Sturge-Weber Syndrome (GNAQ/GNA11 Mosaicism)

Congenital capillary-venous malformation of the face, leptomeninges and eye, caused by postzygotic somatic activating variants in GNAQ (mostly p.Arg183Gln), GNA11 (p.Arg183Cys) and, in one published patient, GNB2; sporadic, non-inherited disease.

Literature and regulatory information updated: 9 October 2026.

0. Initial Reference Table

1. Introduction

1.1. Foundational descriptions and historical overview

Sturge-Weber syndrome (SWS) was delineated in three stages over four decades. In 1879, Sturge described a patient with an extensive port-wine birthmark (PWB) of the face, scalp and trunk, focal seizures contralateral to the skin lesion and intraocular vascular malformation with glaucoma, and proposed that a vascular lesion resembling the skin lesion affected the brain surface [18]. In 1901, Kalischer demonstrated ipsilateral leptomeningeal vascular malformation pathologically, and in 1922 Weber described the radiological signs of cerebral atrophy and intravascular calcification that would define the classical disease image [18]. The alternative eponyms Sturge-Weber-Krabbe and Sturge-Weber-Dimitri remain listed as Orphanet synonyms [1]. Mid-twentieth-century paediatric literature established the disease among congenital childhood neurocutaneous syndromes, with reviews such as Chao's (1959), cited in Smith's textbook [2]. Over time, the definition expanded to cases without a skin lesion, where isolated leptomeningeal malformation is recognised only through neuroimaging [7, 18].

1.2. Nomenclature and reclassifications

Older terminology called the skin lesion a “facial capillary haemangioma” or “naevus flammeus”, and clinical genetics reference texts still retain this wording [2, 13, 32]. However, the lesion is a congenital capillary malformation (CM), present at birth and lifelong, without the proliferative and involutional phases of infantile haemangiomas, which are vascular tumours [17, 33]. In the International Society for the Study of Vascular Anomalies (ISSVA) classification, SWS belongs to capillary malformations associated with central nervous system (CNS) and/or ocular abnormalities [21]. US consensuses and several recent reviews use the term “port-wine birthmark” (PWB), which replaced port-wine stain (PWS) [17, 34]; this article uses PWB as the equivalent of the usual Portuguese term “mancha vinho do Porto” (MVP). In French terminology, made explicit in the 2024 PNDS (French National Diagnosis and Care Protocol), the lesion is called angiome plan (capillary malformation in the ISSVA classification), the term naevus flammeus is reserved for the transient mediofrontal patch of infancy, and “haemangioma” is considered inappropriate because it designates proliferative vascular tumours; for the same reason, the PNDS places “choroidal haemangioma” in quotation marks and records the proposed term “choroidal capillary malformation” [19]. Smith's textbook also calls the disease “Sturge-Weber sequence”, an expression emphasising its origin in a single developmental event with cascading consequences [2]. The Roach classification, cited through indirect attribution, divides the disease into type 1 (facial and leptomeningeal angioma, with or without glaucoma), type 2 (facial angioma without intracranial involvement, with possible glaucoma) and type 3 (isolated leptomeningeal angioma without a facial naevus and typically without ocular involvement) [7, 35].

1.3. Milestones in molecular discovery

Rudolf Happle proposed that sporadic, asymmetric or patchy congenital defects involving skin resulted from mosaic somatic mutations, lethal if present throughout the embryo; SWS fitted this hypothesis through segmental distribution and absence of familial cases [3]. In 2013, Shirley and colleagues sequenced whole genomes from paired affected and unaffected tissue in three patients and identified one shared non-synonymous variant, GNAQ c.548G>A p.(Arg183Gln), subsequently detected in 23 of 26 SWS patients (88%) and 12 of 13 people with apparently non-syndromic PWB (92%), with mutant allele fractions of 1.0% to 18.1% in affected tissue [3]. Identifying the same variant in syndrome and isolated PWB confirmed an allelic spectrum whose extent depends on the timing and cell lineage of the postzygotic event [2, 3]. In 2021, Thorpe and colleagues described a mosaic activating GNA11 variant in SWS with atypical features, and that same year Fjær and colleagues identified mosaic GNB2 p.(Lys78Glu) in a Norwegian patient without a GNAQ variant [4, 6]. Subsequent European and North American series characterised GNA11-associated SWS as having extensive reticulated CM and milder neurological disease [5, 12]. The Brain Vascular Malformation Consortium, funded by the National Institutes of Health (NIH) through the Rare Diseases Clinical Research Network (RDCRN), organised the US working group reviewing diagnostic and therapeutic advances and maintains the natural history study [20, 36, 37].

1.4. Definition of entities within scope

SWS is defined by the presence of at least two of three components: facial PWB, cerebral (leptomeningeal) vascular malformation and ocular vascular malformation [17, 22]. This article addresses classic SWS due to GNAQ p.(Arg183Gln) mosaicism, which accounts for the vast majority of cases; GNA11 p.(Arg183Cys)-associated SWS; SWS with a somatic GNB2 p.(Lys78Glu) variant, known from a single patient; non-syndromic PWB, an allelic GNAQ entity; and Roach type 3 SWS, without a skin lesion [3, 4, 6, 7]. Diffuse GNA11 capillary malformation, pigmentovascular phacomatosis and cutis marmorata telangiectatica congenita are addressed as neighbouring entities within the same G-protein axis [9, 12]. PIK3CA-related overgrowth spectrum (PROS) and Klippel-Trenaunay syndrome, isolated cerebral arteriovenous malformation (AVM) and CM-AVM/Parkes Weber syndrome are summarised differential diagnoses; PHACE syndrome (posterior fossa malformations, haemangioma, arterial lesions, cardiac and eye abnormalities) is addressed in the differential diagnosis [33].

2. Molecular Genetics and Pathophysiology

2.1. Heterotrimeric Gq/11 G proteins

GNAQ (9q21.2) encodes Gαq, the α subunit of heterotrimeric Gq protein. Gαq alternates between active guanosine triphosphate (GTP)-bound and inactive guanosine diphosphate (GDP)-bound states under G-protein-coupled receptor control; when active, it stimulates phospholipase C-β, generating diacylglycerol and inositol trisphosphate, second messengers mobilising calcium and activating protein kinase C and, downstream, the mitogen-activated protein kinase (MAPK) pathway [8]. The protein participates in platelet activation, B-lymphocyte selection, chemotaxis and cardiac development [8]. GNA11 (19p13.3) encodes Gα11, with 359 amino acids and 90% identity to Gαq, and GNB2 (7q22.1) encodes one β subunit (Gβ2) that, with Gγ, forms the Gβγ dimer associated with the α subunit in its inactive state [4, 6]. Inactivation depends on intrinsic α-subunit GTPase activity, hydrolysing GTP and returning the protein to its GDP-bound state [3].

2.2. GNAQ p.(Arg183Gln): molecular mechanism

GNAQ NM_002072.5:c.548G>A p.(Arg183Gln) alters a GTP-binding-pocket residue conserved across 24 human Gαq paralogues and required for GTP hydrolysis; substitution reduces GTPase activity and maintains constitutively active Gq signalling [3, 8]. In HEK293T cells, mutant protein modestly activated extracellular signal-regulated kinase (ERK) and serum response element activity, without activating p38 or AKT and with weak JNK activation; the effect was smaller than that of p.(Gln209Leu), characteristic of uveal melanoma [3]. Whole-genome sequencing showed no mutation accumulation in affected tissues, consistent with a malformative lesion without oncogenic behaviour [3]. GNAQ and GNA11 codon 209 alterations produce stronger activation of the same pathway and drive melanocytic lineage tumours, a biological context distinct from the vascular phenotype [8, 9].

2.3. Mosaicism, cell types and postzygotic event timing

The variant is somatic and postzygotic: in the founding study it was present in PWB skin from 9 of 9 SWS patients and absent from visibly normal skin in 6 of 7, and it was detected in brain from 15 of 18 patients but not in 6 control brains, 4 cerebral cavernous malformations or 664 of 669 samples from the 1000 Genomes project [3]. Variant allele fractions (VAF) in affected tissue ranged from 1.0% to 18.1%, with read depths of 2446 to 93,008 (median 12,947); in PWB skin from SWS patients they ranged from 2.05% to 6.46% and in brain from 0.13% to 11.15% [3]. In Fjær's Norwegian cohort, the variant was detected in endothelial cultures from lesional dermis but not in keratinocytes or fibroblasts, and laser-microdissected endothelium had VAF of 20% to 30%, indicating enrichment of the variant in endothelial cells derived from a subset of angioblasts [6]. The authors of the founding study proposed that non-syndromic PWB results from a late mutation restricted to endothelial cells and that SWS results from an earlier mutation in multipotent progenitors [3]. The classical embryological hypothesis attributes the distribution of the lesions to a defect in a restricted part of the cephalic neural crest, whose cells migrate to the supraocular dermis, the choroid and the pia mater [2]. Waelchli and colleagues related the highest-risk area (the “forehead”, from the midline to a line between the outer canthus and the top of the ear, including the upper eyelid) to the frontonasal prominence and to optic-vesicle-derived skin, the only facial regions formed from forebrain and anterior midbrain neural crest, as are the cortex and the eye; the maxillary and mandibular regions derive from the first branchial arch [18]. This vascular-embryological interpretation explains why PWB distribution follows the embryonic vasculature rather than trigeminal dermatomes [17, 18]. The French PNDS (French National Diagnosis and Care Protocol) describes the same territories as four recurrent facial segments, shared by segmental infantile haemangiomas and probably by PWBs: S1 frontotemporal, S2 maxillary, S3 mandibular and S4 centrofacial (from the forehead to the upper lip); SWS PWB corresponds mainly to S1 and S4 [19]. The same document highlights three consequences of mosaicism: the variant would be lethal if present in every cell, and carrier cells survive only alongside normal cells; any germline involvement has no consequence for offspring, because an embryo with the constitutional variant would not develop; and no hereditary or environmental factor predisposing to the postzygotic event is known [19].

2.4. Tissue mechanisms and cerebral and ocular pathophysiology

Leptomeningeal malformation is capillary-venous, with dilated venules and impaired cortical venous drainage; blood is diverted to enlarging deep medullary and subependymal veins, and venous stasis with recurrent thrombosis is the proposed mechanism of progressive ischaemic injury, atrophy and cortical calcification [22, 23]. Perfusion studies show impaired venous drainage aggravated by seizures, and expansion of deep medullary veins may be compensatory [20, 23]. Neuropathology in a GNA11 case showed thickened congested leptomeninges with dilated venules, degenerated arterial walls containing myxoid material, calcification and macrophages [4]. Ocular glaucoma results from two mechanisms: anterior chamber angle dysgenesis (high iris insertion, hypoplastic scleral spur, ciliary muscle directly attached to trabecular meshwork), predominant in early forms, and increased episcleral venous pressure, predominant in later forms; current ocular hypotensives do not reduce elevated episcleral venous pressure [21, 38]. Levels of pS6, a marker of the mTOR (mammalian target of rapamycin) pathway, are increased in PWB and leptomeningeal endothelium, providing the rationale for sirolimus trials [23, 39]. Recent preclinical studies added two potentially treatable mechanisms: in human dermal endothelial cells expressing Gαq-R183Q, phosphoproteomics identified aberrant calcineurin–NFAT–DSCR1.4 cascade activation, confirmed in patient biopsies and partly normalised by tacrolimus [40]; and variant-bearing endothelial cells, alongside a Gnaq p.R183Q mouse, showed increased endothelial permeability, partly reversed through angiopoietin-2 suppression and MEK1/2 inhibition with trametinib [41]. In a transgenic GNAQ R183Q mouse model, low-dose kainate-induced seizures were more severe and associated with higher mortality, longer and wider cortical microvessels and transient reduction of tight-junction proteins [42]. Fjær and colleagues observed in human umbilical vein endothelial cells (HUVEC) that GNAQ R183Q and GNB2 K78E reduce proliferation and total and phosphorylated YAP levels, and proposed the G-protein–YAP/TAZ (Hippo) pathway as central, with MAPK less essential; this hypothesis rests on in vitro data [6].

2.5. GNA11 p.(Arg183Cys)

Thorpe and colleagues studied five GNAQ p.(Arg183Gln)-negative cases and identified GNA11 c.547C>T through exome sequencing in a 63-year-old man with extensive right-sided facial and body PWB, without seizures or deficits until an acute hypertensive event with left thalamic and right subarachnoid haemorrhages and infarction; VAF was 6.1% on exome and, on targeted validation, 0.7% to 1.9% in brain, 5.9% in affected skin and 0.06% in unaffected skin [4]. The letter text calls the variant “p.R183Q”, but the article table and nucleotide change c.547C>T correspond to p.(Arg183Cys), confirmed through VariantValidator normalisation [4, 11]. Dompmartin and colleagues (Brussels and Nijmegen) described three SWS patients with somatic GNA11 variants and reviewed five published cases: all had widespread CM and limb hypertrophy or hypotrophy, two of three glaucoma, and all neurological symptoms with moderate cerebral atrophy, less severe than classic SWS [5]. In Zhang's multinational series (17 patients with somatic GNA11 p.R183C, VAF 0.7–8%, median age 18 years), CM affected limbs in 100%, trunk in 88% and face in 47%, was bilateral in 71% and reticulated and poorly demarcated in 94%; 82% had limb growth discrepancy, 29% glaucoma (all with bilateral facial CM) and 29% hypertension, including two cases of hyperaldosteronism before 18 years and one bilateral renal artery stenosis [12]. The authors suggest GNA11 leptomeningeal involvement is less frequent and seizures milder and later than in GNAQ disease, a hypothesis based on small retrospective series [12].

2.6. GNB2 p.(Lys78Glu): a single patient

In a Norwegian cohort of six classic SWS patients, five had GNAQ p.(Arg183Gln) in skin biopsy; in the sixth (patient 3, aged 22, left eyelid, nose and eyebrow PWB, mild intellectual disability, refractory epilepsy and left leptomeningeal malformation with atrophy), deep exome sequencing (approximately 400×) of lesional dermis identified GNB2 NM_005273.3:c.232A>G p.(Lys78Glu) in 6% of reads [6]. Amplicon sequencing validation showed VAF of 3.6% in lesional dermis, 21% in endothelial culture and 0.15% in non-lesional dermis; the variant is absent from COSMIC and gnomAD [6]. The Gβ beta-propeller Lys78 residue forms a salt bridge with Gα N-terminal helix residue Asp22 (Asp32 in Gαq), so substitution would disrupt Gαq–Gβγ interaction [6]. GNB2 and GNB1 showed no variants in brain from five additional GNAQ-negative patients or blood and brain from a Japanese patient, and the authors consider the association a rare cause requiring replication [6]. The variant has no ClinVar record [11]. Germline GNB2 variants described in other diseases, such as neurodevelopmental disorder with hypotonia and dysmorphism (OMIM 619503) and familial sinus and atrioventricular dysfunction, are unrelated to SWS [10].

2.7. Variant spectrum and normalisation

Table sources: [3, 4, 6, 11, 12, 43, 44]. All four descriptions were validated using VariantValidator against GRCh38 and MANE Select transcripts [11]. In ClinVar record 548668, classifications were submitted as germline and include CM, segmental undergrowth and a “CLOVES” designation; a separate record (VCV003774519) describes dinucleotide alteration c.546_547delinsTT, also encoding p.(Arg183Cys) [44]. Germline ClinVar classifications apply poorly to somatic variants and cannot estimate recurrence risk [43]. In patients assessed for SWS with atypical features, broad testing identified PIK3CA, PTPRD, RASA1, EPHB4 or KIT variants reclassifying diagnosis (section 6) [45].

