Evaluation of dosing strategy for pembrolizumab for oncology indications

Abstract

Background: Traditionally, most monoclonal antibodies (mAbs) have been dosed based on body weight because of perceived contribution of body size in pharmacokinetic variability. The same approach was used in the initial pembrolizumab studies; however, following availability of PK data, the need for weight-based dosing for pembrolizumab was reassessed.

Methods: A previously established population PK (popPK) model as well as exposure-response results from patients with advanced melanoma or non-small cell lung cancer (NSCLC) were used to evaluate the potential application of a fixed dosing regimen with the aim of maintaining pembrolizumab exposures within the range demonstrated to provide near maximal efficacy and acceptable safety. Individual PK exposures for the selected fixed dosing regimen from recently completed trials with head and neck cancer, NSCLC, microsatellite instability high (MSI-H) in colorectal cancer (CRC) and urothelial cancer were used to confirm acceptability. To determine whether fixed dosing would maintain exposures within the range of clinical experience, the individual AUC distributions with fixed dosing were compared with the range of exposures from the pembrolizumab doses that were evaluated in early studies (2 mg/kg Q3W, 10 mg/kg Q3W/Q2W).

Results: Body-weight dependence of clearance was characterized by a power relationship with an exponent of 0.578, a value consistent with fixed- and weight-based dosing providing similar control of PK variability. A fixed dose of 200 mg Q3W was investigated in trials based on predicted exposures maintained within the established exposure range in all patients. Mean (% CV, n) AUC_{ss, 6-weeks} was 1.87 (37%, 830), 1.38 (38%, 760) and 7.63 (35%, 1405) mg*day/mL in patients receiving 200 mg, 2 mg/kg and 10 mg/kg Q3W pembrolizumab. High-weight patients had the lowest exposures with 200 mg Q3W; however, exposures in this group (>90 kg) were within the range of prior clinical experience at 2 mg/kg Q3W associated with near maximal efficacy.

Conclusions: Doses of 200 mg and 2 mg/kg provide similar exposure distributions with no advantage to either dosing approach with respect to controlling PK variability. These findings suggest that weight-based and fixed-dose regimens are appropriate for pembrolizumab.

Keywords: Clinical dose; Dosing strategy; Exposure-response analysis; Fixed dose; Flat dose; Pembrolizumab; Population pharmacokinetics analysis.

Fig. 1

Simulated distribution of steady-state AUC exposures (2800 replicate simulations) for the weight-based regimens of 2 mg/kg Q3W, 10 mg/kg Q3W, and 10 mg/kg Q2W compared with the simulated distribution of exposures for two potential fixed-dose regimens (log scale): Box: straight middle line = median; edges = 25th and 75th percentiles; whiskers = 10th and 90th percentiles; dots = 5th and 95th percentiles. Horizontal dashed lines represent the range of exposures (5th percentile of 2 mg/kg Q3W and 95th percentile of 10 mg/kg Q2W) from dose regimens demonstrated to have comparable efficacy and tolerability in melanoma and NSCLC trials

Fig. 2

Predicted variation in pembrolizumab AUC exposure by body weight for weight-based (a) and fixed-dose (b) regimens (2800 replicate simulations): Horizontal dashed lines represent the range of exposures (5th percentile of 2 mg/kg Q3W and 95th percentile of 10 mg/kg Q2W) from dose regimens demonstrated to have comparable efficacy and tolerability in melanoma and NSCLC trials

Fig. 3

Consistency of observed concentrations in patients with predictions based on population PK model: Pembrolizumab concentration-time profiles during the first dose (left panels) and at steady state (right panels) of repeated dosing at 2 mg/kg Q3W (top panels) and 200 mg Q3W (bottom panels). Solid markers represent observed pembrolizumab serum concentrations. Solid line represents median predicted concentration time profile, based on the population PK model. Shaded areas represent 90% prediction interval for the prediction

Fig. 4

Distribution of observed pembrolizumab AUC_ss, _0-6weeks: Panel a – Consistency with model predictions (Simulated values shown in gray and observed values in white). Panel b – Variation in exposures with body weight under weight-based versus fixed dosing. Box: straight middle line = median; edges = 25th and 75th percentiles; whiskers = 10th and 90th percentiles; dots = 5th and 95th percentiles. Horizontal dashed lines (------) represent the range of exposures (5th percentile of 2 mg/kg Q3W and 95th percentile of 10 mg/kg Q2W) from dose regimens demonstrated to have comparable efficacy and tolerability in melanoma and previously treated NSCLC trials. Observed data are based on Table 1. In Panel B, distribution of observed AUC_ss, _0-6weeks for light (≤50 kg), middle (between 50 and 90 kg) and heavy (≥90 kg) body-weight patients

Fig. 5

Consistency of pembrolizumab clearance in patients with differing cancer: melanoma from KEYNOTE-001, -002 and -006. NSCLC from KEYNOTE-001, -010 and -024. Other (other cancers) from KEYNOTE-001 in initial cohort. HNSCC (head and neck trial) from KEYNOTE-055. MSIH (MSI-H in CRC) from KEYNOTE-164. UC (urothelial cancer trial) from KEYNOTE-045 and -052