Ekaterina Gibiansky (1), Leonid Gibiansky (1), David Carlile (2), Candice Jamois (3), Vincent Buchheit (3), Nicolas Frey(3)
(1) QuantPharm LLC, North Potomac, MD, USA; (2) Roche Products Ltd, Welwyn, UK; (3) F. Hoffmann-La Roche Ltd, Basel, Switzerland
Objectives: Obinutuzumab (GA101) is a novel, humanized type II anti-CD20 monoclonal antibody (mAb) with a glycoengineered Fc region. The aim of the analysis was to establish a predictive population model that describes PK of GA101 following IV administration and to identify covariate factors that influence its disposition.
Methods: Serum concentrations (12,634) of 678 patients (50.4% with CLL) from 4 Phase I – III studies were analyzed in NONMEM. The full model approach was used for covariate model development. Diagnostics plots and various predictive check procedures were used for model evaluation.
Results: Consistent with other mAbs targeting B-cells, the two-compartment population PK model with time-dependent clearance (CL=CLinf+CLT·exp(-kdest)) described GA101 concentrations. CLT (0.23 L/day [95%CI=0.20-0.27]) was 2.8 times higher than CLinf (0.083 L/day [95%CI= 0.078-0.088]). Both values (shown above for 75 kg female with CLL) depended on diagnosis. They were 17% [95%CI=11-22%] lower for B-cell lymphomas and diffuse large B-cell lymphomas, and 75% [95%CI=25-144%] higher for Mantle cell lymphomas compared to CLL. For patients with CLL and baseline tumor size (BSIZ) > 1750 mm2, decline of time-dependent clearance (t1/2 = 19 days) led to steady-state after approximately 4 months for 1000 mg q4w dosing (with 2 additional doses at weeks 1 and 2 of cycle 1). Clearance declined faster (higher kdes) for patients with NHL (by 108% [95%CI=63-164%]) and patients with BSIZ< 1750 mm2 (by 165% [95%CI: 110-235%]). The results are consistent with target-mediated CL (with higher CL for higher tumor burden and higher CD-20 expression) that decreases with elimination of target cells.
CLinf and CLT were 22% (95%CI: 14-31%) and 49% (95%CI: 23-80%) higher in males, and increased with weight (with power coefficient (α) of 0.62 [95%CI=0.43-0.79]). Central (VC) and peripheral (VP) volumes (2.8 L [95%CI=2.7-2.8] and 1.0 L [95%CI=0.92-1.1]) were typical for mAbs; inter-compartment clearance (Q) was 1.3 L/day (95%CI=1.0-1.6). VC, VP, and Q increased with weight (with α = 0.38 [95%CI=0.29-0.47], 1, and 0.75, respectively); VC was also 18% (95%CI=14-22%) higher in males.
GA101 PK was independent of age, renal function or anti-drug antibodies (detected in 17 subjects).
Conclusions: In CLL patients, the expected differences (<30%) in steady-state exposure based on weight and gender do not warrant a dose modification for the proposed 1000 mg IV q4w dosing regimen.
Reference: PAGE 23 (2014) Abstr 3071 [www.page-meeting.org/?abstract=3071]
Poster: Drug/Disease modeling - Oncology