Population Pharmacokinetic (PK)/Pharmacodynamic (PD) Modeling of Myelosuppression in Patients with Hematologic Malignancies for CPX-351 and Standard-of-Care 7+3 Therapy

CPX-351 (Vyxeos^®), a liposomal encapsulation of cytarabine and daunorubicin at a synergistic ratio, has demonstrated a significant survival benefit vs standard 7+3 in patients (pts) with high-risk/secondary AML. A population PK/PD analysis assessed the correlation between cytarabine and daunorubicin plasma concentrations and myelosuppressive effects (neutropenia, thrombocytopenia) of CPX-351 and 7+3.

The PK/PD population for model development included pts with advanced hematologic malignancies from 3 clinical studies. For CPX-351 and 7+3, respectively, 129 and 79 pts were included in the final neutropenia PK/PD analysis and 137 and 86 pts were included in the final thrombocytopenia PK/PD analysis. For the neutropenia model, median age and body weight were 67 y (range: 23-81) and 78.7 kg (39.5-156.5) for CPX-351 and 68 y (60-75) and 83.0 kg (53.9-136.0) for 7+3.

PK/PD analyses were conducted using nonlinear mixed-effects modeling in NONMEM. Pt-specific PK profiles were simulated using previously developed population PK models for CPX-351 and 7+3. Blood cell dynamics were described by transit-compartment models with proliferating, maturating, and circulating neutrophils or platelets. The effects of CPX-351 or 7+3 were applied to the proliferation phases of the compartment models by a molar composite PK driver (plasma cytarabine + daunorubicin). Inhibition of proliferation of blood cells by CPX-351 and 7+3 is assumed to be similar, via a sigmoidal Imax function. Co-medication of granulocyte colony stimulating factor (GCSF) or platelet infusion was accounted for during model development. Covariates (eg, demographics, clinical laboratory measures, disease status) were evaluated. Model evaluation and selection were assessed using a standard model discrimination process that included statistical criteria (eg, objective function value) and graphical representations of goodness-of-fit.

In the final neutrophil PK/PD models, baseline circulating neutrophil counts were similar for CPX-351 (3.55 × 10⁹/L) and 7+3 (3.76 × 10⁹/L). Mean transit times (MTT) between maturation compartments were estimated at values of 113 h for CPX-351 and 88 h for 7+3. Effects of GCSF on neutrophil production were assumed to be similar for CPX-351 and 7+3. Both treatments had similar maximum inhibition on neutrophil proliferation, with Imax values around 1. However, estimated IC₅₀ values were very different: 24.9 µM for CPX-351 and 0.0286 µM for 7+3.

In the final platelet PK/PD models, baseline circulating platelet counts were the same (98.1 × 10⁹/L) for both CPX-351 and 7+3. The MTTs between each compartment of the maturation processes were 91.2 h for CPX-351 and 120 h for 7+3. Drug-specific parameters for CPX-351 and 7+3, respectively, were as follows: Imax, 0.316 and 1; IC₅₀, 0.324 and 0.0982 µM.

To better understand the behavior of the models and parameter estimates, simulations were conducted to evaluate the temporal events of myelosuppression. Model simulations were conducted for 200 pts with characteristics similar to the PK/PD model population. During simulations, no platelet transfusion or GCSF was administered. Pts received CPX-351 100 units/m² (cytarabine 100 mg/m² + daunorubicin 44 mg/m²) as a 90-min IV infusion on Days 1, 3 and 5 or 7+3 (cytarabine 100 mg/m²/day IV for 7 days continuously + daunorubicin 60 mg/m² IV on Days 1-3).

Median time to initially observe a blood neutrophil count <0.5 × 10⁹/L was longer following CPX-351 (8.3 d) vs 7+3 (7.4 d) treatment. The median duration with neutrophil counts <0.5 × 10⁹/L was longer with CPX-351 (23 d) vs 7+3 (14 d). The median lowest neutrophil counts were well below 0.2 × 10⁹/L for both CPX-351 (0.007 × 10⁹/L) and 7+3 (0.026 × 10⁹/L).

Median time to initially observe a platelet count <50 × 10⁹/L was 6.4 d after CPX-351 and 5.8 d after 7+3, while the median time to an observed platelet count <20 × 10⁹/L was 10.8 d and 8.9 d, respectively. The median duration with platelet counts <20 × 10⁹/L was longer with CPX-351 (18 d) vs 7+3 (8 d), and the median duration of platelet counts <50 × 10⁹/L was 22 d and 15 d, respectively. The median lowest platelet counts were 11.3 × 10⁹/L with CPX-351 and 4.7 × 10⁹/L with 7+3.

In summary, the median duration of myelosuppressive effects was longer with CPX-351 than 7+3, and the median time for initial detection of myelosuppression with CPX-351 was 1 to 2 days later than with 7+3, which might affect the clinical monitoring scheme.