Immune check-point inhibitors are drugs that are markedly different from other anticancer drugs because of their indirect mechanisms of antitumoral action and their apparently random effect in terms of efficacy and toxicity. This marked pharmacodynamics variability in patients calls for reconsidering to what extent approved dosing used in clinical practice are optimal or whether they should require efforts for customization in outlier patients. To better understand whether or not dosing could be an actionable item in oncology, in this review, preclinical and clinical development of immune checkpoint inhibitors are described, particularly from the angle of dose finding studies. Other issues in connection with dosing issues are developed, such as the flat dosing alternative, the putative role therapeutic drug monitoring could play, the rise of combinatorial strategies, and pharmaco-economic aspects.

Pharmacokinetic (PK) parameter estimation is a critical and complex step in the model‐informed precision dosing (MIPD) approach. The mapbayr package was developed to perform maximum a posteriori Bayesian estimation (MAP‐BE) in R from any population PK model coded in mrgsolve. The performances of mapbayr were assessed using two approaches. First, “test” models with different features were coded, for example, first‐order and zero‐order absorption, lag time, time‐varying covariates, Michaelis–Menten elimination, combined and exponential residual error, parent drug and metabolite, and small or large inter‐individual variability (IIV). A total of 4000 PK profiles (combining single/multiple dosing and rich/sparse sampling) were simulated from each test model, and MAP‐BE of parameters was performed in both mapbayr and NONMEM. Second, a similar procedure was conducted with seven “real” previously published models to compare mapbayr and NONMEM on a PK outcome used in MIPD. For the test models, 98% of mapbayr estimations were identical to those given by NONMEM. Some discordances could be observed when dose‐related parameters were estimated or when models with large IIV were used. The exploration of objective function values suggested that mapbayr might outdo NONMEM in specific cases. For the real models, a concordance close to 100% on PK outcomes was observed. The mapbayr package provides a reliable solution to perform MAP‐BE of PK parameters in R. It also includes functions dedicated to data formatting and reporting and enables the creation of standalone Shiny web applications dedicated to MIPD, whatever the model or the clinical protocol and without additional software other than R.

Therapeutic drug monitoring of ibrutinib is based on the area under the curve of concentration vs. time (AUC_IBRU) instead of trough concentration (C_min,ss) because of a limited accumulation in plasma. Our objective was to identify a limited sampling strategy (LSS) to estimate AUC_IBRU associated with Bayesian estimation. The actual AUC_IBRU of 85 patients was determined by the Bayesian analysis of the full pharmacokinetic profile of ibrutinib concentrations (pre-dose T0 and 0.5, 1, 2, 4 and 6 h post-dose) and experimental AUC_IBRU were derived considering combinations of one to four sampling times. The T0–1–2–4 design was the most accurate LSS (root-mean-square error RMSE = 11.0%), and three-point strategies removing the 1 h or 2 h points (RMSE = 22.7% and 14.5%, respectively) also showed good accuracy. The correlation between the actual AUC_IBRU and C_min,ss was poor (r² = 0.25). The joint analysis of dihydrodiol-ibrutinib metabolite concentrations did not improve the predictive performance of AUC_IBRU. These results were confirmed in a prospective validation cohort (n = 27 patients). At least three samples, within the pre-dose and 4 h post-dose period, are necessary to estimate ibrutinib exposure accurately.

Hepatotoxicity is an important concern for nearly 40% of the patients treated with trabectedin for advanced soft tissue sarcoma (ASTS). The mechanisms underlying these liver damages have not yet been elucidated but they have been suggested to be related to the production of reactive metabolites. The aim of this pharmacogenetic study was to identify genetic variants of pharmacokinetic genes such as CYP450 and ABC drug transporters that could impair the trabectedin metabolism in hepatocytes. Sixty-three patients with ASTS from the TSAR clinical trial (NCT02672527) were genotyped by next-generation sequencing for 11 genes, and genotype–toxicity association analyses were performed with R package SNPassoc. Among the results, ABCC2 c.1249A allele (rs2273697) and ABCG2 intron variant c.-15994T (rs7699188) were associated with an increased risk of severe cytolysis, whereas ABCC2 c.3563A allele had a protective effect, as well as ABCB1 variants rs2032582 and rs1128503 (p-value < 0.05). Furthermore, CYP3A5*1 rs776746 (c.6986A > G) increased the risk of severe overall hepatotoxicity (p = 0.012, odds ratio (OR) = 5.75), suggesting the implication of metabolites in the hepatotoxicity. However, these results did not remain significant after multiple analysis correction. These findings need to be validated on larger cohorts of patients, with mechanistic studies potentially being able to validate the functional consequences of these variants.

L'individualisation des doses est essentielle en cancérologie afin de maitriser la variabilité interindividuelle de la pharmacocinétique (PK) des anticancéreux. Le suivi thérapeutique pharmacologique (STP) consiste à mesurer une ou plusieurs concentrations plasmatiques de médicament afin d'estimer l'exposition et adapter la dose en conséquence. Ces concentrations peuvent être interprétées conjointement avec un modèle de PK de population afin d'estimer les paramètres PK individuels et prédire un marqueur d'exposition comme la concentration résiduelle pour les thérapies orales, ou l'aire sous la courbe (AUC) pour les cytotoxiques. Dans le présent travail, le package R mapbayr a été développé pour permettre d'effectuer l'estimation maximum a posteriori (MAP) Bayésienne des paramètres PK, et a été validé en comparaison au logiciel de référence, NONMEM. Puis, l'utilisation d'une approche de population conjointement au STP a été étudiée avec trois applications cliniques. Le premier exemple traite du STP du busulfan. Un modèle de PK de population en pédiatrie a été développé. La meilleure stratégie de calcul de dose a priori a été établie à partir d'un modèle de covariable, et une stratégie de prélèvement optimale (associé à l'analyse MAP-Bayésienne) a été caractérisée avec une étude de simulation-estimation. Le second exemple concerne le STP de l'ibrutinib. Une stratégie de prélèvement optimale sur la base de 3 concentrations a été caractérisée afin d'estimer précisément l'AUC, et l'intérêt de l'analyse conjointe des concentrations de métabolite a été étudiée. Le troisième exemple concerne le STP du sunitinib et du pazopanib. L'impact du choix du modèle de PK de population (parmi ceux disponibles dans la littérature) et deux méthodes alternatives pour améliorer la qualité des prédictions ("model averaging" et "flattened priors") ont été explorés. En conclusion, ces travaux ont permis de transposer en clinique des modèles mathématiques afin de prédire une dose individuelle d'anticancéreux chez les patients traités au sein de notre hôpital.