A message from the editor

Dear Colleagues:

Welcome to the annual CNS Summit Abstracts of Poster Presentations supplement to Innovations in Clinical Neuroscience. We are pleased to provide you with this reference guide to some of the innovative research that was presented during the CNS Summit 2020 event. The supplement is also available online by visiting www.innovationscns.com.

This supplement is just a small representation of the cutting edge research, innovative ideas, and collaborative efforts shared via the CNS Summit platform. Even though a world of physical distancing was not on our radar for 2020, the strong sense of community and collaborative support the Summit continues to foster among its participants enabled us to face the pandemic and ensuing challenges of 2020 confidently, purposefully, and as a united force of individuals committed to developing better, safer, and more accessible therapies for patients around the world. We believe this mission is our ethical duty to the patients we serve. The annual CNS Summit and  its year-round programming are designed to encourage and facilitate open channels of communication and data sharing across all disciplines of medical research—with the ultimate goal of achieving optimal patient outcomes. 

In this abstract supplement, we’ve organized CNS Summit 2020 poster abstracts into the following groups for your convenience and easy reference:

  • Digital Tools and Technology
  • Investigative Drug Compounds and Therapies
  • Patient Assessment
  • Trial Methodology

You will also find an alphabetical index by author and poster title on pages 15 and 16 of this publication.

We hope you find the CNS Summit 2020 Abstracts of Poster Presentations supplement to Innovations in Clinical Neuroscience informative and that it provides you with a useful snapshot of the research being presented at the annual CNS Summit event. Make sure to mark your calendars for CNS Summit 2021, which will take place October 28–31, 2021, at the Boca Raton Resort in Boca Raton, Florida. Visit www.cnssummit.org for more information.

Hope to see you in Boca! As always, we welcome your feedback and participation.


Amir Kalali, MD

Editor, Innovations in Clinical Neuroscience  

Table of Contents

Digital Tools and Technology 

  • Analytical validity of a new algorithm to identify walking periods in continuous wrist worn accelerometer
  • Development and implementation of an innovative remote monitoring approach during a global pandemic
  • A longitudinal web-based survey for delivering speech and language screening tools
  • Smartphones and smartwatches in clinical trials: Smart for all ages
  • Time to complete electronic clinical outcome assessments on handheld versus tablet devices
  • Wearable technologies for improving social functioning, communication, or behavior in autism spectrum disorder

Investigative Drug Compounds/Therapies

  • Correction of hyperglycemia through adenoviral mediated compartmentalized liver transduction of the human insulin gene in a Type 1 diabetes mellitus pig model
  • EEG biomarker development as clinical endpoints
  • [The] effect of a single 50mg/kg dose of propofol on the forced swim test in mice at 24 hours
  • Improvement in quality measures with repository corticotropin injection (Acthar® Gel) in a prospective observational study of multiple sclerosis relapse

Patient Assessment

  • Can blinded, site-independent scoring of audio recorded site-based PANSS interviews of acutely psychotic patients replicate site-based scores and affirm treatment outcome results?
  • Capturing clinical symptoms with ecological momentary assessment: Convergence of momentary reports of psychotic and mood symptoms with diagnoses and standard clinical assessments
  • Comparison of three methods for MADRS administration: Correlation with CGI-S
  • Comparison of baseline severity of schizophrenia negative symptoms in study subjects in the United States versus the rest of the world 
  • Harnessing the immune system for neuroprotection: Development of IBC-Ab002, an anti-PD-L1 monoclonal antibody for Alzheimer’s disease 
  • High diagnostic confidence improves signal detection
  • Improving assessment in pediatric trials: development of an eCOA (electronic clinical outcomes assessment) Positive and Negative Syndrome Scale (PANSS) for use in adolescent schizophrenia trials
  • Marijuana use among potential clinical trial participants with major depressive disorder
  • Proving the equivalence of simplified home sleep testing to polysomnography
  • [A] test-case for deploying clinical measures of neurological function remotely and via local healthcare provider in the state of Wyoming
  • Which cognitive domains contribute most to large MMSE changes in the screening period?

Trial Methodology

  • Considerations for clinical trial design in frontotemporal lobar degeneration (FTLD)
  • Differences in eConsent among diagnosis groups
  • Driving better risk-based quality management (RBQM) using atypical data detection for CNS trials
  • Implementing a fully decentralized clinical trial for major depression
  • Pharmacometric analyses, reformulation of BNC210 and fast track designation, support BNC210 evaluation in a second Phase 2 PTSD trial
  • Pinpointing placebo responders using multi-component vocal analysis
  • Rare neurodegenerative diseases: Early-stage challenges and optimal animal models in orphan drug development
  • Use of euclidean distance heat mapping to identify data of concern from patient-reported outcomes in atopic dermatitis: A potentially useful paradigm for psychiatry trials?  

Digital Tools and Technology

Analytical validity of a new algorithm to identify walking periods in continuous wrist-worn accelerometer

Authors: Kelly P,1 Ellis R,1 Huang C1

Affiliations: 1Koneksa Health Inc., New York, New York

Objective: To determine the analytical validity of a new algorithm to detect walking periods via a wrist worn ActiGraph GT9X

Methods: Eleven healthy volunteers wore the wrist-worn ActiGraph GT9X and walked a marked course, starting and stopping at 10 and 20 second intervals. Study execution was supervised by three raters who noted start and stop times. Analytical validity was determined by the Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE) between corresponding algorithm and rater measures.

Results: Mean subject adherence was 80 percent. Each subject completed a minimum of two attempts at each of 10- and 20-second intervals of walking. The dataset comprised 45 completed assessments. The algorithm captured 100 percent of walking periods. For 10-second walking periods, start time MAE was 0.87 second, end time MAE 0.52 second, and duration MAE 0.49 second (MAPE 5.00%). For 20-second walking periods, start time MAE was 1.05 second, end time MAE 1.16 second, and duration MAE 1.12 second (MAPE 5.60%).

Conclusion: Results suggest that it is possible to accurately identify walking periods from wrist-worn accelerometers. This algorithm may be of use in the study of free-living walking in individuals with neurological disorders, such as Parkinson’s disease.

Funding/disclosures: This study was sponsored by Koneksa Health. The presenters reported no conflicts of interest.

Development and implementation of an innovative remote monitoring approach during a global pandemic

Authors: Frohlich L,1 Versavel S,1 Devis G,1 Whitmore L,1 Wang G,1 Leoni M,1 Perry P,1 Sanchez R1

Affiliations: 1Cerevel Therapeutics, LLC, Boston, Massachusetts

Objective: Remote-source data verification (SDV) ensures oversight of clinical sites and integrity of clinical trial data. Novel approaches can lead to time and cost effective SDVs. In an ongoing Phase Ib multiple-ascending dose trial in patients with schizophrenia, a novel approach for remote SDV enabled acquisition of interim data to inform progression of an adaptive design and mitigate any operational risks related to COVID-19 limitations.

Methods: The remote SDV procedure required identification of a cloud-based Code of Federal Regulations (CFR) Part 11 compliant platform for sharing source documents with clinical monitors (CRAs). Clinical site staff scanned relevant redacted source documents and uploaded into the platform. CRAs securely and remotely accessed the electronic case report forms (eCRFs) and utilized the source assigned to them in platform to perform the remote SDV of the required eCRF pages.

Results: To continue with data cleaning activities for the ongoing Phase Ib trial, development, and implementation of a remote SDV plan allowed for completion of over 2,000 eCRF pages in six weeks. This was accomplished via email notifications in real-time for document reviews, electronic annotations detailing discrepancies between source document and EDC data, and visual adaptations of SDV completeness.

