Empirical Bayesian Research Articles

Distinct Temporal Dynamic Serum Metabolite Profiles Characterize Treatment Response and Early Toxicity after B-Cell Lymphoma Directed CAR-T-Cell Therapy

Introduction: CD19 chimeric antigen receptor T-cell therapy (CD19.CAR-T) has profoundly improved treatment options for patients with relapsed/refractory B-cell lymphomas (r/r BCL), but is also associated with distinct toxicities. To elucidate host-intrinsic immunometabolic factors contributing to clinical benefit or immunotoxicity, we recently demonstrated that an increase of metabolically high-risk adipose tissue mass is associated with an elevated cytokine release syndrome (CRS) risk but also with improved survival (Haematologica 2022, CIR 2023). To gain further insights into the underlying immunometabolic mechanisms, we analyzed temporal dynamic changes of serum metabolites in r/r BCL patients during the first 14 days of CD19.CAR-T treatment. Methods: In this single-center analysis, we studied patients receiving CD19.CAR-T for r/r large BCL or mantle cell lymphoma (MCL) between 01/2019 and 12/2021 in a real-world setting. All patients were consecutively treated in our department (LMU University Hospital) and serum was prospectively collected in the morning hours from fasted patients beginning at the day of CAR-T infusion (D0), and 3-5 and 14 days later. Using liquid chromatography mass-spectrometry, we quantitatively analyzed a total of 305 water-soluble metabolites and 734 lipids. Metabolomic data was integrated with clinical patient records including CRS, 1-year progression-free survial (PFS) and amount of visceral adipose tissue (VAT). Results: In total, we measured 170 serum samples from 55 r/r LBCL and 4 r/r MCL patients, of which 52 had serial samples at all three time points (D0: 55, D3-5: 59, D14: 56). The following CAR-T products were administered: axi-cel (n = 26), tisa-cel (n = 29) and brexu-cel (n = 4). 25 (42.4%) patients developed a CRS grade ≥ 2, while 34 patients had a CRS grade 0-1. The best 90-day overall response rate was 52.5%, with a complete remission rate of 42.4%. At 1 year, 67.8% of patients progressed. From the 1039 metabolites measured, 525 metabolites passed quality control. Metabolite values were normalized and auto-scaled for further analysis. A partial least square differential analysis based on the collected time points led to a clear separation of samples with uric acid, desoxy-ribose-5-phosphate (dRibose-P), essential amino acids and (phospho)lipids having the strongest impact. Whereas 56 metabolites showed a significant increase from D0 to D14 led by uric and essential amino acids, 34 metabolites decreased over the course of treatment with the strongest reduction in dRibose-P and sphingosine phosphate. Performing a metabolic pathway enrichment analysis based on the most altered metabolite levels, we detected 22 significantly affected metabolic pathways and programs including aerobic glycolysis, urea cycle, amino acid and lipid metabolism (FDR < 0.1, Fig. 1A). To further delineate metabolite changes in regards to CRS, 1-year PFS and amount of VAT, we conducted a multivariate empirical bayes analysis of variance to specifically detect metabolites with differential temporal dynamics and considered all metabolites with an FDR < 0.1. Patients who developed a CRS ≥ 2° showed increased levels of kynurenine, phosphatidylethanolamides and lysophosphatidylserines, and decreased levels of ornithine and phosphatidylinositols. 1-year progressors showed a significant increase of several phosphatidylinositols beginning at D3-5 after CAR-T infusion, and decreased levels of phosphatidylcholines. For VAT, we used a threshold of 144 cm² which was derived from previously conducted computer tomography segmentation analyses. As expected, VAT cohorts mostly differed in fatty acid metabolism-associated metabolites including (acyl)carnitines, cholesterol esters, triglycerides and xanthine. In order to identify metabolites being associated with multiple outcomes, we looked for mutual metabolites among distinct cohort profiles. Here, one example is phosphatidylinositol-(22:5/16:0), which was increased in CRS 0-1° and 1y-progressors ( Fig. 1B). Conclusions: Our study highlights that CD19.CAR-T treatment is accompanied by broad changes in the serum metabolome of r/r BCL patients, which can be used to profile patients being affected by CAR-T-related immunotoxicity and their response to treatment. In ongoing analyses, we evaluate the use of these profiles for predictive models, and assess their impact on CAR-T cell function in-vitro.

