Creation of a ustekinumab external control arm for Crohn's disease using electronic health records data: a pilot study

Objectives: The use of external control arms to support claims of efficacy and safety is growing in interest among drug sponsors and regulators. However, experience with performing these kinds of studies for complex, immune-mediated diseases is limited. We sought to establish a method for creating an external control arm for Crohn's disease. Methods: We queried electronic health records databases and screened records at the University of California, San Francisco to identify patients meeting the major eligibility criteria of TRIDENT, a concurrent trial involving ustekinumab as a reference arm. Timepoints were defined to balance the tradeoff between missing disease activity and bias. We compared two imputation models by their impacts on cohort membership and outcomes. We compared the results of ascertaining disease activity using structured data algorithms against manual review. We used these data to estimate ustekinumab's real-world effectiveness. Results: Screening identified 183 patients. 30% of the cohort had missing baseline data. Two imputation models were tested and had similar effects on cohort definition and outcomes. Algorithms for ascertaining non-symptom-based elements of disease activity were similar in accuracy to manual review. The final cohort consisted of 56 patients. 34% of the cohort was in steroid-free clinical remission by week 24. Conclusions: Differences in the timing and goals of real-world encounters as compared to controlled studies directly translate into significant missing data and lost sample size. Efforts to improve real-world data capture and better align trial design with clinical practice may enable robust external control arm studies and improve trial efficiency.

CTNI-40. EVALUATING FEASIBILITY AND EFFICIENCY OF PHASE II ADAPTIVE PLATFORM TRIAL DESIGNS BASED ON THE INDIVIDUALIZED SCREENING TRIAL OF INNOVATIVE GLIOBLASTOMA THERAPY (INSIGhT) EXPERIENCE

BACKGROUND The Individualized Screening Trial of Innovative Glioblastoma Therapy (INSIGhT) is a phase II platform trial with Bayesian adaptive randomization and deep genomic profiling to more efficiently test experimental agents in newly diagnosed glioblastoma and to prioritize therapies for late-stage testing. METHODS In the ongoing INSIGhT trial, patients with newly diagnosed MGMT-unmethylated glioblastoma are randomized to the control arm or one of three experimental therapy arms (CC-115, abemaciclib, and neratinib). The control arm therapy is radiotherapy with concomitant and adjuvant temozolomide, and primary endpoint is overall survival. Randomization has been adapted based on Bayesian estimation of biomarker-specific probability of treatment impact on progression-free survival (PFS). All tumors undergo detailed molecular sequencing, and this is facilitated with the companion ALLELE protocol. To evaluate feasibility of this approach, we assessed the status of this ongoing trial. RESULTS Since INSIGhT was activated 4.3 years ago, it has expanded to include 12 sites across the United States. A total of 247 patients have been enrolled. Randomization probabilities have been repeatedly adjusted over time based upon early PFS results to alter the randomization ratio from standard 1:1:1:1 randomization. All three arms have completed accrual and efficacy estimates are available based upon comparison to the common control arm in context of relevant biomarkers. There are 87 patients alive and in follow-up, and there are ongoing plans to add additional arms to evaluate further treatments in the future. CONCLUSION The INSIGhT trial demonstrates that a multi-center Bayesian adaptive platform trial is a feasible and effective approach to help prioritize therapies and biomarkers for newly diagnosed GBM. The trial has maintained robust accrual, and the simultaneous testing of multiple agents, sharing a common control arm and adaptive randomization serve as features to increase trial efficiency relative to traditional clinical trial designs.

Lessons Learned From Implementing a Novel Bayesian Adaptive Dose-Finding Design in Advanced Pancreatic Cancer

PURPOSE Novel Bayesian adaptive designs provide an effective way to improve clinical trial efficiency. These designs are superior to conventional methods, but implementing them can be challenging. The aim of this article was to describe what we learned while applying a novel Bayesian phase I-II design in a recent trial. METHODS The primary goal of the trial was to optimize radiation therapy (RT) dose among three levels (low, standard, and high), given either with placebo (P) or an investigational agent (A), for treating locally advanced, radiation-naive pancreatic cancer, deemed appropriate for RT rather than surgery. Up to 48 patients were randomly assigned fairly between RT plus P and RT plus A, with RT dose-finding done within each arm using the late-onset efficacy-toxicity design on the basis of two coprimary end points, tumor response and dose-limiting toxicity, both evaluated at up to 90 days. The random assignment was blinded, but within each arm, unblinded RT doses were chosen adaptively using software developed within the institution. RESULTS Implementing the design involved double-blind balance-restricted random assignment, real-time assessment of patient outcomes to evaluate the efficacy-toxicity trade-off for each RT dose in each arm to optimize each patient's RT dose adaptively, and transition from a single-center trial to a multicenter trial. We present lessons learned and illustrative documentation. CONCLUSION Implementing novel Bayesian adaptive trial designs requires close collaborations between physicians, pharmacists, statisticians, data managers, and sponsors. The process is difficult but manageable and essential for efficient trial conduct. Close collaboration during trial conduct is a key component of any trial that includes real-time adaptive decision rules.

