Cinematic 3D Rendering
Photorealistic volume rendering with lighting, tissue colours and cine export. Organ masks come from Local AI.
- WebGPU renderer
- Organ masks: Local AI
- Windows · macOS · Linux
Core 3.0.69+ · AI setup
The download centre · NextRad Workstation V3
Current V3 core updates support Apple silicon Macs, Windows x64 and Linux x64. Add specialist plug-ins and Local AI as you need them. Specialty access depends on your licence and installed plug-ins.
ATLAS STUDIO · NEW
Polished teaching diagrams, annotated synthetic imaging, and PDF / PowerPoint lectures. Works with your Gemini API key or external Gemini Web.
Apple Silicon · Intel Mac · Windows 64-bit · Seven-day trial
02 · Make it yours
Signed packages install from Settings → Plug-ins on any supported system. Your licence decides what each one unlocks.
Photorealistic volume rendering with lighting, tissue colours and cine export. Organ masks come from Local AI.
Core 3.0.69+ · AI setup
Film layouts, image adjustments, preview and PDF composition for DICOM film printers and paper.
Core 3.0.67+
Full-resolution X-ray stitching for long-leg and full-spine runs, PET-CT whole body and multi-station MRI.
Core 3.0.69+
Cephalometric tracing and crown–root ratio; on CBCT, 3D length and angle, implant and endodontic planning, a lateral ceph and airway analysis.
Core 3.0.68+
Cobb angles and sagittal balance, pelvic parameters, leg-length discrepancy and lower-limb alignment, with calibration.
Core 3.0.68+
Phase-contrast CSF flow: velocity and flow curves through a drawn region.
Core 3.0.68+
Vessel extraction for CT angiography. Standard extraction runs in the core; a precise AI extraction is added by Local AI.
Core 3.0.69+ · AI setup
Coronary CT, calcium, chambers and cine MRI function, with AI cine contours, T2* tools and manual LGE segment grading. Automated scar quantification is in development.
Core 3.0.84+ · Research use, not for diagnosis · Test cases
Heart and liver T2* / R2* from multi-echo GRE, with liver and myocardial iron (LIC, MIC) from published calibrations.
Core 3.0.86+ · Research use, not for diagnosis · Test cases
Local AI runtime and model setup for Cinematic, Vascular and Cardiac tools.
Local report editor with templates, PDF export and a verification workflow, launched from a licensed Workstation.
Installs beside the Workstation
Imports DICOM from inserted USB sticks and CDs, and from watched folders.
Installs beside the Workstation
A DICOM node that burns patient CDs and DVDs with an embedded viewer.
Installs beside the Workstation
One package for mammography, breast MRI and ultrasound: structured findings, BI-RADS assessment, kinetic curves and reporting. Requires the Breast licence entitlement.
Core 3.0.86+ · Research use, not for diagnosis
Neuro measurements and planning, with brain segmentation and tractography interfaces. AI workflows require a compatible local runtime and models.
Core 3.0.69+ · Research use, not for diagnosis
Local CT lumen navigation with routes, linked MPR, measurements and cine export.
Core 3.0.72+ · Research use, not for diagnosis
MRI defecography and dynamic perianal / transperineal ultrasound. Review phases, edit measurements and prepare readable reports locally.
Core 3.0.86+ · Research use, not for diagnosis
Open Settings → Plug-ins → Install signed update, choose the downloaded .nrplugin file and restart the Workstation. Plug-ins can be enabled, disabled or rolled back there without a new core.
Signed packages work on every supported Windows, macOS and Linux host, x64 and ARM64. Companion apps (Reporting, CD Burner, Auto-Import) and Local AI use their own system installer.
03 · Process locally
Adds AI segmentation to Cinematic, Vascular and Cardiac. Models run on your machine; images never leave it.
Online setup ≈ 203 MB + ≈ 2.2 GB of runtime packages and models
The public installer is for Windows x64. On macOS and Linux, point Local AI at a compatible local Python runtime.
Cardiac MRI AI model pack ≈ 1.2 GB · Windows, macOS, Linux · core 3.0.85+
Unzip it and run the installer for your system; it checks every file and links the model to Local AI. Signed UK Biobank nnU-Net weights (CC0). Research use, not for diagnosis.
| Workflow | Model / engine | Acceleration | Local AI |
|---|---|---|---|
| Cinematic rendering | WebGPU renderer | GPU | Not required |
| Cinematic organ segmentation | TotalSegmentator | CUDA · Apple Metal · CPU | Required |
| Vascular standard extraction | Local vessel extraction | CPU | Not required |
| Vascular precise AI | TotalSegmentator | CUDA · Apple Metal · CPU | Required |
| Cardiac cine MRI contours | nnU-Net v2 · UK Biobank SAX (model pack) | CUDA · Apple Metal · CPU | Required |
| Neuro brain segmentation Local setup required | SynthSeg | GPU or CPU | Required |
| Neuro tractography Local setup required | DIPY (Diffusion Imaging in Python) | CPU | Required |
04 · Recommended computer
Suggested configurations for the best experience, not minimums or guaranteed processing times.
Windows ARM GPU inference is not verified. CPU fallback works with a compatible runtime but takes longer.
Study size and model settings decide memory use.
