Chiari malformation sits at the seam between bone, brain and CSF dynamics. AURORA-CHIARI treats the cranio-cervical junction as one coupled mechanical-hydrodynamic system — from the pressure phenotype that drives symptoms, to the decompression-vs-watch decision, to long-horizon syrinx surveillance.
A non-functional preview of the AURORA-CHIARI surface. Composable panels, logged overrides, explainable end-to-end. Pilot sites tune this to their own workflow.
CHIARI-ANATOMY Cranio-cervical anatomy88%CHIARI-CSF CSF hydrodynamics74%CHIARI-SYRINX Syringomyelia81%CHIARI-DECIDE Decompression vs watch62%AURORA-CHIARI is an open stack for the most over-imaged, under-decided condition in neurosurgery. Five subsystems treat the cranio-cervical junction as one coupled mechanical-hydrodynamic system — from posterior-fossa morphometry to syrinx surveillance over decades.
Chiari I malformation is defined radiographically — cerebellar tonsils ≥5 mm below the foramen magnum — but the radiographic finding is the easy part. The hard part is deciding which of the dozens of incidentally-discovered cases each year actually need surgery. The literature is full of contradictory thresholds; the same patient gets watched at one centre and decompressed at another. AURORA-CHIARI is built for that decision specifically.
The disease is at least three problems stacked: the bone (a small or shallow posterior fossa that compresses the cerebellum), the brain (downward herniation of cerebellar tonsils, sometimes with brainstem distortion), and the cerebrospinal fluid (impaired pulsatile flow at the cranio-cervical junction that drives symptoms and syrinx formation). Most diagnostic pipelines look only at the first. AURORA-CHIARI couples all three.
Symptoms map to the CSF dynamics, not to tonsillar position. Cough-induced occipital headache, sleep-disordered breathing, dysphagia, scoliosis in children: these correlate with peak systolic velocity and dwell time at the foramen magnum, not with millimetres of tonsillar descent. AURORA-CHIARI reads cine-MRI flow alongside structural imaging and reports both — and reports when they disagree.
Syringomyelia is the long-arc problem. Roughly 40–75% of Chiari I cases develop a syrinx; the natural history is highly variable. AURORA-CHIARI's syrinx module tracks per-vertebral-level morphology over years, not single timepoints — because a syrinx that is stable for five years is a different clinical entity from one that grew 2 mm last year.
We are not building a product around Chiari Malformation. We are seeding an infrastructure — so that any hospital with the will to use it can.
AURORA-CHIARI couples bone, brain and CSF dynamics. Tonsillar position is one input among many — and the worst predictor of who needs surgery. The module is built to make that relationship visible rather than hide it behind a single radiographic threshold.
Every AURORA-CHIARI subsystem can be adopted alone or as part of the bundle. Each ships with its own model cards, eval results and a one-command install.
Posterior-fossa morphometry: tonsillar position, clivus and supraocciput angles, foramen-magnum geometry, peg-like vs rounded tonsils.
Posterior-fossa morphometry. Auto-measures clivus-canal angle, supraocciput length, McRae line, basion-dens interval, and the tonsillar position itself. Distinguishes peg-like from rounded tonsils, which is a meaningful prognostic feature most pipelines miss.
Cine-MRI flow analysis at the foramen magnum; peak systolic velocity, dwell time, regurgitation, phase relationships.
Cine-MRI CSF flow analysis. Peak systolic velocity, dwell time, regurgitation patterns and phase relationships at the foramen magnum. The module also reports whether the flow signal is interpretable on this scanner; flow data from underspec'd protocols is refused rather than under-confidently rendered.
Per-vertebral-level cord-cavity segmentation; syrinx morphology, expansion rate, longitudinal stability index.
Per-vertebral-level cord-cavity segmentation. Tracks holocord vs focal patterns, expansion rate, longitudinal stability index. The syrinx's own time series is the unit of interest — a single timepoint is rarely diagnostic.
Patient-specific risk model for posterior-fossa decompression with or without duraplasty; symptom-driven thresholds explicit.
Patient-specific risk model for posterior-fossa decompression. Compares bone-only vs duraplasty trajectories with calibrated outcome bands. Symptom-driven thresholds are explicit; the module refuses to recommend surgery on radiographic findings alone.
Re-tonsillation, syrinx-recurrence and post-op pseudo-meningocele watch over years.
Long-horizon surveillance. Re-tonsillation, syrinx-recurrence and post-op pseudo-meningocele watch over the years that matter. Plays the same role for Chiari that SB-PROTECT plays for spina bifida — the case never falls off the radar.