2.8. Genotype–phenotype correlations, penetrance and expressivity

Penetrance, a concept pertaining to germline variants, does not apply directly to somatic mosaicism: the same GNAQ variant produces isolated PWB or complete SWS depending on developmental timing and affected lineages [2, 3]. Expressivity depends on malformation extent: number of affected lobes and bilateral involvement are associated with poorer cognitive outcome, and cutaneous CM extent with epilepsy and glaucoma surgery [23]. Comparing genes, GNAQ SWS tends to have homogeneous dark-red CM with soft tissue hypertrophy, whereas GNA11 SWS has reticulated CM darkening over time, naevus anaemicus (in 12 of 12 GNA11 R183C patients versus 4 of 20 with GNAQ R183, according to data cited by Zhang), ipsilateral segmental hypotrophy and milder neurological manifestations [5, 12]. Familial factors may modify the course: in a cohort of 277 patients, a family history of seizures was associated with earlier epilepsy onset and a family history of vascular malformations with more symptomatic strokes [23].

2.9. Allelic diseases and related entities

The GeneGuidance curation consulted links OMIM 163000 to ORPHA:3205 and OMIM 185300 to ORPHA:624, an inverted mapping; Orphanet confirms ORPHA:3205 as SWS (OMIM 185300) and ORPHA:624 as familial multiple nevi flammei (OMIM 163000), the mapping followed here [1, 8]. RASA1 does not cause SWS: no pathogenic variants were identified in nine Chinese patients with sporadic SWS, in whom only a synonymous variant was found, or in 37 patients in a European series [14, 47]. ClinGen has no actionability assertion for GNAQ [8].

3. Clinical Spectrum and Manifestations by System

3.1. Overview and data sources

SWS frequency data come predominantly from US tertiary centres (Kennedy Krieger Institute and Johns Hopkins, Baltimore), a London cohort (Great Ormond Street Hospital), the Italian consensus and Sturge-Weber Foundation surveys; all have referral bias towards more severe disease, and surveys have response bias [18, 22, 23, 48]. The following table presents the most frequently cited frequencies with their relevant populations.

The PNDS frequencies differ from the North American figures mainly for intellectual disability (32–45% versus about 60%) and for the proportion without PWB (15% versus about 10%); the differences reflect the period of the series, recruitment and outcome definitions (clinical diagnosis of disability versus standardised psychometric assessment), and none of the estimates derives from a population-based cohort [19, 21, 23].

3.2. Skin: the port-wine birthmark

PWB is a well-demarcated pink-to-red patch present at birth, unilateral, bilateral or midline, persisting lifelong and rarely increasing in extent [2, 17]. Over time it darkens to red-purple and, particularly on the midface, may develop vascular ectasia, thickening, soft tissue hypertrophy and proliferative nodules that bleed, cause discomfort and rarely become infected; histologically, nodules represent vascular ectasias, pyogenic granulomas or AVMs [17]. Progressive changes are rare before puberty [17]. Meyerson phenomenon (eczematisation of the lesion) may occur in childhood, particularly with atopic dermatitis [17]. CM affects up to 0.5% of newborns, and soft tissue and skeletal hypertrophy of the affected area is described in 65% by the fifth decade [22]. Cutaneous involvement may extend beyond the face to trunk and limbs, and non-facial capillary malformations are among occasional abnormalities in Smith's textbook [2, 6]. Approximately 10% of patients have isolated cerebral or ocular SWS without PWB [23].

The French PNDS considers frontal topography the decisive risk factor and specifies it by segment [19]. SWS PWB is usually unilateral, with no side predominance and a sharp midline boundary; outside the forehead, the risk of SWS is considered exceptional, and even with frontal involvement most patients have neither ocular nor cerebral disease [19]. Studies diverge between an increased risk for any frontal involvement and a risk concentrated in two topographies, the median frontal (segment S4, distinct from the transient naevus flammeus) and the extensive hemifacial, the lateral frontotemporal locations carrying the lowest risk [19]. A PWB limited to the lower eyelid may be accompanied by glaucoma without frontal involvement, and the risk of bilateral PWB is described as higher by some authors and as not increased by others [19]. Distant PWBs, on the trunk or limbs, add neither severity nor ocular or meningeal risk [19]. When the PWB extends to segment S3 it is often associated with capillary malformation of the lip and gingiva with gingival hypertrophy, and in adolescence and adulthood the lesion may be accompanied by soft-tissue hypertrophy and vascular nodules [19]. Figure 1 of the PNDS reproduces the PWB topography of 11 SWS patients from a series of trigeminal-territory PWBs (Dutkiewicz 2015) [19].

3.3. Nervous system: epilepsy

Seizures usually begin in the first year, most frequently between 2 and 7 months, and are often fever-triggered [2, 20]. In Valery's retrospective two-centre series, 90% of infants with cerebral SWS developed seizures before 2 years, and in Day's series 75% of children with cerebral SWS had seizures in the first year [49, 50]. Epilepsy begins earlier in bilateral disease and extensive unilateral hemispheric disease [21]. Focal clonic seizures, potentially generalising, are most common; apnoeic seizures, related to the perisylvian region and amygdala, are more frequent and more concerning in SWS than other epilepsies, and myoclonic-atonic, gelastic and asymmetric epileptic spasms have been described [21]. In infants, focal motor seizures may be subtle, and some present with early hand or gaze preference rather than overt seizures; infantile spasms and myoclonic epilepsy are rare [23]. Approximately 40% have seizure clusters after long seizure-free periods, and febrile illnesses frequently trigger status epilepticus, potentially non-convulsive [21]. Epilepsy becomes drug-resistant in approximately 20–25% [22]. Adult seizure onset occurs [6, 20].

The French PNDS adds semiological and evolutionary details [19]. Seizures are focal in at least two thirds of cases, ipsilateral to the pial angioma and with contralateral clinical expression; status epilepticus, present in about half of patients, may be the presenting event and leave a motor deficit (hemiplegia, hemiparesis or hemineglect), and fever triggers seizures in about 30% [19]. Although the age at onset coincides with that of West syndrome, infantile spasms are rare (1.25%), and presentation with West syndrome is exceptional (about 1%) [19]. Seizures cluster in bouts lasting hours to days separated by long seizure-free periods, semiology may change over time, and reflex seizures triggered by skin stimulation may occur, sometimes with sudden atonia; several authors describe stabilisation after the first decade [19]. The EEG is asymmetric in more than two thirds of patients (low voltage, slow waves or focal spikes), and the asymmetry may persist after stabilisation; longitudinal follow-up shows progressive enrichment in spikes without necessary worsening of the epilepsy, and a French pilot study (Bar 2018) suggests that spikes before the first seizure predict epilepsy [19]. Focal cortical dysplasias, mainly type IIIc, not identifiable on current MRI and possibly contributing to epileptogenesis, have been described in surgical specimens from refractory epilepsy; associated polymicrogyria is rare [19].

3.4. Nervous system: stroke-like episodes, headache and focal deficits

Stroke-like episodes are transient hemiparesis or visual field defects, difficult to distinguish from seizure activity and postictal paralysis, attributed to transient cortical ischaemia or recurrent thrombosis [21]. They are more frequent in young children and often follow falls with apparently minor head injuries [2]. In a retrospective cohort, 54.5% of transient hemiplegia episodes began with seizures, 18.1% after a blow or fall and 6% after brain surgery; diffusion-weighted imaging rarely shows acute infarction [23]. Status epilepticus is frequently associated with these episodes in young patients [23]. Intracranial haemorrhage is rare, and acute thalamic haemorrhage is described in few cases [4, 20]. Headache is the second most common neurological symptom, affecting 30–45%; migraine, under International Headache Society criteria, affects approximately 28%, generally beginning before 10 years; more than half have neurological deficits during attacks (focal weakness, visual loss, paraesthesias), and prolonged visual auras and hemiplegic attacks are frequent [21]. Ictal single-photon emission computed tomography (SPECT) showed occipital hyperperfusion followed by hypoperfusion [21]. Hemiparesis and hemianopia may develop progressively through chronic cortical injury [8, 16].

Stroke-like episodes, often accompanied by headache and also termed “hemiplegic migraines”, were measured in a retrospective cohort of 102 children and young people from a tertiary multidisciplinary clinic (2013–2016, mean age 10.9 years): 32 had transient episodes of hemiparesis, all with epilepsy, and 47 had permanent hemiparesis; recovery to prior function took a median of 24 hours (from 1 minute to 4392 hours, about 6 months), and all recovered fully within 6 months, seizures and head injuries being the associated triggers [51]. The PNDS, which relies mainly on this study, gives a duration of 1 minute to 3 months and an “average” resolution of 24 hours, values that do not correspond to the published maximum or measure of central tendency [19, 51]. According to the PNDS, the episodes occur at any age with two peaks (6 months to 5 years and 9 to 13 years), may recur throughout life and arise with or without seizures; mild head injuries and temperature changes are described triggers [19]. Headache without a stroke-like episode affects at least half of patients and may be associated with seizures or with worsening of hemiparesis [19]. No prognostic criteria for headache or stroke-like episodes have been identified [19]. Permanent motor deficit results from centroparietal lesions and may improve with rehabilitation when epilepsy is controlled; the risk factors are severe epilepsy, extent of the pial angioma, particularly to the frontal lobe, and extent of hypoperfusion on PET [19]. Homonymous hemianopia contralateral to the pial angioma is common, owing to the usual posterior location of the lesion, and is underdiagnosed in young children and in intellectual disability, despite aggravating disability [19].

3.5. Neurodevelopment, cognition and mental health

Intellectual disability affects approximately 60% and severe cognitive impairment approximately 33% [21]. In a sample of children and young adults, intellectual functioning was impaired in 32% and adaptive functioning in 58% [23]. In Bosnyák's longitudinal study (33 children with unilateral SWS, mean age 3.3 years, median follow-up 2 years, without surgery), intelligence quotient (IQ) trajectories varied; lower final IQ was associated with baseline EEG abnormalities, early seizure onset, high seizure frequency and early frontal involvement, and baseline EEG abnormalities were the independent predictor in multivariable analysis [52]. Cognitive outcome is associated with number of affected lobes and bilateral involvement, and patients without PWB have better outcomes and later seizures [23]. ADHD affects approximately 40% [21]. In the English series of 92 children, 24% had ASD, 45% social communication difficulties, 50% behavioural difficulties and 26% sleep problems, and ASD was more likely with bilateral involvement [23]. In a study of 16 patients aged 3 to 34 years, 31% had mood disorder, 25% disruptive behaviour disorder and 25% adjustment disorder; substance use disorder was identified in 67% of adults, and disruptive behaviour was associated with left frontal and parietal involvement [53]. Approximately half of adults have psychological dysfunction (depression, anxiety, low self-esteem, shame, emotional outbursts, isolation), and paranoid disorder and pseudodementia have been described [21]. In the Sturge-Weber Foundation Patient Engagement Network consultation, psychological support was considered the most important unmet need [21]. Visible PWB has its own psychosocial impact: isolation, teasing and intimacy avoidance in patients; time devoted to treatments, strangers' comments and parental guilt [20]. Children aged 7 to 16 with facial differences, including PWB, have reduced health-related quality of life [17].

The French PNDS emphasises the variability of the cognitive profile [19]. Older studies described intellectual disability in more than half of patients and severe psychomotor delay in half of bilateral cases; more recent data (Kavanaugh 2016) show a full-scale IQ of about 80, with one third below 75, one third normal and one third borderline [19]. In the survey of 52 adults aged 18 to 63 years recruited through the Sturge-Weber Foundation (Sujansky 1995), which the PNDS summarises as 45% in independent work and 50% living autonomously or with a family, employability was 46% in adults with seizures and 78% in those without, 39% were financially independent and 55% were or could be married; developmental delay affected 43% of those with seizures and none of the others [54]. In Kavanaugh's study, the values come from 29 participants with neuropsychological assessment, out of a base of 80 patients, with impaired adaptive functioning in 58% [55]. There is no specific cognitive profile, but attention disorders are common and probably underdiagnosed, praxis and visuospatial difficulties should be sought, and children with specific learning disorders without intellectual disability are at risk of not receiving support [19]. In bilateral disease, visual loss may be mixed, ocular and occipital cortical [19]. A study using targeted questionnaires (Sloneem 2022) identified autism spectrum disorder in at least 30%, sometimes masked by good social motivation, without confirmation by other groups [19]. Despite progressive atrophy, progressive cognitive decline has not been demonstrated in the absence of drug-resistant epilepsy; the determinants of deficit are early onset and severity of epilepsy, early status epilepticus, extent or bilaterality of the pial angioma and, for some authors, extent of the PWB [19].

In epilepsy in general, psychopathological disorders affect about 30% of patients with uncomplicated epilepsy and 60% when there is a CNS lesion; in SWS, mood, behavioural and adaptive functioning disorders accompany early and severe epilepsy and intellectual disability, are reportedly more frequent with left-hemisphere involvement, and quality of life relates to cognitive and epileptic severity [19]. In parents, the child's PWB causes worry about severity and guilt, with highly variable coping strategies whose intensity does not correlate with the extent or visibility of the lesion [19]. In children, stigmatisation reduces quality of life and requires vigilance for school bullying; in adults, quality of life measured by the Skindex-29 deteriorates with age and is associated with soft-tissue hypertrophy and depression [19].

3.6. Eye

Approximately half of patients have ocular abnormalities, generally ipsilateral to PWB: eyelid lesions, dilated conjunctival and episcleral vessels, abnormalities of the anterior chamber, cornea, choroid and retina; in dark-skinned children, conjunctival redness may be the only sign [20]. Glaucoma affects 30–70% and has a bimodal distribution: approximately 60% develop in early childhood (up to 2–3 years) through angle dysgenesis, with buphthalmos, corneal oedema, pain, photophobia or simply globe enlargement; approximately 40% develop later in childhood, adolescence or adulthood as secondary open-angle glaucoma from elevated episcleral venous pressure, with onset reported up to 41 years [20, 22, 38]. According to Smith's textbook, glaucoma develops before 2 years when tissues destined to form the anterior chamber angle are affected and often only after 5 years when involvement is limited to conjunctival and episcleral vascular tissues [2]. SWS glaucoma eyes have thicker corneas, less oedema and fewer Haab striae and lower IOP than primary congenital glaucoma, although one series found corneal oedema at diagnosis in 75%; dilated tortuous spiral conjunctival and scleral vessels are considered pathognomonic of the episcleral venous pressure gradient [38]. Glaucoma is rarely bilateral, even with unilateral PWB; in a study of 100 children with PWB, 72% with both eyelids involved and 21% with only upper eyelid involvement developed glaucoma [20, 38]. In Aggarwal's series (55 patients, 1978–2003), 44 (80%) had glaucoma, gonioscopy showed distinct angle abnormalities in 55% of glaucomatous eyes, and iris heterochromia, with a darker iris on the glaucoma side, occurred in 37% with unilateral glaucoma and none without glaucoma [56]. Choroidal haemangioma is typically unilateral, bilateral in 10–30%, appearing as a localised orange lesion or diffuse “tomato ketchup” fundus; complications include subretinal haemorrhage, serous retinal detachment, cystoid macular oedema and neuroepithelial detachment [20, 22]. In an Indian series of 86 patients (62 SWS and 24 pigmentovascular phacomatosis) followed over 27 years, glaucoma affected 49.2% of SWS eyes and 91.6% of pigmentovascular phacomatosis eyes; visual field defects were glaucomatous in 35.5%, neurological in 7.7% and combined in 4.1%, and eyes receiving external beam radiotherapy for diffuse choroidal haemangioma developed localised defects mimicking glaucoma and hemianopia [57]. In an optical coherence tomography (OCT) angiography study with 15 patients and 15 controls, deep retinal capillary plexus density and vascular diameter were increased [58]. Other reported ocular abnormalities include iris colobomas, retinal vascular tortuosity, retinal detachment, strabismus, lens dislocation, optic atrophy and buphthalmos [1, 2].

