- Campbell Arnold
- Mar 4
- 4 min read
Updated: Mar 27
Welcome to RadAccess: Impressions—your quick-read companion to the main RadAccess newsletter. Like a radiology report's Impressions section, here we deliver the essential information concisely, respecting your time. For more details, you can always turn to the full RadAccess newsletter.
Have No Fear, SubtleHD is Here!
Subtle Medical has received FDA clearance for SubtleHD, marking the company's fourth clearance in the past year and completing its ELITE package—a suite of three AI algorithms designed to accelerate MRI scans, enhance image quality, and streamline workflows:
SubtleHD performs denoising and sharpening to improve accelerated scans by up to 80%.
SubtleSYNTH generates synthetic spine STIR images from T1 and T2 sequences for additional time savings.
SubtleALIGN automatically aligns 3D brain images for improved consistency in longitudinal studies.
These vendor-neutral solutions work across different MRI systems to improve scanner efficiency, increase patient throughput, reduce patient backlogs, and extend scanner lifespans.
Foundation Models are Knocking at the FDA’s Door
Aidoc has secured FDA clearance for the first foundation model-powered AI device, a major regulatory milestone that signals the FDA’s evolving approach to AI oversight. Unlike traditional AI models designed for single-use cases, foundation models offer broader applicability, enabling faster development and improved generalization. However, rather than granting broad approvals, the FDA is clearing foundation models for specific use cases, with Aidoc’s CARE1 model approved for CT rib fracture triage. This milestone underscores the growing role of foundation models in healthcare AI while providing insight into the FDA’s regulatory stance. With ongoing discussions at the FDA on generative AI and continuous learning paradigms, the regulatory landscape will continue to shift. Meanwhile, Aidoc is already developing CARE2, its next-generation foundation model, aiming to expand into new clinical applications.
But is Anyone Home at the FDA?
In mid-February, FDA regulators received layoff notices followed by emails reinstating them days later. The staffing turmoil has raised concerns about the stability of AI regulation in healthcare. The Center for Devices and Radiological Health (CDRH), which is responsible for overseeing AI-enabled imaging devices, was heavily impacted by the initial layoffs. The FDA has cleared over 1,000 AI-based medical devices with the number of submissions increasing every year. The agency has been working to adapt its regulatory frameworks to accommodate the rapidly changing AI landscape including generative AI, continuously learning models, and foundation models. However, the combination of staffing instability and evolving policy priorities creates uncertainty for both regulators and developers. The key question now is whether these disruptions will slow AI approvals or hinder the FDA’s efforts to adapt its oversight for the next generation of medical AI.
ASSR 2025: Key Takeaways
The ASSR 32nd Annual Conference showcased groundbreaking advancements in spine imaging, with a strong focus on AI and deep learning. Here are some key takeaways:
Amish Doshi demonstrated how deep learning can improve low-field kinematic spine imaging, revealing pathology missed by standard scans.
Adam Flanders emphasized that AI can assist radiologists in managing rising imaging volumes, but that physicians must remain in control.
My talks showcased 1) SubtleSYNTH and the recently cleared SubtleHD for accelerate spine protocol, and 2) the feasibility of low-dose Gad enhancement.
Tim Amrhein highlighted PCCT’s improved resolution and lower radiation dose, but also noted workflow bottlenecks due to large data sizes.
Larry Tanenbaum explored MRI acceleration tools from various vendors, including Subtle Medical’s SubtleHD, SubtleSYNTH, and AiMIFY.
Johan Van Goethem presented synthetic CT generation from MRI using BoneMRI, helping minimize radiation in pediatric and prenatal cases.
Wende Gibbs urged radiologists to rethink workflows, asking how prioritizing patient outcomes over scan volume could reshape clinical practice.
The conference underscored the growing role of AI in enhancing image quality, improving radiologists workflows, increasing imaging efficiency, and providing better patient care.
TotalSegmentator MRI: Multi-anatomy, Multi-sequence, Multi-modal!
TotalSegmentator continues to push the boundaries of multi-anatomy, multi-sequence segmentation, now extending its capabilities to both MRI and CT. The latest Radiology study expands the model to 80 anatomical structures and more than doubles its training dataset to 1,143 scans (616 MRI, 527 CT). It achieved a Dice score of 0.839 on MRI and outperformed existing segmentation models while maintaining CT performance comparable to TotalSegmentator CT. Notably, training on both MRI and CT enhanced MRI segmentation accuracy, showcasing the power of multimodal learning. The algorithm offers a fully automated, sequence-independent solution that could streamline integration of quantitative imaging into radiology workflows. With its open-source code, dataset, and online segmentation tool, TotalSegmentator is very easy to try out.
First-In-Patient Scan for New Low-field Device
Exciting times for Wellumio, a New Zealand-based low-field device start-up. They scanned their first clinical patient, which marks the beginning of a clinical study evaluating their Axana 0.1T portable MRI for stroke detection. The portable device is a head-only scanner and has a compact cart design. The study is being conducted in partnership with the Australian Stroke Alliance and will assess the feasibility, safety, and usability of Axana in both healthy individuals and stroke patients. Researchers will also compare its diffusion-weighted imaging (DWI) capabilities to standard hospital MRI. The hope is that Axana will streamline stroke triage, allowing for informed decisions and rapid intervention during these emergent cases.
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
https://www.aidoc.com/about/news/aidoc-secures-landmark-fda-clearance/
https://theimagingwire.com/2025/02/19/staff-cuts-have-occurred-at-the-fda/
Akinci D’Antonoli, Tugba, et al. "TotalSegmentator MRI: Robust Sequence-independent Segmentation of Multiple Anatomic Structures in MRI." Radiology 314.2 (2025): e241613.
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.