Precision medicine across endocrine, metabolic, immune, and oncologic systems: an integrated mechanistic review

The pharmacological therapy of neonates is extremely complicated due to the immature organ systems of the body, high water content, and low numbers of clinical studies in children of this age. Conventional dosing strategies, usually based on weight or age, are often inaccurate and may result in either underexposure or toxicity. A structured literature search was performed in PubMed/MEDLINE and Scopus, as well as in Web of Science, from the year 2000 until September 2025. The search was based on neonatal pharmacokinetics, model-informed precision dosing (MIPD), Bayesian dosing, and artificial intelligence/machine learning. Following the removal of duplicates and 2-stage screening (title/abstract and full-text review), 82 studies were used in the end to conduct this review. This review follows the PRISMA 2020 flow selection style. We established that pharmacokinetic and pharmacodynamic (popPK/PD) models, with the help of Bayesian strategies, can offer a context of therapy individualization, but they are still limited by the absence of complete information and high intersubjective differences. Artificial intelligence (AI) can also be used to provide an extra level of support by processing vast and multifaceted data, discovering latent trends, and improving predictions of doses in real time. Our review concluded that applications of AI-driven tools are now emerging in neonatal intensive care, particularly for antibiotics and analgesics, where accurate dosing is critical. Combining AI with the known pharmacometric practices could assist clinicians to minimize adverse drug reactions, enhance treatment outcomes, and get closer to the actual personalized pharmacotherapy in neonates.