Home     |     Our Team     |     News     |     Contact Us     |     Call: 508-231-8842

Growing numbers of medical devices are using artificial intelligence and its associated technology—machine learning—to diagnose patients more precisely and treat them more effectively. Although many devices have already been cleared by the FDA, many regulatory questions remain unanswered.

Artificial intelligence (AI) and machine learning (ML) technologies have the potential to transform healthcare by deriving new and important insights from the vast amount of data generated during the delivery of health care every day. All of this data is being captured, managed, analyzed, and used to help smart medical devices “learn” how to make decisions and perform diagnostic and therapeutic tasks that simulate human intelligence and behavior.

However, the FDA’s traditional paradigm of medical device regulation was not designed for adaptive artificial intelligence and machine learning technologies. There are currently no laws or standards that specifically regulate the use of artificial intelligence in medical devices; these devices, however, still must meet existing regulatory requirements, such as:

  • The manufacturers must demonstrate the benefit and performance of the medical device (for devices that are used for diagnostics purposes, the sensitivity and specificity, for example, must be demonstrated).
  • The devices must be validated against the intended purpose and stakeholder requirements and verified against the specifications.
  • Manufacturers must ensure that the software has been developed in a way that guarantees repeatability, reliability, and performance.
  • Manufacturers must describe the methods they will use for these verifications.
  • If the clinical evaluation is based on a comparator device, this device must be sufficiently technically equivalent, which explicitly includes evaluating the software algorithms.
  • Before development, manufacturers must determine and ensure the competence of the people involved.

In 2019, the FDA published a discussion paper “Proposed Regulatory Framework for Modifications to Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) – Discussion Paper and Request for Feedback” that describes the FDA’s foundation for a potential approach to premarket review for artificial intelligence and machine learning-driven software modifications. The document talks about the challenge of continuously learning systems; however, it observes that previously approved medical devices based on AI procedures worked with “locked algorithms.”

Regarding the two types of algorithm modification, the FDA tries to explain when:

  • It does not expect a new submission, only the documentation of the modification by the manufacturer.
  • It would like to perform a review of the modifications and validation before the manufacturer is allowed to market the modified product.
  • It will insist on a (completely) new submission or approval.

According to its own regulations, the FDA recognizes that a self-learning or continuously-learning algorithm that is in use would need to be inspected and approved again. But that seems too strict even for FDA regulatory compliance parameters. Therefore, it looks at the objectives of a modification to the algorithm and distinguishes between:

  • Improvements to clinical and analytical performance: These improvements may include training with additional data sets.
  • Modification of the “input data” used by the algorithm: This can be additional laboratory data or data from another CT manufacturer.
  • Change of intended use: The FDA gives the example of an algorithm that initially only calculated a “confidence score” intended to aid a diagnosis, but which now provides a definitive diagnosis. A change to the intended patient population would also be considered a change to the intended use.

The FDA wants to use these objectives to decide on the need for new submissions.

Until recently, the regulatory definition of a medical device was relatively narrow; but now AI-based solutions with a medical purpose are being recognized as medical devices. The European Union issued the Regulation EU 2017/745 on Medical Devices (Medical Devices Regulation) describing that software programs created with clear intention to be used for medical purposes are considered medical devices. This broadening of the definition of a medical device affects products that are explicitly intended to prevent or monitor disease without having a diagnostic or therapeutic purpose. Therefore, AI-based health technologies that help to diagnose, predict, monitor, and prevent a disease can now be considered as medical devices.

If your company is making, or considering making, a smart medical device with AI or ML components and you’d like to learn more about medical device compliance and regulatory impact, contact the medical device consulting specialists here at MEDIcept today.

 

MEDIcept … Trusted Solutions, Rapid Response …

 

About Us

MEDIcept is an international consulting firm offering a full portfolio of services to the medical device and IVD industry. For over 25 years, our unique consulting practice and multidisciplinary team of former FDA, Notified Body, and industry experts have assisted hundreds of companies of all sizes with innovative, compliant, trusted, and cost-effective Regulatory, Quality, Clinical, and Engineering solutions.

For additional information, please contact Susan Reilly at SReilly@MEDIcept.com.