The French PNDS attributes ocular disease mainly to intraorbital venous hypertension, which increases secretion and reduces resorption of aqueous humour, in addition to histological abnormalities of the angle; the minimal sign is scleral hyperaemia due to dilatation of the episcleral vessels, and any episcleral dilatation amounts to ocular involvement [19]. It estimates the prevalence of glaucoma at 30–60%, with onset in about half of cases in the first years of life, in one third during childhood and in 20% after the age of 20, and with a higher risk when the PWB affects the eyelids, particularly the lower eyelid [19]. In infants, glaucoma presents as congenital glaucoma (buphthalmos, photophobia, lacrimation, corneal clouding up to the “porcelain cornea”, synonymous with permanent blindness); after early childhood, with an inextensible globe and cornea, it is asymptomatic until the stage of blindness and only screening detects it, as cupped optic neuropathy with raised intraocular pressure, at any age and even without prior glaucoma [19]. The choroidal “haemangioma” corresponds to diffuse thickening of the choroid, quantifiable by ultrasound or by spectral-domain OCT with enhanced depth imaging (EDI), with thickened and tortuous vessels carrying GNAQ variants; it is early and frequent (about 40%) and may never have consequences or may decompensate, particularly from puberty, into exudative chorioretinopathy with serous retinal detachment, field defect and loss of acuity when it reaches the macula [19]. When involvement is strictly unilateral, the risk of visual disability is very low, because the healthy eye preserves function; the risk exists with bilateral involvement, particularly with bilateral congenital glaucoma, and is compounded by cortical hemianopia [19].

3.7. Endocrine system and growth

Growth hormone (GH) deficiency and central hypothyroidism are recognised in SWS [2, 8]. In a registry of 1653 patients, 19 with suspected GH deficiency were reviewed and deficiency confirmed in 9; the Italian consensus cites an 18-fold relative risk in paediatric patients based on this series [22, 59]. In a multidisciplinary centre, 2 of 83 children (2.4%) with cerebral involvement had central hypothyroidism, above the general population proportion; antiseizure medicines also alter thyroid tests [60]. These abnormalities generally occur without specific hypothalamic-pituitary MRI abnormalities, as in a 20-year-old man with SWS, epilepsy and Klippel-Trenaunay who had hypogonadotrophic hypogonadism, GH deficiency and central hypothyroidism [22, 61]. Carbamazepine and oxcarbazepine may be associated with thyroid insufficiency, aggravating hypothalamic-pituitary dysfunction [22]. In a series of 58 patients, 66% had below-normal vitamin D, more frequently African Americans and those with more severe disease; correction was associated with improved hemiparesis, without established causality [23].

3.8. Oral cavity and craniofacial skeleton

Oral manifestations affect 40–50% of patients: vascular mucosal and gingival lesions, gingival hyperplasia (aggravated by antiseizure medicines), pyogenic granulomas, nodules that may cover the dental crowns, spontaneous or chewing-related bleeding, altered tooth eruption and macrodontia [22]. Asymmetric skeletal growth ipsilateral to the CM, with maxillary overgrowth, causes facial asymmetry, occlusal plane canting and malocclusion [2, 22]. The Italian consensus illustrates these findings with an adolescent girl with hemifacial hypertrophy and an 8-year-old child with asymmetric replacement of the deciduous dentition and asymmetry of the frontal sinus, petrous bone and mandible [22]. The French PNDS adds that oral and perioral capillary malformations may affect the lips, gingiva, buccal mucosa, palate, floor of the mouth, tongue, pharynx and larynx, with a red-violaceous colour; gingival hyperplasia is soft, blanches on pressure and ranges from slight thickening to overgrowth that prevents mouth closure and may cause tooth loss [19]. It also describes unilateral bone hypertrophy, asymmetric mandibular growth, botryomycomas, early or delayed tooth eruption, macroglossia and macrocheilia (with malocclusion), periodontitis and pulp lesions; intellectual disability hinders hygiene, and plaque and calculus cause gingivitis that aggravates hyperplasia, while some antiseizure medicines outside first and second line induce gingival hyperplasia and xerostomia [19].

3.9. Other manifestations

Smith's textbook lists occasional abnormalities including microgyria, macrocephaly, hydrocephalus, intracranial haemorrhage, aortic coarctation, enlargement of the pinna when involved by malformation, ear, nose and throat infections, obstructive sleep apnoea and macrodactyly [2]. Sudarsanam's review reports increased prevalence of ear, nose and throat, endocrine and emotional-behavioural problems [16]. Zhang and colleagues cite early-onset hypertension in 17% of patients with GNAQ or GNA11 CM in the group's previous work [12]. The Orphanet table, based on reviews, classifies manifestations by frequency:

Source: [1]. Orphanet classification of choroidal haemangioma and brain calcification as occasional contrasts with clinical series, where cortical calcification is typical of established brain disease and choroidal haemangioma a principal ocular component [2, 22]. HPO “haemangioma” terms correspond to capillary malformations in current terminology.

3.10. Gene-specific and Roach-type forms

GNA11-associated SWS has pink, patch-like CM at birth, extensive across the face, trunk and limbs, darkening to purple, with a reticulated pattern, naevus anaemicus, limb hypertrophy or hypotrophy, glaucoma (2 of 3 patients in Dompmartin's series; 29% in Zhang's, all with bilateral facial CM) and neurological symptoms with moderate cerebral atrophy; in Zhang's series, 29% had psychomotor delay and 6 of 7 patients with facial CM who underwent MRI had abnormalities (engorged choroid plexus in 3, pontine cavernoma, white matter loss, FLAIR/T2 hyperintensities and leptomeningeal angiomatosis in one patient with episodes of left-sided weakness from the age of 3) [5, 12]. Macrocephaly was observed in 3 of 8 patients with GNA11 and facial CM [12]. The French PNDS, based on a 2020 study (Jordan), states that there is no difference in cerebral or ocular phenotype between GNAQ and GNA11 forms; the later Brussels, Nijmegen and Zhang series, by contrast, describe a generally milder neurological picture with GNA11, and the difference may reflect the small number of patients and selection by skin phenotype [5, 12, 19]. The GNB2 patient had a classic presentation: PWB of the eyelid, nose and eyebrow, mild intellectual disability, refractory epilepsy and leptomeningeal malformation with atrophy and gyral calcifications [6]. Roach type 3 SWS, without PWB or ocular involvement, presents with focal seizures, migraine-like headache, progressive hemiparesis, cognitive changes or transient homonymous hemianopia, and diagnosis tends to be delayed [7]. The US consensus review identified 31 cases of SWS diagnosed in adulthood, with unknown incidence, such as that of a 55-year-old woman without PWB with episodes of transient right-sided weakness [20].

3.11. Natural history and treated disease

Brain disease is typically progressive in the first years after symptom onset: leptomeningeal enhancement generally remains stable, whereas atrophy and calcifications progress predominantly during early symptomatic years; progressive atrophy is also described in adults, for example associated with new migraines [20]. Brain calcification is not evident radiographically before the end of the first year, and EEG tends to become more abnormal with time [2, 22]. In the 277-patient cohort, early seizure onset was associated with intellectual disability, learning disorder, stroke-like episodes, symptomatic stroke, hemiparesis, visual field deficit and brain surgery, and bilateral involvement with learning disorder and intellectual disability [23]. These data describe patients receiving antiseizure medicines, frequently ASA and sometimes surgery; no untreated natural history series exist, and presymptomatic treatment series (section 7) show that interventions influence some progression [49, 62]. One patient with small right parieto-occipital involvement, receiving no medication, remained seizure-free after 4 years of follow-up, illustrating limited-disease variability [50].

4. Clinical Recognition and Warning Signs

4.1. Newborn with facial port-wine birthmark

Facial PWB is present at birth and usually initiates SWS recognition, but most facial PWBs are unaccompanied by ocular or brain disease [2, 17]. Facial PWB occurs in approximately 1 in 300 newborns and SWS in 1 in 20,000 to 50,000 live births, making the clinical task identification of risky birthmarks among many [18]. Data cited in the founding study assign a child with facial PWB approximately 6% probability of SWS, rising to 26% when the birthmark occupies the ophthalmic trigeminal division (V1) territory [3]. Older studies estimated 7–28% risk for “V1” PWB, whereas newer studies shifted attention from nerve territory to embryonic vascular territory: the best predictor is PWB involving any part of the forehead, including upper eyelid and midline frontonasal prominence, and risk increases further with hemifacial, midline or bilateral PWB or simultaneous forehead, cheek and jaw involvement; even a small midline forehead birthmark may accompany severe neurological disease [17, 20]. In the London cohort of 192 children with facial PWB, trigeminal branch distribution was not associated with abnormal MRI; central and lateral forehead were the only independently significant territories, and only children with “forehead” PWB (midline to a line between outer canthus and top of ear, including upper eyelid) had seizures, neurodevelopmental abnormalities, glaucoma or abnormal MRI [18]. In that cohort, 83 of 103 children with forehead involvement had SWS; this proportion reflects a tertiary centre and cannot be extrapolated to the general population [18]. A US review places cerebral SWS probability with forehead, eyelid or temple PWB at 20% to 50%, depending on size, location and extent [23].

Consensuses agree that facial PWB should be assessed at birth and high-risk PWB requires early ophthalmological examination to exclude congenital glaucoma and paediatric neurology referral [17, 20, 22]. Ophthalmological examination should occur in the first months without awaiting glaucoma signs (increased corneal diameter, tearing, photophobia, cloudy cornea), and the 2026 US glaucoma consensus recommends assessing all infants with facial PWB within the first month [20, 63]. The 2018 consensus recommends ophthalmological examination for all newborns with eyelid PWB [21]. In a study cited by Wen, 72% of children with both eyelids affected and 21% with only upper eyelid PWB developed glaucoma [38].

PWB differs from two common newborn vascular lesions. Naevus simplex (“salmon patch”), present in approximately 40% of newborns, affects several sites (upper eyelids, forehead, nape, vertex, back), has poorly defined borders, is symmetric and fades from the first weeks [18, 22]. Segmental infantile haemangioma is a tumour proliferating after birth, whereas PWB is present at birth and stable; segmental facial haemangioma may require PHACE syndrome assessment (section 6) [17, 33]. Early recognition also has therapeutic value because laser may begin at an age when treatment is more effective [17].

The French PNDS (French National Diagnosis and Care Protocol) identifies the neonatologist, obstetrician, midwife and paediatrician as first-line professionals and sets the operational rule: any frontofacial PWB of segments S1 or S4 (the territory classically designated V1) prompts investigation of neurological and ophthalmological involvement, with referral to neuroradiology, paediatric neurology, dermatology and ophthalmology with experience of the disease [19]. The physiological transient centrofaciofrontal patch of infancy is more median, not lateralised, often affects the upper eyelids, philtrum and upper lip, and is paler; neonatal polycythaemia may hinder the distinction, in which case observing its evolution is sufficient, without investigations and without alarming the parents [19]. Initial information should be cautious and supported by experts, particularly when the craniofacial PWB does not affect the forehead [19]. Other routes to diagnosis are seizures, in a child whose PWB went unnoticed or who has none, and, exceptionally, the discovery of a choroidal angioma at ophthalmological examination; late diagnosis corresponds to atypical forms or to insufficient initial assessment [19].

4.2. Infants and young children: neurological warning signs

Seizures usually begin in the first year, most frequently between 2 and 7 months, with fever a frequent trigger [2, 21]. Infant focal motor seizures may be subtle, and some infants show only early hand or gaze preference [23]. The Italian consensus lists warning signs parents should recognise: visual field defects, early hand preference, automatisms, nystagmoid movements and developmental delay or stagnation [22]. Apnoeic seizures are more frequent and concerning in SWS than other epilepsies, and approximately 40% have seizure clusters after long seizure-free periods [21]. Transient hemiparesis or visual field loss, often following falls with apparently minor head injuries, indicates brain involvement [2, 21].

In an infant with high-risk PWB, EEG amplitude asymmetry, focal slowing, epileptiform activity or subclinical seizures points to brain involvement, although EEG is not diagnostic [20, 22]. In Valery's two-centre series, slowing, epileptiform discharges or electrographic seizures on EEG before the first clinical seizure were associated with seizure onset by 2 years (p=0.001) [49].

4.3. Emergency signs

The following table presents situations requiring urgent assessment, with consensus-recommended management.

In France, reference centres give the patient an SWS-specific emergency card (ORPHA:3205), drawn up by the DéfiScience filière with reference centres and associations, which the patient carries: it lists the associated disorders (epilepsy, glaucoma, angioma, visual disturbance, motor deficit), the risks (acute motor deficit, blindness, status epilepticus, migraine attacks), current treatment with dosage, difficulty in expressing pain, mode of communication and priority contacts [19, 64]. The PNDS also recommends an information sheet for the attending physician and emergency services on stroke-like episodes [19].

In a retrospective study of 35 patients with acute neurological symptoms presenting to emergency services, with 89 urgent imaging studies, no acute haemorrhagic or ischaemic stroke was identified [20]. Cerebral angiography has little value and substantial risk, and thrombolytics have not been studied in SWS [20].

4.4. Adolescents and adults; SWS without port-wine birthmark

Glaucoma may first appear in adolescence or adulthood, with onset reported up to 41 years, which requires ophthalmological surveillance that does not end in childhood [20]. Adults with facial PWB, with or without glaucoma, who have never undergone neuroimaging should do so [20]. SWS without PWB (Roach type 3) accounts for about 10% of patients with cerebral or ocular SWS and presents with focal seizures, migraine-like headache, hemiparesis or transient hemianopia; without the skin marker, diagnosis is often delayed and depends on neuroimaging [7, 23]. The US consensus review identified 31 cases diagnosed in adulthood [20]. The French PNDS estimates SWS without PWB at 15% and recommends, when MRI performed for epilepsy shows a pial angioma without a skin lesion, a particularly rigorous analysis of the imaging criteria and a search for choroidal angioma [19].

4.5. When to suspect an atypical form

Extensive body and facial CM that is reticulated and poorly demarcated, naevus anaemicus, limb hypertrophy or hypotrophy, macrocephaly, frontal bossing, telangiectasias, absence of neurological signs and family history of vascular malformations should raise suspicion of GNA11-associated SWS or another diagnosis [5, 12, 45]. Dermal melanocytosis with PWB suggests pigmentovascular phacomatosis [3, 12]. Early-onset hypertension has been observed with GNAQ or GNA11 CM, justifying blood pressure measurement (section 11) [12].