Experience rating of risk premium for Esscher premium principle

In this article, a new method is introduced under the Bayesian framework to derive the credibility estimator of risk premiums based on the Esscher premium principle. This new estimator offers desirable statistical properties, making it more useful and practical compared to existing estimators. Additionally, Bayesian models for policy portfolios are established, and empirical Bayes methods are employed to estimate the structural parameters. The empirical Bayesian estimation of risk premiums is also discussed in detail. The convergence rate and goodness of the proposed estimators are verified through simulations. The results demonstrate the effectiveness and accuracy of the new estimator and its superior performance compared to other existing methods. Finally, an empirical analysis is conducted using real insurance data, which further confirms the applicability and reliability of the proposed credibility estimator and its superiority in practical insurance applications.

Robust small area estimation for unit level model with density power divergence.

Unit level model is one of the classical models in small area estimation, which plays an important role with unit information data. Empirical Bayesian(EB) estimation, as the optimal estimation under normal assumption, is the most commonly used parameter estimation method in unit level model. However, this kind of method is sensitive to outliers, and EB estimation will lead to considerable inflation of the mean square error(MSE) when there are outliers in the responses yij. In this study, we propose a robust estimation method for the unit-level model with outliers based on the minimum density power divergence. Firstly, by introducing the minimum density power divergence function, we give the estimation equation of the parameters of the unit level model, and obtain the asymptotic distribution of the robust parameters. Considering the existence of tuning parameters in the robust estimator, an optimal parameter selection algorithm is proposed. Secondly, empirical Bayesian predictors of unit and area mean in finite populations are given, and the MSE of the proposed robust estimators of small area means is given by bootstrap method. Finally, we verify the superior performance of our proposed method through simulation data and real data. Through comparison, our proposed method can can solve the outlier situation better.

Spatial distribution of suicides in Mato Grosso, Brazil, from 2008 to 2018

In the last few decades, it is believed that the state of Mato Grosso has been the stage of a substantial increase in suicide rates. However, the extension of this issue in the health regions of the state is still unknown. Therefore, the goal of this study was to analyze the spatial distribution of suicides in Mato Grosso, Brazil, from 2008 to 2018. This is an ecological study, with spatial-temporal distribution. Data collection took place in 2020 and was carried out in the Information System on Mortality for the period from January 2008 to December 2018. The suicide rates were calculated and smoothed using the empirical Bayes method, aided by the software GeoDa 1.12. To construct the maps, we used the software ArcGis 10.5. There were 1,977 recorded suicides in Mato Grosso, especially in males (79.3%), single (50.6%), from 20 to 29 years old (23.7%), brown (55.4%), with low educational levels (52.2%). Most suicides were carried out by intentional self-harm by hanging, strangulation and suffocation in the residence. The spatial-temporal distribution showed that the mesoregions North and Northeast had the highest rates, especially in cities in the health regions of Mid-Araguaia (for both sexes) and Garças Araguaia (for males). Suicides have followed a random, unequal pattern throughout Mato Grosso, but focused in areas of little development, showing the need to invest in health, and to strengthen and value policies to bring human resources to attend to these demands. The health care network must be structured to do so.

Abstract 16208: Population Pharmacokinetics of Intravenously and Orally Administered Sotalol in Patients With Cardiac Arrhythmias

