HARMONY: An Open-Label, Multicenter, 2-Arm, Dose-Finding, Phase 1b Study of the Combination of Ruxolitinib and Buparlisib (BKM120) in Patients with Myelofibrosis (MF)

BACKGROUND: Ruxolitinib (RUX) is a potent JAK1/JAK2 inhibitor that has demonstrated improvements in splenomegaly, MF-related symptoms, and quality of life measures and prolonged survival in 2 phase 3 studies. JAK signaling can activate the PI3K/mTOR pathway, and constitutive PI3K activation has been implicated in numerous cancers. Preclinical and early clinical results indicate benefit with inhibition of PI3K in MF. Buparlisib is a pan-PI3K inhibitor with specific and potent activity against class I PI3Ks. The aims of HARMONY are to evaluate the safety of co-administration of RUX and buparlisib in patients (pts) with MF and determine a recommended phase 2 dose (RP2D).

METHODS: This is a dose-finding, phase 1b study (NCT01730248) of RUX plus buparlisib in intermediate- or high-risk MF pts. Pts must have palpable splenomegaly ≥ 5 cm below the costal margin and MF symptoms. There are 2 treatment groups (JAK inhibitor–naive pts [arm A] and pts not benefiting from prior JAK inhibition [arm B]) and 2 study phases (dose escalation and expansion). Dose escalation is guided by a Bayesian logistic regression model with overdose control and depends on dose-limiting toxicities in cycle 1 and other safety findings. Each dose level consists of ≥ 3 evaluable pts; the 5 dose levels are (RUX [mg bid]/buparlisib [mg qd]) 10/60, 15/60, 15/80, 20/80, 20/100. 9 evaluable pts are required at the final dose level to determine an RP2D and proceed to the expansion phase; 10 additional pts per arm will be treated at the RP2D. PK of RUX was assessed by LC-MS/MS on day 1 after the first dose of RUX and on days 2 and 8 when administered in combination with buparlisib.

RESULTS: To date, 33 pts (arm A, n = 20; arm B, n = 13) have been treated in the dose-escalation phase. Baseline characteristics were (arm A and arm B) median age, 64 y (range, 48-79 y; 45% ≥ 65 y) and 61 y (53-74 y; 31% ≥ 65 y); male, 50% and 62%. RP2D was selected based on tolerability in the investigated dose levels, with additional confirmation from dose-selection recommendations derived through a Bayesian model.Currently, RUX 15 mg bid/buparlisib 60 mg qd was determined as the RP2D in both arms; the study is proceeding into the expansion phase.

At data cutoff (April 1, 2014), 15 (75%) and 8 (62%) pts were ongoing in arm A and arm B with a median exposure of 11.3 wk (range, 2.3-48.4 wk) and 19.0 wk (range, 0.6-50.9 wk), respectively. The primary reasons for discontinuation included lack of efficacy (10%) in arm A and AEs (15%) in arm B; all other reasons were in 1 pt each. There was 1 death (arm A) from duodenal ulcers/perforation, assessed as not related to study drugs by the investigator.

Despite the short follow-up, a ≥ 50% reduction from baseline in palpable spleen length was achieved by 70% and 54% of pts in arms A and B; 7 pts had a resolution of splenomegaly (6 in arm A; 1 in arm B). All pts with prior JAK inhibitor treatment, including 5 who had no previous spleen length reductions, had reductions with the combination.

Hematologic AEs included anemia (all grade [grade 3/4]: arm A, 30% [15%]; arm B, 31% [23%]) and thrombocytopenia (40% [10%]; 62% [39%]). Nonhematologic AEs were primarily grade 1/2 (Table). Overall, 20% and 23% of pts experienced serious AEs; these occurred in 1 pt each, with the exception of pyrexia (n = 2).

PK parameters of RUX were comparable when administered alone or in combination with buparlisib, suggesting a lack of impact of buparlisib on RUX PK.

CONCLUSIONS: The combination of RUX and buparlisib was well tolerated, with early signs indicating promising efficacy even in pts who have previously failed JAKi; preliminary RP2D was assessed as RUX 15 mg bid/buparlisib 60 mg qd for both arms. Updated efficacy and safety for the dose-escalation phase will be presented (July 1, 2014 data cutoff).