Study Partner Type and Adverse Event Reporting in Mild-to-Moderate Alzheimer's Disease Clinical Trials.

Background: In randomized clinical trials (RCTs), monitoring adverse events (AEs) and serious AEs (SAEs) is critical. All Alzheimer's disease (AD) RCTs require participants to enroll with a study partner. Objective: We examined AE reporting rates in mild-to-moderate AD trials and their associations with study partner type. Methods: We estimated AE reporting rates using placebo data from seven independent RCTs conducted by the Alzheimer's Disease Cooperative Study. We assessed the heterogeneity of reporting rates as a function of visits using generalized estimating equations. In the primary analysis, we tested the hypotheses that the rates of reporting differed by study partner type and time they spent with the participant weekly using Poisson regression with robust variance estimation. In all regression models, log-transformed total patient years was included. Results: The estimated reporting rates were 2.83 (95% CI: 2.66, 3.02), 1.18 (95% CI: 1.09, 1.28), 0.23 (95% CI: 0.19, 0.27), and 0.28 (95% CI: 0.24, 0.33) events per participant year for grade 1-3 AEs and SAEs, respectively. We estimated that greater number of visits per year was associated with increased reporting for grade 1-2 AEs and SAEs. We did not find evidence to suggest that AE reporting differed by study partner type or by time the study partner spent with the participant. Conclusions: Study partner type and time the study partner spent with the participant did not appear to impact AE reporting. Estimated reporting rates may be useful to evaluate safety in future studies, particularly those with no control arm and similar visit frequencies.

Post-disclosure distress among racial and ethnic groups in a preclinical AD trial.

INTRODUCTION: Trialists need a thorough understanding of whether reactions to Alzheimer's disease (AD) biomarker information differ among racial and ethnic groups in preclinical AD trials. METHODS: We used data from the Anti-Amyloid Treatment in Asymptomatic Alzheimer's Disease Study to analyze cognitively unimpaired participants' responses on the Impact of Event Scale (IES) 24 to 72 hours after amyloid disclosure. We fit a linear regression model to test whether mean IES scores differed among participants from specific racial and ethnic groups. We considered potential effect modification by amyloid status. RESULTS: Reactions to disclosure did not significantly differ among participant groups based on self-reported race and ethnicity. Although the results were not significant when stratified by amyloid status, all racial and ethnic groups except for participants self-reporting Hispanic/Latino ethnicity were observed to have higher mean IES in the elevated amyloid group. DISCUSSION: These results support continued use of current disclosure methods in preclinical AD trials.

Retention of American Indian and Alaska Native participants in the National Alzheimer's Coordinating Center Uniform Data Set.

INTRODUCTION: The number of American Indian and Alaska Native (AI/AN) elders is expected to double by 2060. Thus it is imperative to retain AI/AN participants in longitudinal research studies to identify novel risk factors and potential targets for intervention for Alzheimer's disease and related dementias in these communities. METHODS: The National Alzheimer's Coordinating Center houses uniformly collected longitudinal data from the network of National Institute on Aging (NIA)-funded Alzheimer's Disease Research Centers (ADRCs). We used logistic regression to quantify participant retention at 43 ADRCs, comparing self-identified AI/AN participants to non-Hispanic White (NHW) participants, adjusting for potential confounding factors including baseline diagnosis, age, sex, education, and smoking. RESULTS: The odds of AI/AN participant retention at the first follow-up visit were significantly lower than those for NHW participants (adjusted odds ratio [aOR]: 0.599; 95%: 0.46-0.78; p < 0.001). DISCUSSION: These results suggest the need for improved strategies to retain AI/AN participants, perhaps including improved researcher-community relationships and community engagement and education. HIGHLIGHTS: American Indian and Alaska Native (AI/AN) research participants were retained to the first follow-up appointment at lower rates than non-Hispanic White (NHW) participants. AI/AN participants are retained at lower rates than NHW participants for long-term follow-up. The majority of AI/AN participants were not retained to the second follow-up visit.

