Early Response to Therapy Is Significantly Associated with Genetic Subtype of Acute Lymphoblastic Leukemia: A Report from the Children’s Oncology Group.

Improved outcomes for children with acute lympboblastic leukemia (ALL) have been achieved, in part, from adaptation of risk-stratified therapy. The Children’s Oncology Group (COG) has implemented a real-time risk classification system (AALL03B1) using a combination of NCI-Rome risk criteria, blast cell genetic features, and early treatment response to determine the intensity of post-induction therapy. Between December 29, 2003 and June 1, 2007, more than 4,000 children over 1 year of age with B-precursor ALL were enrolled on AALL03B1, including 2293 (62%) with NCI Standard Risk (SR) and 1406 (38%) with NCI High Risk (HR) features who were subsequently enrolled on companion clinical trials. The most favorable genetic features used in AALL03B1 were identified in legacy COG studies and included TEL/AML1(TEL) or triple trisomies of chromosomes 4, 10, and 17 (TT). Unfavorable genetic features included the presence of BCR/ABL, MLL rearrangements, or extreme hypodiploidy (DNA index <.81 or chromosomes <44). Overall, 26% of patients were TEL+ and 24.7% had TT. These genetic subsets occurred more frequently in NCI SR vs. HR patients (30.7% and 30.9% vs. 14.5% and 11.7% respectively). Children achieving an M1 day 15 bone marrow (BM) who also had minimal residual disease (MRD) < 0.1% measured by flow cytometry on day 29 of induction therapy were deemed rapid early responders (RER), while those with either an M2/M3 day 15 marrow or MRD > 0.1% at day 29 were defined as slow early responders (SER). Among the favorable cytogenetic subsets, patterns of early response differed. The presence of TEL was significantly associated with an RER to induction therapy in both NCI SR and HR groups (p< 0.0001), while the presence of TT was not (p=0.058). For NCI SR patients, the presence of TEL was significantly associated with the achievement of an M1 bone marrow by day 8 (50.9% of TEL+ pts vs. 41.2% of TEL- pts, p< 0.0001). Patients with an M1 or M2 BM on day 29 who had MRD >1% received extended induction (EI) for two weeks followed by an additional evaluation of BM morphology and MRD at day 43 of induction. One hundred and nineteen patients received EI, with 40% having NCI SR features at diagnosis. Of the patients who received EI, 63% achieved an M1 marrow with MRD < 1% by day 43 and were eligible to continue on protocol therapy. This was more likely to occur in NCI SR patients (77% vs. 55%, p<0.013). Not surprisingly, 31% of the NCI HR patients receiving EI were BCR/ABL positive, and the presence of BCR/ABL was associated with a slower early response overall. While the presence of the BCR/ABL was associated with a greater likelihood of EI, MLL rearrangements and hypodiploidy were not. These data indicate that early response to induction therapy differs among genetic subsets of pediatric patients with newly diagnosed ALL. In addition a centralized classification system allows for robust collection of data from local and centralized reference laboratories that can be used for real time treatment assignment of ∼2000 patients/year with ALL from > 220 COG institutions.