Futuristic radiology suite with AI interfaces and FDA badge, representing the intersection of technology and regulation in medical imaging.

AI in Radiology: Navigating the New Regulatory Landscape

"How the FDA's innovative approach to AI regulation is reshaping the future of medical imaging and diagnostics."


Artificial intelligence (AI) is rapidly transforming healthcare, and radiology is at the forefront of this revolution. As AI-powered tools become increasingly prevalent in medical imaging, radiologists are keen on understanding the implications of these advancements. The AI health market is experiencing explosive growth, projected to surge from $600 million in 2014 to $6.6 billion by 2021. This growth underscores the urgent need for clear regulatory guidelines to ensure the safe and effective integration of AI into clinical practice.

Radiologists are particularly invested in AI's development because many significant breakthroughs have occurred in image-based technologies. However, the regulatory landscape is still evolving, and the FDA's approach to overseeing AI development is taking shape. Balancing innovation with patient safety remains a key challenge.

While the FDA has decades of experience regulating drug development, evaluating AI technologies requires a new framework. Recently, the FDA has started focusing on this challenge and is exploring innovative programs to handle AI's unique aspects. This article will delve into the FDA's current strategies for regulating AI in radiology, including the Digital Health Software Precertification Program (Pre-Cert), and discuss the potential impact on radiologists and patient care.

The FDA's Evolving Regulatory Approach

Futuristic radiology suite with AI interfaces and FDA badge, representing the intersection of technology and regulation in medical imaging.

Traditionally, the FDA approves medical devices, including AI-based ones, through pathways like the de novo premarket review. This pathway is designed for new devices that pose low to moderate risk and lack a legally marketed predicate device. Between February and May 2018, the FDA approved three AI-based devices for stroke, diabetic retinopathy, and wrist fracture using this route. This demonstrates the FDA's initial steps in addressing AI in healthcare.

However, recognizing the rapid pace of software development, the FDA introduced new programs to streamline the approval process for software, including AI. One notable initiative is the Digital Health Software Precertification Program (Pre-Cert), announced in April 2018 by then FDA commissioner Dr. Scott Gottlieb. Pre-Cert aims to expedite the regulation of software technology.

  • Faster Approval: Unlike traditional processes that focus on specific products, Pre-Cert provides a pathway for approving entire organizations.
  • Organizational Focus: Once an organization is Pre-Cert approved, it can bring lower-risk products to market without premarket review.
  • Flexibility: The FDA suggests that Pre-Cert may allow firms to make minor device changes without requiring new submissions each time.
The Pre-Cert program assesses organizations based on five core principles:

Challenges and the Path Forward

While the Pre-Cert program offers a promising approach, it also raises important questions. Vetting organizations instead of individual products is unprecedented, creating concerns about monitoring safety and efficacy. The FDA is actively working on defining different levels of certification and ensuring that organizations maintain their Pre-Cert status over time.

Establishing a robust framework for risk stratification is critical. The FDA plans to use the International Medical Device Regulators Forum framework, which categorizes risk based on the patient's situation (critical, serious, or nonserious) and how the device's information will be used (inform clinical management, drive clinical management, or treat and diagnose). However, real-world implementation may prove challenging due to the lack of established precedents.

Some healthcare experts worry that the Pre-Cert approach could give software an implicit stamp of approval without sufficient clinical validation. As the FDA continues to refine its regulatory strategies, radiologists must stay informed and actively participate in shaping these changes. The first official version of the Pre-Cert program (Pre-Cert 1.0) is expected in December 2018, necessitating ongoing collaboration with the FDA to ensure a safe, effective, and patient-centered approach to AI in radiology.

About this Article -

This article was crafted using a human-AI hybrid and collaborative approach. AI assisted our team with initial drafting, research insights, identifying key questions, and image generation. Our human editors guided topic selection, defined the angle, structured the content, ensured factual accuracy and relevance, refined the tone, and conducted thorough editing to deliver helpful, high-quality information.See our About page for more information.

This article is based on research published under:

DOI-LINK: 10.1016/j.jacr.2018.09.057, Alternate LINK

Title: The Fda And Artificial Intelligence In Radiology: Defining New Boundaries

Subject: Radiology, Nuclear Medicine and imaging

Journal: Journal of the American College of Radiology

Publisher: Elsevier BV

Authors: Samantha G. Harrington, Monica K. Johnson

Published: 2019-05-01

Everything You Need To Know

1

What is the Digital Health Software Precertification Program (Pre-Cert)?

The Digital Health Software Precertification Program (Pre-Cert) is a program introduced by the FDA to streamline the approval process for software, including AI, in healthcare. It shifts the focus from individual product approvals to evaluating the organization developing the software. If an organization is Pre-Cert approved, it can bring lower-risk products to market without premarket review, offering a faster route to market compared to traditional processes.

2

Why is the FDA changing its regulatory approach to AI in radiology?

The FDA is adapting to the rapid development of AI technologies in radiology, and the Pre-Cert program is a key part of this adaptation. The FDA's traditional pathways, like the de novo premarket review, are being supplemented by programs like Pre-Cert to keep pace with innovation. The Pre-Cert program's organizational focus allows for more agile regulation, which is critical for the fast-paced development cycles of AI software. It allows for faster approval and flexibility in making minor changes to the devices without requiring new submissions.

3

How does the Pre-Cert program work to regulate AI?

The Pre-Cert program is designed to address the challenges of regulating rapidly evolving software technology. Unlike the traditional approach of evaluating individual products, Pre-Cert assesses the organization itself. This approach aims to expedite the approval process, allowing organizations to bring lower-risk products to market more quickly. It emphasizes organizational culture and quality systems, and monitoring, and aims to allow organizations to adapt and make minor changes to software more easily, which is vital for the iterative nature of AI development.

4

What are the implications of the Pre-Cert program for radiologists and patient care?

The implications of the Pre-Cert program are significant for radiologists and patient care. By speeding up the approval process, the program enables faster access to potentially beneficial AI tools for diagnostics and image analysis. This could improve efficiency and accuracy in radiology, ultimately leading to better patient outcomes. The Pre-Cert program also introduces new considerations for radiologists, such as the importance of understanding the organizational quality and processes behind the AI tools they use. It is still evolving, and the FDA is working on defining different levels of certification to ensure that organizations maintain their Pre-Cert status over time.

5

How does the de novo premarket review pathway compare to the Pre-Cert program?

The FDA approved three AI-based devices for stroke, diabetic retinopathy, and wrist fracture using the de novo premarket review pathway between February and May 2018. The Pre-Cert program is a more recent development. Both strategies demonstrate the FDA's effort to address AI in healthcare. While the de novo pathway focuses on individual devices, the Pre-Cert program aims to streamline the process for software by focusing on the organization's overall quality and processes.

Newsletter Subscribe

Subscribe to get the latest articles and insights directly in your inbox.