Introduction: Recently, an intravenous (IV) formulation of sotalol was approved to provide a loading dose before initiating oral sotalol therapy to reduce hospitalization days. However, the population pharmacokinetics (PK) of sotalol after administration of an IV loading dose regimen in patients with atrial fibrillation were not characterized. Understanding the PK of IV sotalol loading dose in patients with atrial fibrillation is required to ensure safety and optimal therapeutic outcomes. Hypothesis: The population PK model of IV and orally administered sotalol can be developed using data from two clinical PK studies with reasonable empirical Bayes estimates (EBEs) of PK parameters absorption rate constant (Ka), clearance (CL), and volume of distribution (Vd). Methods: The PK dataset of sotalol was from: 1) Relative bioavailability study with IV and oral sotalol PK data by sponsor AltaThera Inc., 2) Prospective Evaluation Analysis and Kinetics of IV Sotalol (PEAKS) registry study. Population PK was evaluated using NONMEM® 7.5, Perl-speaks-NONMEM (PsN®) version 5.3.0. Results: 545 sotalol concentrations in 25 patients were analyzed. A one-compartment model with a first-order elimination rate best described the data (Figure 1). The data correlated QTc interval levels within the observed sotalol concentration range (1.2 to 1900 ng/mL) were below 500 ms. In the final model, the EBE’s of sotalol PK parameters were, Ka - 1.44 hr -1 , CL 0.46 - L/h/kg, and the Vd -5.6 L/kg. Further PK/PD studies are needed to establish the correlation between sotalol PK and safety in this population. Conclusions: The PK of sotalol after IV and oral doses were adequately described using a population PK modeling approach. The QTc levels correlated to the plasma concentrations in this study were within the safety limit of 500 ms. This population PK model could be used to correlate the PK with safety (QTc interval prolongation) further to optimize the sotalol IV and oral dosing regimen.

Using Auxiliary Item Information in the Item Parameter Estimation of a Graded Response Model for a Small to Medium Sample Size: Empirical Versus Hierarchical Bayes Estimation.

Marginal maximum likelihood estimation (MMLE) is commonly used for item response theory item parameter estimation. However, sufficiently large sample sizes are not always possible when studying rare populations. In this paper, empirical Bayes and hierarchical Bayes are presented as alternatives to MMLE in small sample sizes, using auxiliary item information to estimate the item parameters of a graded response model with higher accuracy. Empirical Bayes and hierarchical Bayes methods are compared with MMLE to determine under what conditions these Bayes methods can outperform MMLE, and to determine if hierarchical Bayes can act as an acceptable alternative to MMLE in conditions where MMLE is unable to converge. In addition, empirical Bayes and hierarchical Bayes methods are compared to show how hierarchical Bayes can result in estimates of posterior variance with greater accuracy than empirical Bayes by acknowledging the uncertainty of item parameter estimates. The proposed methods were evaluated via a simulation study. Simulation results showed that hierarchical Bayes methods can be acceptable alternatives to MMLE under various testing conditions, and we provide a guideline to indicate which methods would be recommended in different research situations. R functions are provided to implement these proposed methods.

Variation in racial disparities in police use of force

I examine how racial disparities in police use of force vary using new data covering every municipal police department in New Jersey. Along the intensive margin of force severity, I find disparities that disfavor Black subjects and are larger at higher force levels, even after adjusting for incident-level factors and using new techniques to address selection bias. I then extend empirical Bayes methods to estimate department-specific racial disparities and observe significant differences across and within these hundreds of departments. My findings suggest that ignoring heterogeneity in police use of force misrepresents the problem and masks the existence of both departments with very large disparities and those without apparent disparities against Black civilians, but the variation even within departments may make identifying and treating inequitable policing difficult.

The Impact of Social Inclusion and Financial Development on CO2 Emissions: Panel Analysis from Developing Countries

The intricate interplay between the environment and the economy entails numerous multifaceted factors that require thorough investigation. Civic activism, intergroup cohesion, and gender equality are among the pertinent factors that hold the potential to significantly impact CO2 emissions in developing economies. However, these variables have not been explored to the extent that their importance warrants, leaving much to be studied and understood about their complex relationships with carbon emissions. Currently, developing nations find themselves more vulnerable and exposed to a plethora of environmental issues. In response to this pressing matter, the focus of this study is to expound upon the impact of various factors on the environment. To achieve this aim, this study utilizes annual data from 46 developing countries, spanning the extensive period from 1990 to 2014. Using the generalized method of moments and empirical Bayes methods, this study’s results emphasize the significant impact that civic activism, gender equality, intergroup cohesion, and financial development can have on increasing CO2 emissions. However, civic activism reduces CO2 emissions. These findings highlight the crucial importance of adopting a comprehensive approach that accounts for both economic and social cohesion indicators when tackling environmental challenges.

