Characterizing the dynamic interaction among gastric emptying, glucose absorption, and glycemic control in nondiabetic obese adults.

The effects of altered gastric emptying on glucose absorption and kinetics are not well understood in nondiabetic obese adults. The aim of this work was to develop a physiology-based model that can characterize and compare interactions among gastric emptying, glucose absorption, and glycemic control in nondiabetic obese and lean healthy adults. Dynamic glucose, insulin, and gastric emptying (measured with breath test) data from 12 nondiabetic obese and 12 lean healthy adults were available until 180 min after an oral glucose tolerance test (OGTT) with 10, 25, and 75 g of glucose. A physiology-based model was developed to characterize glucose kinetics applying nonlinear mixed-effects modeling with NONMEM7.3. Glucose kinetics after OGTT was described by a one-compartment model with an effect compartment to describe delayed insulin effects on glucose clearance. After the interactions between individual gastric emptying and glucose absorption profiles were accounted for, the glucose absorption rate was found to be similar in nondiabetic obese and lean controls. Baseline glucose concentration was estimated to be only marginally higher in nondiabetic obese subjects (4.9 vs. 5.2 mmol/l), whereas insulin-dependent glucose clearance in nondiabetic obese subjects was found to be cut in half compared with lean controls (0.052 vs. 0.029 l/min) and the insulin concentration associated with 50% of insulin-dependent glucose elimination rate was approximately twofold higher in nondiabetic obese subjects compared with lean controls (7.1 vs. 15.3 µU/ml). Physiology-based models can characterize and compare the dynamic interaction among gastric emptying, glucose absorption and glycemic control in populations of interest such as lean healthy and nondiabetic obese adults.

Characterizing and Forecasting Individual Weight Changes in Term Neonates.

OBJECTIVES: To develop a mathematical, semimechanistic model characterizing physiological weight changes in term neonates, identify and quantify key maternal and neonatal factors influencing weight changes, and provide an online tool to forecast individual weight changes during the first week of life. STUDY DESIGN: Longitudinal weight data from 1335 healthy term neonates exclusively breastfed up to 1 week of life were available. A semimechanistic model was developed to characterize weight changes applying nonlinear mixed-effects modeling. Covariate testing was performed by applying a standard stepwise forward selection-backward deletion approach. The developed model was externally evaluated on 300 additional neonates collected in the same center. RESULTS: Weight changes during first week of life were described as a function of a changing net balance between time-dependent rates of weight gain and weight loss. Males had higher birth weights (WT0) than females. Gestational age had a positive effect on WT0 and weight gain rate, whereas mother's age had a positive effect on WT0 and a negative effect on weight gain rate. The developed model showed good predictive performance when externally validated (bias = 0.011%, precision = 0.52%) and was able to accurately forecast individual weight changes up to 1 week with only 3 initial weight measurements (bias = -0.74%, precision = 1.54%). CONCLUSIONS: This semimechanistic model characterizes weight changes in healthy breastfed neonates during first week of life. We provide a user-friendly online tool allowing caregivers to forecast and monitor individual weight changes. We plan to validate this model with data from other centers and expand it with data from preterm neonates.

Dynamic modeling in ovarian cancer: an original approach linking early changes in modeled longitudinal CA-125 kinetics and survival to help decisions in early drug development.

OBJECTIVE: Early prediction of the expected benefit of treatment in recurrent ovarian cancer (ROC) patients may help in drug development decisions. The actual value of 50% CA-125 decrease is being reconsidered. The main objective of the present study was to quantify the links between longitudinal assessments of CA-125 kinetics and progression-free survival (PFS) in treated recurrent ovarian cancer (ROC) patients. METHODS: The CALYPSO randomized phase III trial database comparing two platinum-based regimens in ROC patients was randomly split into a "learning dataset" and a "validation dataset". A parametric survival model was developed to associate longitudinal modeled CA-125 changes (ΔCA125), predictive factors, and PFS. The predictive performance of the model was evaluated with simulations. RESULTS: The PFS of 534 ROC patients were properly characterized by a parametric mathematical model. The modeled ΔCA125 from baseline to week 6 was a better predictor of PFS than the modeled fractional change in tumor size. Simulations confirmed the model's predictive performance. CONCLUSIONS: We present the first parametric survival model quantifying the relationship between PFS and longitudinal CA-125 kinetics in treated ROC patients. The model enabled calculation of the increase in ΔCA125 required to observe a predetermined benefit in PFS to compare therapeutic strategies in populations. Therefore, ΔCA125 may be a predictive marker of the expected gain in PFS and an early predictive tool in drug development decisions.

