Over the past decade, the number of medical devices equipped with artificial intelligence technology has shown significant growth.

According to the U.S. Food and Drug Administration (FDA) database, as of August 7, 2024, the agency has authorized 950 AI or machine learning (ML) medical devices. Although the FDA approved the first AI device in 1995, submissions have surged in recent years. Data shows that in 2015, the FDA authorized only 6 AI medical devices, while in 2023, this number reached 221 (data reviewed by MedTech Dive).

Experts said in interviews that this trend is driven by increased device connectivity, greater investment in AI and machine learning, and the industry's growing familiarity with regulatory processes for software as a medical device.

"We are definitely seeing a significant increase in investment, there is no doubt about that," said Jennifer Goldsack, CEO of the Digital Medicine Society, a digital health industry organization.

Examples of AI medical device applications

  • AI-Rad Companion:Developed by Siemens Healthineers, this feature is designed to provide quantitative and qualitative measurements of clinical images, as well as clinical data analysis.
  • LumineticsCore:Software developed by Digital Diagnostics that automatically detects diabetic retinopathy by analyzing images. Unlike most AI medical devices, it can make a diagnosis without the need for a specialist.
  • Atrial fibrillation history feature:Apple received FDA clearance in 2022 for this feature, which uses Apple Watch data to show users how often they have shown signs of a common heart arrhythmia over the past week.

Large medical technology companies, including GE Healthcare, Siemens Healthineers, and Medtronic, are integrating AI into devices and developing standalone software tools. Startups such as Aidoc, RapidAI, and Butterfly Network are offering targeted solutions for flagging health conditions and improving ultrasound imaging.

Companies outside the medical device field are also actively involved. Apple has developed features that use watch data to detect heart rhythm abnormalities. Chipmaker Nvidia has partnered with medical device companies such as Medtronic and Johnson & Johnson to help expand their AI applications.

Number of AI device authorizations grows exponentially

AI medical devices can be used to improve image quality, shorten scan times, and assist clinicians in diagnosis or surgical preparation.

MedTech Dive analyzed the FDA's list in late September. The data shows that over the past decade, the number of AI devices cleared through regulatory approval has risen sharply. The data also indicates that the field is dominated by imaging, but companies are gradually expanding into other specialties.

These charts will be updated as new data is released.

Between 2015 and 2023, the number of AI/ML device submissions received by the FDA grew almost exponentially.

So far in 2024, the FDA has authorized 107 devices, on track to match 2023 levels. David Niewolny, director of business development for healthcare at Nvidia, said he has been working with connected medical devices since 2007. At that time, connected devices were still a "vision," and people were still figuring out how to extract data from devices and where to store it. He said, "Now, virtually all devices have some form of connectivity. Everyone feels the pressure to innovate faster." He added that this is "the first time in my career that I have seen all the technology elements come together."

Imaging dominates, cardiovascular follows

More than three-quarters of AI devices authorized to date are in the field of radiology. These devices include features that improve image quality, assist with patient positioning during scans, optimize radiation dose, and flag potential health conditions.

In the FDA list, 55 devices are described as radiology computer-aided triage and notification software, and 24 are radiation therapy planning systems.

Companies are developing AI devices in more other areas. For example, cardiovascular is the second most common specialty, with 98 devices on the list. Examples of AI cardiac devices include electronic stethoscopes and software that uses electrocardiogram data to detect arrhythmias or signs of heart failure.

"These specialties are confident in these tools and have been using them for some time," Goldsack said. "The interesting thing is when we start to see more products entering specialties that may not have 15 years of experience... When will a new wave emerge in new therapeutic areas?"

Nvidia's Niewolny sees robotics as an emerging area for AI applications. Before surgery, software can extract patient history and relevant data; during surgery, augmented reality applications can help surgeons visualize imaging data on the patient; after surgery, AI can read video clips to provide analysis or generate reports. He said, "What started in radiology is migrating to other departments of the hospital."

GE Healthcare and Siemens Healthineers lead the AI device market

Since MedTech Dive first analyzed FDA data in 2022, GE Healthcare and Siemens Healthineers have consistently topped the list in terms of number of AI devices.

As of August 7, 2024, GE Healthcare had 81 authorized AI devices. One of its flagship AI products, Air Recon DL, was launched in 2020. Jan Beger, head of AI advocacy at GE Healthcare, said the algorithm improves image quality and can reduce MRI scan times by up to 50%. As of October, the company had used the software to scan more than 34 million patients.