Effects of informant replacement in Alzheimer's disease clinical trials.

INTRODUCTION: Alzheimer's disease (AD) trials require enrollment with an informant. METHODS: We assessed relationships between informant replacement and Alzheimer's Disease Cooperative Study Activities of Daily Living (ADCS-ADL) scores across four AD trials. Using generalized estimating equations, we examined associations between replacement and change in ADCS-ADL between successive visits. We used analysis of covariance to estimate the association between replacement and 18-month change from baseline, and an F-test to compare the variance of this change. RESULTS: Among 1336 participants, 63 (≈5%) experienced replacement. Between-visit mean change in ADCS-ADL was 2.44 points lower comparing replacement to stable informants (95% confidence interval [CI]: -3.91, -0.98). The difference in between-visit mean absolute change was 2.38 points (95% CI: 1.24, 3.52). Replacement was not significantly associated with an 18-month change from baseline. The ratio of variances (replacement/stable) was 1.80 (95% CI: 1.19, 2.99). DISCUSSION: Informant replacement is associated with bias and increased variability between visits and increased variance for overall ADCS-ADL.

Estimating attrition in mild-to-moderate Alzheimer's disease and mild cognitive impairment clinical trials.

BACKGROUND: Participant retention is a key factor that affects clinical trial integrity. Trial protocols estimate attrition as a function of sample size calculations. Alzheimer's disease (AD) is an area of active treatment development. We aimed to quantify the association between trial duration and completion rates and provide guidance for estimating attrition in AD trial protocols. METHODS: Using the Alzforum and ClinicalTrials.gov databases, we analyzed retention data from 125 mild-to-moderate AD and 12 mild cognitive impairment (MCI) clinical trials. We compared the rates of completion between trial arms (active vs. control) and ran regression models to test the hypothesis that trials with longer study duration have lower trial completion using all available data and restricting to placebo data. Our primary outcome was the odds of trial completion for a 6-month increase in trial duration. From the regression model, we estimated the proportion of participants completing 6-, 12-, and 18-month trials. RESULTS: We found that 21 (17%) mild-to-moderate AD trials and 1 (8%) MCI trial demonstrated greater dropout in treatment compared to placebo arms. For every 6-month increase in trial duration, there was a 27% decrease in the odds of trial completion (OR = 0.73; 95% CI 0.66, 0.81; p < 0.001) among participants in mild-to-moderate AD trials and a 55% decrease (OR = 0.45; 95% CI 0.36, 0.57; p < 0.001) among participants in MCI trials. The proportion of participants in the placebo group completing 6-, 12-, and 18-month trials were estimated to be 85.2%, 80.0%, and 73.3% for mild-to-moderate AD trials and 91.9%, 84.2%, and 71.3% for MCI trials, respectively. CONCLUSIONS: Longer duration trials may be underpowered to demonstrate estimated treatment effects and may suffer from a greater risk of bias than do shorter trials.

Frameworks for estimating causal effects in observational settings: comparing confounder adjustment and instrumental variables.

To estimate causal effects, analysts performing observational studies in health settings utilize several strategies to mitigate bias due to confounding by indication. There are two broad classes of approaches for these purposes: use of confounders and instrumental variables (IVs). Because such approaches are largely characterized by untestable assumptions, analysts must operate under an indefinite paradigm that these methods will work imperfectly. In this tutorial, we formalize a set of general principles and heuristics for estimating causal effects in the two approaches when the assumptions are potentially violated. This crucially requires reframing the process of observational studies as hypothesizing potential scenarios where the estimates from one approach are less inconsistent than the other. While most of our discussion of methodology centers around the linear setting, we touch upon complexities in non-linear settings and flexible procedures such as target minimum loss-based estimation and double machine learning. To demonstrate the application of our principles, we investigate the use of donepezil off-label for mild cognitive impairment. We compare and contrast results from confounder and IV methods, traditional and flexible, within our analysis and to a similar observational study and clinical trial.

Recruitment across two decades of NIH-funded Alzheimer's disease clinical trials.

