22.1 Unruptured Aneurysm, Cavernoma, AVM, and Dural AVF (03.H)

Blueprint item 03.H groups four lesions that reach the neuro ICU as either an incidental imaging surprise or a first hemorrhage. The exam move is not to recite every score point. It is to know which features force treatment, which scores omit those features, and which lesions are observed because the treatment is more dangerous than the natural history.

This is independent OpenExamPrep teaching for ABIM Neurocritical Care candidates. It cites published scores and trials. It is not an American Heart Association, American Association of Neurological Surgeons, or ABIM product.

Unruptured intracranial aneurysm: ISUIA, then PHASES and UIATS

Most saccular aneurysms that reach clinic never rupture. The International Study of Unruptured Intracranial Aneurysms (ISUIA) prospective cohort (Lancet 2003) still supplies the size-and-site frame boards expect. For patients without prior subarachnoid hemorrhage, 5-year rupture risk was essentially 0% for anterior-circulation aneurysms under 7 mm, then rose with size and was higher for the posterior communicating artery and posterior circulation than for the anterior communicating, middle cerebral, or cavernous internal carotid groups. Cavernous carotid aneurysms rarely cause subarachnoid hemorrhage because they are extradural; they present with pain or cranial neuropathy, not cisternal blood.

ISUIA selected patients already chosen for observation, so absolute risks are lower than in an unselected population. That is why later tools exist.

The PHASES score (Greving and colleagues, Lancet Neurology 2014) estimates 5-year rupture risk from six variables you can remember as the acronym:

In the derivation table, 5-year rupture risk ran from about 0.4% at PHASES 2 or lower, through 2.4% at 7 and 5.3% at 10, to about 17.8% at 12 or higher. Two traps matter on a one-best-answer item. First, posterior communicating artery aneurysms score like posterior-circulation aneurysms (4 points), even though many surgeons discuss them with the anterior circle. Second, PHASES does not award points for growth, daughter sacs, or irregular walls. A 5 mm anterior communicating aneurysm that sprouts a daughter sac or enlarges on the 6-month magnetic resonance angiogram is a treatment candidate even if the PHASES integer is still small.

The Unruptured Intracranial Aneurysm Treatment Score (UIATS) (Etminan and colleagues, Neurology 2015) is a two-column consensus tool: aneurysm and patient factors that favor repair versus age, comorbidity, and life-expectancy factors that favor observation. If the treatment column exceeds the conservative column by 3 or more points, the panel favored treatment. If conservative care leads by 3 or more, observe. A difference of 2 points or less is equipoise. You do not need every UIATS row on the exam. You need the idea: PHASES estimates rupture; UIATS asks whether that patient should be treated.

Treat—or at least send urgently to a dual-trained cerebrovascular conference—when any of the following is true:

• Symptomatic mass effect (classic: painful third-nerve palsy from a posterior communicating aneurysm — treat as a warning leak until proven otherwise).
• Documented growth on comparable serial imaging.
• High-risk morphology: daughter sac, multilobulation, irregular wall, high aspect ratio, or a daughter bleb at a branch point.
• PHASES or UIATS in a range where the residual lifetime risk is high (young patient, large posterior lesion, prior SAH from another aneurysm).

Observe many small, smooth, anterior-circulation aneurysms in older patients after smoking cessation and blood-pressure control, with a planned imaging interval. Screening is reasonable in autosomal-dominant polycystic kidney disease and in families with two or more first-degree relatives with intracranial aneurysms; a single affected relative is a weaker indication.

Worked PHASES example: a 72-year-old Japanese patient with treated hypertension and an unruptured 8 mm basilar-tip aneurysm, no prior SAH, scores 3 (Japanese) + 1 (hypertension) + 1 (age) + 3 (size 7.0–9.9 mm) + 0 (no earlier SAH) + 4 (posterior site) = PHASES 12, which sits in the highest published 5-year risk band. That is not an incidental 4 mm middle-cerebral aneurysm in a 45-year-old North American without hypertension (PHASES 2).

Cavernous malformation: hemorrhage clustering and epilepsy

A cavernous malformation (cavernoma, CCM) is a mulberry of thin-walled sinusoids without intervening brain parenchyma. Gradient-echo or susceptibility-weighted MRI shows the classic popcorn lesion with a hemosiderin rim (Zabramski type II). Familial disease (KRIT1/CCM1, CCM2, PDCD10/CCM3) produces multiple lesions; a single lesion is usually sporadic and may sit next to a developmental venous anomaly that you must not sacrifice.

