“Stroke care depends on rapid, accurate diagnosis, yet brain imaging too often remains tied to centralized hospital infrastructure.”
— Adam Caplan, Founder and General Partner of Jumpspace Ventures
Welcome to Radiology Access! Your biweekly newsletter on the people, research, and technology transforming global imaging access.
In this issue we cover:
One Small X-Ray for Man, One Giant Leap for Imaging Access
The Lowest Dose Is the One You Don’t Give: Cerebriu’s SmartGAD
Shedding the Mortal Coil: Investors Bet $9.75M on Wellumio’s Vision for Bedside Imaging
Cortechs.ai Expands NeuroQuant to PET
If you want to stay up-to-date with the latest in Radiology and AI, then don't forget to subscribe!
🎙️ Scanning the Market w/ Jay Gurney
I recently joined Jay Gurney on the Scanning the Market podcast for a wide-ranging conversation about the future of medical imaging. We talked about imaging access, low-field MRI, AI, and where I think the field is headed over the next five years. If you enjoy reading RadAccess, I think you’ll love this conversation, check it out!
One Small X-Ray for Man, One Giant Leap for Imaging Access
The first human radiographs captured in space demonstrate the potential of portable imaging in extreme environments.
It takes Earth 12 months to orbit the Sun, but academic publishing cycles are closer to 16 months. We covered the FRAM2 mission, where the first X-ray in space was collected, in the 12th issue of RadAccess back in April 2025. Now, more than a year later, the results have finally been published in Radiology.
Researchers demonstrated that a commercial, battery-powered portable X-ray system from KA Imaging and MinXray could produce diagnostically adequate human radiographs in orbit, with nonmedical crew members requiring only a few hours of training.
The study compared seven preflight and seven in-flight anatomic radiographs and found no significant differences in overall image quality, spatial resolution, or contrast resolution. The first image shared publicly featured a hand wearing a ring, mirroring the first-ever X-ray taken by Wilhelm Roentgen in 1895. The biggest challenge during the mission was positioning. Imaging of the chest, abdomen, and pelvis was worse in microgravity, highlighting that the hardware may be ready but some workflow challenges remain.
The system also demonstrated a second, unexpected use case: non-destructive testing. The high-resolution detector could visualize internal spacecraft components down to the submillimeter scale, meaning the same technology could potentially monitor both astronauts and spacecraft.
This publication is a powerful reminder that the technologies needed to bring imaging to the most extreme environments can also expand access here on Earth. If a minimally trained astronaut can acquire a diagnostic X-ray in orbit, the same basic approach has obvious implications for rural communities, disaster zones, and other places where conventional imaging infrastructure is hard to reach.
Bottom line: The FRAM2 mission demonstrates that portable X-ray can deliver diagnostic-quality imaging in space.
The Lowest Dose Is the One You Don’t Give
Cerebriu’s Apollo SmartGAD uses AI to turn contrast administration into a real-time imaging decision.
What’s the best way to reduce gadolinium dose? Easy, don’t dose the patient in the first place. That’s the pitch behind a recent Communications Medicine article from Cerebriu, which uses AI to help determine in real time which patients actually need contrast based on images acquired during the scan. Cerebriu’s Apollo SmartGAD system analyzes standard non-contrast brain MR images acquired on scanner and and recommends contrast usage for the protocol, either no contrast, reduced-dose contrast, or standard-dose contrast.
The system uses multiple AI components for lesion detection, T1 contrast-enhanced image synthesis, and segmentation to determine whether contrast should be used. Interestingly, the synthetic image isn't intended for diagnosis. Instead, it serves as a proxy for estimating whether lesions are likely to enhance and therefore whether contrast should be administered. The tool has two potential workflow benefits:
Preventing unnecessary contrast in patients who don't need it.
Avoiding call-backs for patients initially scheduled for non-contrast studies who are subsequently found to have lesions that may require contrast.
