GTE: a graph learning framework for prediction of T-cell receptors and epitopes binding specificity.

The interaction between T-cell receptors (TCRs) and peptides (epitopes) presented by major histocompatibility complex molecules (MHC) is fundamental to the immune response. Accurate prediction of TCR-epitope interactions is crucial for advancing the understanding of various diseases and their prevention and treatment. Existing methods primarily rely on sequence-based approaches, overlooking the inherent topology structure of TCR-epitope interaction networks. In this study, we present $GTE$, a novel heterogeneous Graph neural network model based on inductive learning to capture the topological structure between TCRs and Epitopes. Furthermore, we address the challenge of constructing negative samples within the graph by proposing a dynamic edge update strategy, enhancing model learning with the nonbinding TCR-epitope pairs. Additionally, to overcome data imbalance, we adapt the Deep AUC Maximization strategy to the graph domain. Extensive experiments are conducted on four public datasets to demonstrate the superiority of exploring underlying topological structures in predicting TCR-epitope interactions, illustrating the benefits of delving into complex molecular networks. The implementation code and data are available at https://github.com/uta-smile/GTE.

Normalization of RNA-Seq data using adaptive trimmed mean with multi-reference.

The normalization of RNA sequencing data is a primary step for downstream analysis. The most popular method used for the normalization is the trimmed mean of M values (TMM) and DESeq. The TMM tries to trim away extreme log fold changes of the data to normalize the raw read counts based on the remaining non-deferentially expressed genes. However, the major problem with the TMM is that the values of trimming factor M are heuristic. This paper tries to estimate the adaptive value of M in TMM based on Jaeckel's Estimator, and each sample acts as a reference to find the scale factor of each sample. The presented approach is validated on SEQC, MAQC2, MAQC3, PICKRELL and two simulated datasets with two-group and three-group conditions by varying the percentage of differential expression and the number of replicates. The performance of the present approach is compared with various state-of-the-art methods, and it is better in terms of area under the receiver operating characteristic curve and differential expression.

Phosphorylation in the Ser/Arg-rich region of the nucleocapsid of SARS-CoV-2 regulates phase separation by inhibiting self-association of a distant helix.

The nucleocapsid protein (N) of SARS-CoV-2 is essential for virus replication, genome packaging, evading host immunity, and virus maturation. N is a multidomain protein composed of an independently folded monomeric N-terminal domain that is the primary site for RNA binding and a dimeric C-terminal domain that is essential for efficient phase separation and condensate formation with RNA. The domains are separated by a disordered Ser/Arg-rich region preceding a self-associating Leu-rich helix. Phosphorylation in the Ser/Arg region in infected cells decreases the viscosity of N:RNA condensates promoting viral replication and host immune evasion. The molecular level effect of phosphorylation, however, is missing from our current understanding. Using NMR spectroscopy and analytical ultracentrifugation, we show that phosphorylation destabilizes the self-associating Leu-rich helix 30 amino-acids distant from the phosphorylation site. NMR and gel shift assays demonstrate that RNA binding by the linker is dampened by phosphorylation, whereas RNA binding to the full-length protein is not significantly affected presumably due to retained strong interactions with the primary RNA-binding domain. Introducing a switchable self-associating domain to replace the Leu-rich helix confirms the importance of linker self-association to droplet formation and suggests that phosphorylation not only increases solubility of the positively charged elongated Ser/Arg region as observed in other RNA-binding proteins but can also inhibit self-association of the Leu-rich helix. These data highlight the effect of phosphorylation both at local sites and at a distant self-associating hydrophobic helix in regulating liquid-liquid phase separation of the entire protein.

Strategy for modeling higher-order G-quadruplex structures recalcitrant to NMR determination.