BACKGROUND: Timely accrual of a representative sample is a key factor in whether Alzheimer's disease (AD) clinical trials successfully answer the scientific questions under study. Studies in other fields have observed that, over time, recruitment to trials has become increasingly reliant on larger numbers of sites, with declines in the average per-site recruitment rate. Here, we examined the trends in recruitment over a 20-year period of NIH-funded AD clinical trials conducted by the Alzheimer's Disease Cooperative Study (ADCS), a temporally consistent network of sites devoted to interventional research. METHODS: We performed retrospective analyses of eleven ADCS randomized clinical trials. To examine the recruitment planning, we calculated the expected number of participants to be enrolled per site for each trial. To examine the actual trial recruitment rates, we quantified the number of participants enrolled per site per month. RESULTS: No effects of time were observed on recruitment planning or overall recruitment rates across trials. No trial achieved an overall recruitment rate greater than one subject per site per month. We observed the fastest recruitment rates in trials with no competition and the slowest in trials that overlapped in time. The highest recruitment rates were consistently seen early within trials and declined over the course of studies. CONCLUSIONS: Trial recruitment projections should plan for fewer than one participant randomized per site per month and consider the number of other AD trials being conducted concurrently.

Maternal insulin resistance in pregnancy is associated with fetal fat deposition: findings from a longitudinal study.

BACKGROUND: Newborns exhibit substantial variation in fat mass accretion over gestation. These individual differences in newborn adiposity extend into infancy and childhood and relate to subsequent risk of obesity and metabolic dysregulation. Maternal glucose homeostasis in pregnancy has been proposed as an underlying mechanism; however, the timing in gestation when maternal glucose regulation influences the progression of fetal fat deposition remain unclear. OBJECTIVE: This study aimed to investigate the cross-sectional and longitudinal association of maternal insulin resistance in early, mid, and late pregnancy with fetal fat deposition in uncomplicated pregnancies. We hypothesized that maternal insulin resistance at early, mid, and late gestation is positively associated with fetal fat deposition, and that the magnitude of the association is greater for the mid and late gestation measures than for the early gestation measure. STUDY DESIGN: In a longitudinal study of 137 low-risk pregnancies, a fasting maternal blood sample was obtained and fetal ultrasonography was performed at ≈ 12, 20, and 30 weeks' gestation. Maternal insulin resistance was quantified using the homeostasis model assessment of insulin resistance (fasting insulin×fasting glucose/405). Estimated fetal adiposity was calculated by integrating measurements of cross-sectional arm and thigh percentage fat area and anterior abdominal wall thickness. The associations between maternal homeostasis model assessment of insulin resistance and estimated fetal adiposity and estimated fetal weight were determined by multiple linear regression adjusted for potential confounding factors including maternal age, parity, race and ethnicity, prepregnancy body mass index, gestational weight gain per week, fetal sex, and gestational age at assessments. RESULTS: Maternal homeostasis model assessment of insulin resistance at ≈ 12, 20, and 30 weeks was 2.79±1.79 (±standard deviation), 2.78±1.54, and 3.76±2.30, respectively. Homeostasis model assessment of insulin resistance at 20 weeks was positively associated with estimated fetal adiposity at 20 weeks (r=0.261; P=.005). Homeostasis model assessment of insulin resistance at 20 weeks (r=0.215; P=.011) and 30 weeks (r=0.285; P=.001) were also positively associated with estimated fetal adiposity at 30 weeks. These relationships remained significant after adjustment for confounding factors. There was no significant correlation between homeostasis model assessment of insulin resistance and estimated fetal weight at 20 and 30 weeks' gestation. CONCLUSION: In low-risk pregnancies, maternal insulin resistance at mid and late but not early pregnancy is significantly associated with fetal adiposity but not with fetal weight. Maternal insulin resistance in mid-gestation could provide a basis for risk identification and interventions that target child adiposity.

Post-discharge decolonization of patients harboring methicillin-resistant Staphylococcus aureus (MRSA) USA300 strains: secondary analysis of the CLEAR Trial.

