Computer-based modeling and simulation are increasingly important during the development and authorization of medical devices and for a lot of manufacturers’ market success.
A useful definition of the terms “computer-based modeling” or “computational modeling” comes from the FDA itself:
Computational modeling is the process of representing a real-world system by means of a computer and then running the simulation by implementing a numerical scheme.
Frontiers in Medicine, September 2018, Volume 5, Article 241 by Tina Morrison (FDA) et. al.
One of the possible applications of such modeling and simulation is “in silico medicine.” This can be defined as follows:
It is the direct use of computer simulation in the diagnosis, treatment, or prevention of a disease. More specifically, in silico medicine is characterized by modeling, simulation, and visualization of biological and medical processes in computers with the goal of simulating real biological processes in a virtual environment.
The use of computer models and simulation is widespread in other industries:
Computer-based modeling and simulation can be used by medical device manufacturers in all phases of a device's life cycle.
Laws such as the MDR and the IVDR require manufacturers to comply with the general safety and performance requirements. These include requirements for
Manufactures must justify why they have chosen a particular solution, e.g. a specific design:
“The documentation shall […] include a justification […] of the solutions adopted to meet those [general safety and performance] requirements.
MDR, Annex II, Chapter 4.
This justification will be particularly successful if manufacturers can demonstrate that other solutions are inferior to the chosen solution.
A very efficient way of doing this is to simulate alternative solutions on a computer and select the best option. These options involve the:
However, it’s not just the device itself that is modeled, the people using the device, including their anatomy and physiology, are also modeled.
Some manufacturers use computer-based modeling to simulate production processes, e.g.:
The evidence that a medical device is safe and effective and provides the promised clinical benefits can be generated in part or in full using computer models.
In some cases, the FDA has even made a decision on the authorization of medical devices based exclusively on such models and simulations. This includes a device whose “MRI compatibility” had to be proven.
In addition, the FDA has authorized a device for “3D mammography” whose diagnostic performance was evaluated using synthetic(!) images. For this, the manufacturer simulated both the breast (e.g. size, fat content, geometry, mammary ducts, vessels, lesions) and the medical device itself.
As a result, a large part of the clinical investigation was performed by simulation with in-silico clinical trials.
Simulation can also help manufacturers in the post-market phase:
There is hardly a phase in the device life cycle in which modeling and simulation cannot be used. The FDA agrees.
The advantages of computer-based modeling and simulation are obvious:
Computer-based modeling offers the following advantages:
But these methods also have their price. More on this below.
Even up-to-date laws, such as the MDR and IVDR, do not contain specific requirements for the use of computer models for the development of medical devices. However, they explicitly provide for their use and simulation.
where appropriate, the results of biophysical or modelling research the validity of which has been demonstrated beforehand; (Annex I, 10.1a))
results of tests, such as engineering, laboratory, simulated use and animal tests [...]; (Annex II, 6.1.a))
the pre-clinical testing, for example laboratory testing, simulated use testing, computer modelling, the use of animal models (Annex VII, 4.5.4.a))
ISO 13485 requires manufacturers to:
The FDA was mandated by the US Congress to promote and regulate the use of computer-based modeling and simulation. It had already published the document “Advancing Regulatory Science at FDA” in August 2011. In it, the authority chose to focus on the “use and develop computational methods and in silico modeling.”
In 2018, Tina Morrison, who is driving the issue forward at the FDA, explained in an article in “Frontiers in Medicine” the possibilities offered by “computational modeling for medical devices” and the required “regulatory science.”
Even before that, she had played a key role in the development of the guidance document “Reporting of Computational Modeling Studies in Medical Devices Submission”, a final version of which has been available since 2016.
As the name suggests, this document describes what documentation the FDA expects. But it does not contain very specific guidelines for the validation of the models.
For this, the FDA relies on the “V&V 40” published by the American Society for Mechanical Engineering (ASME) with the title “Verification and Validation of Computational Modeling of Medical Devices”.
The document describes the steps manufacturers should take when validating computer models.
These steps include:
The advantages of computer models for the development and authorization of medical devices are clear. However, in practice manufacturers are faced with numerous challenges:
As with any venture into the unknown, we recommend the following best practices:
In the long run, it may no longer be a question for most manufacturers of whether they use computer-based modeling and simulation for the development, validation and authorization of their devices or not.
Without these methods they will not be able to keep up with the competition. And the competition may not necessarily be their current competitors. Companies and universities with their start-ups who have modeling and simulation tools and expertise will enter the market.
But it is not just competitive pressure that should prompt manufacturers to set out down this path. Exposing patients and animals to unnecessary tests and risks is simply unethical.
The FDA also sees opportunities here:
The all–in silico approach for conducting imaging trials is not intended to replace but rather complement or minimize traditional clinical trials. Incrementally incorporating computational results in regulatory submissions can, for example, help decrease the human trial size and length.
The extent to which Europe is lagging behind many US initiatives when it comes to the regulatory science regarding computer-based modeling and simulation is worrying. More engagement from authorities and notified bodies with this issue is desirable.
Manufacturers can get involved with the Avicenna Alliance in order to actively help shape the regulatory environment.
The Johner Institute can provide even more intensive support for achieving regulatory clarity and making modeling and simulation as fast and easy to use as possible.