Only the imaging lab, 3D CT point-surface exploration, guided anatomy and short checkpoint are functional in this V4 prototype. Later fellowship modules remain planned, not completed.
First learning target
Imaging-based localization before an extended approach
Explore CT-derived thresholded bone geometry, while recognizing it is not a validated surgical segmentation.
4
Pass the anatomy checkpoint
Interpret risk-relevant anatomy, distinguish true imaging observations from illustrative assumptions.
01 · Radiology workstation
CT + MRI multiplanar review
Load example NIfTI volumes, or open your own de-identified NIfTI locally. Three orthogonal views show voxel-aligned image planes.
NO VOLUME LOADED
The CT and MRI public examples are different datasets. Do not treat them as a registered CT/MRI pair or overlay them to infer surgical anatomy. This viewer does not apply the NIfTI affine for clinical orientation; plane labels are nominal voxel directions.
Ready. Choose Load public CT sample to download a published teaching volume in Safari. Internet is needed for samples; local files are processed in your browser only.
* CT presets expect calibrated HU; sample data may not provide reliable HU.
Plane A · nominal axial—
Axial voxel indexDrag slider or scroll image
Plane B · nominal coronal—
Coronal voxel indexDrag slider or scroll image
Plane C · nominal sagittal—
Sagittal voxel indexDrag slider or scroll image
Image dimensions
—
Voxel dimensions
—
Current image
—
You can also load a separate MRI sample using the MRI T1 tab.
02 · CT-derived anatomy exploration
Interactive bone surface from CT voxels
This is a 3D point-surface approximation extracted by intensity threshold and neighborhood boundary testing from a loaded CT volume. It is not a calibrated bone mesh, segmented carotid artery, or surgical navigation model.
LOAD CT TO GENERATE 3D
CT volume required
Open CT sample or a local skull CT, then click Generate. Orbit with one finger; pinch or use zoom controls. No artificial vessel/optic labels are drawn onto this scan.
Reading challenge
Identify the sellar floor, clival curvature, sphenoid sinus, foraminal relationships and limitations of threshold-only bone extraction. Verify findings directly on each 2D plane.
CT alone cannot reliably delineate optic nerves, tumor boundaries, or cavernous ICA course. Those require appropriate imaging, validated segmentation and clinical interpretation.
03 · Expert EEEA anatomy
Interpret the anatomy before choosing a corridor
Tap an advanced topic to expand teaching notes, then examine relevant levels in CT/MRI. No automatic landmark overlay is applied to public scans because no validated segmentation is available.
On CT, distinguish conchal, presellar and sellar pneumatization, Onodi cells and septal insertions toward carotid or optic prominences. Midline orientation on endoscopy is not reliably inferred from a single sinus septum. Confirm bony partitions on multiplanar CT.
The medial and lateral opticocarotid recesses are not interchangeable landmarks. Evaluate optic strut anatomy and its relationship to the optic canal, cavernous-to-clinoid ICA transition and lateral optic canal. Verify canal dehiscence and vascular variants on patient-specific studies.
Evaluate sellar floor, tuberculum, planum and interoptic corridor, and critically assess optic chiasm position, stalk trajectory and suprasellar cisternal relationships. CT establishes bone; MRI delineates neural and soft tissues; the samples here do not belong to the same subject.
CT maps clival bony depth and pneumatization. MRI establishes dorsum, prepontine cistern, basilar artery/brainstem relations and extension of pathology. Bone-only thresholding does not segment the basilar artery or dura.
Interrogate pterygoid base, vidian canal, foramen rotundum and adjacent ICA region; complex pneumatization changes the visible endonasal surface anatomy. The actual vidian nerve and vessel cannot be automatically identified from this generic threshold model.
Before relying on any CT-derived surface, confirm voxel spacing, orientation and coordinate transform; isolate skull base from adjacent structures; inspect bony discontinuities and metal artifacts; verify each landmark against original sections; document segmentation algorithm and limitations. This prototype does not satisfy surgical planning validation.
04 · Anatomy viva / knowledge check
Five advanced imaging decisions
Answer each question. Feedback appears directly under your selection. This is educational assessment, not credentialing.
Your checkpoint score
0 / 5
Choose an answer for each question to complete the lesson.
Data provenance & clinical limitations
Where the scans come from
The viewer is written in plain JavaScript and reads genuine NIfTI voxels. Example scans load over HTTPS from third-party hosts; they are not pre-bundled in the ZIP.
CT teaching volume
CT_Electrodes.nii.gz · from the NiiVue sample repository, attributed to SCI Institute / Seg3DData and distributed under MIT per dataset README. It is a brain/head CT example, not a purpose-built EEEA skull-base atlas.
chris_t1.nii.gz · acquired by Chris Rorden at the McCausland Center for Brain Imaging; described as CC BY-NC 4.0, so this demo use must remain non-commercial. Replace with separately licensed MRI before any paid product.
Independent samples. No coregistration. No patient CT/MRI, no contrast-enhanced skull-base protocol, no CTA, no surgical segmentation, no validated measurements. No patient data should be uploaded to a publicly hosted demo. Local NIfTI files are opened in browser memory without transfer by this app.
Supported local image types: uncompressed NIfTI-1 (.nii) or gzip-compressed NIfTI-1 (.nii.gz). Volumes must fit available phone memory. Some iOS browsers may block direct third-party downloads/CORS; local import remains the fallback.