Explore the vast range of titles available.

Voriconazole (VRC) is used as first line antifungal agent against invasive aspergillosis. Model-based approaches might optimize VRC therapy. This study aimed to investigate the predictive performance of pharmacokinetic models of VRC without pharmacogenetic information for their suitability for model-informed precision dosing. Seven PopPK models were selected from a systematic literature review. A total of 66 measured VRC plasma concentrations from 33 critically ill patients was employed for analysis. The second measurement per patient was used to calculate relative Bias (rBias), mean error (ME), relative root mean squared error (rRMSE) and mean absolute error (MAE) (i) only based on patient characteristics and dosing history (a priori) and (ii) integrating the first measured concentration to predict the second concentration (Bayesian forecasting). The a priori rBias/ME and rRMSE/MAE varied substantially between the models, ranging from −15.4 to 124.6%/−0.70 to 8.01 mg/L and from 89.3 to 139.1%/1.45 to 8.11 mg/L, respectively. The integration of the first TDM sample improved the predictive performance of all models, with the model by Chen (85.0%) showing the best predictive performance (rRMSE: 85.0%; rBias: 4.0%). Our study revealed a certain degree of imprecision for all investigated models, so their sole use is not recommendable. Models with a higher performance would be necessary for clinical use.

Standard tuberculosis (TB) management has failed to control the growing number of drug-resistant TB cases worldwide. Therefore, innovative approaches are required to eradicate TB. Model-informed precision dosing and therapeutic drug monitoring (TDM) have become promising tools for adjusting anti-TB drug doses corresponding with individual pharmacokinetic profiles. These are crucial to improving the treatment outcome of the patients, particularly for those with complex comorbidity and a high risk of treatment failure. Despite the actual benefits of TDM at the bedside, conventional TDM encounters several hurdles related to laborious, time-consuming, and costly processes. Herein, we review the current practice of TDM and discuss the main obstacles that impede it from successful clinical implementation. Moreover, we propose a semi-automated TDM approach to further enhance precision medicine for TB management.

Precision dosing aims to optimize and customize pharmacological treatment at the individual level. The integration of pharmacometric models with Reinforcement Learning (RL) algorithms is currently under investigation to support the personalization of adaptive dosing therapies. In this study, this hybrid technique is applied to the real multiobjective precision dosing problem of givinostat treatment in polycythemia vera (PV) patients. PV is a chronic myeloproliferative disease with an overproduction of platelets (PLT), white blood cells (WBC), and hematocrit (HCT). The therapeutic goal is to simultaneously normalize the levels of these efficacy/safety biomarkers, thus inducing a complete hematological response (CHR). An RL algorithm, Q‐Learning (QL), was integrated with a PK‐PD model describing the givinostat effect on PLT, WBC, and HCT to derive both an adaptive dosing protocol (QLpop‐agent) for the whole population and personalized dosing strategies by coupling a specific QL‐agent to each patient (QLind‐agents). QLpop‐agent learned a general adaptive dosing protocol that achieved a similar CHR rate (77% vs. 83%) when compared to the actual givinostat clinical protocol on 10 simulated populations. Treatment efficacy and safety increased with a deeper dosing personalization by QLind‐agents. These QL‐based patient‐specific adaptive dosing rules outperformed both the clinical protocol and QLpop‐agent by reaching the CHR in 93% of the test patients and completely avoided severe toxicities during the whole treatment period. These results confirm that RL and PK‐PD models can be valid tools for supporting adaptive dosing strategies as interesting performances were achieved in both learning a general set of rules and in customizing treatment for each patient.

Tucuxi, a Swiss‐developed Model‐Informed Precision Dosing (MIPD) software, aims to support clinical dosage decision‐making to achieve therapeutic concentration targets. This study assessed its predictive accuracy compared to NONMEM, a gold‐standard tool for Bayesian PK predictions. A panel of models was created to mimic various pharmacokinetic scenarios following oral, bolus, or intravenous administration. For each scenario, a virtual population of 4000 patients receiving doses ranging from 10 to 120 mg every 24 h was created. Sparse and rich profiles were simulated, with either one or four samples taken per patient. Tucuxi and NONMEM predicted concentrations at sampling times, trough (Cmin) and peak (Cmax) concentrations, and area under the curve (AUC0‐24h) were compared by calculating their relative differences, mean prediction error (MPE) and relative root mean square error (RMSE). The bioequivalence criterion was additionally applied to compare AUC0‐24h, Cmin, and Cmax. All the outcomes predicted by Tucuxi closely matched those predicted by NONMEM. A median of 99.8% of predicted concentrations at sampling times presented relative errors smaller than 0.1%. For all outcomes predicted, MPE and relative RMSE were 0% (−0.09, 0.07) and 0.82% (0%, 18.79%) respectively. The bioequivalence criterion, calculated for AUC0‐24h, Cmin, and Cmax, was verified for all models, with median values of 100%. This project highlights Tucuxi's excellent predictive accuracy compared to NONMEM, demonstrating its reliability and potential for adoption in clinical practice.