Natural history is not AVM natural history. Prospective series (including the Scottish audit of Al-Shahi Salman and colleagues, Lancet Neurology 2012) put the risk of a first-ever hemorrhage from an incidental cavernoma well under 1% per year (5-year first-hemorrhage risk on the order of 2–3%). After a hemorrhage, risk clusters: several percent per year in the first 2 years, then it often settles. Brainstem lesions have a higher rebleed rate and less room for swelling. Seizure, not hemorrhage, is the most common presentation of a supratentorial cavernoma.

Surgery versus observation:

• Operate for recurrent hemorrhage, progressive focal deficit attributable to the lesion, or drug-resistant epilepsy when electroclinical data point at an accessible cavernoma (and often the surrounding hemosiderin-stained cortex).
• Observe incidental deep or brainstem lesions without repeated bleeds. Stereotactic radiosurgery for unresectable brainstem cavernomas remains debated because radiation necrosis can mimic rebleeding.
• After a first seizure, start an antiseizure medication; a single remote hemorrhage without epilepsy is not an automatic indication for resection.

Brain arteriovenous malformation: Spetzler-Martin and ARUBA

A brain AVM is a nidus of arteries shunting into veins without a capillary bed. Annual hemorrhage risk for an unruptured AVM is often quoted at 2–4% per year; the medical arm of ARUBA observed about 2.2% per year. After rupture, early rebleed risk is higher, which is why a ruptured AVM is treated once the patient is salvageable.

The Spetzler-Martin grade (1986) predicts surgical difficulty, not rupture risk:

Grade is the sum (I–V). Compact, non-eloquent, small, superficially draining lesions (I–II) are the ones experienced surgeons still resect with low morbidity. Grade IV–V lesions are often observed unless they have bled repeatedly.

ARUBA (Mohr and colleagues, NEJM 2014) randomized unruptured brain AVMs to medical management versus interventional therapy (any mix of surgery, embolization, and radiosurgery). The trial stopped early after 226 patients and a mean 33 months of follow-up: death or symptomatic stroke occurred in 10.1% of the medical group versus 30.7% of the interventional group. Longer follow-up still favored the medical arm on the original endpoint. The controversy you must be able to state is equally high-yield: short horizon versus lifetime rupture risk, heterogeneous and often incomplete endovascular treatment, under-representation of straightforward grade I–II surgical candidates, and the fact that a young patient with an accessible unruptured AVM is not the average ARUBA enrollee. Exam translation: do not quote ARUBA as a ban on resecting a ruptured or a compact grade I convexity AVM; do not quote it as irrelevant either.

Cranial dural arteriovenous fistula

A cranial dural AVF is a shunt inside the dura, usually into a venous sinus. The feature that changes natural history is cortical venous drainage (CVD) — arterialized flow into pial veins. Borden I drains only into a sinus (antegrade). Borden II is sinus drainage plus CVD. Borden III is CVD only. The Cognard scale splits the same idea more finely (I, IIa, IIb, IIa+b, III, IV with venous ectasia, V with spinal perimedullary drainage). CVD (Borden II–III; Cognard IIb and higher) marks an aggressive fistula: annual intracranial hemorrhage is often cited in the 8–10% range and is higher still with venous ectasia. Those lesions are disconnected (transarterial or transvenous embolization, or surgical interruption of the draining vein). Borden I fistulas can be observed if the only symptom is a tolerable bruit, unless ocular venous hypertension threatens vision (cavernous-region fistulas).

Spinal dural AVF and Foix-Alajouanine

Spinal dural AVF (the common type I spinal shunt) typically affects men in the sixth or seventh decade. A fistula on a radicular sleeve arterializes a radiculomedullary vein. The cord suffers venous congestion, not arterial steal. The bedside story is stepwise or gradually progressive myelopathy: gait, then sphincter, then a sensory level, often mislabeled as spondylotic myelopathy or transverse myelitis. MRI shows longitudinally extensive T2 hyperintensity, a swollen cord, and dorsal flow voids. Catheter angiography finds the fistula; treatment is occlusion of the fistula (endovascular liquid embolic when the anatomy is safe, or surgical ligation of the proximal draining vein). Function can recover over months if congestion is reversed before irreversible necrosis.

Foix-Alajouanine originally described subacute necrotizing myelopathy. Modern reading treats it as the end stage of venous hypertensive myelopathy from an untreated spinal fistula, sometimes with venous thrombosis. The exam action is the same: look for flow voids, do not stop at empiric steroids, and treat the fistula.

Lesion-to-lesion comparison