Contribution of machine learning to tumor growth inhibition modeling for hepatocellular carcinoma patients under Roblitinib (FGF401) drug treatment.

Machine learning (ML) opens new perspectives in identifying predictive factors of efficacy among a large number of patients' characteristics in oncology studies. The objective of this work was to combine ML with population pharmacokinetic/pharmacodynamic (PK/PD) modeling of tumor growth inhibition to understand the sources of variability between patients and therefore improve model predictions to support drug development decisions. Data from 127 patients with hepatocellular carcinoma enrolled in a phase I/II study evaluating once-daily oral doses of the fibroblast growth factor receptor FGFR4 kinase inhibitor, Roblitinib (FGF401), were used. Roblitinib  PKs was best described by a two-compartment model with a delayed zero-order absorption and linear elimination. Clinical efficacy using the longitudinal sum of the longest lesion diameter data was described with a population PK/PD model of tumor growth inhibition including resistance to treatment. ML, applying elastic net modeling of time to progression data, was associated with cross-validation, and allowed to derive a composite predictive risk score from a set of 75 patients' baseline characteristics. The two approaches were combined by testing the inclusion of the continuous risk score as a covariate on PD model parameters. The score was found as a significant covariate on the resistance parameter and resulted in 19% reduction of its variability, and 32% variability reduction on the average dose for stasis. The final PK/PD model was used to simulate effect of patients' characteristics on tumor growth inhibition profiles. The proposed methodology can be used to support drug development decisions, especially when large interpatient variability is observed.

Leveraging Predictive Pharmacometrics-Based Algorithms to Enhance Perinatal Care-Application to Neonatal Jaundice.

The field of medicine is undergoing a fundamental change, transforming towards a modern data-driven patient-oriented approach. This paradigm shift also affects perinatal medicine as predictive algorithms and artificial intelligence are applied to enhance and individualize maternal, neonatal and perinatal care. Here, we introduce a pharmacometrics-based mathematical-statistical computer program (PMX-based algorithm) focusing on hyperbilirubinemia, a medical condition affecting half of all newborns. Independent datasets from two different centers consisting of total serum bilirubin measurements were utilized for model development (342 neonates, 1,478 bilirubin measurements) and validation (1,101 neonates, 3,081 bilirubin measurements), respectively. The mathematical-statistical structure of the PMX-based algorithm is a differential equation in the context of non-linear mixed effects modeling, together with Empirical Bayesian Estimation to predict bilirubin kinetics for a new patient. Several clinically relevant prediction scenarios were validated, i.e., prediction up to 24 h based on one bilirubin measurement, and prediction up to 48 h based on two bilirubin measurements. The PMX-based algorithm can be applied in two different clinical scenarios. First, bilirubin kinetics can be predicted up to 24 h based on one single bilirubin measurement with a median relative (absolute) prediction difference of 8.5% (median absolute prediction difference 17.4 µmol/l), and sensitivity and specificity of 95.7 and 96.3%, respectively. Second, bilirubin kinetics can be predicted up to 48 h based on two bilirubin measurements with a median relative (absolute) prediction difference of 9.2% (median absolute prediction difference 21.5 µmol/l), and sensitivity and specificity of 93.0 and 92.1%, respectively. In contrast to currently available nomogram-based static bilirubin stratification, the PMX-based algorithm presented here is a dynamic approach predicting individual bilirubin kinetics up to 48 h, an intelligent, predictive algorithm that can be incorporated in a clinical decision support tool. Such clinical decision support tools have the potential to benefit perinatal medicine facilitating personalized care of mothers and their born and unborn infants.

Integration of Pharmacokinetics, Pharmacodynamics, Safety, and Efficacy into Model-Informed Dose Selection in Oncology First-in-Human Study: A Case of Roblitinib (FGF401).