Non-static framework for understanding adaptive designs: an ethical justification in paediatric trials

Many drugs used in paediatric medicine are off-label. There is a rising call for the use of adaptive clinical trial designs (ADs) in responding to the need for safe and effective drugs given their potential to offer efficiency and cost-effective benefits compared with traditional clinical trials. ADs have a strong appeal in paediatric clinical trials given the small number of available participants, limited understanding of age-related variability and the desire to limit exposure to futile or unsafe interventions. Although the ethical value of adaptive trials has increasingly come under scrutiny, there is a paucity of literature on the ethical dilemmas that may be associated with paediatric adaptive designs (PADs). This paper highlights some of these ethical concerns around safety, scientific/social value and caregiver/guardian comprehension of the trial design. Against this background, the paper develops a non-static conceptual lens for understanding PADs. It shows that ADs are epistemically open and reduce some of the knowledge-associated uncertainties inherent in clinical trials as well as fast-track the time to draw conclusions about the value of evaluated drugs/treatments. On this note, the authors argue that PADs are ethically justifiable given they (1) have multiple layers of safety, exposing enrolled children to lesser potential risks, (2) create social/scientific value generally and for paediatric populations in particular, (3) specifically foster the flourishing of paediatric populations and (4) can significantly improve paediatric trial efficiency when properly designed and implemented. However, because PADs are relatively new and their regulatory, ethical and logistical characteristics are yet to be clarified in some jurisdictions, the cooperation of various public and private stakeholders is required to ensure that the interests of children, their caregivers and parents/guardians are best served while exposing paediatric research subjects to the most minimal of risks when they are enrolled in paediatric trials that use ADs.

Microbiota or placebo after antimicrobial therapy for recurrent Clostridioides difficile at home: A clinical trial with novel home-based enrollment

Introduction: Clostridiodes difficile infection is the leading cause of infectious diarrhea in the United States, with substantial morbidity and mortality. Recurrent infection is especially challenging, with each recurrence increasing the likelihood of a successive recurrence, leading to cycles of prolonged symptoms, frequent antimicrobial use, and decreased quality of life. Fecal microbiota transplantation to prevent recurrent infection is a promising intervention with a large effect size in observational studies, but with conflicting results from randomized controlled trials. We are conducting a Veterans Affairs-wide randomized controlled trial utilizing centralized case identification, with enrollment and fecal microbiota transplant administration occurring at the participant’s home. This type of trial design significantly improves trial efficiency, greatly decreases trial cost, increases consistency of trial administration, and most importantly makes nationwide clinical trials in less-common diseases possible. Methods: This is a randomized comparison of capsule-delivered fecal microbiota transplant for the prevention of recurrent Clostridiodes difficile infection, administered after successful initial treatment of recurrent C. difficile infection with standard therapy. The primary endpoint is the incidence of recurrent C. difficile infection or death. Cases are identified by searching the Veterans Affairs Corporate Data Warehouse, with central study coordinators then reaching out to potential participants. Individuals meeting inclusion criteria and interested in participation are scheduled for in-home consent, randomization, and capsule administration, followed by telephone follow-up for 6 months. To mitigate risks of COVID-19, enrollment via video visits has been implemented. Results: A total of 102 participants have been enrolled through January 2021. Centralized case identification and in-home enrollment has facilitated enrollment from 34 unique states, with 38% being from rural or highly rural areas. Discussion: Centralized case identification and in-home enrollment is a feasible and innovative method of conducting randomized controlled trials in the Veterans Affairs system, improving access to clinical research for populations who may have difficulty engaging with the traditional model of clinical trials where enrollment is based at large hospitals in major metropolitan areas.

Power prior for borrowing the real-world data in bioequivalence test with a parallel design

Recently, real-world study has attracted wide attention for drug development. In bioequivalence study, the reference drug often has been marketed for many years and accumulated abundant real-world data. It is therefore appealing to incorporate these data in the design to improve trial efficiency. In this paper, we propose a Bayesian method to include real-world data of the reference drug in a current bioequivalence trial, with the aim to increase the power of analysis and reduce sample size for long half-life drugs. We adopt the power prior method for incorporating real-world data and use the average bioequivalence posterior probability to evaluate the bioequivalence between the test drug and the reference drug. Simulations were conducted to investigate the performance of the proposed method in different scenarios. The simulation results show that the proposed design has higher power than the traditional design without borrowing real-world data, while controlling the type I error. Moreover, the proposed method saves sample size and reduces costs for the trial.