4.6. Newborn screening

No biochemical or genetic newborn screening exists for SWS: the variant is somatic, restricted to affected tissues and not reliably detected in blood [8]. Screening is clinical, inspecting newborn skin with the referral pathway described for high-risk PWB [17, 22].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Diagnostic criteria and Roach classification

There are no formally validated SWS diagnostic criteria. Consensuses define disease by at least two of three components (facial PWB, cerebral leptomeningeal vascular malformation, ocular vascular malformation), and intracranial leptomeningeal angiomatosis, generally ipsilateral to PWB, is considered the pathognomonic neuroradiological finding establishing clinical diagnosis [8, 17, 22]. Roach classification, used predominantly to describe incomplete forms, distinguishes three types [7, 35]:

Later articles attribute the classification to Roach; the abstract of the 1992 article does not describe it, and the formulation used here is that of Cheema and colleagues [7, 35]. The French PNDS (French National Diagnosis and Care Protocol) considers it of little practical value but useful for raising the diagnosis when the triad is incomplete, and bases confirmation on identification of the PWB in the V1/S1–S4 territory, ocular involvement and leptomeningeal angioma; when the imaging criteria are present and confirmed by an experienced neuroradiologist, there is no differential diagnosis on MRI [19]. The diagnosis of PWB is exclusively clinical, with no biopsy or other investigation [19]. The International Society for the Study of Vascular Anomalies (ISSVA) classification includes SWS among capillary malformations with central nervous system and/or ocular abnormalities [21].

5.2. Brain magnetic resonance imaging

Gadolinium-enhanced MRI is the reference examination. Characteristic findings are leptomeningeal pial angiomatosis (postcontrast T1 and FLAIR, fluid-attenuated inversion recovery, enhancement), ipsilateral choroid plexus enlargement and enhancement, dilated deep medullary and subependymal veins, cortical atrophy and calcifications [21, 23]. Susceptibility-weighted imaging (SWI) detects calcifications and deep medullary veins and may be positive before enhancement: in the US consensus example, a 4-month-old had abnormal SWI and leptomeningeal enhancement appeared only 8 months later, after the first seizure [20]. Diffusion-weighted imaging (DWI) shows ischaemic lesions and sometimes transient postictal effects [20, 21]. MRI at 3 T is recommended, with 1.5 T acceptable, at specialised paediatric centres [21]. The following tables summarise sequence-specific findings and phase-specific sequences according to the 2018 consensus [21].

Italian protocols additionally include venous MR angiography and high-resolution orbital imaging at 3 T, and three-plane post-gadolinium T1 with FLAIR [22]. Without contrast, indirect involvement signs are lateral ventricular choroid plexus hypertrophy (above 5.6 mm) or asymmetry, early atrophy and white and grey matter abnormalities, present in a high proportion of SWS newborns [22].

The main MRI limitation is false negativity in the first year owing to incomplete myelination. False negatives range from 3% to 23% in retrospective studies; in a series of 32 children with high-risk PWB, screening MRI sensitivity was 25%, specificity 100%, positive predictive value 100% and negative predictive value 77% [20]. Normal neurological examination, absence of seizures and negative contrast-enhanced MRI after 1 year generally exclude brain involvement [22, 23]. Before that age, a negative MRI in an asymptomatic infant does not exclude disease [2, 20].

To reduce sedation requirements, consensuses allow rapid non-sedated non-contrast MRI using feed-and-wrap or low-dose oral midazolam during the first weeks, at tertiary centres while accepting false negatives [20, 22]. In 2026, a University of Wisconsin-Madison group described a targeted abbreviated protocol (Focused Abbreviated Survey Techniques, FAST) with this technique, without sedation, and reported in institutional experience that presymptomatic MRI facilitated referral, parental education and early ASA initiation; the abstract provides no diagnostic performance data [65].

The French PNDS describes its own neonatal protocol and semiology, drafted with paediatric neuroradiology teams from Lyon and Paris [19]:

On the basis of expert opinion, the PNDS advises early MRI in the first month for any S1–S4 frontofacial PWB, combined with EEG, and proposes repeat MRI at about 1 year if the first is normal, to decide whether neurological surveillance should continue; neurological signs, EEG abnormalities or epilepsy are a formal indication to repeat imaging [19]. The document acknowledges the controversy: presymptomatic diagnosis allows parental education, conditional rescue prescription and discussion of preventive treatment, but requires resources, may require anaesthesia and gadolinium, has false negatives, and no study has validated the benefit of treatment before the first seizure [19].

5.3. Computed tomography, nuclear medicine and electroencephalography

Computed tomography (CT) is not routinely indicated owing to poor sensitivity and radiation; it depicts gyral “tram-track” calcifications well and is the reference examination for cortical bone in symptomatic craniofacial overgrowth, and may be used when MRI is unavailable or in emergencies with trauma or haemorrhage [20–22]. Fluorodeoxyglucose PET shows hypometabolism (sometimes early interictal hypermetabolism), and coregistered subtraction ictal and interictal SPECT is useful with discordant presurgical data [21].

EEG is recommended at baseline in children with high-risk PWB, during sleep, ideally recording 30 minutes with at least 8 channels and polygraphic channels, preferably in the first 3 months of life [22]. Quantitative EEG added about 40% diagnostic information to the risk of brain involvement estimated from PWB size and extent [23]. Prolonged video-EEG is indicated for episodic changes of uncertain nature [21]. The French PNDS recommends combining EEG with early MRI, because background rhythm asymmetry strengthens the diagnosis (Zallmann 2018) and an early spike focus has prognostic value for the decision on prophylactic treatment; in a patient with seizures, EEG looks for low voltage, a slow-wave focus and a spike focus, and in drug-resistant epilepsy prolonged EEG serves to record seizures and guide surgical discussion [19].

5.4. Ophthalmological assessment

Baseline assessment includes visual acuity, IOP and dilated fundoscopy to look for choroidal haemangioma and assess the optic nerve [20]. IOP may be measured with rebound tonometry, sometimes without anaesthesia, but where suspicion is high and IOP cannot be measured, examination under sedation or inhalational anaesthesia is performed, with measurement of transverse corneal diameter and axial length; Goldmann applanation tonometry is the reference standard in children [13, 20, 22]. Besides IOP, glaucoma parameters include change in corneal diameter, axial length and cup-to-disc ratio, and suspicious symptoms include blepharospasm, lacrimation and photophobia [22]. For choroidal haemangioma, optical coherence tomography (OCT) is the technique of choice, complemented by indocyanine green angiography, ultrasonography (high internal reflectivity and diffuse choroidal thickening) and gadolinium-enhanced MRI, which shows a crescent-shaped lesion thicker at the posterior pole [22]. In adults, visual fields and optic nerve OCT are added [22]. Interpretation of visual fields requires distinguishing glaucomatous from neurological defects: in an Indian series of 86 patients with SWS or pigmentovascular phacomatosis, 35.5% of eyes had glaucomatous defects, 7.7% neurological and 4.1% combined, and eyes treated with external beam radiotherapy for diffuse choroidal haemangioma developed defects simulating glaucoma and homonymous hemianopia [57]. The French PNDS recommends a specialist ophthalmology consultation in the first month of life, possibly under general anaesthesia, and a specialist consultation at verbal age for retinal photographs and EDI-mode OCT, which show choroidal thickening on the side of the PWB as subclinical ocular involvement [19]. In the treated child, as soon as automated visual fields and/or OCT (retinal nerve fibre layer and ganglion cell complex) can be obtained, the decision to start or intensify treatment should be based on these investigations rather than mainly on IOP, even when corrected for pachymetry [19].

5.5. Molecular testing

Molecular testing is not needed for the diagnosis of typical SWS, and the most experienced US centre does not perform it routinely [8, 23]. It is indicated in atypical forms, where it may change diagnosis, prognosis, care and eligibility for clinical trials: among 12 patients assessed for atypical SWS, 5 had a pathogenic somatic GNAQ or GNA11 variant, 1 a likely pathogenic mosaic PIK3CA variant, 1 a mosaic PTPRD deletion and 5 germline RASA1, EPHB4 or KIT variants [45]. Broad germline and somatic investigation is recommended in these patients [45]. The French PNDS is more restrictive: testing for the mosaic variant in skin has no diagnostic, prognostic (it does not predict cerebral or ocular involvement) or therapeutic value, as there is no targeted treatment, unlike PROS, and in most cases it is not needed [19]. When performed, it uses targeted high-throughput sequencing of a PWB skin biopsy; the variant is always absent from blood, the yield in saliva or buccal swab is unknown and the VAF is generally below 10% [19]. The French and North American recommendations converge for typical SWS and differ for atypical forms, which the PNDS does not single out.

The appropriate sample is affected tissue: PWB skin biopsy or resected brain tissue [8, 23]. Blood is unreliable because the variant is confined to affected tissues and, even there, VAF is low (1.0–18.1% in the founding study) [3, 8]. The variant is enriched in endothelial cells, and microdissected or cultured endothelium may reach 20–30% VAF whereas keratinocytes and fibroblasts are negative [6]. Methods used include high-depth next-generation sequencing (NGS), amplicon sequencing and droplet digital PCR (ddPCR) [3, 4].

Test limitations should be included in reporting and interpretation. Minimum detection thresholds are approximately 1% for NGS, 0.25% for ddPCR and 0.1% for peptide nucleic acid ddPCR, and very low-level mosaicism may yield false negatives: in one SWS case, the GNA11 variant was present in brain at 0.22–0.58% VAF, below the panel's 1% threshold but above background in 10 controls (0.02–0.12%), and was considered pathogenic [4]. In the founding study, 3 of 18 SWS brains were negative and approximately 12% of patients had no variant detected in any tissue [3]. Sensitivity therefore depends on tissue, endothelial cell proportion, read depth and method; a negative result does not exclude SWS, whose diagnosis remains clinical and radiological [4, 8]. Panels targeting GNAQ exon 4 do not detect GNA11 or GNB2, and the GNB2 variant was identified only through deep exome sequencing (approximately 400×) of lesional dermis [3, 6].

The report should specify gene, transcript version (NM_002072.5 for GNAQ, NM_002067.5 for GNA11, NM_005273.4 for GNB2), HGVS description, VAF, tissue and cell type analysed, read depth and detection threshold, and somatic mosaic nature [4, 11]. American College of Medical Genetics and Genomics and Association for Molecular Pathology (ACMG/AMP) classes were designed for germline variants; germline ClinVar classifications apply poorly to somatic variants and cannot estimate recurrence risk [43, 44]. A novel low-VAF mosaic variant found in one patient, such as GNB2, requires replication before being considered an established cause [6].

5.6. Diagnostic algorithm

The following algorithm summarises recommendations; consensus differences are described in section 7.

5.7. Prenatal diagnosis and preimplantation testing

Because this is postzygotic somatic mosaicism with no described transmission, there is no indication for prenatal or preimplantation genetic diagnosis [1, 8]. The sources consulted describe no prenatal ultrasound criteria for SWS; the report of a premature newborn with severe bilateral SWS expressly notes the absence of prenatal diagnosis [66]. The French PNDS summarises the limits: PWB is not visible on routine ultrasound, no changes in combined screening serum markers have been described, leptomeningeal angiomatosis is not visible on fetal MRI, and molecular testing, which requires affected tissue, cannot be performed on amniocytes [19]. The published exception is a case suspected prenatally because of polymicrogyria and white matter signal change limited to one hemisphere, with postnatal confirmation (Cagneaux 2013) [19].

6. Differential Diagnosis

6.1. Comparative table

6.2. Non-syndromic port-wine birthmark, naevus simplex and infantile haemangioma

Non-syndromic PWB shares the GNAQ p.(Arg183Gln) variant with SWS, detected in 12 of 13 lesions studied in the founding article, and is distinguished by the absence of ocular and cerebral involvement; the difference lies in the timing of the postzygotic event and the lineages affected, not in genotype [2, 3]. Molecular testing therefore does not separate isolated PWB from SWS, and the distinction relies on ophthalmological examination, neurological assessment and imaging [8, 20]. Naevus simplex is transient and symmetrical, whereas segmental infantile haemangioma is a proliferative tumour that is absent, or subtle, at birth [17, 18]. The French PNDS specifies the distinguishing criteria: the physiological frontal naevus flammeus, sometimes very extensive, is more median and regresses within a few months, before 18 months; segmental infantile haemangioma at the initial stage has a halo of vasoconstriction at birth and rapidly becomes raised, although abortive forms grow little in thickness, and on the forehead (S1 or S4), particularly in a female newborn, it may be part of PHACE syndrome, whose cerebral vascular disease differs from that of SWS and whose haemangioma responds to propranolol, unlike PWB [19]. In megalencephaly-capillary malformation (MCAP), the PWB is mediofacial (philtrum, upper lip), there are distant reticulated CMs (cutis marmorata) and macrocephaly above +3 standard deviations, with Chiari malformation, developmental delay or epilepsy [19]. Familial occurrence of craniofacial PWB, even frontal, points to CM-AVM syndrome [19]. Other congenital glaucomas are easily excluded by the skin lesion and episcleral venous dilatation [19].

6.3. PHACE syndrome

PHACE syndrome combines segmental infantile haemangioma with posterior fossa malformations, cervical and cerebral arterial lesions, cardiac abnormalities and aortic coarctation, and ocular abnormalities, sometimes with ventral midline defects [33]. In a prospective study, 31% of infants with facial infantile haemangioma of at least 22 cm² (approximately 5 × 4.5 cm) met criteria [33]. The 2009 criteria defined “definite PHACE” as characteristic segmental haemangioma or facial or scalp haemangioma exceeding 5 cm with one major or two minor criteria, and “possible PHACE” as less complete combinations [67]. The revised 2016 criteria retain facial or scalp haemangioma exceeding 5 cm plus one major criterion as “definite PHACE” and add large segmental haemangioma of the neck, upper trunk or trunk with proximal upper limb plus two major criteria [33]. Aortic, cervical and cerebral arterial vascular abnormalities are the most common extracutaneous finding, congenital heart disease occurs in 41–67% and coarctation in 19–30% [33]. Assessment includes full eye and midline examination, echocardiography in all and head and neck MRI with MR angiography, with ischaemic stroke risk stratification [33]. Propranolol, the only US Food and Drug Administration (FDA)-approved drug for problematic infantile haemangiomas, raises concerns in PHACE with arteriopathy, and many children receive daily ASA with unknown long-term impact [33]. SWS and PHACE share segmental facial distribution, which Waelchli relates to the same embryonic territories, and endocrine abnormalities such as hypothyroidism and GH deficiency; the decisive difference is skin lesion nature, a stable malformation in SWS and proliferative tumour in PHACE [18, 33].

6.4. GNA11 forms and pigmentovascular phacomatosis

GNA11 p.(Arg183Cys) CM is reticulated, poorly demarcated, pink-red, extensive (median body surface area 35% in Zhang's series) and bilateral in 71% of patients; naevus anaemicus and ipsilateral segmental hypotrophy are more characteristic of GNA11 than of GNAQ, and limb growth discrepancy occurred in 82% of patients [12]. When there is facial CM, GNA11 disease may include glaucoma, MRI abnormalities and leptomeningeal angiomatosis, with a generally milder and later neurological picture than in GNAQ SWS [5, 12]. The presence of dermal melanocytosis with the PWB defines pigmentovascular phacomatosis, which is also associated with GNAQ and GNA11 and may coexist with SWS; in the Indian series, glaucoma affected 91.6% of eyes with pigmentovascular phacomatosis, compared with 49.2% of eyes with SWS [3, 57]. Coexisting blue naevus and PWB may indicate an increased, albeit rare, risk of uveal melanoma [3]. Pigmentovascular phacomatosis enters the differential diagnosis.

