- Campbell Arnold
- 21 hours ago
- 5 min read
“Not intended to acquire, process, or analyze a medical image.”
— FDA guidance on non-regulated software, Jan 6th 2026
Welcome to Radiology Access! your biweekly newsletter on the people, research, and technology transforming global imaging access.
In this issue, we cover:
FDA Loosens Oversight, But Not for Imaging AI
FDA Clears First Multi-Indication Foundation Model
QT Imaging relists on Nasdaq
If you want to stay up-to-date with the latest in Radiology and AI, then don't forget to subscribe!
🚀 We’re Hiring at Subtle Medical!
Interested in building AI that actually ships into clinical practice and impacts patient care at scale? Subtle Medical is hiring US-based Research Scientists and Machine Learning Software Engineers to work on cutting-edge algorithms used by hospitals around the world.
If you’re excited about deep learning, medical imaging, and translating research into real products, I’d love to hear from you. Email me directly with your resume attached.
FDA Loosens Oversight, But Not on Imaging
How new FDA guidance impacts the regulatory landscape for radiology AI.
Earlier this month, the FDA released updated guidance clarifying how it regulates (and in some cases does not regulate) clinical decision support software, general wellness, and low-risk devices. While the guidance reduces oversight for some digital health tools, it includes an explicit carve-out for medical imaging stating that software “intended to acquire, process, or analyze a medical image” are not eligible for regulatory exemptions. The guidance reinforces continued regulation of image-based AI while simultaneously opening a faster path to market for non-imaging radiology software, such as workflow and guidance tools.
Under the new guidance, clinical decision support software is not regulated as a medical device if it supports clinician decision-making without providing specific diagnostic or treatment recommendations. Despite the medical imaging carve out, there are still notable tools in the radiology space that this would include.
Examples that are not regulated medical devices:
An algorithm that summarizes radiology reports or cleared software findings.
Software that summarizes guidelines to support exam selection or imaging review.
A radiology workflow dashboard that tracks scheduling, turnaround, and image quality.
Examples that are still regulated medical devices:
An algorithm that analyzes chest X-rays and outputs “pneumonia” or “normal”.
Software that segments brain structures and reports volume measurements.
An imaging triage tool that analyzes images and flags high priority cases.
Wearables and software promoting general wellness are now no longer actively regulated as long as they do not make disease claims.
Examples that are not regulated medical devices:
A smartwatch that tracks heart rate, sleep, and activity trends.
A wearable that estimates blood pressure trends and frames results in wellness terms.
A mobile app that encourages exercise, stress reduction, or better sleep.
Examples that are still regulated medical devices:
A wearable marketed to diagnose hypertension or alert users with high blood pressure.
A device claiming to detect cardiac arrhythmias or screen for cardiovascular disease.
A sensor or app marketed to monitor disease progression.
For radiology AI developers, the message is clear: image-based AI remains regulated, regardless of how supportive or assistive the claims may sound. While the FDA’s carve-out draws a bright line around medical imaging, it also allows some radiology clinical decision support tools to innovate with reduced regulatory friction. More broadly, these updates reflect an FDA strategy focused on risk level, with narrowing oversight where patient harm is less likely, while maintaining firm control over technologies that directly influence diagnosis and treatment.
Bottom Line: Image-based AI remains regulated, but non-imaging radiology software now has a clearer path to market under FDA’s updated guidance.
FDA Clears First Multi-Indication Foundation Model
How Aidoc’s new AI triage solution raises the stakes for other players.
In a significant milestone for radiology AI, the FDA has cleared Aidoc’s comprehensive Abdomen CT triage solution, a foundation model covering a record-breaking 14 acute indications. The tool received Breakthrough Device designation in 2025, which Aidoc successfully leveraged to accelerate FDA review and achieve clearance. Unlike traditional narrow AI, which detects a single pathology, this solution simultaneously screens for multiple acute conditions in a single, integrated workflow. The clearance represents a shift away from a fragmented, algorithm-by-algorithm regulatory approach toward validating holistic AI models capable of handling the complexity of clinical cases with nonspecific complaints.
Aidoc claims that by flagging scans associated with any of the 14 indications, the AI prioritizes critical cases, preventing urgent exams from sitting behind routine studies in the radiology queue. In pivotal studies, the model demonstrated a mean sensitivity of 97% and mean specificity of 98%, with Aidoc noting that the foundation model approach reduces false alerts by roughly an order of magnitude compared to earlier-generation tools. This high specificity is essential for radiologist trust, as it helps ensure that the “stat” list is reserved for truly actionable findings.
For the broader radiology AI community, this clearance signals the start of a more comprehensive era of AI. It suggests the FDA is open to reviewing broader, multi-indication models rather than only narrow, single-use algorithms. While it remains to be seen whether clinical adoption will favor broad, platform-based solutions, most experts expect it will. With plans to extend this foundation model approach to chest imaging and eventually pixel-to-report auto-drafting, the clearance sets a new precedent for how AI vendors can scale their offerings. It shifts the conversation from “does this algorithm work?” to “how does this integrate into my workflow?”, creating pressure on other vendors to demonstrate similar breadth, accuracy, and clinical integration.
Bottom line: Aidoc’s FDA-cleared multi-indication foundation model marks a new era in radiology AI, demonstrating that broad tools can meet FDA regulatory standards.
QT Imaging re-lists on Nasdaq
Can financing, a reverse stock split, and renewed strategy pave the way to revitalization?
After a turbulent period on the public markets, QT Imaging has successfully re‑established its listing on the Nasdaq. The company’s stock had been delisted from Nasdaq in early 2025 after failing to meet minimum bid price and market value standards. To regain eligibility, QT Imaging’s leadership undertook a series of strategic actions throughout 2025, including strengthening its balance sheet, securing private financing, and pursuing new revenue growth.
Another key step was the company’s 3:1 reverse stock split in late 2025, which boosted its per‑share price to meet Nasdaq’s minimum listing requirements. With shareholder approval and the reverse split complete, QT Imaging submitted its application and received approval to relist on Nasdaq earlier this month, marking a significant milestone in its organizational turnaround. The relisting hopefully reflects both investor confidence and renewed momentum for the company’s technology and emerging platform strategy.
Bottom line: After a rough initial listing, QT Imaging has completed a turnaround and returns to the Nasdaq.
Feedback
We’re eager to hear your thoughts as we continue to refine and improve RadAccess. Is there an article you expected to see but didn’t? Have suggestions for making the newsletter even better? Let us know! Reach out via email, LinkedIn, or X—we’d love to hear from you.
References
Disclaimer: There are no paid sponsors of this content. The opinions expressed are solely those of the newsletter authors, and do not necessarily reflect those of referenced works or companies.
