Characterization of Comparative Clinical Outcomes and Molecular Features of Black Patients with Acute Promyelocytic Leukemia (APL) Reveals Similar Survival and Distinct Genomic Profiles

Introduction: Although APL, a subtype of acute myeloid leukemia (AML), is an aggressive disease, its treatment was revolutionized with combination all-trans retinoic acid (ATRA) and arsenic trioxide (ATO), making APL a highly curable malignancy. Molecularly, APL is characterized by t(15;17) resulting in PML::RARA gene fusion as disease-initiating event and by several recurrent gene mutations (eg, FLT3) identified in previous studies. In AML, prior studies have established inferior overall survival (OS) of Black versus White AML patients (pts) in both population-based analyses and clinical trials, found a negative prognostic impact of higher social deprivation (SDI), and characterized differences in frequencies and impact of disease-associated gene mutations. Herein, we compared OS of Black and White APL pts and evaluated genomic landscape of Black pts, which has not been comprehensively assessed to date.

Methods: For nationwide population analysis, 2 sets of cancer registries were included in the Surveillance Epidemiology End Results (SEER) Program of the National Cancer Institute to identify 829 adults diagnosed with de novo APL (confirmed by cytogenetic and/or molecular findings) in 1995-2019. To assess survival in the setting of clinical trials, we analyzed 249 pts treated on historic frontline Cancer and Leukemia Group B (CALGB)/Alliance for Clinical Trials in Oncology (Alliance) trials with or without inclusion of ATRA (CALGB 9191, 1992-1995; CALGB 9710, 1999-2005). For Alliance pts, neighborhood SDI was assigned based on pt-reported residence zip code and classified as low (1-25, n=33) or high deprivation (26-100, n=95). OS was analyzed separately for pts diagnosed before and after 2012 to account for treatment advances with ATRA/ATO. To assess molecular features, we performed integrated genomic profiling (paired tumor/normal whole-exome and transcriptome sequencing) on 31 Black APL Alliance pts (9/31 completed, with complete results to be presented at the Annual Meeting).

Results: On a population level using SEER data, 3-year (3y) OS significantly improved from 65% for cases diagnosed in 1995-2012 to 75% in those diagnosed after 2012 (p=<0.001), concordant with ATRA/ATO use. Notably, OS of Black and White pts did not differ significantly before or after 2012. This held true in race-specific findings for median household income, Yost index and urban/rural status; none of which significantly associated with OS. Older age was the only parameter associated with OS (3y rates, 1995-2012: <40y, 79%; 40-59y, 71%; ≥60y, 44%; and 2012-2019: <40y, 92%; 40-59y, 80%; ≥60y, 56%). Again, OS of Black and White pts did not differ within age groups.

Likewise, OS of Black and White APL pts treated in a relatively controlled setting of clinical trials was not significantly different. As in population-based findings, OS improved over time and with ATRA addition (3y OS rates, 1992-1995, 66%; 1999-2005, 85%), and older age was a strong negative OS predictor (p<0.001). When comparing OS of pts treated on clinical trials with population-based OS, trial pts had a longer 3y OS (84% vs 64%, p<0.001). Notably, pts with SDI >25 (n=95) had a worse OS than pts with SDI <25 (n=33, p=0.03).

On a molecular level, preliminary results of whole-exome sequencing (finalized for 9/31 Black pts) showed frequent mutations in FLT3 (5/9 pts), a known APL-associated event, but also revealed several recurrent gene mutations not reported in prior APL sequencing efforts, which were almost exclusively based on White pts, including mutations in DPP6, GOLA6L, NLGN2 and SP8 (found in ≥2 pts). Intriguingly, CALR mutations were found in 3/9 pts. Mutations in FLT3 and CALR were mutually exclusive and found in 8/9 pts, with the remaining pt harboring a SORBS2 mutation.

Conclusions: In contrast to most cancer types, including AML, there was no OS disparity with respect to race in APL pts, both in population-based analyses and in the setting of historic clinical trials. Older age and residence with high SDI (>25) were negative OS predictors. Molecularly, our pilot data reveal existence of ancestry-associated differences in driver mutations, suggesting that FLT3 and CALR mutations represent necessary proliferative signals in addition to PML::RARA-driven differentiation arrest.

Support: U10CA180821, U10CA180882, U24CA196171; Clinicaltrials.gov IDs: NCT00048958, NCT00899223, NCT00900224; https://acknowledgments.alliancefound.org