28th Edition | Dual‑Energy Retrofit, Open AI Platform, and Low-Field Brains

Campbell Arnold
Nov 18
4 min read

“If we can make access more affordable, it could be a turning point in how widely AI is deployed in patient care.”

— Jonathan Whitmore, Director of Global Partnerships, Radiobotics

As the year winds down, I want to thank everyone who has read, shared, and subscribed to RadAccess! Your support helps us to highlight the people, research, and technology transforming global imaging access. With RSNA right around the corner, I’m excited for the opportunity to connect with many of you in person. If you’ll be at the meeting, please reach out, I’d love to say hello and chat with you about the latest in radiology and AI. You can reach me on LinkedIn or by email.

In this issue, we cover:

A Low‑Cost Dual‑Energy Retrofit for Traditional X-Ray
Marketplace Middlemen: Harrison.ai Takes Aim at AI Platform “Innovation Taxes”
Using Low-Field Scanners to Track Brain Changes in At-Risk Populations

If you want to stay up-to-date with the latest in Radiology and AI, then don't forget to subscribe!

A Low‑Cost Dual‑Energy Retrofit for Traditional X-Ray

How a low-cost dual-energy retrofit could expand global access to osteoporosis screening.

A research team from the University of Tokyo and Fujifilm has introduced Q-BONE, a dual-detector add-on that enables dual-energy imaging on a standard X-ray system, providing DEXA-like capabilities. By inserting a dual-layer detector cassette into an existing radiography unit, the system acquires low- and high-energy images in a single shot and computes both a Q-BONE bone mineral density score and high-contrast bone images.

In clinical testing with 200 participants, the Q-BONE score showed strong correlation with DXA (R = 0.91) and closely matched performance in phantom studies. Importantly, Q-BONE images offered significantly clearer visualization of vertebral contours, trabeculae, and fractures than standard radiographs (P < 0.01).

For imaging access, the implications are substantial. DXA scanners remain scarce in many regions, leaving osteoporosis underdiagnosed worldwide. A retrofit like Q-BONE could eliminate the need for separate DXA hardware, simplify workflow, and bring reliable bone health assessment to facilities that currently lack densitometry. If validated more broadly, this approach could meaningfully expand access to osteoporosis detection using existing radiography equipment.

Bottom line: Q-BONE could transform standard X-ray systems into affordable, high-quality osteoporosis screening tools, expanding access where DXA is unavailable.

Marketplace Middlemen: Harrison.ai Takes Aim at AI Platform “Innovation Taxes”

Can a zero-markup model shake up the economics of radiology AI platforms?

Harrison.ai co-founders Dimitry and Aengus Tran. Source: Harrison.ai — *Harrison.ai* *co-founders Dimitry and Aengus Tran. Source:* *Harrison.ai*

In a landscape where radiology AI solutions come from hundreds of vendors, AI platforms have long promised to simplify integration. However, as Harrison.ai’s Chief Growth Officer Josh Duncan puts it, “Traditional platforms promise to solve this but charge a high innovation tax.” To address this, Harrison.ai recently launched its own Open Platform, directly challenging the traditional medical imaging AI platform model built on heavy markup fees. The Open Platform aims to eliminate that tax by offering zero markup on third-party algorithms, emphasizing transparency, vendor neutrality, and clear ROI for customers.

Notably, Harrison.ai says it will host any sufficiently containerized algorithm, even those that compete directly with its own models. For health systems, the pitch is straightforward: a single interface to access multiple AI tools without platform-imposed fees impacting margins. The platform is slated to launch in 2026 with seven third-party vendors already committed:

AZmed (fracture detection)
CoLumbo (lumbar spine MRI)
Lucida Medical (prostate cancer)
Nicolab (stroke)
Radiobotics (fracture detection)
Therapixel (mammography)
Us2.ai (echocardiography)

The timing is striking, with Harrison.ai’s platform launch coming just as another major AI marketplace player, Blackford, shutters its operations. The question now is whether a zero-markup, vendor-neutral model can thrive where traditional platforms have struggled.

Bottom line: Harrison.ai’s Open Platform aims to remove the markup-laden bottlenecks of legacy AI marketplaces, can a vendor-neutral approach succeed and lower deployment costs?

Using Low-Field Scanners to Track Brain Changes in At-Risk Populations

Why lower cost systems could expand brain monitoring to more people living with HIV.

A recent study in Annals of Clinical and Translational Neurology highlights the feasibility of using portable low‑field MRI to monitor brain atrophy in people living with HIV age. Researchers scanned 30 virally suppressed patients with HIV and used a segmentation algorithm to quantify regional brain volumes. They then compared patterns of atrophy to two other age matched groups, patients with vascular comorbidities (n = 25) or mild cognitive impairment due to Alzheimer’s (n = 24).

They found reductions in the caudate, putamen, and white matter, while hippocampal volumes remained similar across cohorts and amygdala volume differences were specific to the mild cognitive impairment group. This demonstrates that even low-field, portable systems can provide clinically meaningful volumetric data in this patient population.

The findings carry significant implications for expanding imaging access. Portable systems can enable biomarker imaging in outpatient or low-resource settings. Additionally, frequent, low-barrier imaging could help track disease progression and intervene earlier in at-risk populations.

Bottom line: Low‑field MRI can deliver clinically meaningful brain imaging, enabling earlier and more frequent monitoring of at-risk populations.

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.