2026 ASH®
Annual Meeting & Exposition December 12-15, 2026 ▪ New Orleans, Louisiana

Paper Details

Abstracts, authors, affiliations, and disclosures.

Saturday, December 12, 2026
09:30 a.m. - 11:00 a.m.
MCCNO - Great Hall AD
Treatment Considerations, Combination Therapy, Biological Therapies, Antibody Therapy, Diseases, Myeloid Malignancies, Acute Myeloid Malignancies, Measurable Residual Disease, Research, AML, Translational Research, Clinical Research, Clinical Trials, Survivorship, Study Population, Human, Adult, Pediatric
PAPER

64 Gemtuzumab ozogamicin efficacy is limited to patients with KMT2A::MLLT3 fusions

Jack H. Peplinski, MSc, Rhonda E. Ries, MA2, Li Fan, PhD3, Danielle C. Kirkey, MD4, Todd A. Alonzo, PhD5, Alan S. Gamis, MD6, Katherine Tarlock I, MD7, Jatinder K Lamba, PhD8, Xiaotu Ma, PhD9, Soheil Meshinchi, MD, PhD10
1 Fred Hutchinson Cancer Center, Seattle, Washington, United States
2 Fred Hutchinson Cancer Center, Seattle, Washington, United States
3 St. Jude Children's Research Hospital, Memphis, Tennessee, United States
4 Fred Hutchinson Cancer Center, Seattle, Washington, United States, Seattle Children's Hospital, Seattle, Washington, United States, University of Washington, Seattle, Washington, United States
5 Children's Oncology Group, Monrovia, California, United States, University of Southern California, Los Angeles, California, United States
6 Children's Mercy Hospitals and Clinics, Kansas City, Missouri, United States
7 Fred Hutchinson Cancer Center, Seattle, Washington, United States, Seattle Children's Hospital, Seattle, Washington, United States, University of Washington, Seattle, Washington, United States
8 University of Florida, Gainesville, Florida, United States
9 St. Jude Children's Research Hospital, Memphis, Tennessee, United States
10 Fred Hutchinson Cancer Center, Seattle, Washington, United States, Seattle Children's Hospital, Seattle, Washington, United States, University of Washington, Seattle, Washington, United States

CD33-directed antibody-drug conjugate Gemtuzumab ozogamicin (GO) has been incorporated into frontline therapy in AML, especially in childhood AML trials. Despite its demonstrated efficacy in CD33+ cases, it continues to be administered without biologically informed patient selection. Previous biomarker efforts have focused primarily on CD33 antigen expression and inherited CD33 polymorphisms. Although both have been associated with differential GO response, their clinical implementation has been limited by the lack of standardized, validated companion diagnostic assays. We assembled a large, molecularly well annotated GO-treated and GO-naïve AML cohort to identify clinically actionable biologic predictors of GO response. We hypothesized that intrinsic leukemia biology, defined by recurrent fusion genes and cooperating mutations, identifies AML subgroups with differential benefit from GO and provides a practical framework for precision patient selection.

We performed an integrated biomarker analysis of 2132 patients enrolled on three consecutive Children’s Oncology Group AML trials: AAML03P1, AAML0531, and AAML1031. The GO-treated cohort consisted of all AAML03P1 patients and patients randomized to the GO-containing arm of AAML0531 (Arm B). The GO-naïve comparator cohort comprised patients randomized to the standard chemotherapy arm of AAML0531 (Arm A) together with all AAML1031 patients.

Patients were assigned to molecular subgroups using recurrent AML-defining fusion genes and somatic mutations and response and outcome was compared between those with and without GO-exposure. Fusion-defined analyses included RUNX1::RUNX1T1, CBFB::MYH11, KMT2A-rearranged AML (KMT2A-r), NUP98-rearranged AML (NUP98-r), CBFA2T3::GLIS2, and other recurrent fusion-defined leukemias. KMT2A-r was further stratified by individual fusion partners. Mutation-defined analyses included FLT3-ITD, NPM1, CEBPA, WT1, NRAS, and KRAS. The primary endpoint was complete remission rate (CR) and MRD rate after course 1, as well as 5-year event-free survival (EFS) and overall survival (OS). Within each molecular subgroup, GO-treated patients were compared with GO-naïve patients using Kaplan–Meier analyses, log-rank testing, and Cox proportional hazards regression.

Clinical characteristics and survival outcomes were comparable between AAML03P1 and arm B of AAML0531, as was outcomes between AAML1031 and arm A of AAML0531 supporting their unification. Initial clinical response and outcomes for those with and without GO treatment in different fusion groups was evaluated. Treatment with GO failed to improve post induction CR or MRD rates in any fusion or mutation group (all p=NS). Evaluation of EFS and OS for patients within different fusion groups demonstrated lack of benefit from GO treatment in all fusion groups except for those with KMT2A fusions. KMT2A-r AML patients treated with GO had a 5-year EFS of 48% vs that of 32% in GO-naïve patients, (p=0.0038) and improved 5-year OS (61.4% vs 51.6%, p=0.0498). Further stratification of KMT2A-r by fusion partner identified KMT2A::MLLT3 as the dominant and only GO-responsive subtype. GO exposure was associated with a marked improvement in 5-year EFS (74.1% vs 34.6%, p=0.0004) and OS (79.6% vs 55.7%, p=0.0176) compared with GO-naïve KMT2A::MLLT3 patients. Multivariable Cox regression demonstrated a substantial reduction in risk of treatment failure among GO-treated KMT2A::MLLT3 patients (HR 0.276, 95% CI 0.13–0.57, padj=0.0111). Other KMT2A fusion partners showed no benefit to the addition of GO. Recurrent AML-associated mutations did not identify GO-responsive populations.

In the largest molecularly annotated GO-treated and GO-naïve AML cohort reported to date, fusion-defined leukemia biology emerged as the dominant predictor of GO benefit. KMT2A::MLLT3 AML represents a uniquely GO-responsive subtype, whereas common AML mutations and other fusion-defined subgroups showed no benefit. Unlike CD33 expression or CD33 polymorphism testing, recurrent fusion genes are already established components of routine AML diagnostic evaluation and require no additional companion diagnostic assay for implementation. These findings support a transition from empiric GO administration toward molecularly guided patient selection and suggest that response to antibody-drug conjugates may be determined not only by target antigen expression, but by the intrinsic molecular state of the leukemia.

Jack H. Peplinski, MSc: No relevant conflicts of interest to declare.