Beyond Population-Level Targets for Drug Concentrations: Precision Dosing Needs Individual-Level Targets that Include Superior Biomarkers of Drug Responses.

The purpose of precision dosing is to increase the chances of therapeutic success in individual patients. This is achieved in practice by adjusting doses to reach precision dosing targets determined previously in relevant populations, ideally with robust supportive evidence showing improved clinical outcomes compared with standard dosing. But is this implicit assumption of translatable population-level precision dosing targets correct and the best for all patients? In this review, the types of precision dosing targets and how they are determined are outlined, problems with the translatability of these targets to individual patients are identified, and ways forward to address these challengers are proposed. Achieving improved clinical outcomes to support precision dosing over standard dosing is currently hampered by applying population-level targets to all patients. Just as "one-dose-fits-all" may be an inappropriate philosophy for drug treatment overall, a "one-target-fits-all" philosophy may limit the broad clinical benefits of precision dosing. Defining individual-level precision dosing targets may be needed for greatest therapeutic success. Superior future precision dosing targets will integrate several biomarkers that together account for the multiple sources of drug response variability.

Development and application of neonatal physiology-based pharmacokinetic models of amikacin and fosfomycin to assess pharmacodynamic target attainment.

Antimicrobial resistance increasingly complicates neonatal sepsis in a global context. Fosfomycin and amikacin are two agents being tested in an ongoing multicenter neonatal sepsis trial. Although neonatal pharmacokinetics (PKs) have been described for these drugs, the physiological variability within neonatal populations makes population PKs in this group uncertain. Physiologically-based pharmacokinetic (PBPK) models were developed in Simcyp for fosfomycin and amikacin sequentially for adult, pediatric, and neonatal populations, with visual and quantitative validation compared to observed data at each stage. Simulations were performed using the final validated neonatal models to determine drug exposures for each drug across a demographic range, with probability of target attainment (PTA) assessments. Successfully validated neonatal PBPK models were developed for both fosfomycin and amikacin. PTA analysis demonstrated high probability of target attainment for amikacin 15 mg/kg i.v. q24h and fosfomycin 100 mg/kg (in neonates aged 0-7 days) or 150 mg/kg (in neonates aged 7-28 days) i.v. q12h for Enterobacterales with fosfomycin and amikacin minimum inhibitory concentrations at the adult breakpoints. Repeat analysis in premature populations demonstrated the same result. PTA analysis for a proposed combination fosfomycin-amikacin target was also performed. The simulated regimens, tested in a neonatal sepsis trial, are likely to be adequate for neonates across different postnatal ages and gestational age. This work demonstrates a template for determining target attainment for antimicrobials (alone or in combination) in special populations without sufficient available PK data to otherwise assess with traditional pharmacometric methods.

Using Machine Learning to Individualize Treatment Effect Estimation: Challenges and Opportunities.

The use of data from randomized clinical trials to justify treatment decisions for real-world patients is the current state of the art. It relies on the assumption that average treatment effects from the trial can be extrapolated to patients with personal and/or disease characteristics different from those treated in the trial. Yet, because of heterogeneity of treatment effects between patients and between the trial population and real-world patients, this assumption may not be correct for many patients. Using machine learning to estimate the expected conditional average treatment effect (CATE) in individual patients from observational data offers the potential for more accurate estimation of the expected treatment effects in each patient based on their observed characteristics. In this review, we discuss some of the challenges and opportunities for machine learning to estimate CATE, including ensuring identification assumptions are met, managing covariate shift, and learning without access to the true label of interest. We also discuss the potential applications as well as future work and collaborations needed to further improve identification and utilization of CATE estimates to increase patient benefit.

Bridging the Worlds of Pharmacometrics and Machine Learning.

Precision medicine requires individualized modeling of disease and drug dynamics, with machine learning-based computational techniques gaining increasing popularity. The complexity of either field, however, makes current pharmacological problems opaque to machine learning practitioners, and state-of-the-art machine learning methods inaccessible to pharmacometricians. To help bridge the two worlds, we provide an introduction to current problems and techniques in pharmacometrics that ranges from pharmacokinetic and pharmacodynamic modeling to pharmacometric simulations, model-informed precision dosing, and systems pharmacology, and review some of the machine learning approaches to address them. We hope this would facilitate collaboration between experts, with complementary strengths of principled pharmacometric modeling and flexibility of machine learning leading to synergistic effects in pharmacological applications.

Digital Therapeutics as a New Therapeutic Modality: A Review from the Perspective of Clinical Pharmacology.