Guanine-rich nucleic acids can form intramolecularly folded four-stranded structures known as G-quadruplexes (G4s). Traditionally, G4 research has focused on short, highly modified DNA or RNA sequences that form well-defined homogeneous compact structures. However, the existence of longer sequences with multiple G4 repeats, from proto-oncogene promoters to telomeres, suggests the potential for more complex higher-order structures with multiple G4 units that might offer selective drug-targeting sites for therapeutic development. These larger structures present significant challenges for structural characterization by traditional high-resolution methods like multi-dimensional NMR and X-ray crystallography due to their molecular complexity. To address this current challenge, we have developed an integrated structural biology (ISB) platform, combining experimental and computational methods to determine self-consistent molecular models of higher-order G4s (xG4s). Here we outline our ISB method using two recent examples from our lab, an extended c-Myc promoter and long human telomere G4 repeats, that highlights the utility and generality of our approach to characterizing biologically relevant xG4s.

A machine learning approach to predict daptomycin exposure from two concentrations based on Monte Carlo simulations.

Daptomycin is a concentration-dependent lipopeptide antibiotic for which exposure/effect relationships have been shown. Machine learning (ML) algorithms, developed to predict the individual exposure to drugs, have shown very good performances in comparison to maximum a posteriori Bayesian estimation (MAP-BE). The aim of this work was to predict the area under the blood concentration curve (AUC) of daptomycin from two samples and a few covariates using XGBoost ML algorithm trained on Monte Carlo simulations. Five thousand one hundred fifty patients were simulated from two literature population pharmacokinetics models. Data from the first model were split into a training set (75%) and a testing set (25%). Four ML algorithms were built to learn AUC based on daptomycin blood concentration samples at pre-dose and 1 h post-dose. The XGBoost model (best ML algorithm) with the lowest root mean square error (RMSE) in a 10-fold cross-validation experiment was evaluated in both the test set and the simulations from the second population pharmacokinetic model (validation). The ML model based on the two concentrations, the differences between these concentrations, and five other covariates (sex, weight, daptomycin dose, creatinine clearance, and body temperature) yielded very good AUC estimation in the test (relative bias/RMSE = 0.43/7.69%) and validation sets (relative bias/RMSE = 4.61/6.63%). The XGBoost ML model developed allowed accurate estimation of daptomycin AUC using C0, C1h, and a few covariates and could be used for exposure estimation and dose adjustment. This ML approach can facilitate the conduct of future therapeutic drug monitoring (TDM) studies.

Multi-variable AUC for sifting complementary features and its biomedical application.

Although sifting functional genes has been discussed for years, traditional selection methods tend to be ineffective in capturing potential specific genes. First, typical methods focus on finding features (genes) relevant to class while irrelevant to each other. However, the features that can offer rich discriminative information are more likely to be the complementary ones. Next, almost all existing methods assess feature relations in pairs, yielding an inaccurate local estimation and lacking a global exploration. In this paper, we introduce multi-variable Area Under the receiver operating characteristic Curve (AUC) to globally evaluate the complementarity among features by employing Area Above the receiver operating characteristic Curve (AAC). Due to AAC, the class-relevant information newly provided by a candidate feature and that preserved by the selected features can be achieved beyond pairwise computation. Furthermore, we propose an AAC-based feature selection algorithm, named Multi-variable AUC-based Combined Features Complementarity, to screen discriminative complementary feature combinations. Extensive experiments on public datasets demonstrate the effectiveness of the proposed approach. Besides, we provide a gene set about prostate cancer and discuss its potential biological significance from the machine learning aspect and based on the existing biomedical findings of some individual genes.

Improvement of differential diagnosis of lung cancer by use of multiple protein tumor marker combinations.