Selecting an appropriate population pharmacokinetic (PK) model for individual patients in model‐informed precision dosing (MIPD) can be challenging, particularly in the absence of therapeutic drug monitoring (TDM) samples. We developed a machine learning (ML) model to guide individualized PK model selection for a priori MIPD of vancomycin based on routinely recorded patient characteristics. This retrospective analysis included 343,636 vancomycin TDM records, each from a distinct adult patient across 156 healthcare centers, along with a priori predictions from six PK models. A multi‐label classification approach was applied, labeling PK model predictions based on whether they fell within 80%–125% of observed TDM values. Various modeling strategies were evaluated using XGBoost as the base algorithm, with binary relevance selected for the final model. At the prediction stage, PK models were ranked and averaged for each patient based on ML‐predicted probabilities that predictions would fall within 80%–125% of the observed concentration. Selecting the highest ranked PK model for each patient and ML‐based model averaging outperformed all single PK models, body mass index‐based selection, and naive averaging. On a population level, these ML approaches resulted in more accurate predictions, a higher proportion of predictions within 80%–125% of observed vancomycin concentrations, and no systematic bias. Predictive performance declined with lower ML‐assigned rankings, and selecting the lowest‐ranked PK model for each patient resulted in worse performance than the worst‐performing single PK model. By guiding the selection of appropriate models and avoiding less suitable ones, ML approaches for a priori MIPD may improve early dosing decisions. Schematic overview of the workflow used to train and apply the multi‐label classification model for PK model selection and averaging in a priori vancomycin precision dosing.

Dose‐dependent hematological toxicity of lenalidomide has been reported previously, and thus, there is a clinical need for dose individualization to manage toxicities. The objectives of this study were to explore optimal individualized dosing regimens for Chinese B‐cell malignancies patients with varying degrees of renal function, and to push forward clinical management of hematological toxicity due to lenalidomide overexposure. A total of 164 plasma concentrations of lenalidomide were obtained from 97 Chinese patients with multiple myeloma (MM) and B‐cell non‐Hodgkin lymphoma (NHL) from a multicenter prospective study. A population pharmacokinetic (PopPK) model for lenalidomide was developed by nonlinear mixed effect modeling. A Monte Carlo simulation was conducted to recommend model‐informed precision dosing (MIPD) for patients with varying degrees of renal function. A one‐compartment model with first‐order elimination best described the pharmacokinetics of lenalidomide. The population typical values of lenalidomide were as follows: absorption rate constant (Ka) of 8.34 h−1, apparent volume of distribution (V/F) of 37.4 L, and apparent clearance (CL/F) of 7.4 L/h. Creatinine clearance (CCr) was identified as a major covariate for CL/F, whereas other demographics or clinical characteristics had no significant effect on the model. When given the identical dose, Chinese patients exhibited a higher exposure than the predominantly non‐Asian population at all dosage regimens, especially in patients with severe renal damage (CCr < 30 mL/min). For Chinese patients with CCr of 15–30 mL/min who do not require dialysis usually, compared to the dosing regimen of 15 mg every other day recommended by drug instructions, there exists a relatively lower risk of hematotoxicity when administered with 5 or 10 mg/day. For Chinese patients with CCr < 15 mL/min requiring dialysis, there was still a certain level of hematotoxicity risk associated with the dosing regimen of 5 mg/day recommended by drug instructions. The PopPK Model‐based simulation suggests that Chinese patients may exhibit a higher exposure than the predominantly non‐Asian population. For patients with severely impaired renal function, compared to dose adjustment in accordance with drug instructions, an individualized dosage strategy based on therapeutic drug monitoring (TDM) and MIPD would be preferable from a safety perspective.

Background: To promote model-informed precision dosing (MIPD) for vancomycin (VCM), we developed statements for therapeutic drug monitoring (TDM). Methods: Ten clinical questions were selected. The committee conducted a systematic review and meta-analysis as well as clinical studies to establish recommendations for area under the concentration-time curve (AUC)-guided dosing. Results: AUC-guided dosing tended to more strongly decrease the risk of acute kidney injury (AKI) than trough-guided dosing, and a lower risk of treatment failure was demonstrated for higher AUC/minimum inhibitory concentration (MIC) ratios (cut-off of 400). Higher AUCs (cut-off of 600 μg·h/mL) significantly increased the risk of AKI. Although Bayesian estimation with two-point measurement was recommended, the trough concentration alone may be used in patients with mild infections in whom VCM was administered with q12h. To increase the concentration on days 1–2, the routine use of a loading dose is required. TDM on day 2 before steady state is reached should be considered to optimize the dose in patients with serious infections and a high risk of AKI. Conclusions: These VCM TDM guidelines provide recommendations based on MIPD to increase treatment response while preventing adverse effects.

