Welcome to the AD/PD™ 2021 Interactive Program
The congress will officially run on Central European Time (CET) - Barcelona Time
To convert the congress times to your local time Click Here
Icons Legend: ![]()
- Live Session | 
- On Demand Session | 
- On Demand with Live Q&A
The viewing of sessions, cannot be accessed from this conference calendar. All sessions are accessible via the Main Lobby.
FOLLOWING THE LIVE DISCUSSION, THE RECORDING WILL BE AVAILABLE IN THE ON-DEMAND SECTION OF THE AUDITORIUM.
DIGITAL TWINS ENABLE SMALLER TRIALS THAT MAINTAIN THEIR POWER: A DEMONSTRATION FOR ALZHEIMER’S DRUG TRIALS
Abstract
Aims
Smaller and faster trials could accelerate drug development in Alzheimer’s Disease (AD). We set out to demonstrate that smaller trials that maintain their power can be designed using a machine-learning model of AD progression trained on placebo subject records from past clinical trials.
Methods
Digital twins are longitudinal, patient-level placebo records with baseline characteristics and treatment duration matched to those of actual subjects randomized into a study. Because they predict outcomes for individual subjects, the outcomes of digital twins may be adjusted for as prognostic covariates to add power while preserving type-I error control (unlike many other methods of historical borrowing). We re-analyzed a placebo-controlled randomized study of docosahexaenoic acid (DHA) in 402 subjects with mild to moderate AD and compared the results to an analysis of a reduced dataset powered using digital twins.
Results
Using digital twins, the same power (80%) was attained with 18% fewer subjects than in the original trial. The estimated standard errors in the results of the digital twins trial were nearly identical to those in the original trial, indicating equal level of confidence in the results despite smaller sample size. Both analyses produced similar (not statistically significant) results on the primary endpoint of change in ADAS-Cog11 over 18 months.
Conclusions
Our retrospective analyses indicate that digital twins enable smaller trials with equal design power and can thus accelerate clinical trials without sacrificing type-1 error control. Our methodology pairs an innovative use of machine learning with proven statistical methods that easily integrate with trial protocols.
ADVANCING DIGITAL TECHNOLOGIES FOR PARKINSON’S: A PATIENT-CENTRIC STRATEGY
Abstract
Aims
To develop a patient-centric strategy to advance the use of digital health technologies in Parkinson’s disease (PD) clinical trials.
Methods
Three source-domains were defined to generate comprehensive inventories to inform the regulatory strategy for PD: Voice of the Patient (VoP), existing clinical outcome assessments (COAs), and studies employing digital health technologies (DHTs). For VoP, 38 publications were selected for assessment from 69 publications after initial screening. For COAs, 172 COAs from 22 publications identified by the Movement Disorder Society (MDS) task force were screened. For studies employing DHTs, 67 distinct studies were identified.
Results
VoP assessment provided information about clinical manifestations and their importance. Quantitative metrics were used to rank the symptoms important to patients. For the identified symptoms, existing COAs were assessed for limitations in addressing signs and symptoms of PD, and for the possibility of utilizing DHTs to improve clinical assessments. Integration of information from the three source-domains resulted in creating a process flow to assess the maturity of DHTs to monitor clinically meaningful features of PD, and for engaging regulatory agencies. This methodology results in identification of gait, sleep, speech, etc. as more mature concepts in comparison to muscle ache, digestive issues, comprehension etc. on the targeted maturity continuum.
Conclusions
The proposed patient-centric strategy provides a methodology to assess gaps that exist in quantifying PD symptom burden and identify areas to facilitate advancement of DHTs in PD clinical trials. This process assists in identifying symptoms to target for future DHT development and regulatory use in PD drug development.
DESIGNING A CLINICAL DECISION SUPPORT SYSTEM (CDSS) FOR PERSONALISED MEDICINE IN PARKINSON’S DISEASE
Abstract
Aims
To design an EHR-agnostic, CDSS for personalized medicine approaches in the management of Parkinson's disease (PD) that complements symptomatic treatment by adopting a holistic strategy, as well as passive (with IoT devices) and active (with diaries) patient monitoring.
