PB1722 IMPACT OF CD33 AND ABCB1 SINGLE NUCLEOTIDE POLYMORPHISMS IN PATIENTS WITH ACUTE MYELOID LEUKEMIA AND HIGH‐RISK MYELODYSPLASTIC SYNDROMES TREATED WITH DECITABINE PLUS GEMTUZUMAB OZOGAMICIN

Abstract

Background:Single nucleotide polymorphisms (SNPs) of CD33 are predictive for relapse risk in children with acute myeloid leukemia (AML) receiving gemtuzumab ozogamicin (GO)‐containing regimens. However, the impact of CD33 SNPs on response to GO is controversial in adults. To our knowledge, no study has evaluated whether SNPs of CD33 or ABCB1, a protein that mediates GO resistance in pre‐clinical studies, predict for outcomes in patients (pts) with AML or high‐risk myelodysplastic syndromes (MDS) receiving low‐intensity therapy in combination with GO.Aims:To evaluate the impact of CD33 expression and of CD33 and ABCB1 SNPs on response, risk of relapse, and survival in pts with AML or high‐risk MDS or myelofibrosis treated in a prospective clinical trial with decitabine plus GO.Methods:Pts received decitabine 20 mg/m2 daily x 5 days plus GO 3 mg/m2 on day 5, followed by up to 5 consolidation cycles of decitabine plus GO, then decitabine alone (Daver et al. Leukemia 2016;30[2]:268‐73). Genotyping of the following SNPs was performed and correlated with clinical outcomes: ABCB1 rs1128503, ABCB1 rs1045642, CD33 rs2455069, CD33 rs35112940, CD33 rs12459419, CD33 rs1803254, CD33 rs61736475, and CD33 rs201074739. An aggregate “6 CD33 SNP score” was determined based on genotypes of the 6 CD33 SNPs, as previously described (Lamba J et al. ASH 2017, abstract 3826).Results:113 pts were treated with decitabine plus GO, 104 of whom underwent CD33 and ABCB1 SNP genotyping. Baseline characteristics of the population is shown in Table 1. For the entire cohort, 36 pts (32%) achieved any response (ORR, i.e. CR + CRi + MLFS), 16 of whom (14%) achieved CR. ORR was 47% and 18%, and CR rate was 23% and 7% for the frontline and relapsed/refractory cohorts, respectively. With a median follow‐up of 107 months, median RFS and OS were 4.8 and 10.0 months in the frontline, and 3.0 and 6.6 months in the relapsed/refractory cohorts, respectively. For the whole cohort, CD33 median fluorescence intensity (MFI) was not significantly associated with ORR, but was associated with CR (OR 1.02; P = 0.02). No individual CD33 or ABCB1 SNP was predictive for ORR or CR; however, among pts with 6 CD33 SNP score ≥0, there was a trend towards higher CR rate (21% vs. 8%, OR 3.08; P = 0.07), but not with ORR (P = 0.94). Neither CD33 MFI nor any individual genotype or the 6 CD33 SNP score were significantly associated with cumulative incidence of relapse (CIR), relapse‐free survival (RFS) or overall survival (OS). In the subgroup of pts with AML (n=91), CD33 MFI was not significantly associated with ORR, but was associated with CR (OR 1.02; P = 0.02). The GG genotype of CD33 rs1803254 (n=67) was associated with lower CIR and better RFS than the CC/CG genotypes (n=15) in the AML cohort (median RFS 5.1 versus 2.5 months, P = 0.02), although this was driven by only 3 pts with the CC/CG genotypes who responded and included in RFS analysis; there was no association with OS. No other SNP, nor the 6 CD33 SNP score, was significantly associated with CR, ORR, CIR, RFS or OS in any subgroup analyses.Summary/Conclusion:In this poor‐risk cohort of pts with MDS or AML treated with decitabine plus GO, neither CD33 MFI nor ABCB1 or CD33 SNPs were meaningfully predictive for outcomes; however, this cohort was heterogeneous and the overall efficacy of this regimen was low, which may have obscured any potential predictive value of these SNPs. Evaluation of these SNPs in pts treated with more effective low intensity regimens (e.g. hypomethylating agent plus venetoclax) in combination with GO is warranted.image