BACKGROUND: Differential diagnosis of non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC) in hospitalized patients is crucial for appropriate treatment choice. OBJECTIVE: To investigate the relevance of serum tumor markers (STMs) and their combinations for the differentiation of NSCLC and SCLC subtypes. METHODS: Between 2000 and 2003, 10 established STMs were assessed retrospectively in 311 patients with NSCLC, 128 with SCLC prior systemic first-line therapy and 51 controls with benign lung diseases (BLD), by automatized electrochemiluminescence immunoassay technology. Receiver operating characteristic (ROC) curves and logistic regression analyses were used to evaluate the diagnostic efficacy of both individual and multiple STMs with corresponding sensitivities at 90% specificity. Standards for Reporting of Diagnostic Accuracy (STARD guidelines) were followed. RESULTS: CYFRA 21-1 (cytokeratin-19 fragment), CEA (carcinoembryonic antigen) and NSE (neuron specific enolase) were significantly higher in all lung cancers vs BLD, reaching AUCs of 0.81 (95% CI 0.76-0.87), 0.78 (0.73-0.84), and 0.88 (0.84-0.93), respectively. By the three marker combination, the discrimination between benign and all malignant cases was improved resulting in an AUC of 0.93 (95% CI 0.90-0.96). In NSCLC vs. BLD, CYFRA 21-1, CEA and NSE were best discriminative STMs, with AUCs of 0.86 (95% CI 0.81-0.91), 0.80 (0.74-0.85), and 0.85 (0.79-0.91). The three marker combination also improved the AUC: 0.92; 95% CI 0.89-0.96). In SCLC vs. BLD, ProGRP (pro-gastrin-releasing peptide) and NSE were best discriminative STMs, with AUCs of 0.89 (95% CI 0.84-0.94) and 0.96 (0.93-0.98), respectively, and slightly improved AUC of 0.97 (95% CI 0.95-0.99) when in combination. Finally, discrimination between SCLC and NSCLC was possible by ProGRP (AUC 0.86; 95% CI 0.81-0.91), NSE (AUC 0.83; 0.78-0.88) and CYFRA 21-1 (AUC 0.69; 0.64-0.75) and by the combination of the 3 STMs (AUC 0.93; 0.91-0.96), with a sensitivity of 88% at 90% specificity. CONCLUSIONS: The results confirm the power of STM combinations for the differential diagnosis of lung cancer from benign lesions and between histological lung cancer subtypes.

Application of peripheral blood routine parameters in the diagnosis of influenza and Mycoplasma pneumoniae.

OBJECTIVES: Influenza and Mycoplasma pneumoniae infections often present concurrent and overlapping symptoms in clinical manifestations, making it crucial to accurately differentiate between the two in clinical practice. Therefore, this study aims to explore the potential of using peripheral blood routine parameters to effectively distinguish between influenza and Mycoplasma pneumoniae infections. METHODS: This study selected 209 influenza patients (IV group) and 214 Mycoplasma pneumoniae patients (MP group) from September 2023 to January 2024 at Nansha Division, the First Affiliated Hospital of Sun Yat-sen University. We conducted a routine blood-related index test on all research subjects to develop a diagnostic model. For normally distributed parameters, we used the T-test, and for non-normally distributed parameters, we used the Wilcoxon test. RESULTS: Based on an area under the curve (AUC) threshold of ≥ 0.7, we selected indices such as Lym# (lymphocyte count), Eos# (eosinophil percentage), Mon% (monocyte percentage), PLT (platelet count), HFC# (high fluorescent cell count), and PLR (platelet to lymphocyte ratio) to construct the model. Based on these indicators, we constructed a diagnostic algorithm named IV@MP using the random forest method. CONCLUSIONS: The diagnostic algorithm demonstrated excellent diagnostic performance and was validated in a new population, with an AUC of 0.845. In addition, we developed a web tool to facilitate the diagnosis of influenza and Mycoplasma pneumoniae infections. The results of this study provide an effective tool for clinical practice, enabling physicians to accurately diagnose and differentiate between influenza and Mycoplasma pneumoniae infection, thereby offering patients more precise treatment plans.

The extent of cytochrome P450 3A induction by antiseizure medications: A systematic review and network meta-analysis.