The study evaluated 1,251 retrospective MRI examinations across two cohorts, patients undergoing first-time brain assessment and patients undergoing glioma monitoring. The results found:
First-time brain scans: The algorithm identified patients who could potentially avoid contrast with 73% sensitivity and 79% specificity. More importantly, it achieved an 86% negative predictive value, meaning that 86% of patients the algorithm recommended for a non-contrast study had no contrast-enhancing lesions.
Glioma monitoring: The algorithm distinguished between non-enhanced, reduced-dose, and standard-dose protocols with 77% sensitivity and 78% specificity.
False negatives: Where things get tricky is when patients would have shown enhancement, but the algorithm misses it. The authors report that false negatives were particularly associated with small lesions, most notably metastases.
The technology is interesting because it leverages synthetic contrast-enhanced image generation in a novel way, as a proxy for determining contrast administration, turning a predetermined protocol into a real-time imaging decision.
The big question is whether this can move from retrospective validation into prospective clinical practice. The algorithm is designed to make a contrast decision while the patient is already in the scanner, which raises practical questions about how it fits into existing protocol ordering, radiologist oversight, and reimbursement and prior-authorization workflows, particularly in the US.
Bottom line: Cerebriu's latest algorithm aims to turn contrast administration from a predetermined protocol into a real-time decision, reducing gadolinium use through elimination.
Shedding the Mortal Coil: Investors Bet $9.75M on Wellumio’s Vision for Bedside Imaging
Can eliminating gradient coils enable a new generation of compact, portable scanners for stroke care?
The New Zealand-based Wellumio has closed an oversubscribed $9.75 million pre-Series A round to accelerate ongoing clinical validation, regulatory submissions, and commercialization of its Axana portable MRI system.
Axana is designed around one of the most compelling applications for portable MRI, point-of-care stroke imaging. While most low-field MRI systems still rely on conventional gradient-based spatial encoding, which adds substantial hardware, power, and engineering requirements, Wellumio has taken a fundamentally different approach with pulsed gradient free mapping, which eliminates pulsed gradient coils entirely. This significantly simplifies the system architecture, enabling a more compact and quieter design specifically targeted for rapid bedside imaging.
The ~110-lb head scanner requires no cryogenics or shielded room and can operate on battery power. The team has deliberately optimized Axana around the target problem and portability rather than trying to replicate the imaging capabilities of a conventional MRI in a smaller package. They're building a purpose-built scanner centered on a single application, rapid stroke assessment.
The latest funding round, led by Nuance Connected Capital, will support expanded clinical evaluation across additional sites and geographies, further product development, regulatory submissions, and commercialization. This is an important milestone for a company I've been following since the early days of RadAccess. What makes Wellumio particularly interesting to me is its laser focus on portability and point-of-care stroke imaging. The questions now are is less truly more, and are the riches in the niches?
Bottom line: Wellumio and its investors up the ante, betting $9.75M that a smaller, simpler, purpose-built bedside scanner can deliver enough clinical value to make up for what it leaves out.
Cortechs.ai Expands NeuroQuant to PET
The company brings its quantitative imaging approach to amyloid PET and Alzheimer’s disease.
Cortechs.ai is expanding NeuroQuant beyond MRI and into PET, following FDA clearance of NeuroQuant PET. The software provides automated quantification of amyloid PET, including regional uptake measurements and Centiloid scores, across commonly used amyloid tracers.
The move extends NeuroQuant’s long-standing approach of turning brain imaging into standardized, quantitative measurements from MRI into molecular imaging. Cortechs has offered NeuroQuant PET in research settings for years, but the new clearance brings the technology into clinical practice.
It’s a logical expansion for Cortechs, particularly given its existing lesion surveillance tools for monitoring Alzheimer’s patients undergoing anti-amyloid therapies. It will be interesting to see whether Cortechs’ established position in quantitative MRI can translate to the increasingly important market for quantitative PET in Alzheimer’s disease.
Bottom line: Cortechs is taking NeuroQuant from MRI into PET, positioning itself to play a larger role in the imaging workflow around Alzheimer’s disease and anti-amyloid therapies.
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.
