Rituximab PK and PD evaluation based on a study in diffuse large B-cell lymphoma: influence of tumor size on PK and assessment of PK similarity
Robin J. Svensson (1), Qing Xi Ooi (1), Lena E. Friberg (1,2), Luis Lopez Lazaro (3)*, Emma Hansson (1)* (*Shared authorship: these authors equally contributed to the work)
(1) Pharmetheus AB, Uppsala, Sweden (2) Department of Pharmacy, Uppsala University, Uppsala, Sweden (3) Dr. Reddy’s Laboratories SA, Basel, Switzerland
Objectives: Addition of rituximab to the standard-of-care chemotherapy cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) leads to significant improvement in the long-term outcomes of patients suffering from diffuse large B-cell lymphoma (DLBCL). DRL-rituximab (DRL_RI, Dr Reddy’s Laboratories SA, Basel, Switzerland) is under development as a rituximab biosimilar. In study RI-01-002 (CTRI/2012/11/003129), 76 patients with DLBCL were randomized to DRL_RI and 75 to the reference medicinal product MabThera® (RMP, Roche, Grenzach-Wyhlen, Germany). In addition to CHOP, all patients received rituximab at the approved dosing regimen. PK equivalence was demonstrated and comparable PD, efficacy, safety, and immunogenicity profiles were shown. In this analysis, the population PK and PKPD of rituximab was characterized; factors influencing the PK similarity assessment between products were explored; and impact of tumor size on rituximab PK was investigated.
Methods: Nonlinear mixed-effects models for PK, tumor size, tumor size-PK, and tumor response were developed using NONMEM version 7.3. The analyses used 3914 PK observations from 151 subjects, 1010 tumor size observations from 141 subjects, and 1003 tumor response observations from 144 subjects, all from study RI-01-002.
The intermediate PK model was based on a previously developed model [1]; tumor size, assessed as sum of products of diameters, was described using the Stein 2008 model [2]; tumor response, assessed using the Cheson 2007 response criteria [3], was characterized using a continuous time Markov model (CTMM). Subsequently, the tumor size-PK model was developed starting from the intermediate PK model, integrating the tumor size predictions obtained from the final tumor size model. The drug product effect was tested using the stepwise covariate model building procedure (SCM) after testing of exploratory covariates in the preceding SCM. Finally, simulations were performed to evaluate the impact of tumor size on rituximab exposure.
Results: Subjects’ demographics were largely comparable between treatment arms, except that baseline tumor size was slightly higher in the DRL-RI arm.
The intermediate PK model was a 2-compartment, target-mediated drug disposition (TMDD) model with quasi-equilibrium assumption. RMP-treated patients were predicted to have a 6.75% higher relative bioavailability. However, after tumor size was included as a covariate on clearance, the drug product effect was no longer observed. Further simulations showed that higher baseline tumor sizes lead to moderate reductions in rituximab exposure.
The final tumor size model included a constant tumor kill rate and a mixture model accounting for 2 sub-populations: one with tumor regrowth (12.9%) and one without. The final tumor response model was a CTMM with 5 states for the 4 tumor response categories and with the dropout state being an absorbing state. Transitions to partial and complete response states were faster during the first 20 weeks of treatment. The model also included a mixture model describing a sub-population of non-responders to treatment (42.4%) with faster transitions to a worse state. No rituximab exposure-response relationship was identified in the tumor size or tumor response models. This may be attributed to single-dose level design of the study and co-administration of CHOP with rituximab being a contributor to an already effective treatment. Differentiation of CHOP and rituximab effects was also not possible due to the lack of PK data for CHOP and the absence of a CHOP-only treatment arm. Drug product was not a significant covariate in either the tumor size or the tumor response models.
Conclusions: Using PK and PKPD analyses, we found that a higher baseline tumor size was associated with lower rituximab exposure. No significant difference was found between drug products in PK, tumor size change, or tumor response. The apparent drug product effect, which was identified on rituximab PK in the intermediate model, was not significant after the difference in baseline tumor size had been accounted for. The current study exemplifies how PK and PKPD analyses can be applied to identify hidden factors generating apparent differences between drug products, thereby contributing to PK similarity assessments.
References:
[1] Candelaria M, Gonzalez D, Fernández Gómez FJ, Paravisini A, Del Campo García A, Pérez L, et al. Comparative assessment of pharmacokinetics, and pharmacodynamics between RTXM83TM, a rituximab biosimilar, and rituximab in diffuse large B-cell lymphoma patients: a population PK model approach. Cancer Chemother Pharmacol. 2018 Mar;81(3):515–27.
[2] Stein WD, Yang J, Bates SE, Fojo T. Bevacizumab reduces the growth rate constants of renal carcinomas: a novel algorithm suggests early discontinuation of bevacizumab resulted in a lack of survival advantage. Oncologist. 2008 Oct;13(10):1055–62.
[3] Cheson BD, Pfistner B, Juweid ME, Gascoyne RD, Specht L, Horning SJ, Coiffier B, Fisher RI, Hagenbeek A, Zucca E, Rosen ST, Stroobants S, Lister TA, Hoppe RT, Dreyling M, Tobinai K, Vose JM, Connors JM, Federico M and Diehl V, 2007, Revised response criteria for malignant lymphoma. Journal of Clinical Oncology vol. 25: 579–586.