Conclusion: Adapting the existing capabilities of the platform enabled preservation of data integrity while mitigating operational risks to maintain the study timeline in lieu of the COVID-19 pandemic. Subsequent adaptation was applied broadly throughout multiple United States-based clinical trials irrespective of development phase and therapeutic indication within the partner global CRO organization.

Funding/financial disclosures:  Lillian Frohlich, Stacey Versavel, Gabrielle Devis, Laura Whitmore, Guang Wang, Matthew Leoni, Pamela Perry, and Raymond Sanchez are employees of Cerevel Therapeutics, LLC.

References: ClinicalTrials.gov Identifier: NCT04136873

A longitudinal web-based survey for delivering speech and language screening tools

Authors: Pissadaki EK,1 Schwoebel JW,1 Schwartz JW,1 Butler M,1 Moss M1

Affiliations: 1Biogen Inc., Cambridge, Massachusetts; Sonde Health, Boston, Massachusetts

Objective: Digital vocal diagnostic tools that identify individuals likely to develop a neurodegenerative disease will improve treatment decisions and early intervention. Despite multiple and significant studies in vocal research for neurodegenerative diseases, there is a lack of standardization and clear description of the normative population for speech and language (S&L) tests routinely used to clinically evaluate individuals. Key cofactors to be accounted for are age, sex, educational level, disparate batteries of speech assessments, and ambient noise. We present a large, remote, normative data study to provide reference for future clinical endeavors.

Methods: To standardize S&L data, we configured a web-based voice survey to generate a normative dataset (N>5000) consisting of speech responses to a battery of neuropsychological assessments such as confrontational naming task, semantic/phonemic verbal fluency, free speech, image description, non-words, sustained phonation, and diadochokinesis. Subjects took different forms of the survey over four successive weeks. We obtained demographic details and medical history per individual to understand the normative distribution.

Results: We focused on aspects of semantic language skills to establish normative distributions initially normalized to age and degree of education toward the goal of creating specific digital semantic language signatures.

Conclusion: By designing a web-based survey, we generated a dataset to characterize normative population and identify candidate language signatures suitable to screen and validate preclinical signs of neurodegeneration.

Smartphones and smartwatches in clinical trials: Smart for all ages

Authors: Yamamoto RT, Dias NR, Ishaque A, Faulkner KG

Objective: To determine current use and familiarity with smartphones and smartwatches and preferences for use in clinical trials across age groups

Methods: Study subject (N=1,046) completed an online survey. Respondents reported current use of smartphones and smartwatches. They were also asked if they were participating in a clinical trial and whether they would prefer to use their own device(s) or one provided solely for use during the trial.

Results: Less than one percent of respondents indicated they did not use a smartphone. Ninety-eight percent endorsed using their smartphone anywhere from a few times a day (20%) to every five minutes (27%). Nintey-seven percent of respondents over 60 years old (60-83 YO) used their smartphones frequently every day. Regardless of age, when asked about using a smartphone during a clinical trial, most (73%) indicated they would be okay using either an app downloaded to their personal smartphone or a smartphone provided to them that could only be used for the research study. While the majority of respondents (69%) did not own a smart/fitness watch, there was a strong willingness to use one in a clinical trial (96% overall; 94% of those aged 60-83), if it were provided during a research study.

Conclusion: Smartphones and smart/fitness watches are rapidly becoming important technologies for collecting clinical trial data. There has been some concern regarding age-related willingness or comfort using technology in clinical trials. The current results suggest that, regardless of age, people are willing to use smartphones and smartwatches when participating in clinical research.

Funding/disclosures: Authors are employees of ERT

Time to complete electronic clinical outcome assessments on handheld versus tablet devices

Authors: Dumais KM,1 Reed DL,1 Faulkner KG1

Affiliations: 1ERT, Boston, Massachusetts

Objective: With multiple options for electronic clinical outcome assessments, it is important to evaluate influence of mode type on usability. We investigated whether mode type (handheld vs. tablet) influences the time to complete assessments, and whether this is impacted by age of the user.

Methods: Time to complete the 36-item Short Form Survey (SF-36), a common quality of life questionnaire, was analyzed in two clinical trials. Among other assessments, the SF-36 was completed during clinic visits on ERT’s tablet in Study A (Phase IIIb severe asthma) and ERT’s handheld in Study B (Phase III severe nasal polyposis).

Results: Study A (N=670) had 1,937 SF-36 completions and Study B (N=461) had 2,080 SF-36 completions. Completion time was significantly faster on the tablet compared with handheld (p<0.01, t-test). For participants under age 35 years, average completion time was 4.57 minutes on the tablet and 5.32 minutes on the handheld. For participants 35 to 65 years of age, average completion time was 6.48 minutes on tablet and 7.40 minutes on handheld. For participants older than 65 years of age, average completion time was 7.29 minutes on tablet and 8.18 minutes on handheld.

Conclusion: Time to complete the SF-36 was faster on the tablet compared to handheld across ages. There was an effect of age, with completion time increasing with age for in both the tablet and handheld groups.

Funding/disclosures: All authors are employees of ERT.

Wearable technologies for improving social functioning, communication, or behavior in austism spectrum disorder

Authors: Bates E,1 Pappadopulos E1

Affiliations: 1Albert Einstein College of Medicine and Upjohn, a division of Pfizer Inc., Bronx, New York

Objective: Core features of autism spectrum disorders (ASD) are chronic impairments in social functioning, communication, and behavior. While there is no cure for autism, early and intensive psychosocial interventions can improve sociability, quality of life, and independence. However, such interventions are often costly and are a burden to families, schools, and communities. Advances in wearable technologies for treating ASD social-behavioral deficits might help overcome treatment barriers; however, their clinical value remains poorly understood. This analysis aims to draw insights from the published evidence on wearables for ASD.

Methods: A literature review using Google Scholar, PubMed, and ClinicalTrials.gov from 2016 to 2020 using the keywords: autism spectrum disorders, digital therapeutics, behavioral, intervention, wearables, and mobile medical application was conducted.

Results: Publications meeting criteria were rated for therapeutic focus, intended user, evidence supporting its use, and plan/progress for FDA approval or registration. Preliminary findings revealed six wearable behavioral therapeutics: one engaged in its first study, four with completed studies submitted for publication, and one undergoing its third study with eight previous publications. No wearable intervention has been FDA approved to date.

Conclusion: Published data for ASD-psychosocial, wearable technologies are limited. Efforts to demonstrate effectiveness and/or achieve FDA verification are lacking. If wearable technologies are to advance social functioning in ASD, regulatory guidance and quality research is needed.

Funding/disclosures: Emily Bates reports no conflicts of interest; Elizabeth Pappadopulos is a full-time employee of UpJohn, a division of Pfizer Inc.

Investigative Drug Compounds and Therapies

Correction of hyperglycemia through adenoviral-mediated compartmentalized liver transduction of the human insulin gene in a Type 1 diabetes mellitus pig model

Author: Cabrera G

Affiliation: Global BioTherapeutics Inc., San Diego, California

Objective: Type 1 diabetes mellitus (T1DM) is a complex metabolic disease characterized by elevated blood glucose levels caused by the lack or insufficient synthesis of the hormone insulin, which results from the autoimmune destruction of insulin-producing pancreatic β-cells. Lifelong daily administration of precise doses of exogenous insulin and monitoring of blood glucose levels to glycemic control and therefore prevention of life-threatening complications remain the cornerstones of T1DM management. Gene therapy strategies aim to ameliorate glycemic fluctuations by the secretion of physiologic levels of insulin from non-pancreatic-β-cells.