Cost free hyper-parameter selection/averaging for Bayesian inverse problems with vanilla and Rao-Blackwellized SMC samplers

In Bayesian inverse problems, one aims at characterizing the posterior distribution of a set of unknowns, given indirect measurements. For non-linear/non-Gaussian problems, analytic solutions are seldom available: Sequential Monte Carlo samplers offer a powerful tool for approximating complex posteriors, by constructing an auxiliary sequence of densities that smoothly reaches the posterior. Often the posterior depends on a scalar hyper-parameter, for which limited prior information is available. In this work, we show that properly designed Sequential Monte Carlo (SMC) samplers naturally provide an approximation of the marginal likelihood associated with this hyper-parameter for free, i.e. at a negligible additional computational cost. The proposed method proceeds by constructing the auxiliary sequence of distributions in such a way that each of them can be interpreted as a posterior distribution corresponding to a different value of the hyper-parameter. This can be exploited to perform selection of the hyper-parameter in Empirical Bayes (EB) approaches, as well as averaging across values of the hyper-parameter according to some hyper-prior distribution in Fully Bayesian (FB) approaches. For FB approaches, the proposed method has the further benefit of allowing prior sensitivity analysis at a negligible computational cost. In addition, the proposed method exploits particles at all the (relevant) iterations, thus alleviating one of the known limitations of SMC samplers, i.e. the fact that all samples at intermediate iterations are typically discarded. We show numerical results for two distinct cases where the hyper-parameter affects only the likelihood: a toy example, where an SMC sampler is used to approximate the full posterior distribution; and a brain imaging example, where a Rao-Blackwellized SMC sampler is used to approximate the posterior distribution of a subset of parameters in a conditionally linear Gaussian model.

Long-term trends of local bird populations based on monitoring schemes: are they suitable for justifying management measures?

Local biodiversity monitoring is important to assess the effects of global change, but also to evaluate the performance of landscape and wildlife protection, since large-scale assessments may buffer local fluctuations, rare species tend to be underrepresented, and management actions are usually implemented on local scales. We estimated population trends of 58 bird species using open-population N-mixture models based on count data in two localities in southeastern Spain, which have been collected according to a citizen science monitoring program (SACRE, Monitoring Common Breeding Birds in Spain) over 21 and 15 years, respectively. We performed different abundance models for each species and study area, accounting for imperfect detection of individuals in replicated counts. After selecting the best models for each species and study area, empirical Bayes methods were used for estimating abundances, which allowed us to calculate population growth rates (λ) and finally population trends. We also compared the two local population trends and related them with national and European trends, and species functional traits (phenological status, dietary, and habitat specialization characteristics). Our results showed increasing trends for most species, but a weak correlation between populations of the same species from both study areas. In general, local population trends were consistent with the trends observed at national and continental scales, although contrasting patterns exist for several species, mainly with increasing local trends and decreasing Spanish and European trends. Moreover, we found no evidence of a relationship between population trends and species traits. We conclude that using open-population N-mixture models is an appropriate method to estimate population trends, and that citizen science-based monitoring schemes can be a source of data for such analyses. This modeling approach can help managers to assess the effectiveness of their actions at the local level in the context of global change.

Utilizing Empirical Bayes Estimation to Assess Reliability in Inverted Exponentiated Rayleigh Distribution with Progressive Hybrid Censored Medical Data

This study addresses the issue of estimating the shape parameter of the inverted exponentiated Rayleigh distribution, along with the assessment of reliability and failure rate, by utilizing Type-I progressive hybrid censored data. The study explores the estimators based on maximum likelihood, Bayes, and empirical Bayes methodologies. Additionally, the study focuses on the development of Bayes and empirical Bayes estimators with balanced loss functions. A concrete example based on actual data from the field of medicine is used to illustrate the theoretical insights provided in this study. Monte Carlo simulations are employed to conduct numerical comparisons and evaluate the performance and accuracy of the estimation methods.

