Outcomes of Children, Adolescents, and Young Adults with Acute Lymphoblastic Leukemia Based on Blast Genotype at Diagnosis: A Report from the Children'S Oncology Group

Mignon Loh,Elizabeth Raetz,Meenakshi Devidas,Yunfeng Dai,Michael Borowitz,Andrew Carroll,I-Ming Chen,Julie Gastier-Foster,Alison Friedmann,Richard Harvey,Nyla Heerema,Eric Larsen,Kelly Maloney,Leonard Mattano,Charles Mullighan,Karen Rabin,Shalini Reshmi,Kathryn Roberts,Cheryl Willman,Brent Wood,Patrick Zweidler-McKay,Jinghui Zhang,Naomi Winick,Stephen Hunger,William Carroll
DOI: https://doi.org/10.1182/blood.v128.22.451.451
IF: 20.3
2016-01-01
Blood
Abstract:Abstract Survival for childhood acute lymphoblastic leukemia (ALL) now approaches 90% with risk adapted therapy based on National Cancer Institute risk group (NCI RG) at diagnosis, somatic lymphoblast genetics, and early response to therapy as measured by minimal residual disease (MRD). The Children's Oncology Group AALL03B1 ALL Classification trial enrolled 11,145 children, adolescents, and young adults less than 31 years of age with newly diagnosed B- or T-lineage ALL between December 2003 and September 2011. Companion therapeutic trials for B-lineage ALL included AALL0331 (n= 5226) for NCI standard risk (SR-ALL)(age 1-10 years and white blood cell count (WBC) < 50,000/uL) and AALL0232 (n=2907] for NCI high risk (HR) ALL (age > 10 years or presenting WBC > 50,000/uL). Assessing outcome by lymphoblast genetics revealed statistically significant distributions of genotype and NCI RG, as well as differences in event-free and overall-survival (EFS, OS) (Table 1). Not surprisingly, favorable genetic groups of Trisomy 4/10/17 (TT) and ETV6/RUNX1 were significantly more common in NCI SR patients (p< 0.0001 for each) while those with unfavorable characteristics (MLL rearranged [MLLr], intrachromosomal amplification of chromosome 21 [iAMP21], BCR/ABL1 and hypodiploidy [n<44]) occurred more frequently in NCI HR patients (p<0.0001 for each). Event-free and OS were correspondingly poorer in NCI HR patients with the exception of iAMP21, where EFS in those treated on AALL0232 was better than that in NCI SR (Table 1). Surprisingly, NCI SR BCR/ABL1 positive patients (N= 64, 1.2%) had a 5-year EFS of 85±5.0%, although these patients came off study at end induction and likely received imatinib with chemotherapy on a companion ALL trial for Ph+ ALL. Notably, hypodiploidy was associated with the worst EFS and OS regardless of NCI RG, with 5-year EFS and OS of 51.3±5.0% and 58.2±5.0%, respectively, suggesting that these patients continue to fare poorly with salvage therapies. Multivariable analysis demonstrated age, WBC, and day 29 MRD as significant independent risk factors for sustained CR, and the individual genetic groups of TT, ETV6/RUNX1, iAMP21, BCR/ABL1 and hypodiploidy, but notably, not MLLr, all retained independent prognostic significance when added to the model individually. A subset of consecutively enrolled (N=605) AALL0232 NCI HR patients underwent additional genomic interrogation, including assessment of Ph-like status, ABL1 class fusions, CRLF2r, JAK mutations (JAKm), and IKZF1 alterations. Based on sample availability, patients studied were younger (p < 0.0001) and had a higher WBC (p < 0.0001) compared to the remainder of AALL0232. There were 85/605 (14.0%) Ph-like patients defined using PAM clustering algorithms and Ph-like status was significantly associated with an IKZF1 alteration (75%) (p < 0.0001) and day 29 MRD > .01% (p < 0.0001). Five-year EFS for Ph-like versus non Ph-like was 62.3±5.8% vs. 83.9±1.7% (p < 0.0001). Outcomes of Ph-like ALL with or without CRLF2r were similar (60.6±8.3% versus 65.4±8.0%, p =0.86). Similarly, 5-year EFS of Ph-like ALL was no different with or without IKZF1 alterations (61.5±7.0% vs. 64.6 ±11.1%). There were 155 (27.1%) IKZF1 alterations, 60 of which occurred in Ph-like ALL with a trend towards concomitant CRLF2r/JAKm (p=0.055). Five-year EFS for patients with IKZF1 alterations was 66.8±4.0% (p < 0.0001) versus 86.4±1.8% for those without IKZF1 lesions. Multivariable analysis demonstrated age, WBC and day 29 MRD as independent risk factors for sustained CR while only Ph-like status, IKZF1 alteration, BCR/ABL1 and ETV6/RUNX1retained independent prognostic significance when added to the model individually. In summary, somatic sentinel cytogenetic alterations are independently prognostic in childhood ALL and are strongly associated with NCI RG and outcome, supporting continued incorporation into risk stratification algorithms. Notably, >44% of all patients have favorable blast cytogenetics with 5-year overall survival rates approaching 100%. In contrast, NCI HR patients with Ph-like ALL have poor outcomes with currently available therapy and this subtype is associated with CRLF2 r and IKZF1 alterations, which do not confer an inferior EFS within the Ph-like subgroup. Novel therapies for genomicallydefined Ph-like ALL may improve outcomes. Disclosures Loh: Bristol Myers Squibb: Membership on an entity's Board of Directors or advisory committees; Abbvie: Research Funding. Borowitz:HTG Molecular: Consultancy; Bristol-Myers Squibb: Research Funding; MedImmune: Research Funding; BD Biosciences: Research Funding.
What problem does this paper attempt to address?