Methods: Our group assessed the glycemic effect on T1DM pigs exerted by the secretion of human insulin dictated by an expression cassette driven by a carbohydrate response element (CRE) carried by E1-E3–deleted Type 5 adenovirus and delivered through compartmentalized liver transduction (CLT). CLT is a novel mode of hepatic vector administration achieved by the intraparenchymal injection of the vector into a blood flow isolated portion of the liver.

Results: At a dose of 2×1010 IFUs/kg (2×1012 VPs/kg), steady levels of insulin were obtained, and correction of fasting and postprandial hyperglycemia was achieved and followed for three and a half years in Vietnamese male pigs (N=2).

Conclusion: Our results demonstrated that glycemic fluctuations are manageable through gene therapy strategies and should be further explored for the treatment of T1DM.

Funding/disclosures: Gustavo Cabrera is founder and CEO of Global BioTherapeutics Inc. and holds company shares. The research was financially supported by private investments.

EEG biomarker development as clinical endpoints

Authors: Waninger S,1 Angelopoulos E,1 Meghdadi A,1 Berka C1

Affiliation: 1Advanced Brain Monitoring, Inc., Carlsbad, California

Objective: Resting state (RS) electroencephalogram (EEG) and event-related potential (ERP) biomarkers are reliable, cost effective, and noninvasive, with potential to assess treatment interventions for cognitive impairment caused by neurodegeneration.

Methods: RS EEG was directly benchmarked against 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) in patients with Alzheimer’s disease (AD) compared to age-matched healthy controls (HCs). ERPs were measured in a mild cognitively impaired (MCI) cohort during a verbal memory scanning (VMS) task and compared to HCs

Results: The hallmark slowing of EEG was observed as an enhancement of power at slow bandwidths and a suppression of power at alpha in the AD cohort compared to healthy controls. Significant correlations were found between EEG variables and hypometabolism in regions identified by FDG-PET. Highest EEG metric correlated was the theta to alpha ratio (TAR), particularly in the precuneus and cingulate gyrus. Significant reduction in N1 peak and P2 power and latency were found in the MCI group compared to HCs during the VMS task.

Conclusion: The data suggest that metrics from RS EEG parallel hypometabolism associated with cognitive decline and Alzheimer’s disease progression. ERPs measured during a VMS task suggest an association between suppression of the N1 and P2 responses and cognitive decline. The results support the use of both RS and ERP EEG pharmacodynamic endpoints as a proxy for neuronal activity to evaluate efficacy of interventions in clinical studies focused on dementia and AD.

[The] effect of a single 50mg/kg dose of propofol on the forced swim test in mice at 24 hours

Authors: Daniel DT,1 Daniel NG,2 Daniel DG,3 Flynn LC, MS,4 Allen MH5

Affiliation: 1Brown University, Providence, Rhode Island; 2Dartmouth College, Hanover, New Hampshire; 3Bioniche Global Development, McLean, Virginia; 4LCF Consulting, LLC, Libertyville, Illinois; 5University of Colorado School of Medicine, Aurora, Colorado

Objective: Propofol is an intravenous anesthetic agent that acts as a gamma-aminobutyric acid type A (GABA-A) agonist and has affinity at the N-methyl-D-aspartate (NMDA) receptor. Multiple anecdotal reports have noted mood elevation upon awakening from propofol anesthesia. In a pilot study, Mickey et al (2018) reported sustained improvement in several patients with treatment-resistant depression after repeated dosages of propofol. We recently reported that immobility time in the FST, a rodent model of depression and stress resiliency, was significantly reduced 45 minutes after administration of 50mg/kg propofol. The current experiment explored the effects of a single dose of 50mg/kg propofol on immobility time in the FST, 24 hours after administration.

Methods: Twenty-four adult male mice (C57/BL6) were given a single intraperitoneal dose of 50mg/kg propofol or saline. Twenty-four hours after dosage the mice underwent the FST and immobility time was recorded.

Results: A two tailed t-test found no significant difference (p>0.05) between immobility time for saline and propofol treated mice.

Conclusion: Although multiple sources of evidence are consistent with an antidepressant-like effect of propofol, a single dose of 50mg/kg propofol was not associated with reduced immobility time in the FST twenty-four hours after dosage.

Funding/disclosures: All authors are inventors or co-inventors of a pending patent that includes propofol for the rapid treatment of depression and suicidality.

Harnessing the immune system for neuroprotection: Development of IBC-Ab002, an anti-PD-L1 monoclonal antibody for Alzheimer’s disease 

Authors: Cedarbaum J, Baruch K, Kertser A, Matalon O, Forsht O, Braiman S, David C, Shochat E, Yoles E, Schwartz M

Affiliation: Immunobrain Checkpoint Ltd., Nes Ziona, Israel

Objective: The ongoing crosstalk between the immune system and the central nervous system (CNS) plays a major role in healthy brain tissue maintenance, with critical implications to brain aging and neurodegeneration. Based on the innovative concept of protective autoimmunity, it was shown that targeting the programmed cell death receptor (PD)-1 / PD-L1 pathway evokes an immune response that supports recruitment of specialized immune cells (primarily of myeloid origin) to the brain territory, where these cells act by enhancing clearance of toxic elements, improving neuronal function, and reducing inflammation in mouse models of Alzheimer’s disease (AD) and dementia. Maintaining a long-lasting effect was found to be dependent on intermittent treatment course. IBC-Ab002 is a novel fully human anti-PD-L1 monoclonal antibody that was designed, by point mutations on its Fc backbone, to meet the pharmacokinetic (PK), pharmacodynamic (PD) and safety characteristics anticipated to be required for treating AD. Antibody optimization and lead selection will be presented.

Methods: Correlation between PK, PD, and antibody efficacy and safety was evaluated using different antibody variants that share the same variable region but differ in their physical properties due to point mutations in their Fc region. These studies, in mouse models of AD and in healthy cynomolgus monkeys, were used for modeling human PK/PD properties of IBC-Ab002 to support design of a first-in-human clinical trial.

Results: Pharmacological studies in mice revealed that treatment efficacy is primarily Cmax-dependent, rather than exposure-dependent. Biomarkers for biological activity were identified in the blood of both treated mice and monkeys, and PK/PD relationships were used to predict IBC-Ab002 PK and effective dose range in humans. The ability to maintain treatment efficacy using intermittent administration regimen was evaluated in a longitudinal rodent study in which the antibody was administered once every six weeks; efficacy was maintained over five months. In addition, using a mouse model that spontaneously develops autoimmune diabetes, we also showed that the chosen antibody modifications reduced the risk of emergence of autoimmune diabetes in these animals.

Conclusion: IBC-Ab002 was advanced to clinical development based on its unique PK/PD profile, tailored for the needs of treating AD, and potentially superior safety profile in terms of immune-related adverse events. We are currently planning the first-in-human study of IBC-Ab002 in AD patients.

Improvement in quality measures with repository corticotropin injection (Acthar® Gel) in a prospective observational study of multiple sclerosis relapse

Authors: Kaplan J,1 Miller T,2 Baker M,3 Due B,4 Zhao E4

Affiliation: 1Kansas City Multiple Sclerosis and Headache Center, Overland Park, Kansas; 2Advanced Neurology of Colorado, LLC, Fort Collins, Colorado; 3Collier Neurologic Specialists, LLC, Naples, Florida; 4Mallinckrodt Pharmaceuticals, Bedminster, New Jersey

Objective: The goal of this study was to assess changes in outcomes that correspond to American Academy of Neurology (AAN) Multiple Sclerosis (MS) Quality Measures in patients receiving repository corticotropin injection (RCI, Acthar® Gel) for MS relapse.