OBJECTIVE: Antiseizure medications (ASMs) are commonly categorized as enzyme-inducers and non-enzyme-inducers based on their propensity to enhance the metabolism of concomitantly administered drugs. This systematic review and network meta-analysis aimed to rank ASMs as cytochrome P450 3A (CYP3A)-inducers based on a comparative assessment of ASM-induced reduction in the concentrations of sensitive substrate drugs. METHODS: The protocol was registered with PROSPERO (International Prospective Register of Systematic Reviews; CRD42022335846), and the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) standards were followed. We searched MEDLINE, Embase, and Cochrane until March 14, 2023 without an initial date restriction. Data were additionally obtained via the US Food and Drug Administration database. Studies had to be prospective, with ASM monotherapy for ≥5 days. The primary parameter was the magnitude of change in the area under the concentration-time curve of CYP3A substrates following treatment with the ASM. The standardized mean difference (SMD) was used as the point estimate for the indirect comparisons between ASMs using the pairwise method. Bias risk was assessed using the PKclin tool. RESULTS: We identified 14 open-label, fixed-sequence studies with 370 participants. The effect size of 600 mg/day carbamazepine did not differ from those of 300 mg/day phenytoin (SMD = -.06, 95% confidence interval [CI] = -.18 to .07) and 200 mg/day cenobamate (SMD = -.11, 95% CI = -.26 to .04). Carbamazepine at 600 mg/day was the strongest CYP3A-inducer (P-score = .88), followed by carbamazepine 400 mg/day (.83), phenytoin 300 mg/day (.79), and cenobamate 200 mg/day (.73). Eslicarbazepine (800 mg/day) ranked higher than cenobamate 100 mg/day and oxcarbazepine 900 mg/day (.60, .39, and .37, respectively). SIGNIFICANCE: Despite the limited number of studies, our network meta-analysis emphasizes that the magnitude of ASM effects on CYP3A substrate metabolism is a dose-dependent continuum. When possible, ASM classification as inducers should apply cutoff values tailored to the outcome. Prescribers should monitor plasma concentrations or clinical effects of CYP3A substrates and consider selecting concomitant medications accordingly.

A Comprehensive Study on Nanoparticle Drug Delivery to the Brain: Application of Machine Learning Techniques.

The delivery of drugs to specific target tissues and cells in the brain poses a significant challenge in brain therapeutics, primarily due to limited understanding of how nanoparticle (NP) properties influence drug biodistribution and off-target organ accumulation. This study addresses the limitations of previous research by using various predictive models based on collection of large data sets of 403 data points incorporating both numerical and categorical features. Machine learning techniques and comprehensive literature data analysis were used to develop models for predicting NP delivery to the brain. Furthermore, the physicochemical properties of loaded drugs and NPs were analyzed through a systematic analysis of pharmacodynamic parameters such as plasma area under the curve. The analysis employed various linear models, with a particular emphasis on linear mixed-effect models (LMEMs) that demonstrated exceptional accuracy. The model was validated via the preparation and administration of two distinct NP formulations via the intranasal and intravenous routes. Among the various modeling approaches, LMEMs exhibited superior performance in capturing underlying patterns. Factors such as the release rate and molecular weight had a negative impact on brain targeting. The model also suggests a slightly positive impact on brain targeting when the drug is a P-glycoprotein substrate.

ISTH bleeding assessment tool and platelet function analyzer in children with mild inherited platelet function disorders.

OBJECTIVES: To evaluate the diagnostic performance of platelet function analyzer (PFA) and The International Society on Thrombosis and Hemostasis bleeding-assessment-tool (ISTH-BAT) in detecting mild inherited platelet function disorders (IPFDs) in children with suspected bleeding disorders. METHODS: Prospective single-center diagnostic study including consecutive patients <18 years with suspected bleeding disorder and performing a standardized workup for platelet function defects including ISTH-BAT, PFA, platelet aggregation testing, blood smear-based immunofluorescence, and next-generation sequencing-based genetic screening for IPFDs. RESULTS: We studied 97 patients, of which 34 von Willebrand disease (VWD, 22 type-1, 11 type-2), 29 IPFDs (including delta-/alpha-storage pool disease, Glanzmann thrombasthenia, Hermansky-Pudlak syndrome) and 34 with no diagnosis. In a model combining PFA-adenosine diphosphate (ADP), PFA-epinephrine (EPI), and ISTH-BAT overall performance to diagnose IPFDs was low with area under the curves of 0.56 (95% CI 0.44, 0.69) compared with 0.84 (95% CI 0.76, 0.92) for VWD. Correlation of PFA-EPI/-ADP and ISTH-BAT was low with 0.25/0.39 Spearman's correlation coefficients. PFA were significantly prolonged in patients with VWD and Glanzmann thrombasthenia. ISTH-BAT-scores were only positive in severe bleeding disorders, but not in children with mild IPFDs or VWD. CONCLUSION: Neither ISTH-BAT nor PFA or the combination of both help diagnosing mild IPFDs in children. PFA is suited to exclude severe IPFDs or VWD and is in this regard superior to ISTH-BAT in children.