Methods
The design is based on the recent literature on CDSS, the findings of previous studies evaluating mhealth for the management of Parkinson’s and the analysis of user needs that defined shared decision making, flexibility that accounts for variation among clinicians and monitoring of information integration from multiple sources as the main design principles.
Results
PRIME is a traditional CDSS in the sense that it is comprised of interoperable, FHIR compliant, software designed to be a direct aid to clinical-decision making; the characteristics of an individual patient derived from EHRs, IoT devices such as Apple iWatch and diary data, and processed with machine learning methods, are matched to a computerized clinical knowledge base (derived from Clinical Guidelines, drug and gene interaction DBs and the PD ontology) and patient-specific assessments or recommendations are then presented to the clinician for a decision through a dedicated, user interface.
Conclusions
PRIME which is co‐financed by the European Regional Development Fund of the European Union and Greek national funds through the Operational Program Competitiveness, Entrepreneurship and Innovation, under the call RESEARCH – CREATE – INNOVATE (project code:Τ2EDK- 05199) will be an evidenced-based CDSS capable of leveraging data and observations otherwise unobtainable or uninterpretable by humans and produce appropriate alerts.
PREDICTING SYMPTOM ONSET IN PARKINSON'S DISEASE WITH LATENT MIXED-EFFECT MODEL
Abstract
Aims
Due to their slow evolution, neurodegenerative diseases can be observed on long periods with several measurements of various biomarkers. Statistical models can fit such longitudinal data to describe the disease's evolution. We show how such a model can predict the onset of diseases' symptoms.
Methods
We used data from patients included in the NS-PARK cohort collecting longitudinal motor and non-motor symptoms from patients with Parkinson’s disease followed in the 25 expert centers in France and updated at each visit at the center. Only patients with at least 3 visits were included into the analysis. We modelled the observations with a non-linear mixed-effect model called Leaspy describing the latent evolution of the disease. The model is composed of population parameters which represent the average disease trajectory. This trajectory can then be personalized for each subject, with an individualized evolution of the disease which we used to predict the future onset of symptoms for new patients.
Results
2821 patients from the NS-PARK cohort were included into the analysis (mean age : 66+/-11; follow-up duration : 2.6+/-1.3). Analysis of the model helps us to understand the variability of each symptom and the prediction task separates symptoms in two groups : symptoms which we can roughly predict (dementia, postural instability...), and symptoms which are unpredictable (insomnia, impulse control disorders...).
Conclusions
Our model allows to decipher between symptoms which are due only to the disease and symptoms which depend on external factors. In the first case we are able to partially predict symptom onset.
3D VIDEO MOTION ANALYSIS AS A METHOD OF PERSONALIZED REHABILITATION ASSESSMENT IN PARKINSON'S DISEASE
Abstract
Aims
Parkinson's disease is one of the most common neurodegenerative diseases, the main clinical manifestation of which is movement disorders. Taking into account the slowly progressing of the disease and the complexity of syndrome that subsequently forms a characteristic movement pattern, the study of innovative objective methods for the diagnosis and rehabilitation of movement disorders in PD is relevant and in demand.
Methods
This article provides an example of a personalized rehabilitation assessment of biomechanical manifestations of the gait function of a patient with a refined diagnosis of stage 3.5 PD according to Hoehn and Yahr, who has postural and gait disorders, using the method of three-dimensional movements video analysis using the Vicon Motion Capture Systems hardware and software complex. The method of movements video analysis was applied after undergoing a rehabilitation course based on the activation of the foot separation from the support surface (“back push”).