Methods: In this multicenter, prospective, observational registry, patients were followed for up to 24 months after initiation of RCI therapy. Measurements included MS Impact Scale (MSIS) physical and psychological subscale scores, Expanded Disability Status Scale (EDSS) scores, Functional Systems Scores (FSS), Clinician Global Impression of Improvement (CGI-I), Work Productivity and Activity Impairment MS (WPAI:MS) questionnaire, and Health Resource Utilization (HRU) items. Statistics were assessed by Wilcoxon signed rank tests.

Results: Significant improvements from baseline were observed with RCI treatment for the mean MSIS physical subscale score (Month 2, -8.0, p=0.0002; Month 6, -9.6, p<0.0001) and psychological subscale score (Month 2, -7.9, p=0.0040; Month 6, -9.9, p=0.0012), EDSS scores (Month 2, -0.4; Month 6, -0.5; both P<0.0001), and FSS (month 2, –1.5; month 6, –1.6; both p<0.0001). CGI-I scores suggested significant improvement in 63.4 percent of patients at Month 2 and 61.4 percent of patients at Month 6 (both p<0.0001). Significant decreases from baseline were reported for the percent work time missed (-27.8; p=0.0255) and percent activity impairment (-11.5; p=0.0003) WPAI:MS subscale scores at Month 6. Decreases from baseline to Month 6 were observed for six of 10 HRU items.

Conclusion: In this real-world registry study, RCI-treated patients experiencing MS relapse showed significant improvements in outcomes that correspond to AAN quality measures for MS.

Funding/disclosures: Professional writing and editorial support were provided by MedLogix Communications, LLC, Itasca, Illinois, under the direction of the authors, and were funded by Mallinckrodt Pharmaceuticals. JK has received speaking fees from Amgen, Allergan, Alexion, Biogen, EMD Serono, Lilly, Lundbeck, Mallinckrodt Pharmaceuticals, Sanofi-Genzyme, Teva, and Viela Bio.

TM has received speaking and/or consulting fees from Acorda Therapeutics, Allergan, Amgen, Biogen, Genentech, Mallinckrodt Pharmaceuticals, Novartis, Reven, Sanofi-Genzyme, and Teva; and has received research support from Allergan, Adamas, Biogen, Elan, EMD Serono, Genentech, Ipsen, Mallinckrodt Pharmaceuticals, Novartis, ONO Pharmaceutical, Sanofi-Genzyme, Sun Pharma, and Teva. MB has received honoraria and consulting fees from Acorda Therapeutics, Alexion, Avanir, Biogen, Celgene, Genentech, Mallinckrodt Pharmaceuticals, Sanofi-Genzyme, and Teva. BD is a former employee of Mallinckrodt Pharmaceuticals. EZ is an employee of Mallinckrodt Pharmaceuticals.

Patient Assessment

Can blinded, site-independent scoring of audio recorded site-based PANSS interviews of patients with acute psychosis replicate site-based scores and affirm treatment outcome results?

Author: Targum SD

Affiliation: Signant Health, Blue Bell, Pennsylvania

Objective: One approach to the confirmation of site-specific research assessments in multicenter trials has been to conduct paired site-independent ratings of audio-recorded site-based assessments such that the independent rating is blind to other clinical and research evaluations conducted on site. We applied this blinded assessment method to confirm data obtained from site-based raters administering the Positive and Negative Syndrome Scale (PANSS) during a clinical trial of acute exacerbation of psychosis in subjects with schizophrenia.

Methods: Data for this analysis came from a five-week, Phase II, randomized, double-blinded study (EMERGENT-1) to assess the safety, tolerability, and efficacy of the investigational antipsychotic KarXT in hospitalized adults with an acute exacerbation of schizophrenia enrolled in a Phase II placebo-controlled, double-blind, randomized, controlled trial (RCT) (ClinicalTrials.gov Identifier: NCT03697252). KarXT is a fixed combination of xanomeline and trospium chloride being investigated for the treatment of subjects with schizophrenia.

The site-independent PANSS ratings were derived from audio recordings of site-based PANSS interviews with accompanying digital notes that provided corroborative informant information. The paired PANSS scoring data were used for quality assurance to independently confirm the site-based PANSS ratings. The study plan anticipated a review all PANSS interviews for completeness and independent scoring of 100 percent of interviews conducted at the baseline and endpoint visits and 20 percent at the other visits.

We conducted concordance analyses of the paired site-based and site-independent PANSS interviews and compared the PANSS treatment outcome at the study endpoint between these two methods of assessment.

Results: There were 553 pairs of paired PANSS scores available for analysis from the full study data. There were 142 subjects with paired site-based and site-independent PANSS data available from both the baseline and end of study (Week 5) visits.

Concordance analyses of the paired site-based and site-independent PANSS ratings revealed a high correlation (ICC=0.746; p<0.001) with minimal scoring discordance. Paired scoring differences were positively correlated with the PANSS total score (Spearman’s rho=0.35, p<0.001). The site-based PANSS total scores revealed a significantly greater improvement from baseline in the KarXT treatment group compared with placebo (p<0.0001). The blinded site-independent PANSS total scores derived from listening to and scoring the recorded site-based PANSS interviews replicated this finding (p=0.0005).

Conclusion: Blinded site-independent scoring of audio-digital recordings of site-based PANSS interviews confirmed the primary PANSS findings in this schizophrenia study and obviated any concerns about possible functional unblinding. This method of blinded assessment via audio-digital recordings might have utility for other studies concerned with ratings precision and/or functional unblinding.

Funding/disclosures: Support for this report was provided by Karuna Therapeutics and Signant Health. Targum, Murphy, and Daniel are employees of Signant Health; Brannan is an employee of Karuna Therapeutics; Breier is chair of the Scientific Advisory Board at Karuna Therapeutics.

Comparison of baseline severity of schizophrenia negative symptom study subjects in the United States versus the rest of the world. 

Authors: Daniel DG,1 Wang X,1 Kott A1

Affiliation: 1Signant Health, Blue Bell, Pennsylvania

Objective: To assess whether baseline characteristics of negative symptoms of schizophrenia in study subjects differ in the United States versus rest of the world (ROW).

Methods: Our dataset consisted of baseline PANSS data collected from six negative symptom schizophrenia clinical trials. Total PANSS score, PANSS subscales, and PANSS factor scores were compared between the United States and the ROW using a t-test.

Results: (N = 2,743 subjects): The PANSS negative and general psychopathology subscales and the PANSS disorganized thought factor were significantly lower in the United States versus the ROW. PANSS positive subscale scores were significantly higher in the United States versus ROW. The PANSS total did not differ between United states and the ROW.

Conclusion: Baseline scores on key inclusion related symptom domains in negative symptom studies differed in the United States versus the ROW. The extent to which this reflects differing patient population, interpretation of symptoms, and rating practices should be explored in future studies.

Funding/disclosures: All authors are full-time employees of Signant Health.

Comparison of three methods for MADRS administration: Correlation with CGI-S

Authors: Machizawa S,1 Brown J,1 Roy M,1 Sachs G,1,2 Tackett ZC,1 Zhou W,1 Feaster T,1 Hong B1

Affiliation: 1Signant Health, Blue Bell, Pennsylvania; 2Massachusetts General Hospital, Boston, Massachusetts

Background: Variation in administration of clinician outcome assessments, such as the MADRS, is recognized as a contributor to clinical trial failure. Structured interview, such as the Structured Interview Guide for the MADRS (SIGMA), offers a solution to this problem. In validation studies, the SIGMA has demonstrated better interrater reliability. Hence, most protocols using the MADRS in multicenter trials require raters to base their scores on a structured interview. Results in actual clinical trials, however, remain unsatisfying.