Nonparametric receiver operating characteristic curve analysis with an imperfect gold standard.

This article addresses the challenge of estimating receiver operating characteristic (ROC) curves and the areas under these curves (AUC) in the context of an imperfect gold standard, a common issue in diagnostic accuracy studies. We delve into the nonparametric identification and estimation of ROC curves and AUCs when the reference standard for disease status is prone to error. Our approach hinges on the known or estimable accuracy of this imperfect reference standard and the conditional independent assumption, under which we demonstrate the identifiability of ROC curves and propose a nonparametric estimation method. In cases where the accuracy of the imperfect reference standard remains unknown, we establish that while ROC curves are unidentifiable, the sign of the difference between two AUCs is identifiable. This insight leads us to develop a hypothesis-testing method for assessing the relative superiority of AUCs. Compared to the existing methods, the proposed methods are nonparametric so that they do not rely on the parametric model assumptions. In addition, they are applicable to both the ROC/AUC analysis of continuous biomarkers and the AUC analysis of ordinal biomarkers. Our theoretical results and simulation studies validate the proposed methods, which we further illustrate through application in two real-world diagnostic studies.

Addressing subject heterogeneity in time-dependent discrimination for biomarker evaluation.

Accurate discrimination has been the central goal in identifying biomarkers for monitoring disease progression and early detection. Acknowledging the fact that discrimination accuracy of biomarkers for a time-to-event outcome often changes over time, local measures such as the time-dependent receiver operating characteristic curve and its area under the curve (AUC) are used to assess time-dependent predictive discrimination. However, such measures do not address subject heterogeneity, although the impact of covariates including demographics, disease-related characteristics, and other clinical information on the discriminatory performance of biomarkers needs to be investigated before their clinical use. We propose the covariate-specific time-dependent AUC, a measure for covariate-adjusted discrimination. We develop a regression model on the covariate-specific time-dependent AUC to understand how and in what magnitude the covariates influence biomarker performance. Then we construct a pseudo partial-likelihood for estimation and inference. This is followed by our establishing the asymptotic properties of the proposed estimators and provide variance estimation. The simulation studies and application to the AIDS Clinical Trials Group 175 data demonstrate that the proposed method offers an informative tool for inferring covariate-specific and time-dependent predictive discrimination.

Predictive Value of Vancomycin AUC24/MIC Ratio for 30-day Mortality in Patients with Severe or Complicated Methicillin-Resistant Staphylococcus aureus Infections: A Multicenter Retrospective Study.

BACKGROUND: Although vancomycin is typically employed against methicillin-resistant Staphylococcus aureus (MRSA) infections, the optimal ratio of 24-h area under the concentration-time curve to minimum inhibitory concentration (AUC24/MIC) for severe or complicated infections lacks clear guideline recommendations. This study aimed to determine the target AUC24/MIC ratio associated with treatment outcomes of infections treated with vancomycin. METHODS: This retrospective multicenter cohort study included adult patients receiving ≥ 5 days of vancomycin for severe/complicated MRSA infections (e.g., osteoarticular, pulmonary, endocarditis, etc.) between January 2018 and December 2023. The primary outcome was 30-day mortality, with secondary outcomes including clinical success, microbiological eradication, and nephrotoxicity. Receiver operating characteristic (ROC) curve analysis was used to identify the AUC24/MIC cutoff for 30-day mortality. Multivariate regression analysis was used to determine association between AUC24/MIC and outcomes. RESULTS: This study included 82 patients. ROC identified a target AUC24/MIC of ≥ 505 for 30-day mortality. The overall 30-day mortality rate (22.0%) was significantly higher for below average AUC24/MIC cutoff (34.1%) than for above AUC24/MIC cutoff group (9.8%). Multivariate analysis confirmed AUC24/MIC of < 505 as an independent predictor (adjusted odds ratio, 5.001; 95% confidence interval, 1.335-18.75). The clinical success rate differed significantly between below- and above-cutoff groups, whereas microbiological eradication tended to favor the above-cutoff group. The nephrotoxicity rates were comparable between groups. CONCLUSIONS: In treating severe/complicated MRSA infections, vancomycin AUC24/MIC ratio ≥ 505 was independently associated with favorable 30-day mortality. Given the retrospective nature of this study, further prospective studies are essential to confirm the reliability of the target AUC24/MIC ratios.