Introduction Estimation of vancomycin area under the curve (AUC) is challenging in the case of discontinuous administration. Machine learning approaches are increasingly used and can be an alternative to population pharmacokinetic (POPPK) approaches for AUC estimation. The objectives were to train XGBoost algorithms based on simulations performed in a previous POPPK study to predict vancomycin AUC from early concentrations and a few features (i.e. patient information) and to evaluate them in a real-life external dataset in comparison to POPPK. Patients and Methods Six thousand simulations performed from 6 different POPPK models were split into training and test sets. XGBoost algorithms were trained to predict trapezoidal rule AUC a priori or based on 2, 4 or 6 samples and were evaluated by resampling in the training set and validated in the test set. Finally, the 2-sample algorithm was externally evaluated on 28 real patients and compared to a state-of-the-art POPPK model-based averaging approach. Results The trained algorithms showed excellent performances in the test set with relative mean prediction error (MPE)/ imprecision (RMSE) of the reference AUC = 3.3/18.9, 2.8/17.4, 1.3/13.7% for the 2, 4 and 6 samples algorithms respectively. Validation in real patient showed flexibility in sampling time post-treatment initiation and excellent performances MPE/RMSE<1.5/12% for the 2 samples algorithm in comparison to different POPPK approaches. Conclusions The Xgboost algorithm trained from simulation and evaluated in real patients allow accurate and precise prediction of vancomycin AUC. It can be used in combination with POPPK models to increase the confidence in AUC estimation.

The increasing availability of healthcare data and rapid development of big data analytic methods has opened new avenues for use of Artificial Intelligence (AI)- and Machine Learning (ML)-based technology in medical practice. However, applications at the point of care are still scarce. Review and discuss case studies to understand current capabilities for applying AI/ML in the healthcare setting, and regulatory requirements in the US, Europe and China. A targeted narrative literature review of AI/ML based digital tools was performed. Scientific publications (identified in PubMed) and grey literature (identified on the websites of regulatory agencies) were reviewed and analyzed. From the regulatory perspective, AI/ML-based solutions can be considered medical devices (i.e., Software as Medical Device, SaMD). A case series of SaMD is presented. First, tools for monitoring and remote management of chronic diseases are presented. Second, imaging applications for diagnostic support are discussed. Finally, clinical decision support tools to facilitate the choice of treatment and precision dosing are reviewed. While tested and validated algorithms for precision dosing exist, their implementation at the point of care is limited, and their regulatory and commercialization pathway is not clear. Regulatory requirements depend on the level of risk associated with the use of the device in medical practice, and can be classified into administrative (manufacturing and quality control), software-related (design, specification, hazard analysis, architecture, traceability, software risk analysis, cybersecurity, etc.), clinical evidence (including patient perspectives in some cases), non-clinical evidence (dosing validation and biocompatibility/toxicology) and other, such as e.g. benefit-to-risk determination, risk assessment and mitigation. There generally is an alignment between the US and Europe. China additionally requires that the clinical evidence is applicable to the Chinese population and recommends that a third-party central laboratory evaluates the clinical trial results. The number of promising AI/ML-based technologies is increasing, but few have been implemented widely at the point of care. The need for external validation, implementation logistics, and data exchange and privacy remain the main obstacles.

Model-informed precision dosing (MIPD) software tools are used to optimize dosage regimens in individual patients, aiming to achieve drug exposure targets associated with desirable clinical outcomes. Over the last few decades, numerous MIPD software tools have been developed. However, they have still not been widely integrated into clinical practice. This study focuses on identifying the requirements for and evaluating the performance of the currently available MIPD software tools. First, a total of 22 experts in the field of precision dosing completed a web survey to assess the importance (from 0; do not agree at all, to 10; completely agree) of 103 pre-established software tool criteria organized in eight categories: user-friendliness and utilization, user support, computational aspects, population models, quality and validation, output generation, privacy and data security, and cost. Category mean ± pooled standard deviation importance scores ranged from 7.2 ± 2.1 (user-friendliness and utilization) to 8.5 ± 1.8 (privacy and data security). The relative importance score of each criterion within a category was used as a weighting factor in the subsequent evaluation of the software tools. Ten software tools were identified through literature and internet searches: four software tools were provided by companies (DoseMeRx, InsightRX Nova, MwPharm++, and PrecisePK) and six were provided by non-company owners (AutoKinetics, BestDose, ID-ODS, NextDose, TDMx, and Tucuxi). All software tools performed well in all categories, although there were differences in terms of in-built software features, user interface design, the number of drug modules and populations, user support, quality control, and cost. Therefore, the choice for a certain software tool should be made based on these differences and personal preferences. However, there are still improvements to be made in terms of electronic health record integration, standardization of software and model validation strategies, and prospective evidence for the software tools' clinical and cost benefits.