Sybrand Zielhuis 1, Karen de Jong 1, Hinke Huisman-Siebinga 1, Lisa van der Heijden 1, Bart Jacobs 1, Jos Beijnen 1, Hilde Rosing 1, Chantal Stockem, Jeroen van Dorp 1, Michiel van der Heijden 1, Alwin Huitema 1,2,3
1 Netherlands Cancer Institute (Amsterdam, The Netherlands), 2 Princess Máxima Center for Pediatric Oncology (Utrecht, The Netherlands), 3 University Medical Center Utrecht (Utrecht, The Netherlands)
Objectives:
In the NABUCCO trial, patients with locoregionally advanced urothelial cancer (UC) received three cycles of pre-operative ipilimumab and nivolumab [1-3]. Across three cohorts, different dosing combinations of both drugs were applied, leading to substantially different treatment outcomes. Patients receiving low-dose nivolumab (1 mg/kg) and high-dose ipilimumab (3 mg/kg) experienced a significantly higher response rate compared to patients receiving high-dose nivolumab (3 mg/kg) and low-dose ipilimumab (1 mg/kg). These findings suggest the existence of a dose-response relationship for ipilimumab, while nivolumab dose does not appear to be related to response. These suggested dose- and exposure-response relationships require additional investigation, since confirmation of the (non)-existence of these relationships would provide strong arguments during decision making in treatment optimization of ipilimumab-nivolumab treatment regimens. Therefore, we aimed to:
(1) Describe the PK of ipilimumab and nivolumab in patients with locoregionally advanced UC.
(2) Assess the existence of clearance (CL)- and exposure-response relationships in this treatment setting.
Methods:
Prior to each treatment cycle and before surgery, serum samples were collected from NABUCCO trial patients. Ipilimumab and nivolumab serum concentrations were quantified using a validated UPLC-MS/MS method [4]. A combined population PK model for ipilimumab and nivolumab was developed by informing PK parameter estimates with prior information from previously published models [5,6] by use of the $PRIOR subroutine in NONMEM (version 7.5) [7]. In both prior models, the PK of ipilimumab and nivolumab was described by a two-compartment model and time-dependent CL. In our pre-operative treatment setting, the follow-up period was too short to assess the existence of time-dependent CL.
Out of the final combined population PK model, individual Bayesian estimates of baseline clearance (CL0) and area under the curve (AUC) from start of the treatment until surgery were derived for both drugs. These estimates were related to pathologic response at resection (classified as responder, non-responder or non-evaluable) in exploratory plots and quantified by a logistic regression model. In this model, combinations of relative inverse CL0 and AUC estimates of both drugs were related to response in a linear function on logit scale.
Results:
Ipilimumab and nivolumab serum levels were measured in 139 samples from 52 patients. Ipilimumab and nivolumab serum concentrations were adequately described by the final combined population PK model. For both drugs, the volume of distribution of the central compartment, the volume of distribution of the peripheral compartment, and the inter-compartmental clearance were informed by informative priors. The CL from the central compartment and all parameters corresponding for inter-individual variability were informed by non-informative priors. For ipilimumab, a lower CL was estimated in the final model in comparison with the prior model (12.0 mL/h (4.4% RSE) vs. 14.1 mL/h (1.7% RSE)). For nivolumab, a higher CL was estimated in comparison with the prior model (13.0 mL/h (4.8% RSE) vs. 10.8 mL/h (1.5% RSE)). Individual CL0 estimates of both drugs were correlated (correlation coefficient = 0.57). Out of the quartile of patients with the lowest CL0 of both drugs, twelve out of thirteen patients were responders. In the logistic regression model, a combination of the relative ipilimumab AUC and inverse nivolumab CL0 was related to probability of treatment response. The model predicted probability of response was higher for high-dose ipilimumab (3 mg/kg) compared to low-dose ipilimumab (1 mg/kg) (54% versus 29%, respectively), which was in line with NABUCCO trial response rates.
Conclusion:
For the first time, the pre-operative PK of ipilimumab and nivolumab were described and, subsequently, related to treatment response in patients with locoregionally advanced UC. In a logistic regression model, we confirmed that individual ipilimumab exposure and nivolumab CL0 were predictive for treatment response in the NABUCCO trial. Our results support the hypothesis that an ipilimumab exposure-response relationship exists, while this is not the case for nivolumab. For nivolumab, a CL0-response relationship possibly exists. Therefore, administration of a relatively high ipilimumab dose (such as 3 mg/kg) should be prioritized during treatment optimization of ipilimumab-nivolumab regimens in locoregionally advanced UC. In the future, a predicted probability of response, calculated with the help of an individual’s PK estimates, could be a part of treatment decision making.
References:
[1] Van Dijk et al. Preoperative ipilimumab plus nivolumab in locoregionally advanced urothelial cancer: the NABUCCO trial. Nat Med. 2020 Dec.
[2] Van Dorp et al. High- or low-dose preoperative ipilimumab plus nivolumab in stage III urothelial cancer: the phase 1B NABUCCO trial. Nat Med. 2023 Mar.
[3] Stockem et al. Final clinical analysis of pre-operative ipilimumab and nivolumab in locally advanced urothelial cancer and exploration of tumor-draining lymph node composition: The NABUCCO trial. Eur J Cancer. 2025 Oct.
[4] de Jong et al. Optimized sample pre-treatment procedure for the simultaneous UPLC-MS/MS quantification of ipilimumab, nivolumab, and pembrolizumab in human serum. J Chromatogr B Analyt Technol Biomed Life Sci. 2022 Apr.
[5] Sanghavi et al. Population Pharmacokinetics of Ipilimumab in Combination With Nivolumab in Patients With Advanced Solid Tumors. CPT Pharmacometrics Syst Pharmacol. 2020 Jan.
[6] Zhang et al. Population Pharmacokinetics of Nivolumab in Combination With Ipilimumab in Patients With Advanced Malignancies. CPT Pharmacometrics Syst Pharmacol. 2019 Dec.
[7] Chan Kwong et al. Prior information for population pharmacokinetic and pharmacokinetic/pharmacodynamic analysis: overview and guidance with a focus on the NONMEM PRIOR subroutine. J Pharmacokinet Pharmacodyn. 2020 Oct.
Reference: PAGE 34 (2026) Abstr 12207 [www.page-meeting.org/?abstract=12207]
Poster: Drug/Disease Modelling - Oncology