Objective: Based on review of taped interviews and feedback from clinical raters, we hypothesized that the performance demands of paper-based structured interviews often led to errors and/or the abandonment of the structured interview in favor of an unstructured clinical interview. To explore this hypothesis, we compared the correlation of Clinical Global Impression of Severity (CGI-S) with the MADRS total scores administered in protocols utilizing the usual SIGMA paper prompted interview (PPI) versus two methods of electronic structured interview administration, which use algorithms to lower or eliminate rater burden: the clinician-administered Computer Prompted Interview (CPI) for the MADRS presents raters with the next question to ask and prompts that follows the logic required for the rater to complete rating for each item; and the Computer Simulated Rater (CSR) administers the structured interview and scores the subject’s responses without the participation of a site rater. We expected the PPI ratings would be associated with greatest variability followed by the CPI and least in the CSR ratings.

Methods: Data were pooled from six MDD studies that administered the CGI-S and MADRS at screen and baseline. A Pearson’s correlation between the MADRS total score and the CGI-S score was used as a proxy indicator of variability. The Fisher Z-Transformation was used for bivariate correlation comparisons.

Results: The data set included 5,550 MADRS ratings. As expected, significant (p<0.01) positive correlation was found between the MADRS and CGI-S scores across all administration modalities, PPI: r(1981)=0.50; CPI: r(2325)=0.64; and CSR: r(1238)=0.64). The Fisher Z-Transformation indicated the correlations for the CPI and CSR were significantly higher than that for the PPI (Z=6.63, p<0.01; Z= -5.74, p<0.01; respectively) but no difference was found between the CPI and CSR (Z= -0.14, p>0.05.

Conclusion: Our results suggest that in actual clinical trial usage, the PPI is associated with more variation than either of the electronic administration methods. This variation occurred prior to randomization and cannot be attributed to treatment effects. Since the CPI and CSR produced similar results, we speculate that the advantage over PPI is driven by the computer-guided logical structure that facilitates a consistent interview without requiring raters split their attention between administrative prompts and interview questions.

Funding/disclosures: The authors are full-time employees of Signant Health.

Improving assessment in pediatric trials: development of an eCOA (electronic clinical outcomes assessment) Positive and Negative Syndrome Scale (PANSS) for use in adolescent schizophrenia trials
Authors: Busner J,1 Daniel DG,1 Atkinson SD,1 and Findling RL1

Affiliation: 1Signant Health, Blue Bell, Pennsylvania

Objective: Assessment challenges in pediatric psychosis trials are numerous, and include developmental limitations in symptom description, the need to integrate and weight information from varied sources, and the lack of pediatric-specific efficacy measures. The most frequently used primary efficacy scale is the (adult) PANSS, which has been shown to pose ratings challenges even when used with the adult patients for whom it was designed. We have previously identified PANSS items with poorest interrater reliability amongst pediatric clinical trial investigators and have developed and delivered training tools and conventions worldwide for adolescent trials resulting in multiple drug approvals across sponsors. To further the work, with the goals of streamlining administration, enhancing consistency, and improving data quality, we are developing a new guided eCOA tool to assist trials raters when administering the PANSS in pediatric trials.

Methods: Item prompts and scoring algorithms for a child/adolescent and parent/caregiver interview are in development based on analysis of expert interviews, quality reviews, and expert consensus. Consistent with our trained conventions, the tool provides guidance for investigators, helps avoid frequently seen administration errors, ensures coverage of symptoms, and provides behind-the-scenes internal scoring quality checks.

Results: The scale will be piloted at identified child psychiatric research sites.

Conclusion: We believe our work represents an advance in the field as the first pediatric-specific eCOA interview for the PANSS. As has been shown repeatedly for other eCOA outcome measures, we expect the pediatric eCOA PANSS to result in more consistent and less error-prone data for clinical trials.

Funding/financial disclosures: Presented in part at the annual meeting of the American Society of Clinical Psychopharmacology, May 26–May 29, 2020, Virtual Meeting. Financial support provided by Signant Health. 

Marijuana use among potential clinical trial participants with major depressive disorder

Authors: Engler J,1 Thorpe D,1 Evans M,1 Marecki S,1 Tucker H,1 Domilici D,1 Sauder C2

Affiliation: 1Adams Clinical, Watertown, Massachusetts; 2Karuna Therapeutics, Boston, Massachusetts

Objective: Central nervous system (CNS) clinical trials often exclude participants who use marijuana due to concerns about its psychoactive properties. However, few studies have directly examined the impact of marijuana on endpoint measures. We examined marijuana use among a Massachusetts Major Depressive Disorder (MDD) clinical trial seeking population to assess the relationship between use and antidepressant therapy (ADT) response.

Methods: Between January 2019 and August 2020, 325 participants completed an ADT prospective lead-in, the TRAIT study. All underwent a urine drug screen and eligibility assessment with the Mini International Neuropsychiatric Interview. Moderate-to-severe substance use disorder was exclusionary. Enrollees received antidepressant therapy (ADT) treatment for 4 to 12 weeks, and differences in treatment response between marijuana users and non-users were assessed using the clinician-rated Montgomery-Asberg Depression Rating scale (MADRS).

Results: About 22 percent of participants tested positive for cannabinoids, with 2.9 percent meeting criteria for marijuana use disorder. At baseline, there were no significant differences in MADRS scores between marijuana users (M=33.2, SD=5.8) and nonusers (M=33.3, SD=5.3). However, there was a significant difference in treatment response for users and nonusers (t(46)=1.87, p=0.03). On average, marijuana users had an 18.1 (SD=12.0) MADRS reduction at end of study compared with 13.5 (SD=13.0) for nonusers.

Conclusion: Our results suggest marijuana use is associated with greater response to ADT. Given its widespread popularity and impact on ADT response, marijuana use exclusions should be carefully considered in future clinical trials.

Funding/disclosures: All presenters are current or former employees of Adams Clinical, an independent CNS research site that conducts industry-sponsored pharmaceutical trials.

Proving the equivalence of simplified home sleep testing to polysomnography

Authors: Dorffner G,1,2 Kemethofer M,2 Gruber G,2 Parapatics S,2 Loretz E2

Affiliation: 1Medical University of Vienna, Section for Artificial Intelligence and Decision Support, Vienna, Austria; 2The Siesta Group, Wien, Austria

Objective: To evaluate a reduced, self-applicable electrophysical montage for measuring sleep in a subject’s home compared to gold standard polysomnography (PSG)

Methods: A retrospective analysis of data from 20 healthy subjects was performed. The montage used two electrodes near the eyes and one at the mastoid behind the right ear. A standard portable PSG was applied in parallel. We employed an equivalence test based on a tolerance interval defining which discrepancy between methods is still negligible. Here the inter-rater variability among expert scorers can be exploited. The average difference between experts was available from another study, the largest of which can be considered as tolerable. The test was then based on calculating the 90-percent confidence interval for the average discrepancy between reduced montage and full PSG. Confidence interval that lies completely within the tolerance interval will demonstrate equivalence.

