Garrett-Bakelman, Francine E.

Abstract Copy Number Alterations (CNAs), including chromosomal gains and losses, shape the genetics and clinical management of acute myeloid leukemia (AML), particularly in complex karyotypes (e.g. CK-AML). Traditionally, AML associated CNAs have been studied using cytogenetic approaches, i.e., karyotyping, that are limited in resolution and sensitivity. Here, low-coverage, sparse whole genome sequencing followed by machine learning applied to over 600 AML patient samples (ECOG-ARCIN trials: E1900, E3999) reveals novel genetic, biologic, and clinical correlates in AML disease. We find that multi-variate Cox-regression analysis applied to the data identifies select patients exhibiting focal and cryptic hemizygous deletions at 17q, 21q, and 3p, in otherwise diploid genomes, associate with inferior survival compared to other patients (median overall survival; OS = 65.5 compared to 421 days, Chi-square test pval=6x10-11), independent of factors such as age, treatment, and karyotype complexity status. In addition, we confirm that amplifications at 11q23, encoding the KM2TA gene, are associated with inferior outcome, and further identify unique breakpoint event features as well as transcriptional signatures suggestive of distinctive KM2TA-translocation independent biology. Applying machine learning approaches to the data, more specifically Non-negative Matrix Factorization (NMF), we develop a prognostic model that captures clinically adverse, cryptic variation such as focal 5q deletions, missed by conventional chromosome analysis and thereby, adds novel, complementary information to traditional, cytogenetic based risk-stratification (e.g. ELN). The model also associates gain of chromosome 8 in Intermediate risk patients with worse outcome compared to other Intermediate risk patients (median OS = 186 compared to 446 days, Chi-square test pval=0.005) and identifies a subgroup of patients within the adverse risk category that lack deletions on chromosomes 5, 7, and 17, and that fare better than the rest of the adverse risk patients (median OS = 188 compared to 97.5 days, Chi-square test pval=0.02), likely reflecting previously identified “atypical” CK-AML patients. Last, we develop a prognostic classifier based on a combination of Neural-Network and Random-Forest based learning, that predicts disease outcome with high accuracy and is based on a three genomic feature list on chromosome bands 5q31.2, 7q31.33, and 8q24, suggesting potential for clinical implementation. Altogether, our results show that low coverage, sparse whole genome sequencing combined with advanced learning methods can be applied clinically to support the diagnosis and clinical management of AML disease as well as serve as a powerful tool to further a basic understanding of the genetics and biology of CNAs in myeloid neoplasms and other cancers.

Abstract Background We performed a prospective study evaluating a baseline GA battery and clinical outcomes in fit older adults (age ≥60 years) with AML and normal cardiac and renal function in the recent E-A NCTN E2906 phase 3 study. A patient-reported outcomes (PRO) assessment of Depression, Activities of Daily Living (ADL), Cognitive Function, Social Support (SS) domains, and comorbidity was performed at study registration to evaluate the prevalence of vulnerabilities and their impact on overall survival (OS) and 30-day induction mortality rates (IM) in this cohort selected for fitness to receive IC. Methods E2906 study design and results have been presented previously [n=727, age ≥60 years, randomized 1:1 to ‘standard 7&3‘ & high dose cytarabine (Arm A) vs. single agent clofarabine (CLO, Arm B), as remission induction (Step 1) and consolidation (Step 2)]. There was no difference in composite complete remission (CCR, 50%) or IM (8.5%), and CLO was inferior for OS. A baseline GA was a key secondary protocol objective, offered to all patients, and n=532 (73%) participated. We assessed Geriatric Depression Scale (GDS, none vs. any depressive symptoms, n=524); cognitive screen using Mini-Mental Exam Section (MMES, score 8/8 vs. <8, n=516); ADL & Instrumental ADL (iADL), (need for assistances vs. not n=526); a Medical Outcomes Study SS Survey (MOS-SS, measured as per unit increase, n=527); a Comorbidity survey ( ≤4 vs. ≥5 patient-reported comorbidities, n=530), patient-reported prescription medications use (PM, 0-3 vs. ≥4, n=519); and a survey of Sexual Desire (n=491) and Sexual Activity (n=477) within 4 weeks of AML diagnosis. Association of GA domains with OS, IM, and achievement of CCR was evaluated using the Chi-squared, Fisher's exact, and Wilcoxon rank sum test. Survival analysis was performed using the Cox proportional hazards models, adjusted for treatment arm. All p-values are 2-sided. Results The median age among those participating in GA battery was 67 yrs (range 60-85), including 38% age ≥70. There were 42% females, and ECOG performance status was 0 (28%), 1 (54%), 2 (16%), or 3 (2%). Central Cytogenetic review was classified (ELN2017) as Favorable (3%), Intermediate (66%), and Adverse (31%) risk. There was a notable rate of GA scores outside ‘normal’ clinical cutoff, including Depression (GDS 31%), cognitive and functional impairment [MMSE (37%), ADL (25%) and iADL (32%)], ≥5 Comorbidities (23%), and ≥4 PM's (44%). 81% & 76%, respectively, reported no sexual activity and low/none sexual desire. MOS-SS tended to be high in this cohort [mean (SD) score 87 (17)]. We observed a significant association of GA domains with OS & IM. For OS, absence of depressive symptoms [GDS, Hazard Ratio (HR) 0.79, 95% confidence Intervals (CI) 0.64-0.97, p=0.025)], MMES 8/8 (HR 0.78, 95%CI 0.64-0.96, p=0.016), requiring no assistance with ADLs (HR 0.65, 95%CI 0.52-0.82, p<0.001) and iADLs (HR 0.69, 95%CI 0.57-0.85, p<0.001), and ≤4 Comorbidities (HR 0.68, 95%CI 0.54-0.84, p=0.001) were significant for superior OS; use of ≥4 PM's was associated with inferior OS (HR 1.32, 95%CI 1.09-1.61, p=0.005). There was no association of sexual activity/desire, BMI ≥30 kg/m2, or MOS-SS with OS. IM was significantly higher among those with any depression symptoms (GDS, p=0.005), any ADL (p=0.010) or iADL (p=0.013) impairment, ≥5 comorbidities (p=0.037) and ≥4 PM's (p=0.001); there was no association with MOS-SS or MMES. There was no significant association of GA scores with cytogenetic risk group or with CCR rates. Only ADL independence (p=0.048), and possibly Comorbidity ≤4 (p=0.071), were associated with proceeding to Step 2. Conclusions In this large prospective cohort of fit older adults with AML selected to receive IC, we observed high rates of PRO-assessed GA vulnerability, which was significantly associated with inferior OS and higher IM rates, and (for ADL impairment) lower rates of proceeding to consolidation therapy. This suggests that functional and cognitive impairments are prevalent and can significantly impact outcome even in fit older adults. Despite potential long-term OS advantages of IC previously observed in E2906, efforts to systematically recognize those at higher risk based on GA PROs are needed to help identify patients who may be more appropriate for study of lower intensity treatment options to mitigate impact on IM and OS. SImilarly we plan to evaluate the potential benefit of GA-driven supportive care to improve patients' outcomes.

