Editorial: Innovative pharmacometric approaches to inform drug development and clinical use Garcia-Cremades et al. 10.3389/fphar.2023.1274139

Abstract

'Hugo Geerts'] A successful Drug Development is driven by many factors, but ultimately it is often dependent upon the weakest part. Therefore, it is necessary to utilize all tools available to support the complete journey for a successful development program. Model-informed drug development (MIDD) in which crucial information is generated from mathematical analysis or modeling, is rapidly becoming a powerful tool in pharmaceutical R&D and in the regulatory environment. Appropriate use of modeling can contribute to more rational and efficient decision-making in drug development, leading to substantial resource savings and shortened timelines. MIDD can broadly be divided in data- and mechanism-driven modeling and the manuscripts in this Research Topic are good examples of the diversity of applied algorithms. In addition, the large range of disease indications is testimony to the impact of these modeling techniques in clinical drug development and clinical practice. Data-driven approaches include more traditional statistical bio-informatics analyses of large datasets or machine learning algorithms to derive predictive insights ([Ribba et al.]()). The quality of these predictions is heavily dependent upon the nature of the training sets and issues of generalizability need to be addressed. Two articles look to derive estimates of drug exposure using more advanced Physiologically-based Pharmacokinetic models (PBPK) to predict drug exposure in other populations ([Zazo et al.]()). An interesting combination of traditional PopPK modeling with machine learning aims to understand the role of covariates ([Zhu et al.]()). Finally, PK modeling can also be used to derive the quantitative pharmacokinetics trajectory of biomarkers which can clarify the contribution of these biomarkers in clinical development ([Michelet et al.]()). Mechanism-driven modeling on the other hand can be helpful in those cases where data are lacking or noisy. This is often the case in CNS disorders, where robust quantitative biomarkers are scarce, functional scales are often based on structured interviews and biomarkers are often not strongly related to clinical outcomes.