Model-informed dose selection has been drawing increasing interest in oncology early clinical development. The current paper describes the example of FGF401, a selective fibroblast growth factor receptor 4 (FGFR4) inhibitor, in which a comprehensive modeling and simulation (M&S) framework, using both pharmacometrics and statistical methods, was established during its first-in-human clinical development using the totality of pharmacokinetics (PK), pharmacodynamic (PD) biomarkers, and safety and efficacy data in patients with cancer. These M&S results were used to inform FGF401 dose selection for future development. A two-compartment population PK (PopPK) model with a delayed 0-order absorption and linear elimination adequately described FGF401 PK. Indirect PopPK/PD models including a precursor compartment were independently established for two biomarkers: circulating FGF19 and 7α-hydroxy-4-cholesten-3-one (C4). Model simulations indicated a close-to-maximal PD effect achieved at the clinical exposure range. Time-to-progression was analyzed by Kaplan-Meier method which favored a trough concentration (Ctrough )-driven efficacy requiring Ctrough above a threshold close to the drug concentration producing 90% inhibition of phospho-FGFR4. Clinical tumor growth inhibition was described by a PopPK/PD model that reproduced the dose-dependent effect on tumor growth. Exposure-safety analyses on the expected on-target adverse events, including elevation of aspartate aminotransferase and diarrhea, indicated a lack of clinically relevant relationship with FGF401 exposure. Simulations from an indirect PopPK/PD model established for alanine aminotransferase, including a chain of three precursor compartments, further supported that maximal target inhibition was achieved and there was a lack of safety-exposure relationship. This M&S framework supported a dose selection of 120 mg once daily fasted or with a low-fat meal and provides a practical example that might be applied broadly in oncology early clinical development.

Pharmacometrics and Machine Learning Partner to Advance Clinical Data Analysis.

Clinical pharmacology is a multidisciplinary data sciences field that utilizes mathematical and statistical methods to generate maximal knowledge from data. Pharmacometrics (PMX) is a well-recognized tool to characterize disease progression, pharmacokinetics, and risk factors. Because the amount of data produced keeps growing with increasing pace, the computational effort necessary for PMX models is also increasing. Additionally, computationally efficient methods, such as machine learning (ML) are becoming increasingly important in medicine. However, ML is currently not an integrated part of PMX, for various reasons. The goals of this article are to (i) provide an introduction to ML classification methods, (ii) provide examples for a ML classification analysis to identify covariates based on specific research questions, (iii) examine a clinically relevant example to investigate possible relationships of ML and PMX, and (iv) present a summary of ML and PMX tasks to develop clinical decision support tools.

Population Pharmacokinetic Modeling in the Presence of Missing Time-Dependent Covariates: Impact of Body Weight on Pharmacokinetics of Paracetamol in Neonates.

Body weight is the primary covariate in pharmacokinetics of many drugs and dramatically changes during the first weeks of life of neonates. The objective of this study is to determine if missing body weights in preterm and term neonates affect estimates of model parameters and which methods can be used to improve performance of a population pharmacokinetic model of paracetamol. Data for our analysis were obtained from previously published studies on the pharmacokinetics of intravenous paracetamol in neonates. We adopted a population model of body weight change in neonates to implement three previously introduced methods of handling missing covariates based on data imputation, likelihood function modification, and full random effects modeling. All models were implemented in NONMEM 7.4, and population parameters were estimated using the FOCE method. Our major finding was that missing body weights minimally affect population estimates of pharmacokinetic parameters but do affect the covariate relationship parameters, particularly the one describing dependence of clearance on body weight. None of the tested methods changed estimates of between-subject variability nor impacted the predictive performance of the model. Our analysis shows that a modeling approach towards handling missing covariates allows borrowing information gathered in various studies as long as they target the same population. This approach is particularly useful for handling time-dependent missing covariates.

Personalized weight change prediction in the first week of life.