Risk of Acute Kidney Injury Based on Vancomycin Target Trough Attainment Strategy: Area-Under-the-Curve-Guided Bayesian Software, Nomogram, or Trough-Guided Dosing.

BACKGROUND: Guidelines support area-under-the-curve (AUC) monitoring for vancomycin dosing which may lower overall doses and reduce acute kidney injury (AKI). OBJECTIVE: The aim of this study was to compare incidence of AKI across 3 vancomycin dosing modalities: AUC-targeted Bayesian pharmacokinetic software, AUC-targeted empiric dosing nomogram, and trough-guided dosing using clinical pharmacists' judgment. METHODS: This retrospective study included adult patients with a pharmacy dosing consult who received ≥1 dose of vancomycin and ≥1 serum vancomycin level documented between January 1, 2018, and December 31, 2019. Patients with baseline serum creatinine ≥2 mg/dL, weight ≥100 kg, receiving renal replacement therapy, AKI prior to vancomycin therapy, or vancomycin ordered only for surgical prophylaxis were excluded. The primary analysis was incidence of AKI adjusted for baseline serum creatinine, age, and intensive care unit admission. A secondary outcome was adjusted incidence of an abnormal trough value (<10 or >20 µg/mL). RESULTS: The study included 3459 encounters. Incidence of AKI was 21% for Bayesian software (n = 659), 22% for the nomogram (n = 303), and 32% for trough-guided dosing (n = 2497). Compared with trough-guided dosing, incidence of AKI was lower in the Bayesian (adjusted odds ratio [OR] = 0.72, 95% confidence interval [CI]: 0.58-0.89) and the nomogram (adjusted OR = 0.71, 95% CI: 0.53-0.95) groups. Compared with trough-guided dosing, abnormal trough values were less common in the Bayesian group (adjusted OR = 0.83, 95% CI: 0.69-0.98). CONCLUSION AND RELEVANCE: Study results suggest that use of AUC-guided Bayesian software reduces the incidence of AKI and abnormal trough values compared with trough-guided dosing.

Limitations of U25 CKiD and CKD-EPI eGFR formulae in patients 2-20 years of age with measured GFR > 60 mL/min/1.73 m2-a cross-sectional study.

BACKGROUND: When applying Pierce U25 formula for estimating glomerular filtration rate (eGFR), we observed a higher proportion of eGFR < 90 mL/min/1.73 m2 (chronic kidney disease (CKD) stage 2). We compared agreement and accuracy of the Pierce U25 (ages 2-25), Pottel (ages 2-100), and CKD-EPI (ages 18-100) formulae to GFR measurements. METHODS: Post hoc analysis of the three eGFRs compared to 367 99m technetium-diethylene-triamine penta-acetic acid (99Tc DTPA) GFR measurements (240 patients) using 3 sampling points and Brockner/Mørtensen correction (body surface area calculation based on ideal weight) on simultaneous serum creatinine and cystatin C measurements. RESULTS: Overall, the U25 formula performed well with a Spearman r of 0.8102 (95% confidence interval 0.7706 to 0.8435, p < 0.0001) while diagnostic accuracy was low in patients with normal mGFR. The U25 formula reclassified 29.5% of patients with normal mGFR as CKD stage 2; whereas the average of the modified Schwartz formula based on serum creatinine and the Filler formula based on cystatin C, only over-diagnosed CKD stage 2 in 8.5%, 24.5% within 10% and 62.7% within 30%. We therefore combined both. The average Schwartz/Filler eGFR had 36.5% of results within 10%, 84.7% within 30%, and normal mGFR accuracy was 26.8%, 63.9% for 10% and 30%, respectively, outperforming the CKD-EPI and Pottel formulae. CONCLUSIONS: The Pierce U25 formula results correlated well with mGFR < 75 mL/min/1.73 m2. Over the entire GFR range, accuracy was better for patients with a higher mGFR, when averaging the combined Schwartz/Filler formulae. More work is needed to prospectively confirm our findings in other centers.

