Nivolumab de-escalation in patients with complete response: model-based simulations of alternative dosing regimens and noninferiority clinical trials
Alicja Puszkiel (1,2), Blaise Pasquiers (1,3), Guillaume Bianconi (2), Mourad Mseddi (2), Jennifer Arrondeau (4), Pascaline Boudou-Rouquette (4), Sixtine De Percin (4), Olivier Huillard (4), Michel Vidal (2,5), Xavier Declèves (1,2), Jérôme Alexandre (4,6), François Goldwasser (4,6), Benoit Blanchet (2,5)
(1) Université Paris Cité, Inserm UMR-S1144, Faculty of Pharmacy, Paris, France (2) Biologie du Médicament - Toxicologie, Cochin University Hospital, Assistance Publique-Hôpitaux de Paris, Paris, France (3) PhinC Developpement, Massy, France (4) Medical Oncology Unit, Cochin University Hospital, Cancer Research for PErsonalized Medicine (CARPEM), Assistance Publique Hôpitaux de Paris, Paris, France (5) University Paris Cité, CNRS UMR8038, Inserm U1268, Faculty of Pharmacy, CARPEM, Paris, France (6) URP 4466 PRETRAM, Assistance Publique Hôpitaux de Paris, Université Paris Cité, Paris, France
Objectives: Nivolumab is a monoclonal antibody restoring anti-cancer immune response by blocking binding of PD-1 receptor on T-cells to PD-L1 and 2 expressed by tumor cells [1]. The clinical benefit is observed in approximately 20% of patients and a half of them present complete response (disappearance of tumor lesions) [1]. Nivolumab is administered at a dose of 240 mg every 2 weeks (Q2W) or 480 mg Q4W until disease progression or unacceptable toxicity [2]. Since no recommendations have been established for patients who responded to nivolumab, they continue to receive treatment for a total duration of two years. Less frequent administration of nivolumab in these patients would increase their quality of life and decrease financial toxicity for the social security (the annual cost per patient in France is approximately 74k €) without decreasing the antitumor efficacy if the efficient plasma trough concentration at steady-state (Cmin,ss) is maintained. The aims of this study were to simulate alternative dosing regimens of nivolumab in patients with complete response using a previously developed population pharmacokinetic (popPK) model. The number of patients to be included in a PK noninferiority clinical trial evaluating the alternative dosing regimens was assessed using simulations.
Methods: A popPK model was previously developed in Monolix 2021R1 based on data from 334 real-world patients (992 plasma concentrations of nivolumab). The model consisted of two compartments with linear clearance (CL) without time-varying PK. The mean estimates of CL, V1, V2 and Q were 0.0091 L/h (IIV = 33%), 4.17 L (31%), 3.38 L (41%) and 0.05 L/h respectively. CL increased by 30% in patients with high C-reactive protein (CRP) values (>50 vs <50 mg/L), by 13% in patients with high body weight (BW, 104 kg vs 71 kg) and increased by 28% in patients with low serum albumin (29 g/L vs 38 g/L). Simulations were performed in Simulx 2021R1. Two dose levels were tested (240 mg and 480 mg), each dose with an increasing dosing interval from 2 to 12 weeks. For each scenario, 2500 patients were simulated. It was hypothesized that patients who achieve complete response improve their health status and do not present cachexia or inflammatory syndrome. Therefore, PK profiles were simulated for patients with CRP < 50 mg/L and serum albumin in the range 38 – 46 g/L. For each simulated dosing schedule,Cmin,ss were extracted and compared to the target Cmin,ss of 1.5 mg/L [3] or 2.5 mg/L [4]. The schedule was considered valid if at least 90% of simulated patients had Cmin,ss above target. The minimum number of patients that should be included in a clinical trial to obtain at least 90% of patients above Cmin,ss (> 1.5 or > 2.5 mg/L) with alpha-risk of 5% was evaluated through simulations. Different sample sizes were tested (n = 30, 50, 75, 100, 150, 300, 1000) and each trial was replicated 1000 times. The number of patients above the Cmin,ss target for each simulated scenario was calculated and the 5th percentile of the 1000 simulations was extracted. The minimum sample size was selected if the 5th percentile of 1000 simulated trials was ≥ 90%.
Results: Simulations showed that dosing intervals could be extended by 2-3 times while Cmin,ss remains above the target in more than 90% of patients. Indeed, 240 mg Q6W and 480 mg Q8W resulted in attainment of target Cmin,ss > 2.5 mg/L in 95.8% and 95.4% of patients, respectively, whereas 240 mg Q8W and 480 mg Q10W resulted in attainment of target Cmin,ss > 1.5 mg/L in 91.0% and 91.8% of patients, respectively. Selection of 240 mg Q6W or Q8W would decrease by 2- to 3-fold the annual costs of treatment and allow saving approximately 50-55k € per patient per year. A minimum sample size to obtain more than 90% of patients with Cmin,ss > 2.5 mg/L, with alpha-risk of 5%, for 240 mg Q6W and 480 mg Q8W would be would be 30 and 50 patients, respectively. For 1.5 mg/L threshold, a minimum of 30 patients would be sufficient for both 240 mg Q6W and 480 mg Q8W. However, for 240 mg Q8W and 480 mg Q10W, a minimum of 1000 patients would be necessary for 1.5 mg/L threshold.
Conclusions: The proposed Cmin,ss targets are based on 0.1 mg/kg Q2W dose which showed similar clinical efficacy in phase 1 studies in advanced melanoma than the approved 3 mg/kg Q2W dose [4]. These results will allow setting up a prospective clinical trial comparing the standard dosing regimen with the best simulated schedule in patients with advanced melanoma who achieved complete response within the first 6 months of treatment.
References:
[1] Desnoyer et al. Eur J Cancer 2020. 128: 119-128.
[2] https://www.ema.europa.eu/en/documents/product-information/opdivo-epar-product-information_en.pdf
[3] Peer et al. J Clin Pharmacol 2022. 62: 532–40
[4] Agrawal et al. J Immunother Cancer 2016. 4: 72