6.5. Other differential diagnoses

CM-AVM syndrome (capillary malformation–arteriovenous malformation) due to RASA1 or EPHB4 is autosomal dominant and presents with multiple small CMs with a pale halo and fast-flow AVMs or arteriovenous fistulas in the face, brain, soft tissues or bone; Parkes Weber syndrome combines an extensive CM with arteriovenous microfistulas and limb hypertrophy [14, 68]. In the differential table of the CM-AVM GeneReviews, SWS shares the intracranial vascular abnormality and is distinguished by segmental facial PWB, seizures and glaucoma [14]. No pathogenic RASA1 variants were identified in 9 Chinese patients with sporadic SWS or in 37 patients in another series [14, 47]. Klippel-Trenaunay syndrome, incorporated into PROS (PIK3CA-related overgrowth spectrum), combines low-flow capillary, venous and lymphatic malformations with limb hypertrophy and may include glaucoma, heterochromia, intracranial calcifications, leptomeningeal enhancement, hemimegalencephaly and seizures, creating overlap with SWS; the absence of arteriovenous fistulas distinguishes it from Parkes Weber syndrome, with which it is frequently confused [68, 69]. The Brazilian textbook Síndromes Genéticas: Reconhecer para Tratar includes SWS in the differential diagnosis of Klippel-Trenaunay-Weber syndrome, describing it as a congenital phacomatosis with predominantly cranial cutaneous, meningeal and ocular involvement [32]. Isolated cerebral AVM is also part of the differential diagnosis.

6.6. Other causes of childhood glaucoma and cortical lesions

Secondary childhood glaucoma includes, besides SWS, PAX6 aniridia, Axenfeld-Rieger anomaly, Peters anomaly and NF1 with plexiform neurofibroma; primary congenital glaucoma birth incidence is approximately 3/100 000 [13]. In the polymicrogyria investigation algorithm, SWS appears among non-genetic or mosaic causes alongside congenital infection and hypoperfusion, and in International League Against Epilepsy (ILAE) epilepsy classification it is a structural-metabolic cause among neurocutaneous syndromes [71, 72]. In type 3 SWS, headaches with prolonged visual aura and transient deficits may be mistaken for primary migraine, and postictal Todd paralysis lasts less than 24 hours [7, 23].

7. Treatments and Risk Management Strategies

7.1. Principles and nature of the evidence

There is no cure or authorised targeted treatment for SWS; management is symptomatic and organ-directed, and early recognition with active symptomatic treatment is its foundation [8]. No randomised trial supports the usual neurological measures; recommendations derive from expert consensuses (Sturge-Weber Foundation in 2018 and 2021, Italian Delphi consensus of 2025, North American glaucoma consensus of 2026), none of them with formal grading of the evidence, and from retrospective tertiary-centre series [8, 20–22, 63]. The only guideline using GRADE (Grading of Recommendations Assessment, Development and Evaluation) methodology that addresses SWS is the 2026 American Epilepsy Society (AES) guideline on infant epilepsy [73]. The French PNDS (French National Diagnosis and Care Protocol) of the HAS, published in February 2024 from a text of July 2023, follows the 2012 HAS method for rare disease protocols, with a PubMed review from 1981 to 2022 and a multidisciplinary drafting and reading group, and classifies the recommendations on early MRI, EEG, antiseizure prophylaxis and ASA as expert opinion [19]. Follow-up should take place at a reference centre with neurology, ophthalmology, dermatology, dentistry and paediatrics, in coordination with the primary care paediatrician, who monitors growth, neurodevelopment and signs of endocrine dysfunction [22].

7.2. Epilepsy: action plan, rescue and antiseizure medicines

All patients with SWS and epilepsy should have an individualised seizure action plan, and families should be taught to use rescue treatment for prolonged or clustered seizures and to manage fever and dehydration [21, 22]. Home benzodiazepines (buccal or intranasal midazolam, rectal diazepam) are first line for acute, prolonged and clustered seizures [22]. In patients with a history of clusters or status epilepticus, intermittent benzodiazepine during febrile illnesses may be considered, in a strategy similar to that used in Dravet syndrome, and EEG should be performed for fever with altered consciousness [21]. The consensuses do not specify doses of antiseizure medicines or rescue benzodiazepines, which follow general paediatric rules [21, 22]. The French PNDS states that any diagnosis of leptomeningeal angioma, even before any seizure, should lead to teaching the family to recognise seizures, apply first-aid measures and administer the conditionally prescribed rescue medicine (rectal diazepam or buccal midazolam), earlier than the usually recommended 5 minutes when the seizure is motor; established status epilepticus follows the general recommendations for children and adults [19]. As the PNDS gives no doses, the reference doses are those in the labels of the rescue medicines, to be used according to the individual plan:

Sources: [74, 75]. The European label of buccal midazolam covers infants from 3 months, whereas the North American rectal gel starts at 2 years; in younger infants, use of any rescue benzodiazepine is an extrapolation and should be defined by the reference centre [74, 75].

The 2018 consensus identified no antiseizure medicine of choice [21]. The Italian consensus recommends focal-acting antiseizure medicines, oxcarbazepine, carbamazepine or levetiracetam, as first line, considering oxcarbazepine possibly preferable for its adverse-effect profile; as alternatives or adjuncts, in patients who do not respond or are not candidates for surgery, it lists lacosamide, lamotrigine, phenobarbital, cannabidiol, topiramate, clobazam and the ketogenic diet [22]. In a multicentre cross-sectional survey, 48% of patients received levetiracetam, 45% low-dose ASA and 40% oxcarbazepine [23]. The French PNDS likewise does not set a protocol for the choice: it proposes levetiracetam, valproate or carbamazepine as first intention, notes that some centres prefer carbamazepine or oxcarbazepine for apparently greater efficacy (Thomas 2003, Kaplan 2015) and points out that, in France, levetiracetam is authorised as monotherapy for focal seizures only from 16 years and as adjunctive therapy from 1 month [19]. In refractory epilepsy, it accepts the combination of stiripentol with carbamazepine, based on the Lyon series (Ville 2002) [19]. Stiripentol (Diacomit®) is authorised in the EU only as an adjunct to clobazam and valproate in Dravet syndrome, with titration of 20 mg/kg/day in week 1 and 30 mg/kg/day in week 2 up to the recommended dose of 50 mg/kg/day in 2 or 3 doses (no clinical safety data above 50 mg/kg/day), always with food and without dairy products; the label advises against combination with carbamazepine in Dravet syndrome and states that stiripentol inhibits the hepatic metabolism of carbamazepine (CYP3A4 and CYP2C19), raising its levels, so its use in SWS is an extrapolation requiring drug level monitoring and carbamazepine dose adjustment [76]. The PNDS also draws attention to weight-based adjustment in the first year, to drugs that may aggravate glaucoma and to visual field monitoring on vigabatrin, and accepts case-by-case discussion of treatment reduction in adolescence, given the frequent stabilisation in the second decade [19]. The choice of valproate in a girl or woman of childbearing potential is constrained, in the EU, by the 2018 measures (contraindication in pregnancy unless there is no alternative and a mandatory pregnancy prevention programme) [77]. The following table brings together the specific indications and doses described in series of patients with SWS; these doses describe the practice of one centre and do not constitute dosing recommendations.

7.3. Presymptomatic treatment

Treatment before the first seizure in infants with demonstrated brain involvement is one of the most debated issues. In Day's series (Baltimore centre, children born between 2006 and 2016), 5 infants treated with presymptomatic ASA alone did not differ from matched controls; 8 infants treated with ASA (mean 3.69 mg/kg/day; recommended dose and maximum in section 7.4) and low-dose antiseizure medicine had lower seizure scores (median 0 versus 2; P=0.02), and 4 remained seizure-free at 14–39 months, compared with onset at 6.0 ± 1.4 months in controls (P=0.03) [50]. In Valery's retrospective two-centre series (92 patients), 32 received off-protocol presymptomatic treatment (ASA alone in 5, ASA with levetiracetam in 16, ASA with oxcarbazepine in 9, valproate in 2); 15 of 32 treated patients (47%) and 7 of 60 untreated patients (12%; p<0.001) were seizure-free at 2 years, despite more bilateral involvement among those treated (38% versus 17%), and hemiparesis at 2 years was less frequent in treated patients [49]. The studies are non-randomised and subject to selection and ascertainment bias (50% bilateral involvement in Day's series, compared with the usual figure of about 15%) [49, 50]. The Italian consensus allows presymptomatic antiseizure medicine to be offered to patients with extensive unilateral or bilateral brain involvement, explaining its empirical nature and adverse effects [22]. Studies of prophylactic valproate or carbamazepine showed no difference in epilepsy onset [23]. A phase 2 trial of presymptomatic cannabidiol began recruiting in September 2026 (section 7.12) [82].

The French PNDS, on the basis of expert opinion, requires every child with SWS, even without neurological symptoms, to be referred to a paediatric epilepsy centre, and allows prophylactic antiseizure medicine to be proposed when there are initial severity criteria associated with a higher risk of presenting status epilepticus: extensive pial angioma and/or interictal paroxysmal abnormalities on early EEG (Comi 2015, Bar 2018) [19]. The electroencephalographic criterion does not appear in the Italian consensus, which relies only on the extent and bilaterality of involvement [19, 22]. The choice of drug follows that for focal epilepsy (levetiracetam, valproate or carbamazepine; carbamazepine or oxcarbazepine in some centres), without a formal recommendation of a specific molecule, and the doses are those in the table in section 7.2 [19]. Regarding preventive ASA from diagnosis in infants, the PNDS acknowledges that some authors discuss it and some teams use it, but considers that current data do not allow a formal recommendation, because the series showing benefit did not validate the indication [19].

7.4. Low-dose acetylsalicylic acid

Low-dose ASA is used to prevent stroke-like episodes and seizures, based on the hypothesis that venous stasis and thrombosis cause progressive ischaemic injury; use is off-label, without placebo-controlled trials [21, 22]. The proposed dose is 3–5 mg/kg/day, with an Italian consensus maximum of 100 mg/day [21, 22]. In Lance's series, infants weighing at least 4 kg started one-quarter of an 81 mg infant ASA tablet, crushed or compounded, to achieve 3–5 mg/kg/day; treatment was tolerated from 1 month, levels were not measured and dose was reduced or spaced for minor effects [62].

Efficacy evidence is observational. In an internet survey, 34 of 98 respondents had used ASA; stroke-like episodes decreased from 1.1 to 0.3 monthly in the following year (n=26; p=0.014) and seizures from a median 3 to 1 monthly (n=21; p=0.002), with self-reported uncontrolled data [83]. Earlier studies reported 65% fewer episodes in 14 patients and fewer seizures and episodes in 6 [62]. In Lance's series, 91% of 35 patients with longitudinal neuroscores had reasonably controlled seizures at last assessment versus 83% initially (P=0.477), but 4 of 5 bilateral patients worsened; authors acknowledge follow-up losses (24 patients), referral bias, no control and antiseizure medicine confounding [62]. In the survey, 7 of 26 stopped having episodes after ASA and 2 began having them [23]. Neurodevelopmental effects are unknown [22].

Adverse effects are generally minor. In the survey, 39% reported complications (mainly bruising, gingival bleeding and epistaxis), with no discontinuations for this reason [83]. In Lance's series, 49 of 58 patients had no significant effects; 6 had minor effects (epistaxis, bruising) and 3 had effects leading to discontinuation: allergic rash, haematemesis in an infant one week after initiation, in the context of severe hand, foot and mouth disease with gastrointestinal lesions, requiring transfusion, and subdural haematoma after minor trauma in a young child, with complete resolution [62]. Subdural haematoma is rare and attributed to atrophy with rupture of bridging veins [23]. Families should be informed of the antiplatelet effect, report bruising, bleeding and rash, and inform other doctors of the treatment [62]. Use is more extensive in the US (49% of patients) than in the Italian experience [22]. The French PNDS points out that ASA increases the risk, intensity and duration of bruising from pulsed dye laser and is a relative contraindication requiring adjustment of laser parameters (section 7.8) [19].

7.5. Epilepsy surgery

Surgery should be considered for drug-resistant seizures, particularly with unilateral involvement, early seizure onset and pre-existing developmental or neurological impairment, always at a tertiary paediatric epilepsy centre [20, 22]. Options are anatomical or functional hemispherectomy (hemispherotomy), lobar or sublobar resections and multilobar disconnections, including posterior disconnection preserving intact frontal lobe and motor cortex [20, 22]. Presurgical assessment includes motor and language functional MRI, DTI tractography with more than 36 directions, PET and SPECT when required [20–22]. In bilateral disease with severe epilepsy, hemispherotomy of the more epileptogenic side or vagus nerve stimulation may be considered with palliative intent in selected cases [22, 23].

The 2026 AES guideline strongly recommends hemispherectomy or hemispherotomy in infants and children younger than 36 months with drug-resistant epilepsy secondary to selected lesional pathologies including SWS, and focal, lobar or multilobar resections and posterior disconnections for drug-resistant focal epilepsy at the same age [73]. The Italian consensus favours hemispheric surgery within the first 4 years or with significant existing impairment because hemiparesis invariably follows hemispherectomy [22].

In the Chinese cohort, persistent seizures after focal resection were associated with partial resection (adjusted hazard ratio 11.50), acute postoperative seizures and generalised seizures, and after hemispheric surgery with acute postoperative seizures, generalised seizures, 30 or more seizures monthly and surgical age of 2 years or older; age below 2 years was associated with cognitive stability, and modified hemispherotomy had outcomes comparable to anatomical hemispherectomy with greater safety [86]. These studies cannot be compared directly: they are retrospective, with surgical selection, differently defined outcomes and, for the survey, response bias [48, 84, 86].

The French PNDS summarises the same literature: 70–100% of patients become seizure-free after hemispherectomy or hemispherotomy and 58% after focal resection, with better outcomes when surgery is early and resection complete; complications reach 47% after hemispheric surgery (mainly haemorrhage and intracranial hypertension) and should be anticipated; in Sugano's series (2014), 8 of 10 patients with angioma limited to the posterior quadrant became seizure-free after posterior disconnection; data in bilateral disease are anecdotal; and the cognitive prognosis is better when the child is operated on at a younger age and with a shorter interval between seizure onset and surgery (Wang 2022) [19]. It recommends referral to an epilepsy surgery centre of every patient with unilateral leptomeningeal angioma and drug-resistant epilepsy, defined as failure of two well-conducted treatments, with a multidisciplinary decision in a specialised paediatric neurosurgery department [19]. In the presurgical assessment, MRI looks for extension of the angiomatosis to the contralateral hemisphere, which contraindicates hemispheric disconnection, and for cortical dysplasias such as polymicrogyria, although the focal dysplasias found in surgical specimens were not retrospectively visible on MRI; long-term EEG records seizures and guides the discussion [19].

7.6. Stroke-like episodes and headache

Stroke-like episodes are treated with fluids, fever, nausea and vomiting management, physiotherapy and occupational therapy; prevention consists of avoiding activities with significant head injury risk, treating intercurrent illnesses and maintaining immunisation [23]. Low-dose ASA is used preventively (section 7.4) [21]. Urgent imaging is necessary only for new, prolonged, worsening or unresolved episodes [20].

No SWS headache guidelines exist. Usual management includes sleep, hydration, ibuprofen, paracetamol and antiemetics; triptans have been used without reported serious effects and with self-reported efficacy, and in young children with seizure clusters and irritability, analgesia with a short benzodiazepine course is used [21, 22]. Prophylaxis uses valproate, lamotrigine, topiramate, gabapentin or flunarizine [21, 22]. Sources provide no SWS-specific doses.