CGAN-EB: A non-parametric empirical Bayes method for crash frequency modeling using conditional generative adversarial networks as safety performance functions

In this paper, a new non-parametric empirical Bayes approach called CGAN-EB is proposed for approximating empirical Bayes (EB) estimates in traffic locations (e.g., road segments) which benefits from the modeling advantages of deep neural networks, and its performance is compared in a simulation study with the traditional approach based on negative binomial model (NB-EB). The NB-EB uses negative binomial model in order to model the crash data and is the most common approach in practice. To model the crash data in the proposed CGAN-EB, conditional generative adversarial network is used, which is a powerful deep neural network based method that can model any types of distributions. A number of simulation experiments are designed and conducted to evaluate the CGAN-EB performance in different conditions and compare it with the NB-EB. The results show that CGAN-EB performs as well as NB-EB when conditions favor the NB-EB model (i.e. data conform to the assumptions of the NB model) and outperforms NB-EB in experiments reflecting conditions frequently encountered in practice, specifically low sample means, and when crash frequency does not follow a log-linear relationship with covariates.

Sequential Bayesian Ability Estimation Applied to Mixed-Format Item Tests.

Large-scale tests often contain mixed-format items, such as when multiple-choice (MC) items and constructed-response (CR) items are both contained in the same test. Although previous research has analyzed both types of items simultaneously, this may not always provide the best estimate of ability. In this paper, a two-step sequential Bayesian (SB) analytic method under the concept of empirical Bayes is explored for mixed item response models. This method integrates ability estimates from different item formats. Unlike the empirical Bayes method, the SB method estimates examinees' posterior ability parameters with individual-level sample-dependent prior distributions estimated from the MC items. Simulations were used to evaluate the accuracy of recovery of ability and item parameters over four factors: the type of the ability distribution, sample size, test length (number of items for each item type), and person/item parameter estimation method. The SB method was compared with a traditional concurrent Bayesian (CB) calibration method, EAPsum, that uses scaled scores for summed scores to estimate parameters from the MC and CR items simultaneously in one estimation step. From the simulation results, the SB method showed more accurate and reliable ability estimation than the CB method, especially when the sample size was small (150 and 500). Both methods presented similar recovery results for MC item parameters, but the CB method yielded a bit better recovery of the CR item parameters. The empirical example suggested that posterior ability estimated by the proposed SB method had higher reliability than the CB method.

Fine-grained cell-type specific association studies with human bulk brain data using a large single-nucleus RNA sequencing based reference panel

Brain disorders are leading causes of disability worldwide. Gene expression studies provide promising opportunities to better understand their etiology but it is critical that expression is studied on a cell-type level. Cell-type specific association studies can be performed with bulk expression data using statistical methods that capitalize on cell-type proportions estimated with the help of a reference panel. To create a fine-grained reference panel for the human prefrontal cortex, we performed an integrated analysis of the seven largest single nucleus RNA-seq studies. Our panel included 17 cell-types that were robustly detected across all studies, subregions of the prefrontal cortex, and sex and age groups. To estimate the cell-type proportions, we used an empirical Bayes estimator that substantially outperformed three estimators recommended previously after a comprehensive evaluation of methods to estimate cell-type proportions from brain transcriptome data. This is important as being able to precisely estimate the cell-type proportions may avoid unreliable results in downstream analyses particularly for the multiple cell-types that had low abundances. Transcriptome-wide association studies performed with permuted bulk expression data showed that it is possible to perform transcriptome-wide association studies for even the rarest cell-types without an increased risk of false positives.

Associations of active-duty mental health trajectories with post-military adjustment: Results from the STARRS Longitudinal Study