The CLEAR Trial recently found that decolonization reduced infections and hospitalizations in MRSA carriers in the year following hospital discharge. In this secondary analysis, we explored whether decolonization had a similar benefit in the subgroup of trial participants who harbored USA300, using two different definitions for the USA300 strain-type.

Effect of Aducanumab Approval on Willingness to Participate in Preclinical Alzheimer's Disease Trials.

BACKGROUND: Clinical trials now test promising therapies in the preclinical stages of Alzheimer's disease (AD). Participant willingness to enroll in different types of preclinical AD trials is understudied and whether the FDA approval of aducanumab affected these attitudes is unknown. OBJECTIVE: To evaluate preferences toward three preclinical AD trial scenarios and whether the FDA approval of aducanumab changed willingness to participate among potential trial participants. METHODS: Through an electronic survey, we asked enrollees in a recruitment registry age 50-79 to rate their willingness (using a 6-point Likert scale) to enroll in three hypothetical preclinical AD trial scenarios: an in-clinic infused monoclonal antibody intervention, a home-infused monoclonal antibody intervention, and an oral BACE inhibitor intervention. We administered the survey before and after the FDA approval of aducanumab. We used a generalized estimating equation model to assess group differences in preference for the trial scenarios. We used a paired t-test to determine if willingness to participate (using total willingness across three scenarios as the outcome) changed after the FDA decision. RESULTS: At baseline, the mean participant willingness was highest in the in-clinic infusion scenario. There was no significant change in willingness to participate, overall, after the FDA decision. Participants who were independently aware of the FDA's decision (prior to the second survey) demonstrated reduced willingness to participate; participants unaware of the FDA decision demonstrated no change. CONCLUSION: Willingness to participate in preclinical AD trials may have been negatively affected by the FDA's decision to approve aducanumab among those aware of the decision.

Strategies Associated with Retaining Participants in the Longitudinal National Alzheimer's Coordinating Center Uniform Data Set Study.

BACKGROUND: Best approaches for retaining research participants in Alzheimer's disease cohort studies are understudied. OBJECTIVE: Using data from the National Alzheimer's Coordinating Center Uniform Data Set, we evaluated the associations of unique strategies with participant retention across Alzheimer's Disease Research Centers and explored potential effect modification by race, ethnicity and diagnostic group. METHODS: We examined retention at the first follow-up visit among participants enrolled during 2015-2017. Structured surveys ascertained 95 retention tactics among 12 strategies. Strategy-specific summary scores were created based on the number of implemented tactics for each strategy and grouped into tertiles. Generalized estimating equations were constructed to evaluate associations between strategy scores and the odds of retention, controlling for age, sex, education, study partner type, marital status, visit length, battery length, diagnostic group, race and ethnicity. Separate models were stratified by race, ethnicity and diagnostic group. Effect modification was formally tested with interaction terms. RESULTS: Among 5,715 total participants enrolled, 4,515 were Non-Hispanic White (79%), 335 were Hispanic/Latino (6%), 651 were Non-Hispanic Black (11%), and 214 were Non-Hispanic Asian (4%). Compared to the lowest tertile of scores, the highest tertile of scores involving improvement in study personnel and communication of study requirements and details were associated with 61% higher odds of retention in fully adjusted models (adjusted Odds Ratios [aOR] = 1.61, 95% Confidence Interval [CI] = 1.05-2.47 and aOR = 1.55, 95% CI = 1.03-2.35, respectively). We did not find evidence for effect modification. CONCLUSION: In the setting of limited resources, specific retention strategies may be more valuable than others.

Dyadic Enrollment in a Phase 3 Mild Cognitive Impairment Clinical Trial.