The French PNDS recommends informing the patient and family about these episodes and their signs and giving them an information sheet for the attending physician and the emergency department (section 4.3) [19]. In the acute phase, imaging is not systematic but may be needed to exclude stroke, thrombophlebitis or another cause of acute motor deficit, particularly in a first or atypical episode, and associated status epilepticus should be sought clinically and electroencephalographically [19]. It recommends neither anticoagulation in the acute phase nor formal preventive restriction of physical activity, unlike the North American recommendation to avoid activities with a significant risk of head injury, and recommends ASA 3–5 mg/kg/day after the first episode (section 7.4) [19, 23]. Headache is treated as common migraine, both as background prophylaxis and for attacks, and the PNDS stresses that it is often undervalued; it calls for caution with triptans, owing to the risk of aggravating motor deficit, whereas the North American consensus reports their use without serious effects being reported [19, 21].

7.7. Neurodevelopment, mental health and rehabilitation

Developmental and neurocognitive assessment with standardised scales (e.g. Bayley or Griffiths) should occur at diagnosis, seizure onset and suspected delay or behavioural change, with rehabilitation and an individual education plan [22]. Formal neuropsychological testing is possible from 3–4 years [21]. In ADHD, stimulants were considered subjectively effective with few adverse effects [21]. Adult psychological dysfunction and suicide risk should be assessed, including when choosing medication, and diagnostic uncertainty warrants psychodiagnostic assessment [22]. An observational SWS-adapted neuropsychological rehabilitation study began recruiting in Rome in June 2026 [87].

The French PNDS considers epilepsy control, with attention to the efficacy and adverse effects of antiseizure medicines, the first measure to prevent cognitive and motor deficit [19]. It recommends longitudinal follow-up of cognitive functions at the stages of the French school pathway (entry to nursery school at 3 years; before the first year of primary school, at about 5–6 years; end of primary school, at about 10 years; and end of lower secondary school, when there is a question of educational or vocational orientation) and whenever warning signs arise, using the Wechsler scales (WPPSI-IV, WISC-V and, in adults, WAIS) and targeted speech and language therapy, psychomotor, occupational therapy and orthoptic assessments [19]. Deficits lead to rehabilitation (speech and language therapy, psychomotor therapy, occupational therapy, orthoptics, physiotherapy), school adaptations and, if necessary, adapted educational or vocational orientation; special attention is warranted for children with specific learning disorders and borderline IQ, whose difficulties tend to be underestimated, for praxis and visuospatial disorders, and for visual disturbance and hemianopia, whose detection requires sufficient cooperation and which should be communicated to the school [19]. Attention tests are available only from 6 years, and psychostimulants may be offered, with special monitoring if epilepsy is not stabilised [19].

For hemiparesis, the PNDS recommends assessment and follow-up in physical and rehabilitation medicine, which evaluates functional disability, contractures, orthopaedic deformities and pain and prescribes motor physiotherapy (the priority), psychomotor therapy, occupational therapy, orthoses and botulinum toxin injections; the deficit may improve with rehabilitation when epilepsy is controlled, and in adulthood rehabilitation continues, if necessary, in an adult facility [19]. The PNDS gives no botulinum toxin doses; for comparison, the US label of onabotulinumtoxinA (Botox®), authorised for spasticity from 2 years, specifies for paediatric spasticity 3–6 U/kg in the upper limb (maximum 200 U) and 4–8 U/kg in the lower limb (maximum 300 U), not exceeding the lower of 10 U/kg or 340 U per 3-month interval; use for hemiparesis in SWS is an extrapolation from paediatric spasticity of other causes [88].

On the psychological level, the PNDS proposes a meeting with a psychologist for parents and child from school age, in joint consultations with a physician or nurse or in individual interviews, and discussion groups for affected children, parents and siblings [19]. Anxiety, depression and relational or behavioural difficulties should be sought, particularly in drug-resistant epilepsy and cognitive impairment, with referral to psychology, psychiatry or child psychiatry; the impact of PWB on quality of life can be measured with the CDLQI (Children's Dermatology Life Quality Index), and school phobia should raise suspicion of bullying [19]. Cognitive behavioural therapy techniques for children with visible differences in appearance have not been evaluated for efficacy (Jenkinson 2017), but have been integrated into the French therapeutic education programme «E… change de regard», of the ANDDI-RARES filière, aimed at school-age children who suffer from how others look at them, whatever the disease; the materials of the British association Changing Faces exist only in English [19]. Psychological support also helps acceptance of the birthmark [19].

7.8. Port-wine birthmark: laser and other therapies

PWB treatment aims to reduce psychosocial impact, decrease nodularity and potentially hypertrophy, weighing costs; there is no clear evidence that laser definitively prevents hypertrophy because it does not reach deep vessels [17]. Pulsed dye laser (PDL) is first-line for all PWBs of any size, location or colour and is safe at all ages with an experienced operator [2, 8, 17]. It acts through selective haemoglobin photothermolysis, targeting dermal capillaries and venules 1.2–1.5 mm deep [22]. Complete clearance is rare: PDL achieves 50–90% lightening, most exceed 50%, and typically 8–10 or more sessions with frequent touch-ups are necessary [2, 17]. Better responses occur with light skin, face and neck (versus extremities), lateral face (versus central), pink, red or reticulated lesions (versus purple or geographical) and flat lesions (versus hypertrophic); in the US consensus authors' opinion, SWS PWB may be more resistant [17].

Early initiation, particularly in the first year, is associated with better outcomes in retrospective short-follow-up studies, and efficacy is greatest in early childhood and lesions up to 20 cm² [17, 22]. In a Chinese series of 848 patients, response was 93.9% in children treated within the first year versus 25% in adults older than 50 [17]. Session intervals are unestablished: infant studies used 2 weeks to 3 months, the Italian consensus mentions 4–6 weeks, and more pigmented skin benefits from longer intervals until purpura resolves [17, 22].

In a Chinese multicentre randomised trial not specific to SWS, in patients aged 14–65, photodynamic therapy with intravenous hemoporfin 5 mg/kg and 532 nm laser produced at least 20% improvement at 8 weeks in 89.7% (95% CI 85.9–92.5) versus 24.5% with placebo, with hyperpigmentation in 22.9 per 100 treatments [89]. A network meta-analysis of randomised trials through 2023 concluded hemoporfin photodynamic therapy is more effective for lightening than PDL with greater adverse-effect risk, and mostly unclear risk of bias [90]. In the Navarra intraindividual double-blind randomised phase II trial in 23 adults with SWS (median age 33), PDL plus topical rapamycin produced greatest lightening and lowest vessel percentage with good tolerability; rapamycin concentration is absent from the consulted abstract [91].

The Italian consensus recommends early-childhood treatment, preferably under sedation or general anaesthesia in the first year to reduce distress and session numbers; sedation is always necessary around the orbit in children [22]. The US consensus allows PDL without general anaesthesia in young infants, notes FDA advice for caution with repeated anaesthesia or sedation below 3 years (warning based predominantly on preclinical data) and proposes shared decision-making (sections 7.16 and 7.18) [17]. In a study cited by the 2018 consensus, patients receiving multiple laser treatments under general anaesthesia had no more neurodevelopmental disorders than the general population [21]. Topical lidocaine is safe but carries a methaemoglobinaemia risk, particularly in infants [17].

The French PNDS considers birth the ideal time to assess the indication for laser and PDL the reference treatment for any PWB, with a favourable benefit-risk ratio at any age, and treatment may begin in the first days of life; the earlier, the better the results [19]. In France, the classification of medical procedures (CCAM) allows sessions to be billed only from 1 year of age, and reimbursement requires prior approval by the health insurance and a minimum interval of 3 months between billed sessions, although the optimal interval has not been established and closer sessions may be needed at the start of treatment [19]. General anaesthesia is decided case by case, with the parents or with the patient when of an age to consent, and leptomeningeal angioma or prior seizures require caution because of the risk of triggering a seizure during the session [19]. Circular impacts are 5–12 mm in diameter and require eye protection for the operator (filtering goggles) and the patient (metal eye shields); short pulse durations (0.45–1.5 ms) produce almost constant purpura, an immediate marker of efficacy; ASA increases the risk and duration of bruising and is a relative contraindication requiring adjustment of pulse duration, fluence and cooling; epidermal necrosis is rare with adapted fluence and cooling; and in pigmented skin, with more dyschromia and atrophic scarring, reduced fluences, less overlap and more cooling are recommended [19]. Without general anaesthesia, lidocaine-prilocaine cream (EMLA®; also Anesderm® in France) is applied, with caution in infants because of the risk of methaemoglobinaemia [19]. The US label of the cream sets maximums by age and weight: 1 g over 10 cm² for 1 hour from 0 to 3 months or below 5 kg; 2 g over 20 cm² for 4 hours from 3 to 12 months and above 5 kg; 10 g over 100 cm² from 1 to 6 years and above 10 kg; and 20 g over 200 cm² from 7 to 12 years and above 20 kg [92]. The efficacy of a session can be assessed only after several weeks, and long-term efficacy after several sessions; there is late redarkening (35% at 10 years with older-generation dye lasers, Huikeshoven 2007) [19]. In the event of insufficient efficacy, refusal or contraindication, camouflage with corrective make-up is offered, possibly with a socio-aesthetician [19]. Vascular nodules (pyogenic granulomas) and soft-tissue hypertrophy are treated with PDL, long-pulsed Nd:YAG or KTP, or with surgery; there is no targeted pharmacological treatment for the hypertrophy, and the PNDS recommends keeping the family informed about research in this area [19].

7.9. Glaucoma and choroidal haemangioma

SWS glaucoma has two mechanisms: angle dysgenesis, in early-onset forms, and raised episcleral venous pressure, in late forms; current ocular hypotensive agents do not reduce episcleral venous pressure [21, 38]. In children younger than 3–4 years, glaucoma treatment is almost always surgical, and medical therapy serves only as a bridge to surgery; in older children and adults, topical drugs are used first (beta-blockers, carbonic anhydrase inhibitors, prostaglandin analogues, alpha-2 agonists), following general glaucoma guidelines, with surgery if IOP is not controlled on maximal therapy or if there is intolerance or poor adherence [20–22]. The 2026 North American glaucoma consensus recommends initial medical intervention from 4 years of age and, below that age, primary intervention that is probably surgical, with goniosurgery as the first choice [63]. Latanoprost lowers IOP in late but not in early glaucoma; netarsudil 0.02%, a Rho-associated kinase inhibitor that also lowers episcleral venous pressure, showed efficacy in a small cohort; oral propranolol 2 mg/kg (study dose; frequency and maximum dose not given in the source) lowered IOP in 4 children in the first week, but the effect waned after one month [38]. Laser trabeculoplasty is not applicable, and retroequatorial drainage devices are the last choice [22].

The French PNDS diverges for young children: it considers that SWS glaucoma usually has an insidious onset and that the first line is medical, often with dual topical therapy from the outset in infants, preferring beta-blockers and carbonic anhydrase inhibitors, avoiding prostaglandin analogues because of the risk of uveal effusion (Gambrelle 2008) and requiring punctal occlusion after each instillation to prevent systemic absorption [19]. The PNDS gives no doses; for comparison, the US labels specify timolol 0.25% (or 0.5% if the response is insufficient), 1 drop twice daily, and dorzolamide 2%, 1 drop three times daily, or the fixed dorzolamide-timolol combination twice daily, with safety and efficacy established from 2 years of age for timolol and the combination, and in a 3-month paediatric trial for dorzolamide; below 2 years use is an extrapolation, which reinforces the need for punctal occlusion [93, 94]. The PNDS reserves surgery for the rare severe acute forms, sometimes neonatal, and for failure of medical treatment, with filtering surgery as first intention (penetrating or non-penetrating trabeculectomy, combined by some with trabeculotomy), repeatable if efficacy is lost, and considers that angle surgery should not be recommended in SWS (Yeung 2020); second-line procedures belong to the subspecialty [19]. Once the child allows automated visual fields and/or OCT (nerve fibre layer and ganglion cell complex), the decision to treat or intensify is based on these investigations and not mainly on IOP, even when corrected for pachymetry; surveillance is lifelong, because recurrence or decompensation may occur at any age [19]. The North American and Italian consensuses prefer surgery, with goniosurgery as first choice, below 3–4 years and include prostaglandins among topical drugs (section 7.16) [21, 22, 63].

SWS eyes have increased intraoperative and postoperative complication risk, particularly suprachoroidal haemorrhage and recalcitrant serous retinal detachment [20]. Described perioperative measures include slow globe decompression (high-viscosity viscoelastic, tight flap closure), B-scan ultrasonography and fundoscopy to assess choroidal malformation, off-label oral propranolol 2 mg/kg/day from 1 week before to 6–8 weeks after surgery under cardiology guidance to reduce choroidal effusion, postoperative topical corticosteroids and non-steroidal anti-inflammatory drugs, and a benzodiazepine emergency plan for seizure risk; observation is the main treatment for exudative detachment [38]. Some groups use prophylactic sclerotomy [38]. Amblyopia, the second treatable cause of visual loss, requires rigorous treatment, and ptosis should be reassessed at 3–6 months [20, 22, 38].

Laser treatment of choroidal malformations is indicated when there is a risk of subretinal haemorrhage, serous retinal detachment, cystoid macular oedema or neuroepithelial detachment; argon photocoagulation is associated with poor final acuity, recurrent exudation, retinal detachment and irreversible scotoma near the disc or macula [22]. Photodynamic therapy and intravitreal anti-vascular endothelial growth factor (anti-VEGF) agents have been described for exudative detachment, and oral sirolimus was used off-label in a case of exudative detachment due to diffuse choroidal haemangioma in an 8-year-old girl [38, 95]. External beam radiotherapy for diffuse choroidal haemangioma was associated with visual field defects simulating glaucoma [57]. For exudative chorioretinopathy due to diffuse choroidal haemangioma, the method most used in France is low-dose whole-choroid irradiation by external photon radiotherapy, 20 Gy in 10 fractions, feasible in childhood and under general anaesthesia if cooperation is insufficient; the retina reattaches in about 90% and vision stabilises or improves in about 75% (Randon 2018), with refractory cases and exudative recurrences [19]. Proton therapy (indicated for circumscribed haemangiomas), photodynamic therapy, gamma knife radiosurgery and intravitreal anti-VEGF agents are alternatives of limited success: the globe may be preserved, but the visual prognosis is poor, particularly with macular involvement [19].

7.10. Oral cavity and maxillofacial surgery

The first dental assessment should take place soon after diagnosis, with follow-up every 6 months without oral lesions and every 3–4 months with lesions (table in section 11) [22]. Gingival hyperplasia results from the synergistic effect of dental plaque, antiseizure medicines and mucosal hypertrophy with CM, and is managed with repeated non-surgical periodontal therapy and hygiene [22]. Laser is the reference treatment for vascular lesions of the oral mucosa and for gingivectomy (CO₂ 10,600 nm; Nd:YAG 1064 nm, gingivectomy at 40 Hz, 130 mJ and 4 W; diode 808–980 nm, 1.5–3 W in continuous mode), with less bleeding and pain and no suturing [22]. For extractions, prior selective embolisation of the maxillary arteries, perilesional anaesthesia with a needle change for each infiltration, systematic suturing and the possibility of transfusion in extensive surgery have been described [22]. Orthognathic surgery has outcomes comparable to those in patients without SWS, with the same indications, in patients with adequate neurocognitive development [22]. The French PNDS insists that the dentist be familiar with the disease, because routine treatments may be complicated by intra- and postoperative bleeding, and makes prevention (oral and dietary hygiene, fluoride toothpaste) the basis of care [19]. Before treating areas with a vascular component, MRI or MR angiography may be indicated, in consultation with the medical team, particularly to exclude an AVM that a surgical procedure would aggravate; the stress of care is managed with a behavioural approach and with conscious sedation by inhalation of an equimolar mixture of oxygen and nitrous oxide (EMONO) or with sedative drugs; periodontal disease requires a detailed initial assessment and regular monitoring; gingivectomy and extractions may require hospital admission, general anaesthesia and haemostatic measures agreed with the medical team; orthodontics and orthognathic surgery are performed if hygiene is adequate; and speech and language therapy should be considered [19]. The PNDS proposes dental visits 3 times a year in children and at least twice a year in adults, intervals different from those of the Italian consensus (sections 7.16 and 11.2) [19, 22].

