Long-Term Follow-up Results: A Phase 2 Trial of Imatinib Mesylate As Maintenance Therapy for Patients with Newly Diagnosed c-Kit Positive Acute Myeloid Leukemia (AML)

The c-kit (CD117) receptor is expressed on > 10% blasts in 64% of de novo AMLs and mediates proliferation and anti-apoptotic effects.  High c-kit levels [defined as mean fluorescent intensity (MFI) > 20] correlate with a shorter time to relapse and decreased overall survival (OS).  Imatinib mesylate (IM), a c-kit inhibitor, has activity against relapsed/ refractory AML.  The primary objective of this study was to determine whether adding maintenance IM for 1 yr after completion of standard induction (IT) and post-remission therapy (PRT) in pts with newly diagnosed c-kit + AML improves progression-free survival (PFS) compared to historical controls.  We previously presented our toxicity and correlative data at ASH 2012 (Abstract 3597).  Here, we present our long term follow-up results.  

Methods:   Pts were treated at Cleveland Clinic, Duke, Roswell Park, and University Hospitals of Cleveland from 2008 to 2012.  IM was supplied by Novartis.  Eligibility criteria: pts age ≥ 18 yrs, AML in first complete remission (CR1), ≥ 20% c-kit+ blasts at diagnosis (dx), ECOG performance status 0-2.  Cytogenetics (CG) were classified per CALGB 8461.  Pts must have received IT (7+3 [continuous infusion cytarabine and an anthracycline] or ADE [cytarabine, daunorubicin, etoposide]) and PRT (≥ 1 course for pts ≥ 60 yrs; ≥ 2 courses for pts < 60 yrs).  CR was confirmed by bone marrow analysis prior to study enrollment.  MDR expression was analyzed by IHC on diagnostic samples (n=19); AF1q gene expression was analyzed by RT-PCR on RNA from available diagnostic pt samples (n=9).  C-kit MFI was calculated as the mean channel number (MCN) of the blasts/ MCN auto fluorescence using a CD45/ orthogonal light scatter gate to isolate blasts and lymphocytes.  All pts received IM 600 mg/day for 12 months (mos) unless they experienced toxicity or disease progression.  Dose modifications were made for Grade 2-4 non-hematologic toxicity and Grades 3-4 neutropenia and thrombocytopenia.  PFS was measured from the CR date to the time of relapse or death.  Primary endpoints:  Based on historical data from the Cleveland Clinic and SWOG, the median PFS for all AML pts undergoing IT < 60 yrs of age is 13 mos and for pts ≥ 60 yrs of age is 8 mos.   The goal of this study was to see a 30% improvement in PFS at these time points in the respective age groups (i.e. 65% PFS at 13 mos for  pts < 60 yrs; 65% PFS at 8 mos for pts ≥ 60 yrs).  

Results:  Of 32 pts enrolled, the median age was 54 yrs (range 19-81), median WBC at dx 22.13 K/ uL (1.55-98.44), median peripheral blood blasts at dx 23.6% (range 0-85), and 44% were male.   CG risk included:  16% (5) good, 66% (21) intermediate, 16% (5) poor, 3% (1) miscellaneous.  Of the pts with normal CG, 10 were NPM1+, FLT3 ITD negative; and 1 pt was FLT3 ITD+.  The median c-kit+ blast % was 79.9, and median c-kit MFI 39.8 (range 6.5-120.1).  Median AF1q expression was 9.59 (range 1.83-161.85) (> 9 is considered high and is associated with a poor prognosis; high AF1q is also associated with high c-kit expression).  Eight-four percent of pts had moderate or high levels of drug resistance factors (GST1, MDR1, LRP1, and/or MRP1); almost half (47%) had high expression.  There was no correlation between MDR and c-kit MFI.  Pts received IM for a median of 4.0 mos (range 0.1-12.2) and the median daily dose was 600 mg.  Twelve pts (38%) were dose reduced to 400 mg.  Forty-five percent of pts experienced Grade 3 reactions possibly related to treatment, with the majority (31%) being myelosuppression.  With a median follow-up time of 56.3 mos, the estimated median OS was 51.3 mos and estimated median relapse-free survival (RFS) 18.9 mos.  The estimated PFS at 13 mos for pts < 60 yrs of age was 71 ± 10% (p=0.017, compared to the null hypothesis); and the estimated PFS at 8 mos for pts ≥ 60 yrs of age was 64 ± 15% (p=0.166, compared to the null hypothesis).   Predictors of worse RFS included: age, WBC at dx, % peripheral blasts at dx, CG risk, and MDR expression.  C-kit MFI and Af1q were not associated with RFS or OS.   

Conclusions:  Use of IM maintenance therapy appeared to be associated with improved PFS compared to historical controls in pts < 60 yrs of age.  In addition to a high c-kit MFI, these pts had other adverse characteristics (moderate to high levels of MDR. high AF1q).  Though previous studies have demonstrated that c-kit MFI > 20.3 was an independent adverse prognostic factor for RFS and OS (median RFS 10.7 months) in AML, use of IM maintenance therapy in this study appeared to mitigate this, supporting further investigation.