Results
Changes in the tempo-rhythm parameters of walking in a patient with PD in comparison with a healthy person were revealed: a decrease in the time of double support, an acceleration in the moment of leg separation and a decrease in walking speed, a decrease in the amplitude of flexion in the hip joint in the phase of double support and the phase of its transfer, insufficient extension and flexion of the lower limb of the most affected side, and a decrease in the amplitude of dorsal flexion of the foot.
Conclusions
The 3D VIDEO MOTION ANALYSIS is a valuable diagnostic tool to objectively identify the targets of rehabilitation.
AN INTERPRETABLE COMPUTER VISION MODEL FOR ALZHEIMER’S DISEASE TO IDENTIFY BRAIN BIOMARKERS LINKED TO COGNITIVE DECLINE
Abstract
Aims
The last few years have observed tremendous advancements in the application of computer vision algorithms for Alzheimer’s disease (AD). Although models have reached high accuracy, they fail to connect brain imaging biomarker changes and AD-related cognitive decline. In this study, we tested the ability of medical images (MRI, FDG-PET or AV45-PET) to predict diagnosis and cognitive decline as measured by ADAS-Cog13, using 3D-CNNs. We also aimed to identify brain biomarkers that are most sensitive to AD-related cognitive decline.
Methods
Briefly, the core 3D-CNN is composed of 14 convolutional layers with inception and skip connections, followed by two fully connected layers and a final output layer for each of ADAS-Cog13 sub-scores. A Feature Importance algorithm is applied to identify brain regions most important for model predictions. We included two extensions to the model: A diagnostic extension (nAD: AD), and an interpretable AI extension (Grad-RAM).
Results
Only MRI and FDG-PET models could predict cognitive performance with R2 of ~80%, whereas all 3 modalities achieved similar diagnosis accuracies of >90%. The MRI model diagnostic extension was validated on an external database (RADC, 84%). The MRI and FDG-PET 3D-CNNs paid attention to different brain regions when predicting cognitive abilities. The hippocampus was the most important brain region for MRI images, whereas FDG-PET used a collection of cortical regions.
Conclusions
In summary, our work demonstrates the link between cognitive decline and brain biomarker changes detected by various modalities of medical images and generates insights into the underlying disease & symptomatic status of AD.
FREE-LIVING GAIT CHARACTERISTICS AMONG THOSE WITH DEMENTIA AND NEURODEGENERATIVE DISEASE DURING CONTINUOUS MONITORING USING WEARABLE SENSORS
Abstract
Aims
Objectives: Quantifying mobility, via gait analysis, in neurodegenerative disease (NDD) can indicate cognitive decline and predict incidence of dementia. Assessing gait in free-living continuous monitoring environments provides insight to naturally occurring gait, dual-tasking and infrequent events (e.g. falls). The objective of this study is to analyze free-living gait to identify key differences in daily gait (e.g. bout lengths, time of day) between people with NDD and healthy controls (CN).
Methods
Methods: Data were collected from 40 participants with NDD and 44 CN enrolled in a remote monitoring feasibility study as part of the Ontario Neurodegenerative Disease Research Initiative (ONDRI). Participants wore accelerometers mounted bilaterally on the ankles and wrists, as well as the chest 24h per day for 7-days. A novel application of existing step detection algorithms was used to define walking periods for analysis.
Results
Results: Initial findings indicate differences in volume and distribution of daily gait bouts over the testing period. Individuals with NDD show a smaller number of long gait bouts (>300 steps/bout) (NDD: 1.21 (0.88), CN: 3.75 (1.59)) and more short bouts (<41 steps/bout) (NDD: 198.02 (55.2), CN: 114.45 (44.69)).
Conclusions
Conclusions: Analysis reveals a shift in bout duration between CN and NDD groups. The distribution of walking throughout the day, as opposed to total step counts or single clinical indices, provides unique evaluation of daily walking in NDD. It is expected that in-depth analysis of gait quality (e.g. variability), with multi-modal sensors (e.g. chest and wrist-worn), may differentiate between neurodegenerative diseases and improve fall-risk detection and cognitive decline.