7.11. Endocrine abnormalities and blood pressure

GH deficiency and central hypothyroidism require endocrinology referral; routine thyroid tests are recommended, particularly with clinical manifestations, recognising that antiseizure medicines also alter these tests [22, 60]. In Hadid's case, hypogonadotrophic hypogonadism and central hypothyroidism were treated with testosterone and levothyroxine and GH was considered [61]. There are no immunisation contraindications, and some authors suggest annual influenza vaccination [22]. For CM suggestive of GNAQ or GNA11 mosaicism, Zhang and colleagues propose, based on a retrospective series, blood pressure measurement at diagnosis and every consultation, renal Doppler if hypertensive, and investigation of calcium metabolism abnormalities and hyperaldosteronism (aldosterone/renin ratio) at diagnosis and, if normal, every 1–2 years [12].

7.12. Investigational therapies: cannabidiol, sirolimus and others

Cannabidiol and mTOR (mechanistic target of rapamycin) pathway inhibitors are being evaluated as preventive or disease-modifying treatments, with initial small-study findings, but neither is authorised for SWS [8]. Proposed cannabidiol mechanisms include mTOR pathway blockade, inhibition of Gαq-coupled receptors and action on GPR55; the sirolimus rationale rests on increased phosphorylated ribosomal S6 protein in PWB skin and leptomeningeal endothelium [23, 39].

Cannabidiol was studied in an expanded access programme in 5 patients with SWS and refractory epilepsy: among 4 with data to week 14, 2 at that week and 3 with bilateral involvement at last visit had greater than 50% seizure reduction and improved quality of life, maintained over 63–80 weeks, with mild adverse effects in 3 [96]. The same study's registry documents seizure-frequency reductions of 12%, 100%, 83% and 64% in the 4 assessed patients and serious adverse events in 3 of 5 [97]. In a 6-month open-label trial of 10 patients with cerebral SWS, controlled seizures and cognitive impairment, oral cannabidiol (Epidiolex®) 5–20 mg/kg/day was well tolerated; 6 had mild to moderate effects and 1 withdrew, and improvements occurred in neuroscore, self-reported quality of life, anxiety, emotional regulation and bimanual use [80]. In the corresponding registry, the primary outcome (working memory T-score) increased from 37.14 ± 10.92 to 39.14 ± 13.78, and non-serious adverse events occurred in 7 of 10 (diarrhoea, dizziness, fatigue, headache, transaminase elevation in 2, nausea, PWB darkening in 1), without serious events [98]. One author of these studies (A. Comi) declares patents on cannabidiol use in SWS [23, 99].

Sirolimus was studied in an open-label two-centre pilot trial (2017–2018) in 10 patients with cerebral SWS and cognitive impairment (mean age 12 years and 7 months), using oral sirolimus solution (Rapamune®) twice daily for 6 months, starting dose calculated by height and weight and adjusted through Bayesian pharmacokinetics, with a maximum 2 mg/day and target trough 4–6 ng/mL; non-oestrogen contraception was mandatory [39]. Six reached target levels at some point [39]. Processing speed improved (p=0.031), as did neurological quality-of-life scales for anger, depression and cognitive function, without significant neuroscore, PWB score or EEG changes; the 3 with previous severe stroke-like episodes had episodes during the study but shorter recovery [39]. Nine of 10 had possibly or probably related events, almost all grade 1: hypertriglyceridaemia (4, managed by diet), aphthous ulcers (3), headaches, elevated liver tests (1, with dose reduction), low HDL cholesterol and personality change (2) [39]. A retrospective series of 6 with levels of 10–15 ng/mL suggested safety, PWB lightening and seizure benefit, according to trial authors [39]. The Italian consensus mentions possible sirolimus efficacy in drug-resistant seizures based on case series [22]. Sirolimus with ASA in a premature newborn with severe bilateral SWS, seizure-free at 23 months, and oral sirolimus for diffuse choroidal haemangioma are also described [66, 95].

The SWAC (Sturge-Weber Acute Crisis) index, published in 2026, quantifies acute crises (seizures, stroke-like episodes, headaches) and was applied retrospectively to a 10-patient open-label trial published in 2020, showing improved composite severity, stroke-like episodes and headache; it is intended as an outcome for future trials [100]. QLS-101 eye drops, a prodrug of an ATP-sensitive potassium channel opener designed to lower episcleral venous pressure, reduced IOP in mice and were tested in two phase 2 safety trials in adolescents and adults with SWS glaucoma, without published registry results [101–103]. MEK (mitogen-activated protein kinase kinase) inhibitors such as trametinib have only a preclinical SWS basis [41].

7.13. Regulatory table

No drug has a specific SWS authorisation in the US or EU; all described uses are off-label [24, 25, 104, 105].

Sources: [24, 25, 39, 80, 82, 104, 105]. Low-dose ASA is used off-label at doses described in section 7.4 [21, 22]. The US cannabidiol label warns of dose-dependent transaminase elevation (ALT above 3 times the upper limit in 13% with 10–20 mg/kg/day and 12% with 25 mg/kg/day versus 1% placebo), greater with concomitant valproate, somnolence, suicidal behaviour and ideation, hypersensitivity and status epilepticus risk with abrupt withdrawal, and states safety and efficacy are unestablished below 1 year [24].

7.14. Clinical trials

Available trials are small, mostly open-label and uncontrolled, with multiple outcomes without adjustment for multiple comparisons, and do not permit indirect comparison [39, 80]. The only randomised trials concern PWB rather than neurological outcomes [89, 91]. Negative or unreported trials (adjunctive everolimus epilepsy trial withdrawn without patients; topical timolol PWB trial terminated with 3 patients; QLS-101 without results) form part of the evidence and demonstrate recruitment difficulties in rare disease [102, 103, 107, 108]. Observational natural history and biomarker cohorts are ongoing, including the Brain Vascular Malformation Consortium and Kennedy Krieger Institute longitudinal cohort, with 600 participants planned through 2027 [20, 37, 109].

7.15. Developments in the last 12 months

Consultation of FDA, EMA and ClinicalTrials.gov databases identified no SWS-specific authorisation, application, opinion or human trial of a MEK inhibitor; this negative search does not establish absence of ongoing processes [24, 25, 41].

7.16. Guideline recommendations and consensus differences

7.17. Agents and procedures to avoid or use cautiously

Topiramate is associated with glaucoma in the general population and should be used with caution, although one SWS study showed no increase in glaucoma [21, 22, 38]. Carbamazepine and oxcarbazepine may worsen thyroid dysfunction [22]. Cannabidiol with valproate increases the risk of transaminase elevation [24]. For sedation, drugs without proconvulsant action should be preferred [22]. Topical lidocaine requires adherence to limits because of the risk of methaemoglobinaemia in infants [17]. Repeated CT, routine contrast at follow-up and cerebral angiography should be avoided [20, 21]. Thrombolytics have not been studied in SWS [20]. During sirolimus treatment, the trial required oestrogen-free contraception [39]. The French PNDS adds caution with antiseizure medicines that may aggravate glaucoma, visual field monitoring on vigabatrin, which adds to hemianopia, caution with triptans owing to the risk of aggravating motor deficit, avoidance of prostaglandin analogues owing to the risk of uveal effusion, no anticoagulation in the acute phase of stroke-like episodes, avoidance of CT in the newborn, caution with lidocaine-prilocaine cream in infants and adjustment of laser parameters in patients on ASA [19].

7.18. Anaesthesia, sedation and perioperative care

In patients without neurological symptoms in the preceding 3 months, anaesthetic management is the same as in the healthy paediatric population; with uncontrolled neurological symptoms, sedation requires post-procedure observation for at least 24 hours, and clinical stability without seizures for at least 3 months is the main criterion for outpatient sedation [22]. A recent EEG (within 3 months) is recommended in patients on antiseizure medicines [22]. Care includes the airway, which may have a vascular malformation, spontaneous ventilation and maintenance of normal intracranial, intraocular and arterial pressures, to avoid stroke-like episodes or acute ocular injury [22]. The proposed regimen combines lidocaine and prilocaine cream with oral or intranasal midazolam and/or dexmedetomidine, without doses specified by the consensus [22]. The cardiorespiratory risk of general anaesthesia is greatest in the neonatal period [17]. Pain regulatory pathways develop only at 12–18 months, and randomised trials showed no neurodevelopmental differences between general and locoregional anaesthesia in the neonatal period and in infancy [22]. In eye surgery, the perioperative plan includes benzodiazepines for seizure risk and the measures in section 7.9 [38]. The French PNDS performs MRI in the newborn after a bottle feed, without sedation, and reserves sedation or general anaesthesia for after 4–6 months and, after a seizure, for children younger than 4–5 years; it proposes combining ophthalmological examinations under general anaesthesia in infants with glaucoma with a laser session, to reduce the number of anaesthetics; and it accepts conscious sedation with EMONO for dental care [19].

7.19. Pregnancy and lactation

A review of 15 published cases and 1 own case of pregnancy in women with SWS, some with refractory epilepsy, described neurological complications in 4 of 14 reviewed cases and uncomplicated delivery in 7 [110]. Another report, of a 38-year-old woman with SWS and epilepsy, had a safe maternal and fetal outcome and proposed a preconception and pregnancy care plan, with controversy over delivery management [111]. Management follows the principles of epilepsy in pregnancy, and the sources consulted describe no data on lactation in SWS or on the use of ASA, cannabidiol or sirolimus during pregnancy in patients with SWS. The French PNDS asks that, from adolescence, girls' antiseizure medicine be adjusted to contraception and to the prospect of a future pregnancy (section 7.20), which in the EU includes the restrictions on valproate [19, 77].

7.20. Transition to adulthood

Annual ophthalmological surveillance, neurological follow-up with MRI only for new or progressive symptoms, and mental health attention continue in adulthood; the Patient Engagement Network considered mental health the most important unmet need [20–22]. PWB may darken and develop nodules after puberty, justifying ongoing dermatological follow-up [17]. Pregnancy and contraception should be planned with the epilepsy team [39, 111].

The French PNDS recommends preparing the transition, which usually takes place at about 18 years, by raising the subject between 14 and 16 years, identifying with the family an adult neurologist experienced in a reference centre and offering a joint transition consultation [19]. These years serve to build a career plan compatible with cognitive difficulties, hemiparesis and epilepsy, to provide information on driving licence legislation (epilepsy, hemianopia, cognitive and motor aspects), to adjust girls' antiseizure medicine to contraception and a future pregnancy and to support medico-social and legal procedures (adult application to the MDPH, disabled worker status, adapted facilities for the non-autonomous young adult, guardianship or curatorship); platforms or therapeutic education sessions dedicated to transition may be used [19]. In adulthood, seizures are usually less frequent, epilepsy follow-up follows the general HAS recommendations for focal epilepsies, treatment withdrawal may be discussed when epilepsy is controlled, without formal recommendations, and rehabilitation may continue in adult physical and rehabilitation medicine; literature on adults is scarce, as shown by the French series of 10 adults followed in an epileptology department, with heterogeneous presentations (Veillieux 2022) [19].

8. Reproductive Options and Family Counselling

8.1. Sporadic nature and recurrence risk

SWS results from a postzygotic somatic variant confined to the affected tissues; it is not inherited, Orphanet classifies inheritance as “not applicable”, and most cases are sporadic [1, 8, 13, 23]. Counselling should therefore not present recurrence risks as though the disease were autosomal dominant [8]. Happle's hypothesis, revisited in the founding study, proposes that the variant would be lethal in the constitutional state and is compatible with life only in mosaic form, which explains the absence of described familial transmission [3]. The sources consulted provide no empirical estimate of recurrence risk for siblings or offspring; the absence of familial cases in the series does not amount to a risk demonstrated to be zero. In a survey of 52 adults with SWS, 10 had liveborn children, with no mention of recurrence [23]. ClinVar germline classifications for GNAQ and GNA11 variants cannot be used to estimate recurrence risk [43, 44]. The French PNDS (French National Diagnosis and Care Protocol) makes the reasoning explicit: the variant may also affect the germline, but an embryo receiving it in every cell would not be viable; it therefore concludes that the risk of recurrence in a subsequent pregnancy is no higher than in the general population and that there is no screening measure to offer (Kinsler 2019), and it knows of no predisposing hereditary or environmental factor [19]. This conclusion rests on the mechanism and on the absence of familial cases, not on an empirical estimate of recurrence.

8.2. Family testing, prenatal diagnosis and preimplantation testing

Neither cascade testing of relatives nor prenatal or preimplantation diagnosis is indicated, because the variant is not present in the patient's germline and is not reliably detected in blood [8, 23]. A family history of vascular malformations is, by contrast, a warning sign for an alternative dominantly inherited diagnosis: germline RASA1, EPHB4 and KIT variants have been identified in patients with atypical SWS, implying a 50% risk for offspring and justifying investigation of relatives [14, 45]. In CM-AVM syndrome, about 70% of patients with a RASA1 variant and about 80% of patients with an EPHB4 variant have an affected parent, with penetrance of 90–99% and 93%, respectively [14]. The French PNDS asks that a craniofacial PWB, even frontal, in a familial context should raise the possibility of CM-AVM syndrome due to a heterozygous RASA1 or EPHB4 variant, and points out that molecular testing, as it requires affected tissue, cannot be performed on amniocytes (section 5.7) [19].

8.3. Counselling families and minors

Family counselling should explain the mosaic mechanism, the non-hereditary nature, the spectrum from isolated PWB to complete disease and the unpredictability of cerebral evolution in the first years [2, 3, 20]. Molecular testing is not needed for routine diagnosis, and in minors it is justified only when it changes the diagnosis or management, as in atypical forms [23, 45]. Decisions on anaesthesia for laser, presymptomatic imaging and presymptomatic treatment are shared with parents, supported by low-level evidence (section 7.16) [17, 22]. Psychological support for parents, who report guilt, comments from strangers and the time burden of treatments, is part of counselling [20]. Pregnancy in women with SWS and epilepsy should be planned before conception (section 7.19) [111]. The French PNDS describes a dedicated disclosure consultation, held after the synthesis of investigations and with a secure diagnosis, by a physician experienced in the disease, at one of two moments: before any seizure, when the diagnosis follows the discovery of a V1/S1–S4 PWB, or at disease onset with a neurological symptom, usually the first seizure [19]. The consultation provides information on the disease and the risk of epilepsy, teaches recognition and management of a seizure, prescribes rescue medicine if necessary and possibly maintenance treatment, organises paediatric neurology follow-up and referral to ophthalmology and dermatology, and may involve a psychologist and an advanced practice or therapeutic education nurse [19]. Initial information should be cautious and supported by experts, without alarming parents when the craniofacial PWB does not affect the forehead [19].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Prevalence and incidence

Birth prevalence of 3.5/100 000 and the estimate of 1/20 000 to 1/50 000 are of similar magnitude but measure different things and derive from no population-based studies identified in consulted sources [1, 17]. SWS proportions among children with PWB come from tertiary centres and overestimate general population risk [18]. Prevalence includes clinical SWS regardless of gene and excludes non-syndromic PWB [8]. No geographical or ethnic correlation is known [38].

