Contact us
Back to Insights

Regulatory

Understanding FDA software guidance for AI medical devices

Share this article

Aug 10, 2023

8 min read

samd series | author

Weronika Michaluk

Digital Health Principal
SaMD Lead at HTD Health

The US Food and Drug Administration (FDA) issued draft guidance in early April 2023 on “predetermined change control plans” for AI medical devices, allowing medtech companies to more easily update their devices’ machine learning components within a predetermined scope reviewed by regulators. Previously, medical algorithms were “locked” and unable to change after FDA clearance or approval. The new FDA software guidance will allow for modifications to be made without changing the efficacy or safety of the medical device. According to the agency, the goal is to “protect and promote public health.”

The agency has outlined elements device-makers should submit as part of these plans, including a description of the planned modifications, how the company plans to validate and implement the changes, and an impact assessment. Medical device companies had until July 3 2023 to comment on the draft.

Free consultation

Still have questions? Reach out to our SaMD expert and chat

We will develop an executive plan for your medical device software

Contact Us

Regulations & AI medical device

AI medical devices, including machine learning algorithms, have become increasingly popular in healthcare. The FDA has granted authorization to over 500 AI/ML-driven medical devices, with numerous others in the pipeline. The FDA’s core public health mission revolves around guaranteeing the safety and efficacy of these advanced devices, ensuring they can maximize their benefit to individuals.

AI-driven technologies aid physicians in identifying detrimental alterations like tumors and arteriosclerosis while delivering precise assessments of organ dimensions and blood circulation. In numerous medical imaging tasks, AI algorithms have outperformed human capabilities. Forward-thinking companies have crafted AI tools to interpret radiology and digital pathology visuals. AI-infused devices have emerged as essential tools for decision-making in the medical field.

As the number of AI medical devices increases, the FDA has been considering how to best regulate these software components. The proposed predetermined change control plan is an important step to facilitate the rapid and regular improvement of AI/ML-enabled device performance across diverse populations.

The FDA software guidance will encourage innovation and the timely delivery of new medical technologies, but developers will need to prepare for new challenges. Beyond the highly involved planning process ahead of alterations, when the changes are actually made, the new medical device guidance necessitates a tremendous amount of documentation. The submissions won’t be filed with the FDA, but organizations should have the paperwork on hand in the event that the FDA comes for an inspection.

What are the examples of AI in medical devices?

The FDA software guidance for medical devices is a significant development that will enhance the ongoing development, validation, implementation, and monitoring of AI/ML-enabled devices. The capabilities for machine learning in medical devices are increasingly being used to automate clinical decisions, reducing human error and improving patient outcomes. Additionally, with AI, medical practitioners can use predictive models to anticipate patient outcomes and tailor treatments.

FDA software guidance

AI-enabled Software as a Medical Device (SaMD) has shown tremendous promise in improving patient care, diagnosis, and treatment. Here are some notable examples of such software solutions:

Radiology imaging

  • An AI-based platform designed to scan and interpret various medical images, capable of detecting anomalies in regions like the brain, c-spine, chest, and abdomen.
  • A machine learning-driven diagnostic tool that can process a wide array of medical images and subsequently identify numerous medical conditions based on those images.
  • Intelligent imaging software that automates the recognition of disease markers in radiological scans, enhancing the speed and accuracy of diagnosis.

Cardiology

  • An AI-powered cardiac diagnostic tool that takes ECG readings and, through advanced algorithms, identifies patterns indicative of atrial fibrillation.
  • A digital health solution tailored for cardiology specialists, offering real-time AI analysis of ECG data to pinpoint irregular heart rhythms.
  • A software application that utilizes machine learning models to instantly assess and categorize electrocardiograms, particularly focusing on signs of atrial fibrillation.

Oncology

  • An AI-driven clinical platform that dives deep into both clinical and molecular patient data, furnishing doctors with insights to optimize cancer treatment plans.
  • A digital oncology companion that leverages machine learning to shift through extensive patient data, enabling more individualized and effective treatment strategies.
  • A comprehensive AI solution tailored for oncologists, which facilitates a deeper understanding of tumor profiles by analyzing a vast array of patient data, from genetic markers to clinical histories.

Ophthalmology

  • A specialized AI system devised for ophthalmology that scrutinizes retinal images to uncover early indications of diabetic retinopathy.
  • An intelligent retinal imaging tool that, with the help of AI, can rapidly and accurately flag potential cases of diabetes-induced eye complications.
  • A diagnostic platform that integrates advanced AI capabilities to evaluate retinal photographs and spot the subtle changes typical of diabetic retinopathy.

Neurology

  • A cutting-edge software application that utilizes AI to analyze CT scans of the brain, sending instant alerts to specialists upon detecting potential stroke indicators.
  • An AI-enabled neurology tool that streamlines the process of stroke care, scanning CT images for signs of large vessel occlusion and rapidly informing relevant medical experts.
  • A machine learning-driven platform designed for swift and precise analysis of CT scans, targeting early detection of severe neurological events, such as potential strokes.

 

It’s important to note that any AI-based software used in a medical context needs rigorous validation and regulation. Regulatory bodies like the FDA in the US provide software guidance and oversight to ensure the safety and effectiveness of these technologies.

Get the latest news about SaMD from our experts.

Newsletter terms

Building trust with AI in the medical field

One of the biggest challenges facing widespread acceptance of healthcare AI is consumer trust. As artificial intelligence and its potential biases can impact clinical decisions, there is a need to build more trust in machine learning models for healthcare.

The different types of machine learning – supervised, unsupervised, and reinforcement learning – each have their own strengths and weaknesses. Medical device makers and health IT leaders need to develop robust quality management systems to monitor and document an algorithm’s purpose, data quality, development process, and performance. Continuous monitoring is necessary to ensure fair and bias-free performance, as regulators have also recognized.

Looking to the future

In conclusion, the FDA’s draft guidance on “predetermined change control plans” is an important development for the medical device industry. The proposed framework provides medical device manufacturers with a process to develop, validate, implement, and monitor modifications to their AI / Machine learning devices while maintaining the device’s efficacy and safety.

The new framework will enable manufacturers to quickly make modifications that are important for patient care without the need for a lengthy regulatory process. Ultimately, the guidance for FDA software will encourage innovation in the development of new medical technologies while ensuring patient safety and improved healthcare outcomes.

SaMD industry is growing rapidly. If you were ever thinking about entering the SaMD space, now is the time for you to explore opportunities in the medical device industry.

At HTD, we’re more than just a service provider; we’re your dedicated partner. Our expertise in Software as a Medical Device ensures your product aligns with the latest guidelines, meets regulatory standards, and delivers top-tier value swiftly — all tailored to help you achieve your unique business aspirations. Interested in bringing your SaMD vision to life? Reach out to us at info@htdhealth.com and let’s start with a FREE consultation. Your journey into the future of healthcare starts here!

Other content you may be interested in

View all articles
New FDA guidance

May 31, 2023

5 min read

Regulatory

FDA cybersecurity guidance: New law all SaMD manufacturers must follow

Read more
documentation for compliant software

Sep 17, 2020

6 min read

Regulatory

How does FDA documentation fit into agile development?

Read more

Jul 05, 2024

SaMD

Sprint to success: Enhancing productivity with Agile methodologies in Med Tech

Read more