Inputs the module reads from your existing systems, outputs it returns to them, the protocols it speaks, and the lifecycle of one case as it moves through AURORA-CHIARI.
Posterior-fossa morphometry. Measures clivus-canal angle, supraocciput length, McRae line, basion-dens interval, tonsillar position and tonsil shape (peg-like vs rounded — a real prognostic axis that single-number tonsillar-descent measurements miss). Each measurement reports its uncertainty and the slice it came from; no opaque single 'severity score'.
Cine-MRI CSF flow analyser. Phase-contrast, PSIR and PC-VIPR sequences supported. Reports peak systolic velocity, dwell time, regurgitation patterns and phase relationships at the foramen magnum. If the cine sequence is missing or under-spec'd, the module refuses inference and surfaces what would be needed — it does not render confident-looking numbers from bad data.
Per-vertebral-level cord-cavity segmentation. Distinguishes holocord, focal and presyrinx patterns. Stability index is computed across all available prior imaging — a syrinx that has been stable for five years is a different clinical entity from one that grew last year, and the module is explicit about which it is looking at.
Decompression-vs-watch decision model. Symptom-panel-gated: refuses to produce a decompression recommendation on imaging alone. Compares bone-only vs duraplasty trajectories with patient-specific outcome bands derived from comparative-effectiveness cohorts; the surgical team makes the call.
Long-horizon surveillance. Re-tonsillation watch, syrinx-recurrence early warning, pseudo-meningocele detection on post-op imaging. The case stays in CHIARI-PROTECT for the patient's life; the surveillance cadence is calibrated by the patient's stability index, not by calendar.
5 drafts by the module team and pilot collaborators. Each documents one slice of the substrate — methods, evaluation protocol, lessons. Slot PDFs in as they're ready; the entries point to placeholders until then.
These are the endpoints AURORA-CHIARI is built to measure across pilot deployments. The targets below are pilot goals, not retrospective results.
CCOS-aligned, across pilot-site cohorts (target lift vs current standard).
Reduction in operations on radiographic-only criteria; pre-AURORA baseline carries many of these.
12-month expansion vs stability on held-out longitudinal cohorts.
Selecting bone-only vs duraplasty with patient-specific bands.
Paediatric cohorts, polysomnography-confirmed.
Every clinician override of an AURORA-CHIARI recommendation is logged with reason.
pip install to a hospital deploy in one afternoon.Available to pilot partners today on private registries. At public alpha (Q4 2026) the same images, weights and signatures ship under MIT on public registries. For the full per-product download grid, see the unified download page.
# 1 · install pip install aurora-chiari # just AURORA-CHIARI # or pip install aurora-neuro[chiari] # paediatric+adult bundle
# 2 · verify aurora doctor chiari # → AURORA-CHIARI ✓ python ✓ models ✓ datasets # → CHIARI-ANATOMY · live · CHIARI-CSF · live · CHIARI-SYRINX · beta # → cine-MRI support: PSIR / phase-contrast / PC-VIPR # → ready in target <140ms
# 3 · try a public case
from aurora.chiari import load_case
case = load_case("bench/chiari/case-0042")
print(case.tonsillar_descent_mm) # → 9.4
print(case.psv_at_fm_cm_s) # → 3.6
print(case.syrinx.stability_index) # → 0.92 (stable)
print(case.symptom_panel.required) # → TrueAURORA-CHIARI is designed to run on hardware that already exists inside hospital networks. CPU-only inference is supported for the lighter subsystems; the heavier ones benefit from a recent GPU but do not require one. A single modern workstation handles routine cases under the latency target.
Federated deploys do not require open inbound ports. The runtime opens an outbound mTLS connection to the federation control plane; model updates are signed, audited and pulled. Patient data never traverses the federation. Consent metadata is a first-class object — subsystems with an explicit consent dependency refuse to run on records that lack the appropriate cohort policy.
Most pilot sites land on AURORA-CHIARI with an existing system in place. The recommended path is: install alongside, compare outputs on a held-out cohort for one quarter, then move read-only surfaces to AURORA, and decide on the rest. Reversibility is a design goal — nothing in the install creates lock-in.
For the full per-product, per-platform download buttons (macOS / Windows / Linux installers + pip / docker / helm / rust commands), see the unified download page. Each subsystem of AURORA-CHIARI is independently downloadable there.
AURORA is in private pilot today and opens to the world at public alpha in Q4 2026 under MIT. If you carry the weight of these diseases — as a clinician, scientist, builder, patient or advocate — there is a seat at the table.