The promise of transforming digital technologies into treatments is what drives the development of digital therapeutics (DTx), generally known as software applications embedded within accessible technologies-such as smartphones-to treat, manage, or prevent a pathological condition. Whereas DTx solutions that successfully demonstrate effectiveness and safety could drastically improve the life of patients in multiple therapeutic areas, there is a general consensus that generating therapeutic evidence for DTx presents challenges and open questions. We believe there are three main areas where the application of clinical pharmacology principles from the drug development field could benefit DTx development: the characterization of the mechanism of action, the optimization of the intervention, and, finally, its dosing. We reviewed DTx studies to explore how the field is approaching these topics and to better characterize the challenges associated with them. This leads us to emphasize the role that the application of clinical pharmacology principles could play in the development of DTx and to advocate for a development approach that merges such principles from development of traditional therapeutics with important considerations from the highly attractive and fast-paced world of digital solutions.

Synthetic Model Combination: A new machine-learning method for pharmacometric model ensembling.

When aiming to make predictions over targets in the pharmacological setting, a data-focused approach aims to learn models based on a collection of labeled examples. Unfortunately, data sharing is not always possible, and this can result in many different models trained on disparate populations, leading to the natural question of how best to use and combine them when making a new prediction. Previous work has focused on global model selection or ensembling, with the result of a single final model across the feature space. Machine-learning models perform notoriously poorly on data outside their training domain, however, due to a problem known as covariate shift, and so we argue that when ensembling models the weightings for individual instances must reflect their respective domains-in other words, models that are more likely to have seen information on that instance should have more attention paid to them. We introduce a method for such an instance-wise ensembling of models called Synthetic Model Combination (SMC), including a novel representation learning step for handling sparse high-dimensional domains. We demonstrate the use of SMC on an example with dosing predictions for vancomycin, although emphasize the applicability of the method to any scenario involving the use of multiple models.

Artificial intelligence assessment of the potential of tocilizumab along with corticosteroids therapy for the management of COVID-19 evoked acute respiratory distress syndrome.

Acute respiratory distress syndrome (ARDS), associated with high mortality rate, affects up to 67% of hospitalized COVID-19 patients. Early evidence indicated that the pathogenesis of COVID-19 evoked ARDS is, at least partially, mediated by hyperinflammatory cytokine storm in which interleukin 6 (IL-6) plays an essential role. The corticosteroid dexamethasone is an effective treatment for severe COVID-19 related ARDS. However, trials of other immunomodulatory therapies, including anti-IL6 agents such as tocilizumab and sarilumab, have shown limited evidence of benefit as monotherapy. But recently published large trials have reported added benefit of tocilizumab in combination with dexamethasone in severe COVID-19 related ARDS. In silico tools can be useful to shed light on the mechanisms evoked by SARS-CoV-2 infection and of the potential therapeutic approaches. Therapeutic performance mapping system (TPMS), based on systems biology and artificial intelligence, integrate available biological, pharmacological and medical knowledge to create mathematical models of the disease. This technology was used to identify the pharmacological mechanism of dexamethasone, with or without tocilizumab, in the management of COVID-19 evoked ARDS. The results showed that while dexamethasone would be addressing a wider range of pathological processes with low intensity, tocilizumab might provide a more direct and intense effect upon the cytokine storm. Based on this in silico study, we conclude that the use of tocilizumab alongside dexamethasone is predicted to induce a synergistic effect in dampening inflammation and subsequent pathological processes, supporting the beneficial effect of the combined therapy in critically ill patients. Future research will allow identifying the ideal subpopulation of patients that would benefit better from this combined treatment.

Model enhanced reinforcement learning to enable precision dosing: A theoretical case study with dosing of propofol.

Extending the potential of precision dosing requires evaluating methodologies offering more flexibility and higher degree of personalization. Reinforcement learning (RL) holds promise in its ability to integrate multidimensional data in an adaptive process built toward efficient decision making centered on sustainable value creation. For general anesthesia in intensive care units, RL is applied and automatically adjusts dosing through monitoring of patient's consciousness. We further explore the problem of optimal control of anesthesia with propofol by combining RL with state-of-the-art tools used to inform dosing in drug development. In particular, we used pharmacokinetic-pharmacodynamic (PK-PD) modeling as a simulation engine to generate experience from dosing scenarios, which cannot be tested experimentally. Through simulations, we show that, when learning from retrospective trial data, more than 100 patients are needed to reach an accuracy within the range of what is achieved with a standard dosing solution. However, embedding a model of drug effect within the RL algorithm improves accuracy by reducing errors to target by 90% through learning to take dosing actions maximizing long-term benefit. Data residual variability impacts accuracy while the algorithm efficiently coped with up to 50% interindividual variability in the PK and 25% in the PD model's parameters. We illustrate how extending the state definition of the RL agent with meaningful variables is key to achieve high accuracy of optimal dosing policy. These results suggest that RL constitutes an attractive approach for precision dosing when rich data are available or when complemented with synthetic data from model-based tools used in model-informed drug development.