BACKGROUND: Dyadic enrollment of a participant and study partner is required in mild cognitive impairment (MCI) clinical trials, despite participants being functionally independent. Research examining how the study partner requirement impacts MCI trials remains limited. METHODS: Using the Alzheimer's Disease Cooperative Study donepezil and vitamin E MCI trial data, we quantified the proportions of enrolled spouse, adult child, and other dyads. We used multinomial regression to identify which baseline participant characteristics (age, sex, race and ethnicity, apolipoprotein E ε4 status, education, residence type) were associated with dyad type. RESULTS: Among 769 randomized dyads, 73% were spousal, 14% adult child, and 13% other dyads. Adjusting for multiple comparisons, underrepresented racial and ethnic background (eg, comparing Hispanic to non-Hispanic White participants: adult child vs. spouse odds ratio = 5.86; 95% confidence interval: 2.09, 16.5; other vs. spouse odds ratio = 4.95; 95% confidence interval: 1.83, 13.4), female sex, age, nonhouse residence, and apolipoprotein E ε4 noncarriage were each associated with a higher odds of having an adult child, as well as an other, study partner at enrollment. DISCUSSION: Increasing participation among nonspousal dyads may facilitate more inclusive and representative MCI trial samples.

A generalized interrupted time series model for assessing complex health care interventions.

Assessing the impact of complex interventions on measurable health outcomes is a growing concern in health care and health policy. Interrupted time series (ITS) designs borrow from traditional case-crossover designs and function as quasi-experimental methodology able to retrospectively analyze the impact of an intervention. Statistical models used to analyze ITS designs primarily focus on continuous-valued outcomes. We propose the "Generalized Robust ITS" (GRITS) model appropriate for outcomes whose underlying distribution belongs to the exponential family of distributions, thereby expanding the available methodology to adequately model binary and count responses. GRITS formally implements a test for the existence of a change point in discrete ITS. The methodology proposed is able to test for the existence of and estimate the change point, borrow information across units in multi-unit settings, and test for differences in the mean function and correlation pre- and post-intervention. The methodology is illustrated by analyzing patient falls from a hospital that implemented and evaluated a new care delivery model in multiple units.

Censoring-robust time-dependent receiver operating characteristic curve estimators.

Time-dependent receiver operating characteristic curves are often used to evaluate the classification performance of continuous measures when considering time-to-event data. When one is interested in evaluating the predictive performance of multiple covariates, it is common to use the Cox proportional hazards model to obtain risk scores; however, previous work has shown that when the model is mis-specified, the estimand corresponding to the partial likelihood estimator depends on the censoring distribution. In this manuscript, we show that when the risk score model is mis-specified, the AUC will also depend on the censoring distribution, leading to either over- or under-estimation of the risk score's predictive performance. We propose the use of censoring-robust estimators to remove the dependence on the censoring distribution and provide empirical results supporting the use of censoring-robust risk scores.

RITS: a toolbox for assessing complex interventions via interrupted time series models.

BACKGROUND: Various interacting and interdependent components comprise complex interventions. These components create difficulty in assessing the true impact of interventions designed to improve patient-centered outcomes. Interrupted time series (ITS) designs borrow from case-crossover designs and serve as quasi-experimental methodology able to retrospectively assess the impact of an intervention while accounting for temporal correlation. While ITS designs are aptly situated for studying the impacts of large-scale public health policies, existing ITS software implement rigid ITS methodology that often assume the pre- and post-intervention phases are fully differentiated (by a known change-point or set of time points) and do not allow for changes in both the mean functions and correlation structure. RESULTS: This article describes the Robust Interrupted Time Series (RITS) toolbox, a stand-alone user-friendly application researchers can use to implement flexible ITS models that estimate the lagged effect of an intervention on an outcome, level and trend changes, and post-intervention changes in the correlation structure, for single and multiple ITS. The RITS toolbox incorporates a formal test for the existence of a change in the outcome and estimates a change-point over a set of possible change-points defined by the researcher. In settings with multiple ITS, RITS provides a global over-all units change-point and allows for unit-specific changes in the mean functions and correlation structures. CONCLUSIONS: The RITS toolbox is the first piece of software that allows researchers to use flexible ITS models that test for the existence of a change-point, estimate the change-point (if estimation is desired), and allow for changes in both the mean functions and correlation structures at the change point. RITS does not require any knowledge of a statistical (or otherwise) programming language, is freely available to the community, and may be downloaded and used on a local machine to ensure data protection.

