An Evidence-Based Rationale for Dose De-escalation of Subcutaneous Atezolizumab

Atezolizumab is a programmed death-ligand 1 (PD-L1) checkpoint inhibitor for the treatment of different forms of cancer. The subcutaneous formulation of atezolizumab has recently received approval. However, treatment with atezolizumab continues to be expensive, and the number of patients needing treatment with this drug continues to increase.

We propose two alternative dosing regimens for subcutaneous atezolizumab to reduce drug expenses while ensuring effective exposure; one may be directly implemented in the clinic.

We developed two alternative dose interval prolongation strategies based on pharmacokinetic modeling and simulation. The first dosing regimen was based on patients' weight while maintaining equivalent systemic drug exposure by adhering to Food and Drug Administration (FDA) guidelines for in silico dose adjustments. The second dosing regimen aimed to have a minimum atezolizumab concentration above the 6 µg/mL threshold, associated with 95% intratumoral PD-L1 receptor saturation for at least 95% of all patients.

We found that, for the weight-based dosing regimen, the approved 3-week dosing interval could be extended to 5 weeks for patients < 50 kg and 4 weeks for patients weighing 50-65 kg. Besides improving patient convenience, these alternative dosing intervals led to a predicted 7% and 12% cost reduction for either the USA or European population. For the second dosing regimen, we predicted that a 6-week dosing interval would result in 95% of the patients above the 6 µg/mL threshold while reducing costs by 50%.

We have developed and evaluated two alternative dosing regimens that resulted in a cost reduction. Our weight-based dosing regimen can be directly implemented and complies with FDA guidelines for alternative dosing regimens of PD-L1 inhibitors. For the more progressive alternative dosing regimen aimed at the intratumoral PD-L1 receptor threshold, further evidence on efficacy and safety is needed before implementation.