PII-057 - INTEGRATED PHARMACOKINETIC (PK)/PHARMACODYNAMIC MODELING OF PK, SAFETY, AND EFFICACY DATA TO SUPPORT DOSE OPTIMIZATION OF BMS-986158, A BROMODOMAIN AND EXTRA-TERMINAL INHIBITOR (BETI), IN PATIENTS (PTS) WITH MYELOFIBROSIS (MF).

S. Das¹, S. Du², M. Kozicki², H. Kandoussi³, O. Esposito⁴, S. Coker⁴, C. Tehlirian⁴, Z. Nikolova⁴, Y. Liu⁴, M. Lamba⁵, F. Wu⁶; ¹Bristol Meyers Squibb, Princeton, New Jersey, United States, ²Bristol Meyers Squibb, Lawrenceville, NJ, United States, ³Bristol Myers Squibb, Lawrenceville, NJ, United States, ⁴Bristol Meyers Squibb, Princeton, NJ, United States, ⁵Bristol Myers Squibb, Summit, NJ, United States, ⁶Bristol Meyers Squibb, Summit, NJ, United States.

Background: BMS-986158 is a potent, selective BETi being evaluated in phase 1/2 study in pts with DIPPS-Intermediate or High-Risk MF (NCT04817007). In MF, spleen volume reduction (SVR) is the primary clinical efficacy response marker. Reversible thrombocytopenia (TTP) is a primary on-target safety signal of BETi. Integrated population PK-platelet (PK-PLT) model and exposure-SVR (ER) models were developed with emerging data from the phase 1b/2 study to support recommended phase 2 doses of BMS-986158.

Methods: Three oral BMS-968158 doses (2mg n=5; 3mg n=2; 3.75mg n=4; 5 days on/2 days off) were evaluated in combination with ruxolitinib (RUX) 15mg BID in RUX naïve pts. BMS-986158 PK was well described by a preliminary 2-compartment Population PK model with first order absorption and linear elimination. Using BMS-986158 exposures, a preliminary semi-mechanistic PK-PLT model was developed to describe drug induced platelet reduction, and an ER model with exponentially decaying drug effects was developed to describe temporal profiles of SVR.

Results: PK results showed similar BMS-986158 exposures as shown from a monotherapy study (Hilton J et al. Cancers (Basel) 2022;14(17):4079), indicating no apparent DDI effects of RUX. The BMS-986158 PK data appeared dose proportional across the three doses. The PK-PLT model suggested the TTP rate and treatment-related PLT kinetics were associated with BMS-986158 doses and baseline PLT counts. Simulations from the PK-PLT model given 2mg (5 days on 2 days off) predicted a median nadir of 72 x 109 PLT/L from a baseline of 100 to < 200 x 109 PLT/L, ~2-fold lower than a median nadir of 135 x 109 PLT/L predicted from a baseline of >200 x 109 PLT/L ~21 days post first dose. At baseline PLT of 100-200 and ≥200 x 109 PLT/L, the 2 and 3mg doses were expected to maintain median PLT ≥ CTCAE Grade 3 TTP threshold of 50 x 109 PLT/L. The ER model predicted mean %SVR from baseline of -43.0% and -48.8% with 2 and 3mg doses, respectively, at week 12, and of -52.6% and -55.7%, respectively, at week 24 suggesting improvements in SVR over time.

Conclusion: Similar BMS-986158 PK was observed between monotherapy and with RUX. The PK-PLT and ER models suggested median PLT ≥ Grade 3 TTP threshold at both 2 and 3mg doses, albeit modest differentiation between the two doses over time. BMS-986158 doses of 2 and 3mg given at 5 days on 2 days off were recommended for further exploration.