In the second part of a two-part interview, Nina Kottler, M.D., says the transparency emphasis of the recent FDA guidance on AI-enabled software is welcome but needs to go beyond additional documentation to clarify how adjunctive AI is making its decisions.
Increased transparency is a cornerstone of the recent draft guidance from the Food and Drug Administration (FDA) on submissions for AI-enabled software. While this will result in an increased amount of documentation on the training as well as limitations for adjunctive use of AI models, Nina Kottler, M.D., maintained in a recent interview that this well-intended effort at transparency may fall short of having an impact in daily practice.
“ … Any of this documentation that the FDA is requiring right now is not so helpful at the point of care. It is good that they're doing it. It's not a bad thing. It's a very good thing, but we need a way now to translate that information to be helpful at the point of care,” posited Dr. Kottler, the associate chief medical officer of clinical AI at Radiology Partners.
Whether it comes in the form of heat mapping or a color-coded assessment by the adjunctive AI, Dr. Kottler says the capability of the AI tool to express a degree of confidence in the evaluation can be beneficial for end-user radiologists.
Additionally, Dr. Kottler maintained that there should be clear transparency on how the adjunctive AI is making its assessment. She cautioned that most people don’t recognize that adjunctive AI doesn’t consider every single piece of input data in its interpretation.
“ … Another thing that the FDA should be requiring is to tell the end user, what are the images, or what are the components of the language, or whatever that the AI is using so radiologists know when they're looking at a study, ‘Oh, this series wasn't even evaluated by the AI.’ That is really important information,” emphasized Dr. Kottler.
(Editor’s note: For related content, see “How Will the New FDA Guidance Affect AI Software in Radiology?: An Interview with Nina Kottler, MD, Part 1,” “Mammography Study Shows Merits of AI for Improving Breast Cancer Detection and Effectiveness of Recalls” and “Study Assesses Lung CT-Based AI Models for Predicting Interstitial Lung Abnormality.”)
For more insights from Dr. Kottler, watch the video below.
Stay at the forefront of radiology with the Diagnostic Imaging newsletter, delivering the latest news, clinical insights, and imaging advancements for today’s radiologists.
The Reading Room Podcast: A Closer Look at Remote MRI Safety, Part 3
August 1st 2025In the third of a three-part podcast episode, Emanuel Kanal, M.D. and Tobias Gilk, MRSO, MRSE, discuss strategies for maintaining the integrity of time-out procedures and communication with remote MRI scanning.
The Reading Room Podcast: A Closer Look at Remote MRI Safety, Part 2
August 1st 2025In the second of a multi-part podcast episode, Emanuel Kanal, M.D. and Tobias Gilk, MRSO, MRSE, share their perspectives on remote MRI safety protocols for ensuring screening accuracy and adherence to conditional implant guidelines as well as a rapid and effective response to adverse events.
FDA Clears Point-Of-Care Ultrasound Platform and AI Software for Neuraxial Procedures
August 1st 2025The dual FDA clearances for the Accuro 3S point-of-care ultrasound device and the SpineNav-AI machine learning-based software may enhance precision and safety with ultrasound-guided neuraxial procedures.
Stroke MRI Study Assesses Impact of Motion Artifacts Upon AI and Radiologist Lesion Detection
August 1st 2025Noting a 7.4 percent incidence of motion artifacts on brain MRI scans for suspected stroke patients, the authors of a new study found that motion artifacts can reduce radiologist and AI accuracy for detecting hemorrhagic lesions.