BACKGROUND & AIMS: Almost all neonates show physiological weight loss and consecutive weight gain after birth. The resulting weight change profiles are highly variable as they depend on multiple neonatal and maternal factors. This limits the value of weight nomograms for the early identification of neonates at risk for excessive weight loss and related morbidities. The objective of this study was to characterize weight changes and the effect of supplemental feeding in late preterm and term neonates during the first week of life, to identify and quantify neonatal and maternal influencing factors, and to provide an educational online prediction tool. METHODS: Longitudinal weight data from 3638 healthy term and late preterm neonates were prospectively recorded up to 7 days of life. Two-thirds (n = 2425) were randomized to develop a semi-mechanistic model characterizing weight change as a balance between time-dependent rates of weight gain and weight loss. The dose-dependent effect of supplemental feeding on weight gain was characterized. A population analysis applying nonlinear mixed-effects modeling was performed using NONMEM 7.3. The model was evaluated on the remaining third of neonates (n = 1213). RESULTS: Key population characteristics (median [range]) of the whole sample were gestational age 39.9 [34.4-42.4] weeks, birth weight 3400 [1980-5580] g, maternal age 32 [15-51] years, cesarean section 26%, and girls 50%. The model demonstrated good predictive performance (bias 0.01%, precision 0.56%), and is able to accurately predict individual weight change (bias 0.15%, precision 1.43%) and the dose-dependent effects of supplemental feeding up to 1 week after birth based on weight measurements during the first 3 days of life, including birth weight, and the following characteristics: gestational age, gender, delivery mode, type of feeding, maternal age, and parity. CONCLUSIONS: We present the first mathematical model not only to describe weight change in term and late preterm neonates but also to provide an educational online tool for personalized weight prediction in the first week of life.

Pharmacometrics-based decision tools facilitate mHealth implementation.

INTRODUCTION: The healthcare system is experiencing a paradigm shift in delivering its services, evolving from a reactive 'one size-fits-all' structure to a patient-centric model focusing on individualized medicine. This change is driven by scientific progress, including quantitative evaluation and optimization of treatment strategies through pharmacometric approaches, harnessing the power of the digital revolution. Areas covered: This review describes four main steps to apply pharmacometrics-based decision support tools, consisting of validated scientific components, available technical options, consideration of regulatory aspects, and achievement of efficient commercialization. Examples of pharmacometrics-based decision tools that support monitoring of patients and individualization of treatment strategies in neonates, children and adults are presented. Expert commentary: We envision that user-friendly decision support tools will facilitate implementation of mobile health approaches (mHealth) realizing benefits to paediatric and adult patients and their caregivers.

Quantitative clinical pharmacology practice for optimal use of antibiotics during the neonatal period.

INTRODUCTION: For safe and effective neonatal antibiotic therapy, knowledge of the pharmacokinetic parameters of antibacterial agents in neonates is a prerequisite. Fast maturational changes during the neonatal period influence pharmacokinetic and pharmacodynamic parameters and their variability. Consequently, the need for applying quantitative clinical pharmacology and determining optimal drug dosing regimens in neonates has become increasingly recognized. AREAS COVERED: Modern quantitative approaches, such as pharmacometrics, are increasingly utilized to characterize, understand and predict the pharmacokinetics of a drug and its effect, and to quantify the variability in the neonatal population. Individual factors, called covariates in modeling, are integrated in such approaches to explain inter-individual pharmacokinetic variability. Pharmacometrics has been shown to be a relevant tool to evaluate, optimize and individualize drug dosing regimens. EXPERT OPINION: Challenges for optimal use of antibiotics in neonates can largely be overcome with quantitative clinical pharmacology practice. Clinicians should be aware that there is a next step to support the clinical decision-making based on clinical characteristics and therapeutic drug monitoring, through Bayesian-based modeling and simulation methods. Pharmacometric modeling and simulation approaches permit us to characterize population average, inter-subject and intra-subject variability of pharmacokinetic parameters such as clearance and volume of distribution, and to identify and quantify key factors that influence the pharmacokinetic behavior of antibiotics during the neonatal period.

Pharmacometric Approaches to Personalize Use of Primarily Renally Eliminated Antibiotics in Preterm and Term Neonates.

Sepsis remains a major cause of mortality and morbidity in neonates, and, as a consequence, antibiotics are the most frequently prescribed drugs in this vulnerable patient population. Growth and dynamic maturation processes during the first weeks of life result in large inter- and intrasubject variability in the pharmacokinetics (PK) and pharmacodynamics (PD) of antibiotics. In this review we (1) summarize the available population PK data and models for primarily renally eliminated antibiotics, (2) discuss quantitative approaches to account for effects of growth and maturation processes on drug exposure and response, (3) evaluate current dose recommendations, and (4) identify opportunities to further optimize and personalize dosing strategies of these antibiotics in preterm and term neonates. Although population PK models have been developed for several of these drugs, exposure-response relationships of primarily renally eliminated antibiotics in these fragile infants are not well understood, monitoring strategies remain inconsistent, and consensus on optimal, personalized dosing of these drugs in these patients is absent. Tailored PK/PD studies and models are useful to better understand relationships between drug exposures and microbiological or clinical outcomes. Pharmacometric modeling and simulation approaches facilitate quantitative evaluation and optimization of treatment strategies. National and international collaborations and platforms are essential to standardize and harmonize not only studies and models but also monitoring and dosing strategies. Simple bedside decision tools assist clinical pharmacologists and neonatologists in their efforts to fine-tune and personalize the use of primarily renally eliminated antibiotics in term and preterm neonates.

