1082P - A rational approach to dose optimisation of pembrolizumab and nivolumab using cost analysis and pharmacokinetic modelling and simulation

Date 09 October 2016
Event ESMO 2016 Congress
Session Poster display
Topics Cancer Immunology and Immunotherapy
Presenter Kayode Ogungbenro
Citation Annals of Oncology (2016) 27 (6): 359-378. 10.1093/annonc/mdw378
Authors K. Ogungbenro1, A. Patel2, R. Duncombe3, J. Clark3, P. Lorigan4
  • 11centre For Applied Pharmacokinetic Research, The University of Manchester, Manchester, United Kingdom, M13 9PL - Manchester/GB
  • 2Experimental Cancer Medicine Team, The Christie NHS Foundation Trust, Manchester/GB
  • 3Pharmacy, The Christie NHS Foundation Trust, Manchester/GB
  • 4Medical Oncology, University of Manchester - Christie NHS Foundation Trust, M20 4BX - Manchester/GB



Pembrolizumab and nivolumab are PDL1 inhibitors approved for the treatment of advanced malignancies. Both are approved based on body size (mg/kg) dosing, which can be associated with significant wastage due to available vial sizes and acquisition cost of the drugs. Dose banding is using a defined set of ranges (±10%) offers an alternative approach to body weight dosing. The aim of this work is to evaluate cost and dosing using pharmacokinetic (PK) simulation and data from 42 patients entered into the MHRA Early Access to Medicines Scheme for pembrolizumab and 24 patients enrolled in checkmate-172 trial for nivolumab.


Consecutive patients receiving pembrolizumab (n = 42) or nivolumab (n = 24) and with a median weight of 79.5kg (range 44-130kg) and 73.3kg (range 52-103kg) respectively were analysed. The costs were based on 12 weeks of treatment which was the time take to first assessment for both drugs. This was 4 cycles of pembrolizumab every 21 days (Q3W) and 6 cycles of nivolumab every 14 days (Q2W). 1000 random individuals were evaluated in a simulated study for pharmacokinetic and pharmacodynamic assessment. Published population PK models were used to simulate exposure and probability of trough levels achieving target level (10mg/L) required for maximum target engagement or receptor occupancy for pembrolizumab and nivolumab respectively.


The costs of different strategies are illustrated in Table 1. The table also include simulated area under the curve (over the last two cycles) and mean probability of trough levels (after the first cycle of treatment) achieving 10mg/L or receptor occupancy for the different dosing strategies.

Pembrolizumab Q3W Cost of 4 cycles* (n = 42) Relative Drug Cost Saving (-) /Expense (+) Simulated AUC Mean mg.day/L (SD) Mean prob of Target Engagement
2mg/kg £ 836,340 0% 1365 (385) 0.52
Dose Band (10% variance) £ 707,470 -16% 1360 (369) 0.53
150mg Fixed dose £662,760 -19% 1387 (406) 0.56
200mg Fixed dose £883,680 +8% 1849 (542) 0.92
PK derived £694,320 -17% 1437 (383) 0.64
Nivolumab Q2W Cost of 6 cycles* (n = 24) Mean prob of Receptor Occupancy
3mg/kg £ 404,135 0% 2098 (879) 0.91
Dose Band (10% variance) £ 372,541 -8% 2036 (845) 0.90
240mg Fixed dose £ 379,123 -6% 2280 (989) 0.94
PK derived £ 363,326 -10% 2068 (820) 0.91

*UK cost excluding VAT; company's submission


Banded fixed dose strategies result in comparable levels of exposure and target engagement or receptor occupancy and can offer significant cost reduction. Vial size availability can contribute to the level of savings to be gained.

Clinical trial identification


Legal entity responsible for the study

Kayode Ogungbenro




R. Duncombe: Paid consultant to Bayer; Janssen; Amgen; Pfizer; MSD; Roche; Sanofi Support for travel from Novartis. No funding received for this study. P. Lorigan: Paid consultant to BMS, Merck, Amgen, Novartis, Roche, GSK and Chugai. Support for travel from BMS and Merck. No funding received for this study. All other authors have declared no conflicts of interest.