Individualized Delivery of Vancomycin by Model-Informed Bayesian Dosing Approach to Maintain an AUC24 Target in Critically Ill Patients.

INTRODUCTION: Monitoring of AUC24 was updated recommendation in the guideline for the therapeutic drug monitoring (TDM) of vancomycin in Chinese pharmacological society published in 2020. Vancomycin pharmacokinetic profiles are diverse and unique in critically ill patients because of the drastic variability of the patients' physiological parameters, while the study for population pharmacokinetic (PPK) models in Chinese critically ill patients has been rarely reported. The objectives of this study were to construct a PPK model to describe the pharmacokinetic characteristics of vancomycin in critically ill patients and to individualize vancomycin dosing by model-informed Bayesian estimation for maintenance of AUC24 target at 400-650 mg h/L recommended by the 2020 guideline. METHODS: Vancomycin with different dosing was administered intravenously over 1 h for critically ill patients, TDM was started at 48 h or 72 h since initiation of vancomycin therapy for patients. Blood samples were collected from patients for trough concentrations or Cmax. Vancomycin concentrations were determined by high-performance liquid chromatography method with ultraviolet detection. PPK model was performed using the nonlinear mixed-effect model (NONMEM®). Individual PK parameters for critically ill patients treated with vancomycin were estimated using a post hoc empirical Bayesian method based on the final PPK model. AUC24 was calculated as the total daily dose divided by the clearance (L/h). RESULTS: The PPK of vancomycin was determined by a one-compartment model with creatinine clearance as fixed effects. The PK estimates in the final model generally agreed with the median estimates and were contained within the 95% CI generated from the bootstrap results, indicating good precision and stability in the final model. The visual predictive check plots showed the adequate predictive performance of the final PK model and supported a good model fit. The model-informed Bayesian estimation was used to predict the AUC24 of critically ill patient by the acquired TDM results, and the dosing adjustment by maintenance of AUC24 at 400-650 mg h/L had made a great therapeutic effect for the case. CONCLUSION: This study established a PPK model of vancomycin in Chinese critically ill patients, and individualized dosing of vancomycin by model-informed Bayesian estimation to maintain an AUC24 target at 400-650 mg h/L has been successfully applied in clinic. This result supports the continued use of model-informed Bayesian estimation to vancomycin treatment in critically ill patients.

Potential risk factors for early acute kidney injury in patients treated with vancomycin.

PURPOSE: The acute kidney injury (AKI) onset owing to vancomycin (VCM) is reported that depend on the area under the blood concentration-time curve (AUC) and occur comparison early phase (early AKI). This study aimed to investigate the occurrence of early AKI in patients treated with VCM and new indicators to avoid early AKI. METHODS: Adult patients who received VCM treatment for more than 4 days and whose trough values measured at least once on or after day 4 and serum creatinine before day 7 from the initiation of VCM administration between August 2021 and September 2022 at the Yamanashi Prefectural Central Hospital were enrolled. Early AKI (defined as AKI occurring within day 7 from VCM administration) and the association between each AUC (0-24, 24-48, 48-72, 0-48, 24-72, 0-72) were investigated. Furthermore, each AUC cut-off value for early AKI was calculated. RESULT: In total, 164 patients were enrolled; early AKI developed in 21 patients and most frequently occurred on day 4. All stratified AUC were associated with early AKI development. The AUC cut-off values were AUC0-24: 470.8 µg/mL⋅h; AUC24-48: 473.0 µg/mL⋅h; AUC48-72: 489.7 µg/mL⋅h; AUC0-48: 910.2 µg/mL⋅h; AUC24-72: 1039.2 µg/mL⋅h; and AUC0-72: 1544.0 µg/mL⋅h. CONCLUSION: The possibility of AKI development owing to the AUC accumulation of VCM was observed (accumulation toxicity). Concentration control through early-phase blood concentration measurements and a transition to AUC0-48 <910.2 µg/mL⋅h may reduce the early-phase AKI onset.