Abstract Background: Population registries and the Alliance co-operative group database have highlighted racial disparities in AML survival attributed to differential prevalence and impact of oncogenic drivers on disease biology. However, the small number of Black patients with well-annotated genomic profiles limits the validity of these findings. The current study leverages ECOG-ACRIN cytogenetic and outcomes data from newly diagnosed adult AML patients treated on intensive chemotherapy-based clinical trials over three decades. Methods: Self-reported race was categorized as White, Black or Other. Descriptive statistics were used to summarize baseline characteristics using the Wilcoxon rank-sum and Fisher's exact tests. Multivariable Cox proportional hazards regression models evaluated the association between race and survival outcomes adjusting for age, hemoglobin, platelet count, white blood cell count, blast percentage, performance status and ELN cytogenetic risk score (RS). All Cox models were stratified by protocol. Subgroup analyses of overall survival (OS) and disease-free survival (DFS) were performed using Kaplan-Meier curves stratified by race using log-rank tests. All statistical tests were two-sided. Analyses included only patients with complete data for the variables of interest. Results: Data was collected across 10 phase II and III interventional trials for newly diagnosed adult AML that accrued patients between 1984 and 2019. The study population included 3469 White, 184 Black, and 156 Other patients with a Black representation of 1.1-11% across trials. Blacks were significantly younger at diagnosis with a median age of 47.9 years compared to 53.5 years in Whites (p<0.001). Racial groups were well matched for gender and performance status. Black race was an independent prognostic factor for inferior OS (HR 1.212, CI 1.01 – 1.453; p = 0.0383) and DFS (HR 1.313, CI 1.05-1.641; p = 0.017). Complete cytogenetic results were available for 117 Black and 2162 White patients and revealed no significant racial differences in the distribution of ELN 2017 cytogenetic RS, although core binding factor AML was non-significantly enriched in Black patients (12.8% vs 8.3%). No difference was observed between Blacks and Whites in the prevalence of complex karyotype (CK), CK with -5/del(5q), -7/del(7q), or -17/del(17p), and AML-MRC including +8 and del(20q). Notably, the ELN 2017 cytogenetic RS effectively predicted OS in Black (57.8 vs 14.2 vs 5.8 months in favorable, intermediate and adverse RS, p = 0.018), and White patients (60.1 vs 15.2 vs 7.1 months, p < 0.0004 for Whites) but only predicted DFS for White AML patients (p < 0.0001 for Whites, p=0.48 for Blacks). In patients with molecular data for FLT3-ITD (n=1218), CEBPA (n=576), TP53 (n=572), or NPM1(n=683), no racial difference in the prevalence of these mutations was observed. However, NPM1 mutations were associated with inferior OS in Blacks compared to Whites (19.1 vs 8.9 months; p = 0.0095). Black and White AML patients had similar responses with complete remission rates of 58.2% vs 57.5%, p=0.55, and treatment tolerability with 30-day mortality of 7.1% vs 8.9%, p=0.50. However, consolidation approaches differed significantly. Among patients that underwent stem cell transplant (SCT), a significantly higher percentage of White patients received allogenic SCT compared to Blacks (48.5% versus 37.1%, p <0.01). Conclusions: This study represents one of the largest race-based comparisons of cytogenetic abnormalities and clinical outcomes in prospective trials of intensive chemotherapy in AML . Our data confirms the independent prognostic relevance of race in AML survival and validates the predictive ability of the ELN 2017 cytogenetic RS for OS across racial groups. There is ancestry-based attenuation of the survival benefit of NPM1 mutations as previously reported, underscoring the need to tailor the current treatment approach to NPM1 mutated AML based on race. The availability of NGS data limited the genomic analysis to a subset of patients. Future studies should examine the depth of treatment response incorporating MRD biomarkers and explore the effect of race on co-occurring mutations and the leukemic stem cell transcriptome to better explain the observed disparities. Finally, persistent disparities in access to allo-SCT even within the context of clinical trials reinforces the need to address diversity in the stem cell donor pool and social barriers to transplant.