Results: Most of the sleep variables demonstrated equivalence. Tolerance intervals were defined for sleep efficiency: [-4.1–4.1] percent points, percentage of N1 sleep: [-5.6–5.6] percent, N2 sleep: [-7.5–7.5] percent, N3 sleep [-7.7–7.7] percent, REM sleep [-2–2] percent. With one exception, 90 percent confidence intervals for the mean deviation between reduced montage and full PSG were within those tolerance intervals: efficiency: [-0.55–1.49], N1: [-1.52–2.01], N2: [-2.24–3.45], and N3: [-2.19 –3.36]. Only for REM sleep they were not [-3.21–0.33].

Conclusion: Results suggest that for most variables of sleep architecture, the measurement by a self-applicable montage is statistically equivalent to a full PSG.

Funding/disclosures: The original study was partly funded by the Vienna Center for Innovation and Technology (ZIT). Georg Dorffner, Manuel Kemethofer, Silvia Parapatics, Erna Loretz and Georg Gruber are employees and shareholders of The Siesta Group, a service provider for measuring electrophysiological signals including sleep in clinical trials.

[A] test-case for deploying clinical measures of neurological function remotely and via local healthcare provider in the state of Wyoming

Author: Remington E,1 Kangarloo T,1 Erb K,1 Scott N,1 Butts C,1 Fisher K,1 Gee A,2 Bergethon P1

Affiliation: 1Biogen, Cambridge, Massachusetts; 2Wyoming Health Innovation Living Laboratory, Cody, Wyoming

Objective: Distributed clinical trials (DCTs), in which some or all of trial activities are conducted via mobile or local healthcare providers, telemedicine, and/or mobile technologies, can expand the geographic reach of clinical trials, accelerating enrollment, providing access to patients with rare diseases, and increasing participant diversity. For neurological disorders, a challenge is that standard clinical assessments require a neurologist’s judgement, extensive training, or both. This effort explores how clinical and mobile technologies can help realize the potential of DCTs in neurological disease by enabling objective measurement of patient function without requiring the expertise and infrastructure of a traditional trial site.

Methods: A neurology clinic in rural Wyoming without clinical research experience was selected to evaluate the feasibility of conducting a study involving in-clinic and at-home instrumented motor assessments and app-based cognitive tests. We assessed challenges pertaining to general logistics of human-subjects research, as well as those associated with the specific technologies and assessments.

Results: Prior to enrollment of study participants, challenges included funding study-specific staff, delivery and collection of equipment, complexity of specific assessments (hardware setup, procedure training, and software configuration), and data management (no site experience with electronic data capture solutions).

Conclusion: Setting up a site to conduct clinical research is a significant undertaking, but early evidence suggests that digital tools can provide value to the DCT model. Future efforts could focus on more efficiently transferring human subjects research expertise to local health care providers, as well as streamlining all aspects of technologies enabling remote clinical assessments.

Which cognitive domains contribute most to large MMSE changes in the screening period?

Authors: Kott, A,1 Miller D,1 Daniel DG1

Affiliation: 1Signant Health, Blue Bell, Pennsylvania

Objective: Instability of cognition measurements (e.g., MMSE changes 5 points) during the screening period is a risk factor for poor drug-placebo separation in Alzheimer’s clinical trials. We have previously reported such instability in as many as 12.5 percent of study subjects. [Kott, Miller, 2020]. In the current analysis, we assessed which cognitive domains contributed most to the observed instability.

Methods: MMSE changes of 5 points or more (“large changes”) between screening and baseline were identified in data pooled from eight randomized, double-blind clinical trials in prodromal-to-moderate severity Alzheimer’s disease. While correcting for screening MMSE score, we estimated the association of change in individual domains with the presence of the large changes using logistic regression.

Results: Our dataset consisted of 6,480 pairs of screening and baseline MMSE scores. Large changes were identified in 470 subjects (7.25%). The three domains most associated with the unexpectedly large MMSE changes were orientation to time OR=2.98(2.65–3.64), orientation to place OR=2.71(2.35–3.12), and attention and concentration OR=2.37(2.15–2.61), all p<0.001. The association of the other domains with large MMSE changes, while still significant, was weaker.

Conclusion: Our analysis reconfirmed our prior finding that unexpectedly large MMSE changes in the screening period affect a substantial number of subjects. The cognitive domains of orientation to time and place were most frequently impacted, followed by attention and concentration.

Funding/disclosures: All authors are full time employees of Signant Health.

Trial Methodology

Capturing clinical symptoms with ecological momentary assessment: Convergence of momentary reports of psychotic and mood symptoms with diagnoses and standard clinical assessments
Authors: Harvey PD,1 Pendergrass JC,1 Depp CA,1 Pinkham A2

Affiliation: 1University of California, San Diego; 2University of Texas at Dallas School of Behavioral and Brain Sciences

Objective: The development and deployment of technology-based assessments of clinical symptoms are increasing. This study uses Ecological Momentary Assessment (EMA) to examine clinical symptoms and relates these sampling results to structured clinical ratings.

Methods: Three times a day for 30 days, participants with bipolar disorder (n=71; BPI) or schizophrenia (n=102; SCZ) completed surveys assessing five psychosis-related and five mood-related symptoms, in addition to reporting their location and who they were with at the time of survey completion. Participants also completed PANSS interviews with trained raters. Mixed-model repeated-measures (MMRM) analyses examined diagnostic effects and the convergence between clinical ratings and EMA sampling.

Results: There were 12,406 EMA samples collected, with 80-percent adherence to prompts. EMA-reported psychotic symptoms manifested substantial convergence with equivalent endpoint PANSS items. Patients with SCZ had more severe PANSS and EMA psychotic symptoms. There were no changes in symptom severity scores as a function of the number of previous assessments.

Conclusion: EMA surveyed clinical symptoms converged substantially with commonly used clinical rating scales in a large sample, with high adherence. This suggests that remote assessment of clinical symptoms is valid and practical and was not associated with alterations in symptoms as a function of reassessment, with additional benefits of “in the moment” sampling such as eliminating recall bias and the need for informant reports.

Considerations for clinical trial design in frontotemporal lobar degeneration (FTLD)
Authors: Koschalka L,1 Bell J,1 Katz S1

Affiliation: 1Syneos Health, Morrisville, North Carolina

Objective: Frontotemporal lobar degeneration (FTLD), once considered rare, is the second most prevalent form of early-onset neurodegenerative dementia. Three main clinical variants are recognized within the FTLD spectrum: behavioral variant frontotemporal dementia (bvFTD), semantic dementia (SD), and progressive non-fluent aphasia (PNFA). Due to the often early age of onset and prominent behavioral and personality symptoms, it might be mistaken for a primary psychiatric disorder. There has been a marked increase in the number of Phase II and III industry-sponsored clinical trials in FTLD in recent years. As might be expected in an emerging field of study with a challenging population, there have been clear differences in methodological approaches to study design.

Methods: This literature review will seek to provide reflections for trial design that address the challenges of patient selection—only 20 percent of patients in a recent behavioral variant FTD trial subjects were appropriate for study enrolment.

Results: Recommendations will be provided and mitigation strategies discussed from the perspective of experienced clinical trialists.

Conclusion: Biomarker selection will be opined, as well as cognitive and behavioral scale selection with tabulated data regarding the psychometric qualities of key tests.

Funding/disclosures: All are employees of Syneos Health, a CRO providing contracted services for Clinical Trials

Differences in eConsent among diagnosis groups

Authors: Profit D,1 Koenen C,1 Carson W,1 Saldarini C,1 Gant J,1 Belete M,1 Glenny H1

Affiliation: 1Otsuka Pharmaceutical Development and Commercialization, Inc., Tokyo, Japan

Objective: To identify and provide insights on the differences of the eConsent process among diagnosis groups

Methods: eConsent data from 27 clinical trials, totaling 2,364 randomized participants, were analyzed.