Reasons for undergoing amyloid imaging among cognitively unimpaired older adults.

OBJECTIVES: Preclinical Alzheimer's disease (AD) clinical trials screen cognitively unimpaired older adults for biomarker criteria and disclose their results. We examined whether participants in the Anti-Amyloid Treatment in Asymptomatic Alzheimer's disease Study with "elevated" and "not elevated" amyloid differed in scores on the "Views and Perceptions of Amyloid Imaging" questionnaire. We hypothesized that, prior to disclosure, those with elevated amyloid would score higher than those with not elevated amyloid. We also quantified how responses changed after result disclosure. METHODS: We assessed data from 4327 individuals who completed the questionnaire at screening visit 1 and after amyloid disclosure. We used linear regression models to assess the relationship between questionnaire category scores and amyloid status. We also quantified the relationship between category score changes and amyloid status. RESULTS: Overall, participants scored altruism and contribution to research as the strongest motivations for undergoing amyloid imaging. Those with elevated amyloid scored 0.23 points higher in the Perceived Risk category, on average, than those who had not elevated amyloid prior to disclosure; this effect attenuated towards zero after adjusting for Cognitive Function Instrument score. After disclosure, participants with elevated amyloid demonstrated less within-subject change in Perceived Risk, on average, compared to those with similar pre-disclosure scores who had not elevated amyloid, while demonstrating greater changes in the altruism and planning categories. INTERPRETATION: Altruism and learning disease risk motivated enrollment in this preclinical AD trial. Participants with elevated amyloid differed from their not elevated counterparts in their perceptions of amyloid imaging, even before undergoing the procedure.

On the design of early-phase Alzheimer's disease clinical trials with cerebrospinal fluid tau outcomes.

BACKGROUND/AIMS: The focus of Alzheimer's disease studies has shifted to earlier disease stages, including mild cognitive impairment. Biomarker inclusion criteria are often incorporated into mild cognitive impairment clinical trials to identify individuals with "prodromal Alzheimer's disease" to ensure appropriate drug targets and enrich for participants likely to develop Alzheimer's disease dementia. The use of these eligibility criteria may affect study power. METHODS: We investigated outcome variability and study power in the setting of proof-of-concept prodromal Alzheimer's disease trials that incorporate cerebrospinal fluid levels of total tau (t-tau) and phosphorylated (p-tau) as primary outcomes and how differing biomarker inclusion criteria affect power. We used data from the Alzheimer's Disease Neuroimaging Initiative to model trial scenarios and to estimate the variance and within-subject correlation of total and phosphorylated tau. These estimates were then used to investigate the differences in study power for trials considering these two surrogate outcomes. RESULTS: Patient characteristics were similar for all eligibility criteria. The lowest outcome variance and highest within-subject correlation were obtained when phosphorylated tau was used as an eligibility criterion, compared to amyloid beta or total tau, regardless of whether total tau or phosphorylated tau were used as primary outcomes. Power increased when eligibility criteria were broadened to allow for enrollment of subjects with either low amyloid beta or high phosphorylated tau. CONCLUSION: Specific biomarker inclusion criteria may impact statistical power in trials using total tau or phosphorylated tau as the primary outcome. In concert with other important considerations such as treatment target and population of clinical interest, these results may have implications to the integrity and efficiency of prodromal Alzheimer's disease trial designs.

Evaluation of the Safety and Efficacy of a Novel Thrombin Containing Combination Hemostatic Powder Using a Historical Control.