OXA-48-Like ß-Lactamases: Global Epidemiology, Treatment Options, and Development Pipeline.

Modern medicine is threatened by the rising tide of antimicrobial resistance, especially among Gram-negative bacteria, where resistance to ß-lactams is most often mediated by ß-lactamases. The penicillin and cephalosporin ascendancies were, in their turn, ended by the proliferation of TEM penicillinases and CTX-M extended-spectrum ß-lactamases. These class A ß-lactamases have long been considered the most important. For carbapenems, however, the threat is increasingly from the insidious rise of a class D carbapenemase, OXA-48, and its close relatives. Over the past 20 years, OXA-48 and "OXA-48-like" enzymes have proliferated to become the most prevalent enterobacterial carbapenemases across much of Europe, Northern Africa, and the Middle East. OXA-48-like enzymes are notoriously difficult to detect because they often cause only low-level in vitro resistance to carbapenems, meaning that the true burden is likely underestimated. Despite this, they are associated with carbapenem treatment failures. A highly conserved incompatibility complex IncL plasmid scaffold often carries blaOXA-48 and may carry other antimicrobial resistance genes, leaving limited treatment options. High conjugation efficiency means that this plasmid is sometimes carried by multiple Enterobacterales in a single patient. Producers evade most ß-lactam-ß-lactamase inhibitor combinations, though promising agents have recently been licensed, notably ceftazidime-avibactam and cefiderocol. The molecular machinery enabling global spread, current treatment options, and the development pipeline of potential new therapies for Enterobacterales that produce OXA-48-like ß-lactamases form the focus of this review.

Drug Dosing Recommendations for All Patients: A Roadmap for Change.

Most drug labels do not contain dosing recommendations for a significant portion of real-world patients for whom the drug is prescribed. Current label recommendations predominately reflect the population studied in pivotal trials that typically exclude patients who are very young or old, emaciated or morbidly obese, pregnant, or have multiple characteristics likely to influence dosing. As a result, physicians may need to guess the correct dose and regimen for these patients. It is now feasible to provide dose and regimen recommendations for these patients by integrating available scientific knowledge and by utilizing or modifying current regulatory agency-industry practices. The purpose of this commentary is to explore several factors that should be considered in creating a process that will provide more effective, safe, and timely drug dosing recommendations for most, if not all, patients. These factors include the availability of real-world data, development of predictive models, experience with the US Food and Drug Administration (FDA)'s pediatric exclusivity program, development of clinical decision software, funding mechanisms like the Prescription Drug Users Fee Act (PDUFA), and harmonization of global regulatory policies. From an examination of these factors, we recommend a relatively simple, efficient expansion of current practices designed to predict, confirm, and continuously improve drug dosing for more patients. We believe implementing these recommendations will benefit patients, payers, industry, and regulatory agencies.

A Real-World Evidence Framework for Optimizing Dosing in All Patients With COVID-19.

Potential treatments for coronavirus disease 2019 (COVID-19) are being investigated at unprecedented speed, and successful treatments will rapidly be used in tens or hundreds of thousands of patients. To ensure safe and effective use in all those patents it is essential also to develop, at unprecedented speed, a means to provide frequently updated, optimal dosing information for all patient subgroups. Success will require immediate collaboration between drug developers, academics, and regulators.

Model-Informed Artificial Intelligence: Reinforcement Learning for Precision Dosing.

The availability of multidimensional data together with the development of modern techniques for data analysis represent an exceptional opportunity for clinical pharmacology. Data science-defined in this special issue as the novel approaches to the collection, aggregation, and analysis of data-can significantly contribute to characterize drug-response variability at the individual level, thus enabling clinical pharmacology to become a critical contributor to personalized healthcare through precision dosing. We propose a minireview of methodologies for achieving precision dosing with a focus on an artificial intelligence technique called reinforcement learning, which is currently used for individualizing dosing regimen in patients with life-threatening diseases. We highlight the interplay of such techniques with conventional pharmacokinetic/pharmacodynamic approaches and discuss applicability in drug research and early development.