Results: Participants had a diagnosis of healthy, autosomal dominant polycystic kidney disease (ADPKD), attention deficit hyperactivity disorder (ADHD), post-traumatic stress disorder (PTSD), bipolar disorder, major depressive disorder (MDD), or schizophrenia. Those with bipolar disorder spent significantly less time on the instructional video while those with schizophrenia spent significantly more. Engagement was defined as participants who watched at least 75 percent of the video and were in the 50th percentile of their specific consent form. Participants with ADPKD had an engagement rate of 27 percent, making them the least engaged, while participants with schizophrenia were the most engaged at a rate of 49 percent. Knowledge check scores for participants with ADHD averaged 93 percent, significantly lower than the other diagnosis groups.

Conclusion: The data available through eConsent provides crucial insights into the consenting differences among participants diagnosis. Understanding these differences will support tailoring the eConsent process to a more patient centric design and ensure study participants understand what they are consenting to.

Funding/disclosures: This research was funded by Otsuka Pharmaceutical Development and Commercialization, Inc. 

Driving better risk-based quality management (RBQM) using atypical data detection for CNS trials

Authors: Roberts R, Barag J, Armstrong K

Affiliation: Remarque Systems, Premier Research in North Carolina. 

Objective: Demonstrate how to operationalize RBQM to drive clinical trial execution with the highest standard possible 

Methods: RBQM is a huge paradigm shift and is effectively the opposite of the historical way we assessed data quality in clinical trials—reviewing a patient visit observation, such as each temperature visit by visit. Instead of starting at the bottom and working up to review the data, we start at the top and work down. However, RBQM has to be thoughtfully implemented. By using two tools in conjunction, a risk-assessment combined with atypical data detection (a data-driven statistical approach), we provide a team with a holistic means to methodically review data. 

Results: The guidance from our Regulatory Agencies is clear: use RBQM. Our front line of defense is performing a mindful risk assessment allow the risks to drive the decisions. In addition, and complementary to risk-based approaches, utilizing a high-level data-driven approach allows for a wider net to be thrown. We’re in search of the approximate 3 percent data errors—to preserve the quality and safety of our trials. This combined approach provides a roadmap to find those most effectively, efficiently, and respectively.

Conclusion: The Clinical Research Industry is at the precipice of a revolution. We have the opportunity to choose, together, to move forward in attending to the data integrity of clinical trials in the most effective way and depart from the traditional way of monitoring trial data. A methodology combining both risk-based and data-driven approaches respects the practice of clinical research to the highest possible standard.

High diagnostic confidence improves signal detection

Authors: Sachs G, Siu C, Pikalov A, Loebel A

Objective: Does enrollment of subjects with low diagnostic confidence contribute to poor signal detection?

Introduction: Eligibility criteria for clinical trials testing treatments for subjects with major depressive episodes typically require subjects meet both the DSM diagnostic criteria for MDE and exceed a threshold score on a depression severity scale, such as the Hamilton depression rating scale (HAM-D) or the Montgomery Asberg Depression Rating scale (MADRS).

These diagnostic and severity assessment techniques should produce congruent results. Notably, the 10 items of the MADRS roughly map into eight of the DSM-V nine criteria for major depressive episode. All MADRS items are scored on a 0 to 6 scale. An item scored 4 or higher represents clear cut clinically significant pathology. The DSM 5 criteria requires at least five symptoms most of the day nearly every day. Since the MADRS only covers eight of the nine criteria, we expect that subjects with a confident diagnosis of MDE will have a confirmatory MADRS with at least four items scored 4 or higher.

Our prior analysis of blinded data from 13 placebo-controlled RCTs accessing treatments for acute depressive episode suggested the percentage of subject enrolled with a high confidence diagnosis might be a useful quality indicator. None of the trials in which less than 70 percent of the sample met the definition for high confidence were successful.

Methods: A post-hoc analysis was conducted using data from a double-blind RCT in which subjects meeting DSM 5 criteria for bipolar depression were randomized to receive monotherapy with lurasidone 20 to 60mg/day (n=161), lurasidone 80 to 120mg/day (n=162), or placebo (n=162). The primary and key secondary study endpoints in that study were changes from baseline to Week 6 on MADRS and Clinical Global Impression Bipolar version, Severity of Illness (CGI-BP-S), respectively.

We defined high diagnostic confidence as scoring in the clinically significant range on at least four MADRS items at both the screening and pretreatment baseline visits. Low diagnostic confidence was defined as not meeting the high confidence criteria.

Results: A total of 320 (66%) of patients meets the criteria for high diagnostic confidence. High diagnostic confidence was associated with a significantly higher drug-placebo difference. in MADRS total change score at Week 6 (P<0.05, statistical interaction test). We found a placebo-corrected effect size of 0.64 (lurasidone -16.15 vs. placebo -10.34, n=320) in the high diagnostic confidence group compared with 0.17 in the low confidence group (lurasidone -13.66 vs. placebo -12.13, n=165). Furthermore, the placebo-corrected effect size for improvement of CGI-BP-S (depression) from baseline to Week 6 endpoint was 0.72 (-1.79 for Active vs. -0.99 in the high diagnostic confidence group compared with 0.17 in the low confidence group (-1.71 for Active vs. -1.52 for placebo), indicating significant moderating effect by diagnostic confidence (P<0.05, diagnostic confidence by treatment interaction effect at Week 6)

Conclusion: This post-hoc analysis suggests that limiting enrollment to subjects with high diagnostic confidence can increase the detectable treatment effect size. We observed subjects with high diagnostic confidence have both increased response to active drug, and a reduced response to placebo. Our findings have implications for patient selection to improve the likelihood of detecting a treatment signal in bipolar depression trials and perhaps other indications.

Implementing a fully decentralized clinical trial for major depression
Author: Reist C

Affiliation: Science 37, Inc., Los Angeles, California

Objective: The steps for bringing a decentralized clinical trial (DCT) for major depressive disorder (MDD) from concept to reality will be described.

Methods: Science 37 recently launched the first full interventional DCT for MDD. A technology platform supports interactions with the patient including video communication, eConsent, medication adherence, and entry of patient reported outcomes (PRO). Other components include a centralized rater team, refining messaging for digital recruitment, development of digital versions for PROs, and deploying a mobile nursing team for conducting in-home assessments and data collection. Two projects provided valuable experience: a pilot study to compare face-to-face assessments with those done remotely in home and a four-week, non-interventional feasibility study enrolling 57 patients with MDD.

Results: Key findings from study of remote administration of assessments:

  • Face-to-face assessment showed high concordance with the home telemedicine approach (MADRS ICC: 0.98 (.93, .99).
  • Subjects successfully completed PROs and preferred the mobile interface (compared with paper).
  • Subjects were comfortable with interviews conducted with a mobile device.
  • Key findings from the four-week, non-intervention trial:
  • Appropriate patient population can be identified using social engagement
  • Very high retention (>90%)
  • Reliable data collection
  • Remote administration of ClinROs successful
  • Subjects reliably used devices and ender PRO data
  • High level of satisfaction

Conclusion: The DCT model is suitable for studies focused on major depression. The platform allows clinical operations to be carefully coordinated between investigators, raters, coordinators, mobile nurse team, and the trial participant. Information can be reliably collected in a way that is patient-centric, resulting in a high level of satisfaction.

Funding/disclosures: All authors are employees of Science 37, Inc.

