p = 0.082 and 83% vs. 60%; p = 0.066 for typical Ph+ ALL and CML-like, respectively). To shed more light on the nature of the two subtypes, we now collect additional data (including transcriptome profiles, single cell characteristics, genomic breakpoint features) in order to reveal biological background, cell of origin and further potential differences between the subtypes. Such data could help to find optimal treatment strategies, and also to answer the attractive question whether CML-like disease is in fact a regular CML in LBC or whether there is a wider spectrum of BCR::ABL1-positive leukemias. If these studies lead to a conclusion that CML-like is indeed indistinguishable from LBC-CML, it will be necessary to re-define CML as a disease manifesting - at least in childhood - very often in LBC and with the minor-BCR::ABL1/p190 fusion variant. In conclusion, our data unequivocally confirm that although clinically presenting at diagnosis as one entity, there are two biologically distinct subtypes of BCR::ABL1-positive ALL - typical Ph+ ALL and CML-like disease. The different biology is reflected in treatment response and although the overall survival rates of both subtypes are similar on current protocol, the key reasons of treatment failure are different - toxicity of intensive ALL treatment in CML-like patients and relapses in typical Ph+ ALL. Thus, early distinguishing of the two subtypes is essential to enable optimal treatment approach and therapy adjustments in upcoming trials. For the typical Ph+ ALL, TKI and concurrent chemotherapy with risk-directed intensity should be recommended; in the CML-like disease, representing one-fourth to one-third of childhood patients diagnosed as BCR::ABL1-positive ALL, no relevant risk factor applicable for therapy tailoring was found so far.
Read full abstract