Beger divides GE Healthcare's AI strategy into three groups: products for improving imaging efficiency (such as Air Recon DL), AI that integrates multi-source data to support clinical decisions, and enterprise-level systems for planning. These features are sometimes built into the imaging devices sold, and sometimes sold as standalone subscriptions. "Overall, the idea of AI is: how do we automate those redundant, repetitive, monotonous tasks?" Beger said.

Two shoulder X-ray images shown side by side. The left image is a shoulder scan, and the right image is the same image sharpened using an AI feature. Air Recon DL is one of GE Healthcare's flagship AI products, designed to improve image quality and shorten MRI scan times.

Image source: GE Healthcare

GE Healthcare has also recently acquired several AI product companies, including Caption Health, which develops ultrasound imaging guidance software, MIM Software, which develops cancer treatment and radiation dose software, and BK Medical, which manufactures ultrasound devices for surgical guidance.

Beger said he is not focused on point solutions like triage for specific conditions, but rather on "foundation models" that can be adapted for specific uses.

Siemens Healthineers has 70 AI devices on the FDA list. Peter Shen, head of digital and automation for Siemens Healthineers in North America, said its AI products include features built into imaging machines and diagnostic algorithms. For example, the company has developed features that help patients position themselves on MRI scanners for optimal images, as well as algorithms that help clinicians diagnose conditions such as coronary artery calcification.

Shen noted that radiation therapy planning is a key focus area for Siemens Healthineers. The goal is to precisely target tumors while avoiding radiation to surrounding healthy tissue. AI can help clinicians find the best contours to target radiation, enabling faster treatment planning. Looking ahead, Shen is most excited about multimodal AI, which integrates different data such as imaging, lab results, and patient history, and can support decisions such as whether to biopsy a tumor or how much radiation to give a specific tumor. "It won't replace any decisions, but rather respects the doctor-patient relationship by providing more information or data so that doctors can make appropriate decisions for patients," Shen said.

Regulatory pathway: 510(k) dominates overwhelmingly

The FDA has cleared the vast majority of AI devices through the 510(k) pathway, which is more lenient, faster, and less costly than other market authorization options. As of August 2024, approximately 97% of AI devices on the list were 510(k) clearances.

The 510(k) pathway is for moderate-risk devices, where applicants must demonstrate that their device is "substantially equivalent" to an already authorized "predicate" device.

Another 22 devices were cleared through the de novo classification (for low-to-moderate risk devices with no predicate). Only 4 AI devices have received premarket approval (PMA), the most stringent pathway for high-risk devices.

The FDA says all devices on the list must be validated and the diversity of study populations must be assessed based on the device's intended use and technical characteristics. Nevertheless, patient advocates continue to call for stronger regulation.

Many AI tools are also exempt from FDA regulations. The Pew Charitable Trusts notes that the FDA does not regulate software intended to assist with administrative tasks such as scheduling, inventory management, and financial processing.

Software intended to assist clinical decision-making has been in a gray area. In 2022, the FDA issued guidance clarifying that AI intended to provide specific recommendations on diagnosis or treatment (for example, using patient information to flag potential sepsis cases) should be considered a medical device, while software that matches patient data with current treatment guidelines for common diseases is exempt.

Methodology

MedTech Dive downloaded the FDA's list of AI/ML medical devices on September 6, 2024. The FDA last updated the database on August 7, 2024. The FDA compiles the list using product codes and device summaries. The list is not exhaustive but is intended to represent devices integrating AI/ML across medical disciplines, the agency said. Device definitions include both hardware and software functions.

MedTech Dive collected information on each applicant and analyzed whether companies were acquired. If a company was acquired multiple times, the most recent parent company or majority shareholder is listed. One exception is Siemens Healthineers, which is majority-owned by Siemens. Parent companies are defined as the companies that manufacture the products; private equity and investment firms are not listed. MedTech Dive used this parent company information to count the companies with the most AI devices. Submission types were extracted from device submission numbers, which contain prefixes such as "P," "K," and "DEN," representing "premarket approval," "510(k) clearance," and "de novo classification," respectively. MedTech Dive also collected classification information from each file to describe the devices. In a few cases, the classification field was empty because the file did not include that information. We made edits to reflect MedTech Dive's editorial style.