BackgroundMany servicemembers experience difficulties transitioning from military to civilian life. We examined whether changes in mental health observed during active duty were associated with indices of post-military adjustment. MethodsSurvey data from the multi-wave Army STARRS Pre/Post Deployment Study (PPDS; conducted 2012–2014) were linked to follow-up data from wave 1 of the STARRS Longitudinal Study (STARRS-LS1; conducted 2016–2018). Empirical Bayes estimates of intercepts and slopes of posttraumatic stress, problematic anger, and depressive symptoms during the PPDS were extracted from mixed-effects growth models and evaluated as predictors of life stress among 1080 participants who had separated or retired from the Army at STARRS-LS1; and of job satisfaction among 586 veterans who were employed at STARRS-LS1. ResultsHigher average levels and larger increases in posttraumatic stress, anger, and depression over the deployment period were each associated with increased stress and (in the case of anger and depression) reduced job satisfaction. Posttraumatic stress and anger slopes were associated with overall stress (b = 5.60, p < 0.01 and b = 15.64, p = 0.04, respectively) and relationship stress (b = 5.50, p = 0.01 and b = 22.86, p = 0.01, respectively) beyond the average levels of those symptoms. LimitationsSome transition-related difficulties may have resolved before outcome assessment; some measures were not previously validated. ConclusionsLarger increases in posttraumatic stress and anger over a deployment period were associated with increased stress after leaving the Army, even after controlling for average symptom levels during the same period. Monitoring changes in mental health during active duty may help identify personnel who need additional support to facilitate the military-to-civilian transition.

Safety Effectiveness of Rectangular Rapid Flashing Beacons Pedestrian Enhancement

A rectangular rapid flashing beacon (RRFB) is a pedestrian crosswalk countermeasure system that cautions drivers by providing them with real-time warning about the presence of pedestrians in an upcoming crosswalk. This paper investigated the safety effectiveness of existing RRFBs installed in the state of Florida between 2013 and 2018. Additionally, traffic volume, roadway characteristics, and land-use mix were included as explanatory variables. The study evaluated 154 treated sites with RRFB installations and 158 control sites with similar characteristics. Safety performance functions (SPF) were developed, and crash modification factors (CMF) were calculated using the empirical Bayes (EB) methodology for various crash types. A CMF of 0.31 for total pedestrian crashes was estimated, showing that RRFBs have the potential to reduce 69% of total pedestrian-involved crashes. CMF for total vehicular crashes and rear-end crashes were also investigated to ascertain whether RRFBs have adverse effects on rear-end crashes. CMF for total vehicular crashes was found to be 1.14, showing that RRFBs do not have positive effect in reducing total vehicular crashes. Similarly, it was observed that CMF for the rear-end crashes was 1.11, indicating RRFBs might have adverse effect on rear-end crashes because vehicles needs to slow down or stop due to the presence of pedestrians. CMFunctions for significant variables for pedestrian, rear-end, and total crash models were also estimated in this study. An economic analysis indicated that the implementation of RRFBs will yield a benefit to cost ratio (calculated over 3 years) of approximately 85.5 with a 4% discount rate.

A Propensity-Score Integrated Approach to Bayesian Dynamic Power Prior Borrowing

Use of historical control data to augment a small internal control arm in a randomized control trial (RCT) can lead to significant improvement of the efficiency of the trial. It introduces the risk of potential bias, since the historical control population is often rather different from the RCT. Power prior approaches have been introduced to discount the historical data to mitigate the impact of the population difference. However, even with a Bayesian dynamic borrowing which can discount the historical data based on the outcome similarity of the two populations, a considerable population difference may still lead to a moderate bias. Hence, a robust adjustment for the population difference using approaches such as the inverse probability weighting or matching, can make the borrowing more efficient and robust. In this article, we propose a novel approach integrating the propensity score for the covariate adjustment and Bayesian dynamic borrowing using power prior. The proposed approach uses Bayesian bootstrap in combination with the empirical Bayes (EB) method using quasi-likelihood for determining the power prior. The performance of our approach is examined by a simulation study. We apply the approach to two Acute Myeloid Leukemia (AML) studies for illustration.