This clinical study compares 2 hemostatic agents, a novel combination powder (CP) (HEMOBLAST™ Bellows) and an established polysaccharide starch powder (PP) (Arista™ AH) to assess the usefulness of CP. Retrospective comparative analysis of CP (July 2018 to July 2019, 68 patients) to PP (January 2011 to January 2013, 94 patients) in cardiothoracic patients was performed using linear regression models adjusting for age, sex, and procedure type for the endpoints: blood loss; protamine to skin closure time (hemostasis time); chest tube output and blood products required 48 hours postoperatively; ICU stay; postoperative comorbidities; and 30 day mortality. 162 patients (108 M: 54 F) underwent 162 cardiothoracic surgical procedures including: transplantation (n = 44), placement of ventricular assist device (n = 87), and others (n = 31). Use of CP compared to PP (Estimated Mean Difference [95% CI], P-value) produced significant reductions: blood loss (mL) (-886.51 [-1457.76, -312.26], P = 0.003); protamine to skin closure time (min) (-16.81 [-28.03, -5.59], P = 0.004); chest tube output (48 hrs, mL) (-445.76 [-669.38, -222.14], P < 0.001); packed red blood cell transfusions (units) (-0.98 [-1.56, -0.4], P = 0.001); and postoperative comorbidities (-0.31 [-0.55, -0.07], P = 0.012). There were no differences in the ICU stay (4.07 [-2.01, 10.15], P = 0.188) or 30-day mortality (0.57 [0.20, 1.63], P = 0.291). The use of CP in complex cardiothoracic operations resulted in improved hemostasis and significant clinical benefits in blood loss, transfusion requirements, morbidity, and time in operating room.

The ANTsX ecosystem for quantitative biological and medical imaging.

The Advanced Normalizations Tools ecosystem, known as ANTsX, consists of multiple open-source software libraries which house top-performing algorithms used worldwide by scientific and research communities for processing and analyzing biological and medical imaging data. The base software library, ANTs, is built upon, and contributes to, the NIH-sponsored Insight Toolkit. Founded in 2008 with the highly regarded Symmetric Normalization image registration framework, the ANTs library has since grown to include additional functionality. Recent enhancements include statistical, visualization, and deep learning capabilities through interfacing with both the R statistical project (ANTsR) and Python (ANTsPy). Additionally, the corresponding deep learning extensions ANTsRNet and ANTsPyNet (built on the popular TensorFlow/Keras libraries) contain several popular network architectures and trained models for specific applications. One such comprehensive application is a deep learning analog for generating cortical thickness data from structural T1-weighted brain MRI, both cross-sectionally and longitudinally. These pipelines significantly improve computational efficiency and provide comparable-to-superior accuracy over multiple criteria relative to the existing ANTs workflows and simultaneously illustrate the importance of the comprehensive ANTsX approach as a framework for medical image analysis.

Recruitment and retention of participant and study partner dyads in two multinational Alzheimer's disease registration trials.

BACKGROUND: Early study exit is detrimental to statistical power and increases the risk for bias in Alzheimer's disease clinical trials. Previous analyses in early phase academic trials demonstrated associations between rates of trial incompletion and participants' study partner type, with participants enrolling with non-spouse study partners being at greater risk. METHODS: We conducted secondary analyses of two multinational phase III trials of semagacestat, an oral gamma secretase inhibitor, for mild-to-moderate AD dementia. Cox's proportional hazards regression model was used to estimate the relationship between study partner type and the risk of early exit from the trial after adjustment for a priori identified potential confounding factors. Additionally, we used a random forest model to identify top predictors of dropout. RESULTS: Among participants with spousal, adult child, and other study partners, respectively, 35%, 38%, and 36% dropped out or died prior to protocol-defined study completion, respectively. In unadjusted models, the risk of trial incompletion differed by study partner type (unadjusted p value = 0.027 for test of differences by partner type), but in models adjusting for potential confounding factors, the differences were not statistically significant (p value = 0.928). In exploratory modeling, participant age was identified as the primary characteristic to explain the relationship between study partner type and the risk of failing to complete the trial. Participant age was also the strongest predictor of trial incompletion in the random forest model. CONCLUSIONS: After adjustment for age, no differences in the risk of incompletion were observed when comparing participants with different study partner types in these trials. Differences between our findings and the findings of previous studies may be explained by differences in trial phase, size, geographic regions, or the composition of academic and non-academic sites.