Standard atezolizumab leads to severe overexposure in real-world patients with lung cancer: How far could we go in extending dosing intervals and saving money?

person Sophie Marolleau COMPO Centre de Recherche en Cancérologie de Marseille Inserm U1068 CNRS, Marseille, France info_outline Sophie Marolleau, Alice Mogenet, Clara Boeri, Mourad Hamimed, Joseph Ciccolini, Laurent Greillier

Authors person Sophie Marolleau COMPO Centre de Recherche en Cancérologie de Marseille Inserm U1068 CNRS, Marseille, France info_outline Sophie Marolleau, Alice Mogenet, Clara Boeri, Mourad Hamimed, Joseph Ciccolini, Laurent Greillier Organizations COMPO Centre de Recherche en Cancérologie de Marseille Inserm U1068 CNRS, Marseille, France, Aix-Marseille University, APHM, INSERM, CNRS, CRCM, and Hôpital Nord, Marseille, France, COMPO Centre de Recherche en Cancérologie de Marseille Inserm U1068, Marseille, France, SMARTc Centre de Recherche en Cancérologie de Marseille Inserm U1068, Marseille, France, Assistance Publique Hôpitaux de Marseille (AP-HM), Marseille, France, Multidisciplinary Oncology and Therapeutic Innovations, Hôpital Nord, Marseille, France Abstract Disclosures Research Funding No funding sources reported Background: Immune checkpoint inhibitors are given as fixed doses, which may lead to drug exposure exceeding the plasma concentrations required for target engagement. If confirmed, this could pave the way for de-escalating treatments without compromising efficacy while reducing drug-costs. Methods: We monitored plasma concentrations of Atezolizumab in 40 real-world patients. A final set of 33 patients with SCLC or NSCLC were available for full PK/PD analysis. Individual PK parameters were derived using a population pharmacokinetics approach. Individual PK parameters allowed to simulate the full concentration-time profiles and to further simulate for each patient the time to reach the target threshold after the first dose of Atezolizumab. In addition, efficacy (RECIST criteria) and immune-related adverse events, (IRAEs, CTCAE grading) were monitored and statistical analysis was performed to search for possible exposure-response relationships. Results: Overall Response Rate was 55% and no severe IRAEs were observed. All patients had Cmin levels significantly above the target threshold on the first cycle of Atezolizumab. The mean Cmax after the first dose on cycle 1 was 426.1 µg/ml and the mean Cmin was 68.5 µg/ml, i.e. 11-time higher than the target threshold associated with target engagement (i.e., 6 µg/ml). The mean AUC at the first cycle was 4763 µg.d/ml. Exposure metrics and PK parameters at baseline and simulated prior RECIST evaluation were tested together with other covariates (i.e. age, sex, smoking, NLR, number of metastatic sites, line of treatment) as possible predictive markers of response or toxicity. None of the parameters related to Atezolizumab pharmacokinetics or exposure levels were associated with efficacy or toxicity, most likely because all patients had plasma levels well above pharmacologically active concentrations. First-line therapy was associated with better response on univariate analysis, but this effect was lost on multivariate analysis. Further simulations based on time required to fall below the efficacy threshold suggest that dosing intervals could be extended from Q7W to Q19W (mean: Q12W). Considering the annual budget for Atezolizumab at our institution, further simulations showed that the annual drug cost could thus be reduced from 2.5 M€ to 0.6 M€ while ensuring that all patients remain above active concentrations. Conclusions: Our study demonstrates in a real-world setting that standard Atezolizumab given as 1200 mg Q3W infusion leads to severe overexposure of the drug. This strong overexposure was not associated with more severe treatment-related toxicities. Finally, our study suggests that PK-guided dosing with Atezolizumab could help customizing the dosing interval while maintaining efficacious trough levels and that Q12W could be an alternate scheduling leading to significant cut in drug costs.