Luyuan Qi (1) and Peter McCormack (2)
(1) Novartis Pharma AG, Basel, Switzerland; (2) Novartis Pharmaceuticals UK Limited
Introduction:
The use of radioligand therapy (RLT) is gaining ground in oncology emerging as an essential pillar in cancer care, with clinical benefits already established for some advanced cancers [1, 2].
RLT offers great potential as a treatment option, focusing on specifically targeting the tumor (or its microenvironment) while minimizing damage to the surrounding non-tumor tissue. The targeting is achieved by e.g., selective tumor receptor expression, binding of an isotope linked ligand and radiation induced breaks in the tumor DNA leading to cell death. However, heterogeneity within targeted patient populations (e.g., receptor expression profiles, tumor perfusion rates, plasma and tumor decay rates, binding and internalization rates, etc.), has made variability a competing factor in the understanding of the dose response relationship, and difficulty in identifying the ‘ideal’ dose and patient population early in the development process. Additional factors leading to difficulty in defining the ideal treatment options include the interrelation between the dose, the regimen and schedule, the role of cell repair rate differences across individual and the non-tumor External Beam Radiation Therapy (EBRT) thresholds. One significant challenge in early oncology trials is to define, balance and understand these features within the Phase I population for support of the Phase II protocol design and selected patient population in addition to elucidation of the potential optimal treatment schedules. The challenge therefore is to identify the appropriate dosing regimens and patient subgroups early and assess potential treatment options all within the context of the high variability and the current safety (EBRT) thresholds.
Modelling and simulation-based approaches can provide a tool to support decision making regarding patient selection and dose regimen optimization in RLT. This approach can be applied to the initiation of clinical trials, continuously leveraging available data from multiple sources. As new data arrives from dosed cohorts, the established models can be continuously refined to guide decision making. Here, we present a generic modelling and simulation framework and illustrate how this can be used to support RLT development via simulation scenarios.
Objectives:
To illustrate how the proposed modelling and simulation framework can aid in the decision criteria for both patient selection and dose regimen optimization in RLT development.
Methods:
Date available from the Phase 1 part of LuMIERE trial (A Study of 177Lu-FAP-2286 in Advanced Solid Tumors) [3] was used to develop the initial models and perform simulation scenarios.
We aimed at establishing models that can characterize the event chain from dose to response (dose-exposure- efficacy/toxicity relationship). Various models including PKPD, empirical regression models, and decision tree model were developed for investigating the relationship from dose to dosimetry (a surrogate of efficacy). We explored patient factors (e.g., age, weight, renal function, cancer type, etc.) that may impact efficacy and safety within the developed models. The developed models were used to perform simulations and illustrate how patient selection and dosing optimization could be guided by anticipated high tumor uptake and lower non-tumor organ uptake within EBRT limit.
Results:
Starting from data collected in Phase 1, when multiple indications were included in the dose escalation, the proposed approach could aid the decision making regarding the identification of potential indications to move forward, e.g., by investigating the impacts of indication on Emax (indicating maximum radiation uptake) and EC50 (indicating drug efficiency) of the PKPD model and selecting those with a relatively high Emax and reasonable EC50. The results from PKPD model could be compared with those from other models such as the decision tree model. Results from one simulation scenario also indicated that higher dose per cycle did not necessarily lead to higher tumor dosimetry.
Conclusions:
We highlight the value of model-based approaches in best supporting patient selection and dosing regimen optimization decisions during RLT development. Using this approach from the early stage of development space is expected to play an increasingly important role.
References:
[1] Updated treatment recommendations for prostate cancer from the ESMO Clinical Practice Guideline considering treatment intensification and use of novel systemic agents.
eUpdate – Prostate Cancer Treatment Recommendations (esmo.org)
[2] EMA. A summary of the European public assessment report (EPAR) for Lutathera.
Lutathera, INN-lutetium (177Lu) oxodotreotide (europa.eu)
[3] A Study of 177Lu-FAP-2286 in Advanced Solid Tumors (LuMIERE).
A Study of 177Lu-FAP-2286 in Advanced Solid Tumors (LuMIERE) – Tabular View – ClinicalTrials.gov
Reference: PAGE 32 (2024) Abstr 11099 [www.page-meeting.org/?abstract=11099]
Poster: Drug/Disease Modelling - Oncology