Pharmacometric analyses, reformulation of BNC210, and fast track designation support BNC210 evaluation in a second Phase III PTSD trial

Authors: Doolin E,1 Paul D,1 Crossman J,1 Odontiadis M,1 Rolan P,1 O’Connor S1

Affiliation: 1Bionomics Limited, Adelaide, Australia

Objective: BNC210, a novel negative allosteric modulator of the alpha 7 nicotinic acetylcholine receptor, was evaluated in 2018 in a Phase II Post-Traumatic Stress Disorder (PTSD) trial (RESTORE) in 193 patients but did not meet the primary endpoint [(change in Clinician-Administered PTSD Scale (CAPS-5) scores]. Trial participants were administered a liquid suspension formulation of BNC210 that was required to be given with food; subsequent pharmacokinetic (PK) modelling showed that participants had much lower exposures than expected. A pharmacometric analysis established an exposure-response (PK-PD) model between BNC210 exposure and CAPS-5 scores (p<0.01), indicating the potential for significant benefit from BNC210 if adequate exposures are achieved. Our objective was to develop a solid dose formulation that provides adequate exposure to support further evaluation of BNC210 in a second Phase 2 PTSD trial.

Methods: A spray-dried dispersion tablet formulation of BNC210 was developed to overcome the food- dependence for maximal absorption of the suspension formulation used in the RESTORE trial.

Results: In healthy volunteer PK studies, the BNC210 tablet overcame the “food effect” of the liquid suspension and achieved exposures predicted by the PK-PD model to give clinically meaningful and statistically significant changes on CAPS-5 scores. These data and the PK-PD modelling supported a successful application for FDA Fast Track designation for BNC210 in the treatment of PTSD.

Conclusion: Pharmacometric modeling, improved drug formulation, demonstrated favorable safety profile in patients with PTSD, and Fast Track designation for BNC210, support Bionomics’ plans for a second Phase II PTSD trial planned to commence in mid-2021.

Funding/disclosures: Funded by Bionomics Limited

Pinpointing placebo responders using multicomponent vocal analysis
Author: Begel D

Affiliation: AP Systems and Chase Love, University of LaVerne, LaVerne, California

Objective: To study the feasibility of using acoustic analysis of speech to predict placebo responders in a clinical trial

Methods: Twenty subjects satisfying clinical criteria for depression and scoring greater than 24 on the HAM-D or MADRS were randomized to receive either placebo (N=10) or bupropion SR 100mg (N=10) for three weeks. Prior to treatment, two-minute recordings of free speech were obtained. Twenty-second specimens were extracted. In PRAAT, multiple acoustic features were measured periodically at 10ms intervals. In MATLAB, simultaneous deltas of all features of phonated speech were computed at 10ms intervals. Dense (core) neighborhoods of paired feature delta values were distinguished from less dense (border) neighborhoods. At three weeks, subjects classified as placebo responder (PR) reported clinical improvement and scored 15 or less on the HAMD/ MADRS. Non-responders (PNR) reported minimal improvement and scored 20 or greater. A principal component analysis was performed yielding 97 variables correlating at greater than 60 percent with both PR and PNR distributions. Significance tests were performed.

Results: PR=4 PNR=6. Initial HAMD/MADRS: mean(SD)PR= 31.5 (2.6), mean(SD)NPR= 31.8 (7.5) Final HAMD/MADRS: mean(SD)PR=12.3(2), mean(SD)PNR=25.8 (4.5). Two variables were significant. Variable 1 is the inverse of core pitch/intensity covariance. Variable 2 is the inverse of border pitch/intensity covariance. Lower values reflect tighter linkage of pitch and intensity activity.Variable 1, mean (SD) PR=154.5 (102.3); Variable 2, mean (SD) PR=211.8 (143.2), mean (SD) PNR= 39.6 (51.7), mean (SD) PNR=54.6 (62.8;) p<0.045, p<0.043; Cohen’s d=1.42, Cohen’s d=1.42.

Conclusion: Acoustic analysis might be a feasible way to predict placebo response. With larger sample sizes and segmented data extraction the method described here lends itself to machine learning techniques offering greater predictive power.

Rare neurodegenerative diseases: Early-stage challenges and optimal animal models in orphan drug development
Author: Duarte D

Affiliation: INFARMED, National Authority of Medicines and Health Products, Lisbon, Portugal

Objective: Investigate the clinical/nonclinical nature of data to support medical plausibility of orphan designations in European Union (2001–2019) for a group of rare and pediatric neurological diseases and critical revision of preclinical models that may be used to support medical plausibility in rare neurodegenerative diseases, with orphan designations based only in nonclinical data

Methods: Observational, retrospective study

Results: Many disease-specific animal models are used to test emergent medicines in neurology, but there is a substantial difficulty in choosing an optimal model or choosing measurements which would be truly informative of the product’s efficacy. From our sample of 30 diseases, 70 percent are rare with pediatric onset and 37 percent have approved orphan designations. The use of nonclinical data was significantly higher than clinical data (65% vs. 35%, p=0.013) to support medical plausibility. Appropriate animal models are presented for two conditions, based on the knowledge of the molecular pathology of the human disease, are a significant element to support the medical plausibility of an orphan designation during the development of orphan medicines for rare neurological diseases.

Conclusion: Identification of models with best predictive value as well as those acceptable based on their face value. Nonclinical appropriate models, assessing disease relevant endpoints, might contribute to increase the translational value of animal models, in pediatric and rare neurological area, to accelerate research and the effective development of treatment options.

Use of euclidean distance heat mapping to identify data of concern from patient reported outcomes in atopic dermatitis: A potentially useful paradigm for psychiatry trials?
Authors: Everhart AT,1 Daniel DG,1 Wang X,1 Kott A1

Affiliation: 1Signant Health, Blue Bell, Pennsylvania

Objective: Psychiatry clinical trials have relatively low success rates (24% and 56% in Phases II and III, respectively). The relatively modest drug-placebo differences observed in psychiatry trials are multi-factorial in origin but appear in part to be a product of the subjective rating scales utilized by the field. Ongoing monitoring of data quality indicators is frequently employed during psychiatry clinical trials to identify measurement error and fraud in the application of clinician reported outcomes. Data quality monitoring techniques are rarely applied to patient reported outcomes (PROs) in psychiatry trials with the same rigor as they are applied to clinician reported outcomes. We described the use of Euclidean Distance (ED) and heat mapping to identify data of concern from PRO measures in a dermatology clinical trial as a potential model for monitoring PROs in psychiatry.

Methods: Data from 1,799 patients participating in an ongoing atopic dermatitis study were evaluated for atypical data patterns in PRO measures from four different instruments. ED was calculated for each patient versus every other patient within the same investigative site using the baseline score for each of 23 questions: six questions from the EQ-5D-5L, 10 questions from the Dermatology Life Quality Index, (DLQI), two questions from the Scoring Atopic Dermatitis (SCORAD), and seven questions from the Patient Oriented Eczema Measure (POEM). A heat map was then generated where darker shading corresponded to smaller EDs.

Results: Visual evaluation of the heat map identified a single site with extremely low EDs for its patients. Further evaluation of EQ-5D-5L completion times for this site revealed significantly shorter durations when compared with the remaining sites in the study.

Conclusion: The combination of low EDs and short completion times for PROs at a single site raises concern for possible data manipulation, which would need to be evaluated further. The techniques employed in this dermatology study could and should be used in other therapeutics areas such as neurology and psychiatry where patient reported, and clinician reported outcome measures are also susceptible to possible manipulation.