05 · On the horizon
Automated scar quantification and T1 / T2 mapping are in development. The available Cardiac plug-in includes manual LGE segment grading and T2* tools.
HL7 .hl7 / .txt and Paxera .txt / .rep parsing.
06 · Explore
Version 1.0.0 · 3,203 DICOM instances · DTI covers a limited parietal slab · facial anatomy removed from the brain study · no cephalometry or CBCT.
69e5f8bf09f58cf492b08b46028a71b897f5b964cc2fe73f894c62810c24c614Links to public research datasets for trying the Cardiac and Iron plug-ins. The cardiac CT sources are NIfTI / NRRD, so convert them to DICOM first. They are not hosted here: download from the source, under its own terms, and cite it as it asks. For testing and evaluation, not diagnosis.
1,000 adult coronary CTAs (ECG-gated acquisitions, one reconstructed phase, no RR timing) with expert coronary-artery labels. NIfTI: convert to DICOM before importing. CC BY-NC 4.0, non-commercial research and education; the images are on Kaggle (free account).
Zeng A et al. Comput Med Imaging Graph 2023;109:102287 · ImageCAS authors
Open source pagegithub.com ↗A 4D cardiac CT with 21 phases (truncal valve, 1.3 GB) and a single-phase structural CTA (55 MB). Phase indices only, no ECG timing. NRRD: convert to DICOM (for example in 3D Slicer) before importing.
SlicerHeart project, public TestingData release on GitHub
Open source pagegithub.com ↗45 SSFP cine short-axis studies with expert LV contours. DICOM in five batches of 0.4–0.8 GB. Public domain (CC0), direct download, no account.
Radau P, Lu Y, Connelly K, Paul G, Dick AJ, Wright GA. MIDAS J 2009, hdl:10380/3070 · Cardiac Atlas Project
Open source pagecardiacatlas.org ↗15 healthy volunteers: SSFP cine in short-axis, 2- and 4-chamber views, plus 3D tagging. DICOM, about 0.3–0.6 GB per case. Open access with attribution.
Tobon-Gomez C et al. Med Image Anal 2013;17:632–648, doi:10.1016/j.media.2013.03.008 · Cardiac Atlas Project
Open source pagecardiacatlas.org ↗124 thalassaemia patients, 12-echo GRE of the mid-ventricular heart and the liver, with reference T2* values. DICOM (one archive linked from the record). CC BY 4.0.
Abedi I, Zamanian M, Bolhasani H, Jalilian M. Mendeley Data V1, doi:10.17632/3d8bwn2fsh.1
Open source pagedata.mendeley.com ↗07 · Infrastructure
The viewer on its own, reporting, and the services that connect sites to a PACS.
The viewer alone: local DICOM folders and any DICOMweb server, as a desktop app or a PACS web build.
—
Standalone reporting: worklist by QIDO / C-FIND or HL7, signed PDFs back to the PACS.
—
Modality worklist server and RIS bridge for the modalities on site.
—
Sends studies from a site to the cloud PACS.
—
Patients view and download their studies and reports. Linux build upcoming.
—
A DICOM print server that drives film and paper printers. Linux build upcoming.
—
Receives studies and forwards them between DICOM nodes. Linux build upcoming.
—
The Broker built for Windows in C#. Linux build upcoming.
—
Independent signed plug-in: Fast light uses WebGL2; Path traced uses WebGPU. Version 1.2.7 adds natural tissue colours and named appearance presets. Separate organ hues require masks from Local AI; CT density colours work without masks.
Needs core 3.0.85 or newer, Local AI and the signed cine MRI model pack, downloaded from the Local AI section (about 1.2 GB). Model weights CC0: Ugurlu D et al., doi:10.5281/zenodo.15649643.
Public cardiac cases are linked further down this page.
Research use, not for diagnosis. Results must be confirmed by a qualified reader.
The CHMMOTv1 dataset has heart and liver multi-echo series with reference values.
Under review. Research use, not for diagnosis.
Install the signed package in Settings → Plug-ins and restart NextRad. Open Specialty → Breast. Current cores require a key explicitly including breast; installing the package does not add its licence entitlement.
Under review. Research use, not for diagnosis. Reader review is required.
Neuro measurements and planning, with brain segmentation and tractography interfaces. Install the signed package from Settings → Plug-ins and restart.
Brain segmentation and tractography depend on separately installed compatible local Python runtimes and model files. The plug-in download does not bundle these runtimes or weights. AI/model availability varies by platform.
Under review. Research use, not for diagnosis. Numeric and AI results require reader review.
Install from Settings → Plug-ins and restart NextRad. Open Specialty → Virtual Scope. Appropriate source data and licence access are required.
Under review. Research use, not for diagnosis.
Editable measurements with undo / redo, independent phase records and automatic local saving. Send to Reporting copies the formatted report and opens the same case; paste it into the editor. HTML and Excel / CSV exports are also available. Images and processing stay on your computer.
Requires core 3.0.86 or newer and the defecography licence entitlement. In Settings → Plug-ins, install the signed .nrplugin file and restart NextRad. Open an MR or US series, then choose Specialty → Anal canal to use the side panel. No core reinstall is required on a compatible version.
Research use, not for diagnosis. Manual measurements require reader review; clinical measurement accuracy has not been validated.
Install the signed plug-in update and restart NextRad. A core reinstall is not required for this print update.