Identification of Most Aggressive Carcinoma Among Patients Diagnosed With Prostate Cancer Using Mathematical Modeling of Prostate-Specific Antigen Increases.

BACKGROUND: Tools for differentiating aggressive and indolent prostate carcinoma (PCa) are needed. Mathematical modeling is a promising approach for longitudinal analysis of tumor marker kinetics. PATIENTS AND METHODS: The prostate-specific antigen (PSA) increases from patients with PCa and those with benign prostatic hyperplasia (BPH) were retrospectively analyzed using a mathematical model. Using the NONMEM program, individual PSA kinetics were fit to the following equation: [d(PSA)/dt = (PROD.K × exp [RHO1 × t]) × (1 - BPH) + PROD.NK × exp (RHO2 × t) - KELIM × (PSA)], where RHO1 is the PSA production increase rate by PCa cells (PROD.K), RHO2 is the PSA production increase rate by non-PCa cells (PROD.NK), and KELIM is the PSA elimination rate. The comparative value of the modeled kinetic parameters, estimated for each patient, for predicting the D'Amico score and relapse-free survival (RFS) were tested using logistic regression analysis and multivariate survival tests. RESULTS: The PSA kinetics from 62 patients with BPH and 149 patients with PCa before radical prostatectomy were successfully modeled. We identified statistically significant relationships between the PSA growth rate related to cancer cells (RHO1) and the probability of D'Amico high-risk group (less than the median RHO1 vs. at the median or greater: odds ratio, 2.15; 95% confidence interval [CI], 1.00-4.77; P = .05). RHO1 was also a significant prognostic factor for RFS on univariate analysis and against other reported prognostic factors using multivariate Cox tests. Three independent prognostic factors of RFS were found: RHO1 (hazard ratio [HR], 2.71; 95% CI, 1.25-5.84; P = .01), Gleason score (HR, 8.54; 95% CI, 4.19-17.40; P < .01), and positive surgical margins (HR, 2.04; 95% CI, 1.05-3.97; P = .03). CONCLUSION: Using a few PSA time points analyzed with a mathematical model (easily manageable in routine practice), it could be possible to determine before surgery whether a patient has presented with aggressive PCa.

Effect of Kidney Function on Drug Kinetics and Dosing in Neonates, Infants, and Children.

Neonates, infants, and children differ from adults in many aspects, not just in age, weight, and body composition. Growth, maturation and environmental factors affect drug kinetics, response and dosing in pediatric patients. Almost 80% of drugs have not been studied in children, and dosing of these drugs is derived from adult doses by adjusting for body weight/size. As developmental and maturational changes are complex processes, such simplified methods may result in subtherapeutic effects or adverse events. Kidney function is impaired during the first 2 years of life as a result of normal growth and development. Reduced kidney function during childhood has an impact not only on renal clearance but also on absorption, distribution, metabolism and nonrenal clearance of drugs. 'Omics'-based technologies, such as proteomics and metabolomics, can be leveraged to uncover novel markers for kidney function during normal development, acute kidney injury, and chronic diseases. Pharmacometric modeling and simulation can be applied to simplify the design of pediatric investigations, characterize the effects of kidney function on drug exposure and response, and fine-tune dosing in pediatric patients, especially in those with impaired kidney function. One case study of amikacin dosing in neonates with reduced kidney function is presented. Collaborative efforts between clinicians and scientists in academia, industry, and regulatory agencies are required to evaluate new renal biomarkers, collect and share prospective pharmacokinetic, genetic and clinical data, build integrated pharmacometric models for key drugs, optimize and standardize dosing strategies, develop bedside decision tools, and enhance labels of drugs utilized in neonates, infants, and children.