NEW

GE HealthCare Aims to Advance Women’s Health Imaging

GE HealthCare recently launched the Voluson Signature 20 and 18 ultrasound systems, which comprehensively integrate artificial intelligence (AI), advanced tools and an ergonomic design to speed exam time for clinicians while delivering a clearer picture of various...

FDA Releases Two Companion Reports on Medical Device Safety and Innovation

Today, the U.S. Food and Drug Administration’s (FDA) Center for Devices and Radiological Health Center (CDRH) is releasing two reports on medical device safety and innovation – the core pillars that help protect and promote public health for all. The “CDRH 2024 Safety...

Tri-Imaging, RTI Group Work Together

In a LinkedIn post, Tri-Imaging states, “We are honored to be the first ISO in North America to have the opportunity to try RTI’s new Mako X-Ray Testing Meter. The Mako meter is the most accurate and efficient testing meter that covers the broadest application range...

Detection Technology announces global availability of TFT flat panel detectors

Detection Technology, a global leader in X-ray detector solutions, announces the global availability of a comprehensive range of TFT (thin-film transistor) flat panel detectors. The portfolio includes IGZO (indium gallium zinc oxide) and a-Si (amorphous silicon)...

FDA Sends AI Warning Letter

PACS/ITI am of the option that artificial intelligence (AI) in imaging can improve accuracy, productivity and safety if used correctly. 

The use of radiation for treatment and diagnosis was voted by the American Medical Association as one of the top 10 most useful innovations in health care.

As radiology technologists, we are charged with the safe use of this potentially dangerous tool. First do no harm. The industry has set up stringent limits and complex monitoring systems to insure safe best practices in radiation use. 

A vendor of a radiation-generating device cannot insure the safe administration of the proper dose. This is why, as an open advocate of medical imaging AI, I feel obligated to share with you this letter from the U.S. Food and Drug Administration.

“The U.S. Food and Drug Administration (FDA) is reminding health care providers about the intended use of radiological computer-aided triage and notification (CADt) devices for intracranial large vessel occlusion (LVO). LVO is an obstruction of one of the large arteries in the brain and is a common cause of acute ischemic strokes. LVO CADt devices are software devices intended to aid in prioritization and triage of time-sensitive suspected findings of LVO based on the analysis of radiological exams of the brain. LVO CADt devices do not provide diagnostic information or remove any cases from the imaging physician’s reading queue. Information from real-world use suggests that providers may not be aware of the intended use of these devices. If LVO CADt devices are not used as intended, there is the potential for misdiagnosis resulting in patient injury or death. The FDA is bringing this information to your attention as a reminder and to ensure that providers are aware of the intended use of the software as a triage and prioritization tool.”

Recommendations

The FDA recommends that health care providers:  

  • Be aware that LVO CADt devices only flag radiological exams with suspected findings and should never be used as a replacement for informed interpretation by an imaging physician.
  • Recognize that LVO CADt devices cannot rule out the presence of an LVO. If a radiological exam is not flagged by a LVO CADt device, an LVO may still be present.
  • Recognize that when the device is used as intended (as a prioritization and triage tool and not a diagnostic device), it can improve workflow by prioritizing suspected cases. 
  • Recognize that the device does not remove any radiological exams from the queue for interpretation by an imaging physician. When used as intended, exams that are not flagged by the device are still interpreted by an imaging physician according to the standard of care.
  • Be aware of the design of LVO CADt devices. This includes an understanding of the vessels (arteries) for which the device was designed and tested to detect LVO. LVO CADt devices may not be designed and tested to evaluate all intracranial vessels. If you have questions about which vessels the LVO CADt device is designed to evaluate, or are unsure, review the most recent product labeling or contact the device manufacturer for further information or questions on device performance.

Background

Acute ischemic stroke caused by LVO contributes to stroke-related disability and death. Timely intervention may improve health outcomes. Many CADt devices are based on artificial intelligence and machine learning (AI/ML) technology. LVO CADt devices are classified under 21 CFR 892.2080 Radiological computer-assisted triage and notification software (eCFR :: 21 CFR 892.2080 — Radiological computer aided triage and notification software). 

Health care providers may not be fully aware that LVO CADt devices are intended for prioritization and triage only. These devices should not be relied on when making any diagnostic decisions. In addition, health care providers should ensure that they understand which vessels the device was designed and tested to evaluate. If the device is used for diagnostic decisions or to evaluate vessels for which it was not designed, delayed or missed diagnoses may occur and result in worsened outcomes for patients. When used as intended, CADt devices can improve the workflow by prioritizing suspected exams. The device does not remove any radiological exams from the queue for interpretation by an imaging physician. When used as intended, exams that are not flagged by the device are still interpreted by an imaging physician according to the standard of care. 

FDA Actions

The FDA is working with LVO CADt device manufacturers to ensure that health care providers are informed about the intended use and design of these devices. More generally, because postmarket performance in clinical practice may not match the premarket data, the FDA is also engaged in continuing efforts to evaluate the real-world performance of AI/ML imaging devices, including CADt devices. The FDA will continue to keep health care providers and the public informed if new or additional information becomes available. 

Reporting Problems to the FDA

The FDA encourages health care providers to report any adverse events or suspected adverse events experienced with imaging software for LVO.

  • Voluntary reports can be submitted through MedWatch, the FDA Safety Information and Adverse Event Reporting program.
  • Device manufacturers and user facilities must comply with the applicable Medical Device Reporting (MDR) regulations.
  • Health care personnel employed by facilities that are subject to the FDA’s user facility reporting requirements should follow the reporting procedures established by their facilities.

Prompt reporting of adverse events can help the FDA identify and better understand the risks associated with medical devices.”

The vendors will tell you their product “can” do this, “over 100 finding.”

It is our obligation to the patient and our health care organizations to understand and apply AI in imaging in a safe and compliant manner.

Mark Watts is an experienced imaging professional who founded an AI company called Zenlike.ai.

Previous

Next

Submit a Comment

Your email address will not be published. Required fields are marked *

Open