Abstract Submission

Abstract Submission Now Closed

Due to a high volume of requests for those needing additional time to submit their abstract, we are providing an extended deadline of Wednesday, November 28, 2018.  

Abstracts are now being solicited that highlight the latest developments in best practices in computational modeling and simulation of medical devices. Deadline for submission is November 9, 2018.  


  Track Descriptions
Track Title Track Description Track  Chair 
Clinically Relevant Models
In-silico medicine holds great promise to enable clinical procedures that are guided by reliable, patient specific outcome prediction.  The first step to effectively predict clinical outcomes is to develop digital models that mimic reality to the best of our knowledge and capabilities, and continuously improve as our understanding grows. Examples for this track include development of computer models that accurately represent the medical device, patient’s anatomy, patient’s physiology, or the interaction between the device and human body. Application areas such as computationally assisted medical imaging, surgical simulation and neuromodulation (both invasive and non-invasive) are of particular interest. 

Laleh Golestani Rad (Northwestern University)
Steve Levine (Dassault Systèmes)
Real World Data as Model Input
Using real world data for model construction and input is essential for patient specific simulation and a big step forward for ensuring that computational models appropriately represent reality.  However, challenges exist in data collection and applying the real world data to model intended context of use.  This track includes topics of real world data collection and interpretation, data infrastructure and security, sensitivity and uncertainty analyses, and validation of models with real world data inputs.  Applications of using real world data to demonstrate clinical safety are particularly welcomed.
Ethan Kung (Clemson University)
Tina Zhao (Edwards Lifesciences)
Artificial Intelligence and Machine Learning in Medical Devices
Artificial intelligence (AI) and machine learning hold big promise in transforming healthcare in the coming years.  This track includes applications of AI in medical device development in different areas including natural language processing (NLP) for mining medical data, imaging diagnostics algorithms for detection or classification, recommender systems, and pattern discovery from genomics data. Methods for overcoming the small sample size problem in training and testing of AI systems for medical applications, as well as ones that address interpretability and explainability of algorithm outputs are also of great interest.
Dave Hoadley (MathWorks)
Aria Pezeshk (FDA)
Digital Twins in Healthcare
A digital twin is defined as a simulation of a product or process that interfaces with real-world information to mirror the performance of its corresponding physical twin.  A digital twin typically relies on mechanistic models, data analytics, and machine learning to represent its’ physical counterpart.  Sources of real-world data are equally diverse, including sensor data, historical information, and expert opinion.  The resulting digital twin can be used to ana­lyze and diagnose operational states and to optimize performance under real-world operating conditions.  This enables companies to make predictions about future performance, improve product operation and productivity, and reduce the cost and risk of unplanned downtime.
Building on the success of other industries, the digital twin concept is now taking hold in the healthcare industry.  Unique to healthcare, digital twins for patients provide a platform for optimizing device performance during patient use.  Wearable medical devices are a current example of how patient data collected through the IoT combined with physiological models can optimize treatment for individual patients.  This track invites submissions that use digital twin concepts to improve the ability of medical devices and other therapies to treat patients more effectively and accurately.

Marc Horner (ANSYS, Inc)
Drew Pruett (University of Mississippi Medical Center)
Assessing Credibility of Models
Credibility is the trust in the predictive capability of a computational model for a specific context of use. Trust is gained by collecting evidence through verification, validation, and uncertainty quantification (VVUQ) activities. This track includes studies undertaken to demonstrate the credibility of computational modeling applied across the product lifecycle. Topics may include pre-clinical evaluation of medical devices, patient-specific modeling for clinical decision support, use of “software as a medical device,” post-market analyses, or any other area where VVUQ is leveraged to assess model credibility.
Payman Afshari (DePuy Synthes J&J)
Brent Craven (FDA)