Multi-Modality Machine Learning Predicting Parkinson’s Disease

SUMMARYBackgroundPersonalized medicine promises individualized disease prediction and treatment. The convergence of machine learning (ML) and available multi-modal data is key moving forward. We build upon previous work to deliver multi-modal predictions of Parkinson’s Disease (PD).MethodsWe performed automated ML on multi-modal data from the Parkinson’s Progression Marker Initiative (PPMI). After selecting the best performing algorithm, all PPMI data was used to tune the selected model. The model was validated in the Parkinson’s Disease Biomarker Program (PDBP) dataset. Finally, networks were built to identify gene communities specific to PD.FindingsOur initial model showed an area under the curve (AUC) of 89.72% for the diagnosis of PD. The tuned model was then tested for validation on external data (PDBP, AUC 85.03%). Optimizing thresholds for classification, increased the diagnosis prediction accuracy (balanced accuracy) and other metrics. Combining data modalities outperforms the single biomarker paradigm. UPSIT was the largest contributing predictor for the classification of PD. The transcriptomic data was used to construct a network of disease-relevant transcripts.InterpretationWe have built a model using an automated ML pipeline to make improved multi-omic predictions of PD. The model developed improves disease risk prediction, a critical step for better assessment of PD risk. We constructed gene expression networks for the next generation of genomics-derived interventions. Our automated ML approach allows complex predictive models to be reproducible and accessible to the community.FundingNational Institute on Aging, National Institute of Neurological Disorders and Stroke, the Michael J. Fox Foundation, and the Global Parkinson’s Genetics Program.RESEARCH IN CONTEXTEvidence before this studyPrior research into predictors of Parkinson’s disease (PD) has either used basic statistical methods to make predictions across data modalities, or they have focused on a single data type or biomarker model. We have done this using an open-source automated machine learning (ML) framework on extensive multi-modal data, which we believe yields robust and reproducible results. We consider this the first true multi-modality ML study of PD risk classification.Added value of this studyWe used a variety of linear, non-linear, kernel, neural networks, and ensemble ML algorithms to generate an accurate classification of both cases and controls in independent datasets using data that is not involved in PD diagnosis itself at study recruitment. The model built in this paper significantly improves upon our previous models that used the entire training dataset in previous work1. Building on this earlier work, we showed that the PD diagnosis can be refined using improved algorithmic classification tools that may yield potential biological insights. We have taken careful consideration to develop and validate this model using public controlled-access datasets and an open-source ML framework to allow for reproducible and transparent results.Implications of all available evidenceTraining, validating, and tuning a diagnostic algorithm for PD will allow us to augment clinical diagnoses or risk assessments with less need for complex and expensive exams. Going forward, these models can be built on remote or asynchronously collected data which may be important in a growing telemedicine paradigm. More refined diagnostics will also increase clinical trial efficiency by potentially refining phenotyping and predicting onset, allowing providers to identify potential cases earlier. Early detection could lead to improved treatment response and higher efficacy. Finally, as part of our workflow, we built new networks representing communities of genes correlated in PD cases in a hypothesis-free manner, showing how new and existing genes may be connected and highlighting therapeutic opportunities.

Plasma neurofilament light predicts mortality in patients with stroke

Given the heterogeneity of stroke brain injury, there is a clear need for a biomarker that determines the degree of neuroaxonal injury across stroke types. We evaluated whether blood neurofilament light (NFL) would fulfill this purpose for patients with acute cerebral infarction (ACI; N = 227), aneurysmal subarachnoid hemorrhage (aSAH; N = 58), or nontraumatic intracerebral hemorrhage (ICH; N = 29). We additionally validated our findings in two independent cohorts of patients with ICH (N = 96 and N = 54) given the scarcity of blood biomarker studies for this deadliest stroke type. Compared to healthy individuals (N = 79 and N = 48 for the discovery and validation cohorts, respectively), NFL was higher for all stroke types. NFL associated with radiographic markers of brain tissue damage. It correlated with the extent of early ischemic injury in patients with ACI, hemorrhage severity in patients with aSAH, and intracranial hemorrhage volume in patients with ICH. In all patients, NFL independently correlated with scores from the NIH Stroke Scale, the modified Rankin Scale, and the Mini-Mental State Examination at blood draw, which respectively assess neurological, functional, and cognitive status. Furthermore, higher NFL concentrations independently associated with 3- or 6-month functional disability and higher all-cause mortality. These data support NFL as a uniform method to estimate neuroaxonal injury and forecast mortality regardless of stroke mechanism. As a prognostic biomarker, blood NFL has the potential to assist with planning supportive and rehabilitation services and improving clinical trial efficiency for stroke therapeutics and devices.