High Deliverability of a Midostaurin Triplet Regimen Incorporating Venetoclax and Low Dose Cytarabine in Non-Adverse Cytogenetic Risk Acute Myeloid Leukaemia: A Sub-Analysis of the Australasian Leukaemia Lymphoma Group (ALLG) Intervene Study

Abstract

Background Adaptive resistance mechanisms leading to treatment failure in older patients (pts) receiving frontline venetoclax (VEN) combination therapies include emergence/expansion of kinase activating mutations (principally, FLT3-ITD) in pts with non-adverse (NON-ADV) karyotype AML. In both the RATIFY and QuANTUM-First trials, the FLT3 inhibitor was scheduled sequentially after chemotherapy, with no evidence that this approach augmented marrow toxicity. We conducted the Phase 1b/2 study, INTERVENE, to explore the deliverability and preliminary efficacy of a novel triplet regimen incorporating the FLT3/multi-kinase inhibitor midostaurin (MIDO) delivered sequentially after low dose cytarabine (LDC), in tandem combination with VEN. We hypothesized that this triplet schedule would suppress emergence of kinase-activating mutations on VEN-based therapy. MethodsEligibility: Pts with treatment naïve AML (excluding APL), aged ≥60 years and unfit for intensive chemotherapy. Prior hypomethylating agents for antecedent myeloid neoplasms were permitted. Pts with NON-ADV karyotype were enrolled to receive VEN-LDC-MIDO. Treatment (safety run-in): LDC 20 mg/m2 SC D1-10, MIDO 50 mg BD commenced on D11 and continued to D28. After a dose ramp-up in cycle 1 (C1), VEN D1-28 was administered at two dose levels (1= 400 mg, 2= 600 mg). Azole antifungals were prohibited during C1 but allowed from C2 with VEN dose reduced to 100 mg. Molecular studies: NGS using a custom 48-gene Roche KAPA HyperCap panel (sensitivity 1%) and FLT3-ITD PCR-NGS (sensitivity 10-4-10-5) were performed on serial bone marrow samples. RT-qPCR was used to detect NPM1 measurable residual disease (MRD) (sensitivity 10-6). Results First pt enrolled: 7SEP2020. Data cut-off: 31MAY2022. Eighteen pts were enrolled to receive VEN-LDC-MIDO in the run-in phase. Median age was 77 years (range 73-87, 84% age≥75). 44% had secondary AML, including 11% t-AML. 2 pts withdrew consent within the first 7 days due to non-therapy related reasons and were not response evaluable. 30-day mortality was 0%. There were no dose limiting toxicities observed. With a median follow up of 18 months (m), median overall survival (OS) was not reached (95% CI 8.02-NR). Overall response rate (CR+CRi+CRh) by intention-to-treat was 77.8%, with 44.4% CR. Median relapse free survival was 11.7m. To assess treatment deliverability, we evaluated the inter-cycle time for the first 12 cycles among pts achieving CR/CRi/CRh (Fig 1A). The median inter-cycle times ranged from 28-35 days, with majority commencing the next cycle in <42 days. There was no evidence of cumulative lengthening of inter-cycle time with successive treatment cycles, suggesting that MIDO did not promote marrow toxicity when combined with VEN-LDC. Pts received a median number of 9 cycles (range 1-16), with 7 completing ≥12 cycles (Fig 1B). Of these, 3 electively ceased therapy (at C12, C14, C16), with a treatment-free remission (TFR) lasting 3.3-5.7m, from which one relapsed. For pts opting for therapy continuation (TC) >12m (4/7), one has relapsed. Of the remaining 5 (2 TFR, 3 TC) pts with durable remissions, 4 had either NPM1 or IDH2 mutations at enrolment and all achieved undetectable MRD in CR. One additional patient with NPM1 mutationat diagnosis relapsed after C5 with NPM1 negative AML. Molecular data and treatment outcomes are shown in Fig 1B. Two of 16 response-evaluable pts had FLT3 mutations at baseline (1=FLT3-ITD, 1=TKD). Using PCR-NGS to enhance detection of FLT3-ITD microclones, an additional 3 FLT3-ITDs were identified in 2 pts. During VEN-LDC-MIDO, all 4 FLT3-ITDs were suppressed to undetectable levels by PCR-NGS. No pts relapsed with FLT3 variants at the time of progression. Other kinase variants extinguished by VEN-LDC-MIDO include NRAS G12D and KRAS G12D. In contrast, 4 pathways of adaptive resistance were observed at time of treatment failure: emergence of 1) TP53 hotspot variants (N=2), 2) PPM1D frameshift variants (N=2), 3) BAX inactivating variant (N=1), and 4) evolution of KRAS Q61 mutations (N=2), Conclusion After a median follow up of 18m, we report that VEN-LDC-MIDO, with MIDO delivered sequentially after LDC, is tolerable with excellent treatment deliverability. Preliminary efficacy of this triplet is promising, with molecular evidence of FLT3-ITD suppression in all cases so far. A randomized phase 2 comparison of VEN-LDC-MIDO versus VEN-LDC is ongoing. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal