AUTHORS
Lukas Ronner, MD MSCR, Lacey Williams, MD2, Pamela J. Sung, MD PhD3, Akmaljon Salimov4, Paige Pammer, MD5, Eesha Balar6, Katie Moffitt7, Leslie Martinez-Guitierrez8, Rahul Kumar Thakur, MD9, Matthew P Connor, MD10, Ximena Jordan Bruno, MD11, Shannon R. McCurdy I, MD12, Elizabeth Hexner, MD13, Alexander E. Perl, MD14, Keith W. Pratz, MD15, Selina Luger, MD, FRCPC16, Andrew Matthews, MD17, Martin Carroll, MD18, Sarah Skuli, MD, PhD19, Catherine Lai, MD, MPH20
AFFILIATIONS
1 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
2 University of North Carolina Medical Center, Chapel Hill, North Carolina, United States
3 Roswell Park Comprehensive Cancer Center, Buffalo, New York, United States
4 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
5 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
6 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
7 University of North Carolina Medical Center, Chapel Hill, North Carolina, United States
8 Roswell Park Comprehensive Cancer Center, Buffalo, New York, United States
9 Roswell Park Comprehensive Cancer Center, Buffalo, New York, United States
10 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
11 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
12 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
13 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
14 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
15 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
16 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
17 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
18 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
19 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States, University of Cincinatti, Cincinatti, Ohio, United States
20 Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States
ABSTRACT
Background: TP53-mutated AML is a large unmet need in malignant hematology. Exhibiting severe genomic instability and high rates of treatment resistance, optimal therapy is actively debated. This debate is further complicated by evolving knowledge of genetic heterogeneity of TP53-AML. While TP53 allelic status clearly associates with outcomes, the effect of co-mutations in targetable drivers is poorly understood. IDH1 or IDH2 are co-mutated in up to 12% of TP53-AML, and multiple retrospective cohorts (COMMAND, Flatiron Health) link IDH co-mutation to improved overall survival (OS). Whether this reflects an effect of the IDH mutation or the cytomolecular context in which IDH mutations occur remains unknown. We therefore assembled a multi-institutional, cytomolecularly-annotated TP53-AML cohort enriched for IDH co-mutations to compare the genetic landscape and outcomes of these subgroups.
Methods: Subjects with TP53-AML (including concomitant IDH1 or IDH2 mutation, “IDHm”) diagnosed and treated at the Hospital of the University of Pennsylvania (HUP) 2013-2023 were identified by EMR query and enriched with similarly-identified TP53/IDH co-mutated subjects from the University of North Carolina (UNC) and Roswell Park Comprehensive Cancer Center (RPCCC). TP53 variants were referenced against the IARCTP53 database to assign mutation type and affected domain. Demographics, induction approach, multi-hit status (WHO definition), and AML predisposition (de novo, antecedent myeloid neoplasm, or therapy-related) were compared between the IDHm and IDHwt subgroups using standardized mean difference (SMD). An oncoprint visualized the cytomolecular landscape among subjects with complete profiling (karyotype and NGS). A Cox model controlling for age, AML predisposition, and multi-hit status assessed the independent effect of IDH co-mutation on OS. Among treated subjects with ≥1 available response evaluation, best response was compared using SMD, and time to best response was modeled with death as a competing risk (Fine-Gray subdistribution hazard model, same covariates).
Results: Of 234 TP53-AML subjects collated, 32 had co-mutated IDH1 or IDH2. Age and AML predisposition did not differ meaningfully between groups. Intensive induction was more common in the IDHm cohort (34% vs 13%, SMD=0.741). Complete karyotype and NGS were available for 208 (89%) subjects (29 IDHm, 179 IDHwt). Within this subcohort, the incidence of complex karyotype was high across both groups (85% IDHm, 95% IDHwt), though less common in the IDHm group (SMD=0.344). A similar pattern of TP53 multi-hit status was observed (72% vs 85%, SMD = 0.415). An oncoprint demonstrated enrichment of RAS-pathway/signaling mutations (NF1, NRAS, PTPN11, CBL) in the IDHwt cohort. TP53 variant location was reported for 211 (90%, 12 IDHm, 199 IDHwt) subjects. Most were missense mutations affecting the DNA-binding domain (74%), though the IDHwt group exhibited a more heterogeneous mutational repertoire. In a Cox model, IDH status was not associated with overall survival (HR 0.86, CI 0.56 - 1.30, p = 0.465); only age and AML predisposition retained significance. Response evaluation included 157 subjects (67%; 22 IDHm, 135 IDHwt): complete responses were less common in IDHm (21% vs 40%, SMD=0.727), and IDH status showed no independent association with time to best response.
Conclusion: In this multi-institutional TP53-AML cohort, IDH co-mutant status conferred no difference in OS or time to best response. Though meaningful differences in the incidence of complex karyotype, multi-hit status, and affected TP53 domain were noted, the cytomolecular landscape of the two subgroups was heavily enriched for high-risk features, possibly reflecting the subspecialized quaternary-care systems from which they were collated. The lack of an IDHm survival benefit in this cohort differs from prior reports from COMMAND and our own analysis of Flatiron Health. Acknowledging that definitions of “multi-hit” differ across analyses, IDHm cases in COMMAND and Flatiron were less likely to have multi-hit disease (21% and 33%) compared to this cohort (83%). Interpreted in context, our analysis suggests that IDH mutations do not confer therapeutic sensitivity or a more indolent disease course independent of their cytomolecular context, though we could not statistically verify this within this dataset. A larger analysis, including a wider range of TP53 allelic states amongst an IDH co-mutated cohort, is ongoing.
DISCLOSURES
Lukas Ronner, MD MSCR: Verona Pharma - Stock (publicly traded)