A novel estimator for the two-way partial AUC.

BACKGROUND: The two-way partial AUC has been recently proposed as a way to directly quantify partial area under the ROC curve with simultaneous restrictions on the sensitivity and specificity ranges of diagnostic tests or classifiers. The metric, as originally implemented in the tpAUC R package, is estimated using a nonparametric estimator based on a trimmed Mann-Whitney U-statistic, which becomes computationally expensive in large sample sizes. (Its computational complexity is of order [Formula: see text], where [Formula: see text] and [Formula: see text] represent the number of positive and negative cases, respectively). This is problematic since the statistical methodology for comparing estimates generated from alternative diagnostic tests/classifiers relies on bootstrapping resampling and requires repeated computations of the estimator on a large number of bootstrap samples. METHODS: By leveraging the graphical and probabilistic representations of the AUC, partial AUCs, and two-way partial AUC, we derive a novel estimator for the two-way partial AUC, which can be directly computed from the output of any software able to compute AUC and partial AUCs. We implemented our estimator using the computationally efficient pROC R package, which leverages a nonparametric approach using the trapezoidal rule for the computation of AUC and partial AUC scores. (Its computational complexity is of order [Formula: see text], where [Formula: see text].). We compare the empirical bias and computation time of the proposed estimator against the original estimator provided in the tpAUC package in a series of simulation studies and on two real datasets. RESULTS: Our estimator tended to be less biased than the original estimator based on the trimmed Mann-Whitney U-statistic across all experiments (and showed considerably less bias in the experiments based on small sample sizes). But, most importantly, because the computational complexity of the proposed estimator is of order [Formula: see text], rather than [Formula: see text], it is much faster to compute when sample sizes are large. CONCLUSIONS: The proposed estimator provides an improvement for the computation of two-way partial AUC, and allows the comparison of diagnostic tests/machine learning classifiers in large datasets where repeated computations of the original estimator on bootstrap samples become too expensive to compute.

A limited sampling model to estimate the area under the curve of mycophenolic acid in hematopoietic stem cell transplantation recipients.

Relationship between the areas under the curve (AUC) of mycophenolic acid (MPA) and the likelihood of rejection is well-established in solid organ transplantation recipients. In hematopoietic stem cell transplantation (HSCT), MPA AUC is also linked to graft versus host disease. This study aimed to develop a simplified method to estimate MPA AUC0-12 in Chinese patients undergoing allogeneic HSCT (allo-HSCT). Intensive sampling was conducted in 22 patients who were orally administered mycophenolate mofetil. Plasma concentrations of total MPA were measured, and a model predicting AUC0-12 using data from these 22 patients was constructed through regression analysis. The accuracy of the most suitable model was assessed in an additional 20 patients. None of the individual MPA concentrations showed a strong correlation with AUC0-12 (r2 < 0.7). Models utilizing 4 or more concentrations were found to effectively estimate MPA AUC0-12 (r2 > 0.87). The most operationally feasible model demonstrated good predictive performance with a mean absolute percentage error (APE%) < 20%. Single MPA concentrations showed poor correlation with MPA AUC0-12. A model utilizing 4 oral concentrations (0, 0.5, 1, and 4 h postdose) over a 12-h period could effectively estimate MPA AUC0-12 with precise results and minimal bias.