Corridor-level network screening and modeling of fatal and serious injury crashes on urban and suburban arterial corridors in Florida

Network screening is a key aspect of the safety management process. The current standard methods in the Highway Safety Manual assume that intersections and roadway segments are independent, even though interactions between sites likely exist. One way to account for these interactions is by studying roadway corridors made of consecutive intersections and segments. This paper developed a roadway corridor definition to identify urban and suburban arterial corridors in Florida with context classification and lane count as defining features. Applying this definition to 10 Florida counties resulted in 1105 corridors with 14,097 fatal or serious injury crashes from 2017 to 2021. Historic crashes and a developed regression model were used to estimate the expected crash frequency for fatal and serious injury crashes with the empirical Bayes method. After performing k-means clustering on expected crash frequency per mile, 25 high-risk corridors were identified. The difference between the expected and predicted crash frequencies (potential for safety improvement) for each high-risk corridor provided a list of priority corridors for safety improvements. Countermeasures focused on speed reduction in suburban corridors would likely lead to the greatest reductions in fatal and serious injury crashes. This corridor approach allows agencies to quickly identify high-risk areas for priority treatment.

Population Pharmacokinetics and Exposure-Response Analysis of Tremelimumab 300mg Single Dose Combined with Durvalumab 1500mg Q4W (STRIDE) in Patients with Unresectable Hepatocellular Carcinoma.

A novel single-dose regimen of 300mg tremelimumab in combination with durvalumab (STRIDE) has demonstrated a favorable benefit-risk profile in the phase1/2 Study22 trial (in patients with unresectable hepatocellular carcinoma, uHCC) and in the phase3 HIMALAYA study. The current analysis evaluated the population pharmacokinetics (PopPK) of tremelimumab and durvalumab, and the exposure-response (ER) relationship for efficacy and safety of STRIDE in patients with uHCC. Previous PopPK models for tremelimumab and durvalumab were updated using data from previous studies in various cancers combined with data from Study22 and HIMALAYA. Typical population mean parameters and associated inter- and intra-individual variability were assessed, as was the influence of covariates. Individual exposure metrics were derived from the individual empirical Bayes estimates as drivers for ER analysis related to efficacy and safety from HIMALAYA. The observed pharmacokinetics of tremelimumab in uHCC were well described by a 2-compartment model with both linear and time-dependent clearance. All identified covariates changed tremelimumab PK parameters by <25%, and thus had minimal clinical relevance; similar results were obtained from durvalumab PopPK analysis. None of tremelimumab or durvalumab exposure metrics were significantly associated with overall survival (OS), progression-free survival (PFS), or adverse events. Baseline aspartate aminotransferase and neutrophil-to-lymphocyte ratio (NLR) were associated with OS (P<.001) by the Cox proportional hazards model. No covariate was identified as a significant factor for PFS. No dose adjustment for tremelimumab or durvalumab is needed based on PopPK covariate analyses or ER analyses. Our findings support the novel STRIDE dosing regimen in patients with uHCC.

Pharmacodynamic Model-Based Safety Management of Eribulin-Induced Myelosuppression in Patients With Breast Cancer.

Neutropenia is a major dose-limiting toxicity of cancer chemotherapy. Semimechanistic mathematical models have been applied to describe the time course of neutrophil counts. The objectives of this study were to develop a mathematical model describing changes in neutrophil counts during eribulin treatment, to apply the empirical Bayes method to predict the probability of developing neutropenia ≥ grade 3 during eribulin treatment in each patient, and to propose the implementation of this mathematical tool in clinical practice for individual safety management. The present model analysis and subsequent external evaluation were performed using the data of 481 patients with breast cancer, previously obtained from a postmarketing surveillance (training set) and a phase 2 clinical study (validation set). The model we previously reported (Kawamura et al 2018) was modified to improve its predictive capability. The individual time course of neutrophil changes during the treatment period was predicted by the empirical Bayes method using the observed neutrophil counts at baseline and the first measurement after the first eribulin dose. To evaluate the predictability of this method, the predicted neutrophil counts were compared with those of the observed values. The developed model provided good individual predictions, as indicated by the goodness-of-fit plots between the predicted and observed neutrophil counts, especially for a lower neutrophil count range. Days required to reach the nadir after the dose were also well-predicted. The sensitivity, specificity, and accuracy of the prediction of neutropenia grade ≥3 were 76%, 53%, and 71%, respectively. We developed a mathematical method for predicting and managing the risk of neutropenia during eribulin treatment. This method is generally applicable to other cases of chemotherapy-induced neutropenia and can be a new practical tool for individual safety management.