> To follow up with some research I did on this:
>
> No free DICOM viewer on the market carries FDA clearance or CE marking. Not 3D Slicer [1], not Horos [2], not OHIF. Even RadiAnt — which is paid and arguably the most popular Windows viewer — explicitly states "our viewer is not a medical product" and has no FDA/CE certification or timeline for one [3]. Interestingly, RadiAnt is also a solo-developer operation (Medixant, Poznań) — so the indie medical imaging developer is not a new phenomenon.
>
> The only DICOM viewer with both FDA 510(k) and CE Class IIa is OsiriX MD — paid (~$69/mo), Mac-only.
>
> What's interesting is how the open-source ecosystem handles this. As @shoo noted, license disclaimers provide baseline liability protection. But beyond that, both 3D Slicer and OHIF have served as the
foundation for FDA-cleared commercial products — Kitware built an FDA-cleared lung imaging tool on top of Slicer [4], and OHIF's GitHub explicitly notes it has "served as the basis for many FDA Cleared medical imaging viewers" [5].
>
> So the pattern seems to be: the platform stays uncertified, companies build certified products on top of it. The certification barrier appears to be fundamentally an enterprise-level compliance investment (ISO 13485 QMS, 510(k) submission, post-market surveillance) — not a reflection of software quality.
>
> OrthoRay sits in the same regulatory space as every other free viewer. The difference is the intended audience — orthopedic surgeons doing preoperative planning, teaching, and research — not primary radiological diagnosis.
>
> What I'm more curious about is the performance side. I built OrthoRay's rendering pipeline in Rust/wgpu because I kept hitting walls with existing viewers when doing real-time 3D volume rendering and MPR reconstruction on large datasets. The bottleneck wasn't the UI — it was the engine. I think there's an underserved gap in academic and research workflows where real-time GPU-accelerated rendering matters — intraoperative planning, surgical simulation, 3D reconstruction from CT/MRI for teaching.
>
> @verdverm — from your position in the commercial space, do you see demand for this kind of native, GPU-first rendering performance? Or is the industry moving entirely toward cloud/server-side rendering and the local viewer is becoming irrelevant?
>
> [1]
https://www.slicer.org/commercial-use.html
> [2]
https://groups.google.com/g/horos-project/c/TksXEbIDDek
> [3]
https://www.radiantviewer.com/dicom-viewer-forum/ce-certific...
> [4]
https://www.kitware.com/kitware-supports-development-and-lau...
> [5]
https://github.com/OHIF/Viewers