9.2. Regional data

Most clinical data come from the US: Kennedy Krieger Institute and Johns Hopkins centres (Baltimore), Boston Children's Hospital, the Sturge-Weber Foundation Clinical Care Network (25 centres) and its registry, assembling 1653 patients in the GH deficiency study [21, 49, 59]. In Europe, key sources are the London Great Ormond Street Hospital cohort of 192 children with facial PWB, the English national centre with 92 children assessed for ASD, the three Italian Delphi consensus reference centres, Brussels and Nijmegen centres describing GNA11 SWS, the Norwegian National Centre for Rare Epilepsy-Related Disorders cohort identifying the GNB2 variant, and the Navarra topical rapamycin trial [5, 6, 18, 22, 23, 91]. Outside Europe and the US, data include a Chinese surgical cohort from 16 centres (214 patients), Chinese, Korean and Indian laser series, an Indian glaucoma and visual field series (86 patients) and the Pakistani case in section 10 [7, 17, 57, 86]. Consulted sources describe no founder populations, consistent with the somatic mechanism. They contain no Portuguese or Brazilian epidemiological data; the Brazilian genetic syndromes textbook mentions SWS only in the Klippel-Trenaunay-Weber differential [32].

9.3. Special populations

In dark-skinned children, conjunctival redness may be the only ocular sign, and laser requires adapted parameters, with more dyschromia and longer intervals [17, 20]. Vitamin D insufficiency was more frequent in African Americans [23]. Late-diagnosed adults, including type 3 SWS, and pregnant women with epilepsy have specific needs (sections 4.4 and 7.19) [20, 110]. Patients with GNA11 CM risk early-onset hypertension and limb growth discrepancy [12].

9.4. European Reference Networks and patient associations

Italian consensus institutions belong to ERN EpiCARE (rare and complex epilepsies), and Ospedale Pediatrico Bambino Gesù also belongs to ERN-Skin, VASCERN (vascular anomalies working group, VASCA) and ERN-ITHACA [22, 26–29]. The US Sturge-Weber Foundation maintains Clinical Care Network and Patient Engagement Network and funded the 2018 and 2021 consensuses [17, 21, 30]. Associazione Sturge-Weber Italia participated through a non-voting representative and funded the Italian consensus [22, 31]. The National Institutes of Health Rare Diseases Clinical Research Network Brain Vascular Malformation Consortium conducts the natural history study [20, 37]. Consulted sources identified no patient associations in Portugal or Brazil.

In France, the PNDS was promoted by the rare disease reference centre (CRMR) Epilepsies Rares (CRéER), of the DéfiScience filière (rare neurodevelopmental diseases), coordinated at the Assistance Publique-Hôpitaux de Marseille (Hôpital de la Timone), with its lead site at the Hospices Civils de Lyon (Hôpital Femme Mère Enfant and Hôpital Neurologique Pierre Wertheimer, Bron), 12 constituent sites and 17 competence centres, and by the reference centre for rare genetic skin and mucosal diseases (MAGEC), of the FIMARAD filière (rare dermatological diseases), coordinated by MAGEC Nord at the Hôpital Necker-Enfants Malades (Paris) and MAGEC Sud at the Hôpital Larrey (Toulouse), with its lead site at the CHU de Dijon, 8 constituent sites and 12 competence centres [19]. These centres can be contacted by the patient, the family or any professional at any stage; the PNDS also refers to the Alliance Maladies Rares, the Maladies Rares Info Services helpline and Orphanet, and mentions no SWS-specific French association [19]. On the medico-social level, the diagnosis justifies an application for recognition as a long-term condition (ALD, with exemption from co-payment) and, depending on needs, an application to the Maison Départementale des Personnes Handicapées (MDPH) for the disabled child education allowance (AEEH), the disability card, school adaptations with a support assistant (AESH) and school orientation (ULIS, IME, SEGPA), as well as the disability compensation benefit (PCH) and the daily parental presence allowance (AJPP) [19]. Laser sessions are reimbursed, whereas psychomotor therapy and occupational therapy, not reimbursed by the health insurance, may be funded by MDPH notification; rehabilitation takes place in private practice or in dedicated facilities (CAMSP, CMP, SESSAD), and the orientation and coordination platforms (PCO) give children younger than 6 years early access to rehabilitation, without waiting for a complete diagnosis [19]. The attending physician, registered with the health insurance from 16 years, directs the child to experts, provides cautious initial information, ensures specialty follow-up and participates in psychological and social support [19].

10. Illustrative Clinical Case

10.1. Roach type 3 SWS in an adult with epilepsy and postictal headache

Cheema and colleagues described a 30-year-old man presenting to a Lahore (Pakistan) hospital emergency department in November 2024 after two right-sided focal seizures lasting 2–3 minutes, 20 minutes apart, preceded by aura and followed by postictal confusion and weakness, then severe frontal headache, photophobia and vomiting [7]. He had no PWB or ocular involvement. Learning and memory difficulties had been lifelong despite normal developmental milestones, and his first seizure occurred at 4 years; carbamazepine was stopped for allergic rash and replaced by valproate, and right-sided focal seizures continued approximately once yearly or every two years. At 18 years, seizures became more severe with several days' postictal confusion, and headaches, present from childhood and always seizure-associated, lasted 3–4 days after each episode. There was no family history of neurocutaneous disease or epilepsy [7].

He was disorientated and drowsy, scoring 14 of 15 on the Glasgow Coma Scale, with right-sided muscle strength 3 of 5 [7]. Random glucose was 19.2 mmol/L and glycated haemoglobin 11.2%, with glycosuria and ketonuria, revealing undiagnosed diabetes; thyroid, liver and renal tests and full blood count showed no other relevant abnormalities [7]. Unenhanced and contrast-enhanced CT showed thick predominantly left posterior parietal gyral “tram-track” calcifications, cortical atrophy, mild ipsilateral lateral ventricular dilation and enhancing left parietal leptomeningeal angiomas; contrast MRI confirmed left parietal leptomeningeal enhancement with volume loss and dilated venous structures, and CT angiography showed parieto-occipital calcifications and left hemispheric atrophy [7]. Article figure 1, an axial postcontrast T1 MRI, shows enlarged lateral ventricular choroid plexus, leptomeningeal enhancement and cortical atrophy; figure 2, a coronal CT angiogram, shows left hemispheric gyral cortical-subcortical “tram-track” calcifications (figures consulted through captions and full text at https://pmc.ncbi.nlm.nih.gov/articles/PMC12683189/) [7]. No genetic testing was performed.

During admission he received intravenous levetiracetam 500 mg twice daily and intravenous valproate 500 mg three times daily; for analgesic-refractory headache, oral lacosamide 50 mg twice daily and oral topiramate 12.5 mg twice daily were added, with ondansetron as needed, flunarizine 10 mg at bedtime for migraine prophylaxis and ASA 75 mg/day intended to reduce stroke risk [7]. Levetiracetam was reduced for mood changes, and metformin 500 mg twice daily started for diabetes. After one month he had no seizures or headaches, normal strength and controlled blood glucose [7].

10.2. Commentary

This case illustrates SWS without a skin marker, recognised only through neuroimaging and diagnosed late, and prolonged postictal headache with deficit that can be confused with migraine [7, 23]. Imaging findings (leptomeningeal enhancement, enlarged choroid plexus, dilated veins, atrophy and gyral calcifications) correspond to consensus descriptions [20, 21]. Reported management is not a routine model: four-drug antiseizure polytherapy after two brief seizures reflects the report's acute situation; topiramate warrants caution for glaucoma risk, although this patient had no ocular involvement; ASA was used in an adult whereas consensus doses and indications predominantly concern children; and one month's follow-up cannot assess outcome [21, 22, 38]. Absence of genetic testing is consistent with usual non-testing in typical SWS but precludes identifying the causative gene [23].

11. Prognosis and Long-Term Follow-up

11.1. Prognosis

Prognosis depends on the extent of cerebral involvement, age at onset and seizure control. Early onset, high seizure frequency, bilateral involvement, number of lobes affected, frontal involvement and interictal epileptiform abnormalities are associated with poorer cognitive outcome [2, 21, 23, 52]. Intellectual disability affects about 60% of patients, epilepsy is drug-resistant in 20–25%, and patients without PWB have more favourable outcomes [21–23]. In Smith's textbook, 39% of patients have normal intelligence, in an unidentified series [2]. After epilepsy surgery, 70.7% of patients in the Chinese cohort remained seizure-free at 5 years and 78.7% of patients in the individual patient data meta-analysis were seizure-free at last follow-up [85, 86]. In Valery's presymptomatic treatment series, 47% of treated and 12% of untreated patients were seizure-free at 2 years, in a non-randomised study [49]. Glaucoma and amblyopia are the main causes of visual loss and are treatable if detected early [20]. The French PNDS (French National Diagnosis and Care Protocol) summarises the prognosis with estimates from series: epilepsy controlled in 40–85% depending on the series and age, intellectual disability in 32–45% (compared with about 60% in the North American estimate, a difference attributable to definitions and recruitment; section 3.1), motor deficit in 40–50%, generally mild to moderate, and no progressive cognitive decline without drug-resistant epilepsy; it identifies no prognostic factors for headache or stroke-like episodes, and considers the risk of visual disability very low when ocular involvement is strictly unilateral [19].

Adult data are scarce. In a survey of 52 adults, seizures were associated with greater delay, more emotional and behavioural problems and less employment; 39% were financially independent, 55% married or considered capable of marriage and 10 had liveborn children, based on self-report [23]. Approximately half of adults have psychological dysfunction [21]. Cerebral atrophy may also progress in adulthood, and glaucoma can emerge up to the fifth decade [20]. PWB darkens and may become hypertrophic and nodular, with soft tissue and skeletal hypertrophy in 65% by the fifth decade [17, 22]. Consulted sources provide no survival or mortality data.

11.2. Surveillance table

The intervals in the table derive from expert consensuses, including the French PNDS, and, for blood pressure and the calcium-aldosterone axis, from author recommendations based on a retrospective series; none has been prospectively validated [12, 19, 22]. The French PNDS rows diverge from the Italian ones in the pace of initial neurological follow-up (quarterly in the first year versus six-monthly in the first 2 years), in dental follow-up and in the staging of ophthalmological surveillance (section 7.16) [19, 22].

12. Summary and Final Considerations

Sturge-Weber syndrome is a congenital capillary-venous malformation of the face, leptomeninges and eye, caused in most cases by somatic GNAQ NM_002072.5:c.548G>A p.(Arg183Gln), constitutively activating Gαq and present at low allele fractions in affected tissue endothelial cells. The same variant causes isolated port-wine birthmark, with the conditions differing through postzygotic event timing and affected lineages. A minority have GNA11 p.(Arg183Cys), with extensive reticulated capillary malformation, altered limb growth and generally milder neurological manifestations, and one published patient has GNB2 p.(Lys78Glu). As a somatic disease, it is not inherited, molecular testing requires affected tissue and a negative result does not exclude it.

Recognition relies on the distribution of the birthmark: involvement of the forehead, including the upper eyelid, and hemifacial, midline or bilateral birthmarks concentrate the risk of brain disease and glaucoma, and about 10–15% of patients have no birthmark. Epilepsy, usually beginning in the first year, stroke-like episodes, headache, hemiparesis, intellectual disability, behavioural disorders, bimodal glaucoma, choroidal haemangioma and oral and endocrine abnormalities make up the picture, whose severity depends on the extent and laterality of cerebral involvement and on early seizure onset. Contrast MRI is the reference investigation, but it is falsely negative in some infants, which explains the differences between the North American, Italian and London consensuses on presymptomatic imaging.

Treatment is symptomatic and multidisciplinary: seizure action plan, focal-acting antiseizure medicines, early surgery for drug-resistant epilepsy, with a strong American Epilepsy Society recommendation below 36 months, off-label low-dose ASA, early pulsed dye laser for the birthmark, angle surgery for glaucoma in young children and medical treatment for late glaucoma, lifelong ophthalmological surveillance and dental, endocrine and psychological care. Presymptomatic treatment, cannabidiol and sirolimus are supported by small series and open-label trials, and ongoing trials and preclinical studies on the MAPK, calcineurin-NFAT and endothelial permeability pathways define the next steps in research. The 2024 French national protocol converges with the consensuses on most points, but advises MRI and EEG in the first month, prefers medical treatment and filtering surgery for glaucoma in young children, recommends ASA only after the first stroke-like episode and proposes its own schedules for neurological, ophthalmological and dental follow-up, and it is the reference for the organisation of care in France.

13. References

1. Orphanet. Sturge-Weber syndrome (ORPHA:3205). Available from: https://www.orpha.net/en/disease/detail/3205 (epidemiological, natural history and phenotypic data dated 2026-06-23; accessed 2026-10-10).
2. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. Sturge-Weber Syndrome; p. 698–9.
3. Shirley MD, Tang H, Gallione CJ, Baugher JD, Frelin LP, Cohen B, et al. Sturge-Weber syndrome and port-wine stains caused by somatic mutation in GNAQ. N Engl J Med. 2013;368(21):1971-9. doi:10.1056/NEJMoa1213507. PMID: 23656586. PMC3749068.
4. Thorpe J, Frelin LP, McCann M, Pardo CA, Cohen BA, Comi AM, et al. Identification of a Mosaic Activating Mutation in GNA11 in Atypical Sturge-Weber Syndrome. J Invest Dermatol. 2021;141(3):685-688. doi:10.1016/j.jid.2020.03.978. PMID: 32771470. PMC8483769.
5. Dompmartin A, van der Vleuten CJM, Dekeuleneer V, Duprez T, Revencu N, Désir J, et al. GNA11-mutated Sturge-Weber syndrome has distinct neurological and dermatological features. Eur J Neurol. 2022;29(10):3061-3070. doi:10.1111/ene.15452. PMID: 35715928.
6. Fjær R, Marciniak K, Sundnes O, Hjorthaug H, Sheng Y, Hammarström C, et al. A novel somatic mutation in GNB2 provides new insights to the pathogenesis of Sturge-Weber syndrome. Hum Mol Genet. 2021;30(21):1919-1931. doi:10.1093/hmg/ddab144. PMID: 34124757. PMC8522634.
7. Cheema ZZ, Jahan A, Siddiq N, Mumtaz A, Usman M. Type 3 Sturge-Weber Syndrome Presenting With Concurrent Epilepsy and Migraine. Cureus. 2025;17(11):e96262. doi:10.7759/cureus.96262. PMID: 41367481. PMC12683189.
8. GeneGuidance. GNAQ — gene and condition curation. Version of 2026-09-26.
9. GeneGuidance. GNA11 — gene curation. Version of 2026-09-26.
10. GeneGuidance. GNB2 — gene curation. Version of 2026-09-26.
11. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: Accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-8. doi:10.1002/humu.23348. PMID: 28967166. Available from: https://variantvalidator.org (GRCh38, MANE Select; accessed 2026-10-10).
12. Zhang D, Sánchez-Espino LF, Ivars M, Pope E, Nopper AJ, Arkin LM, et al. Phenotypic Spectrum of GNA11 R183C Mosaicism. Pediatr Dermatol. 2025;42(3):475-480. doi:10.1111/pde.15802. PMID: 39654261. PMC12118530.
13. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Glaucoma; p. 428–9.
14. Bayrak-Toydemir P, Stevenson DA. Capillary Malformation-Arteriovenous Malformation Syndrome. 2011 Feb 22 [updated 2019 Sep 12]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–. Available from: https://www.ncbi.nlm.nih.gov/books/NBK52764/ PMID: 21348050.
15. Carter MT, Mirzaa G, McDonell LM, Boycott KM. Microcephaly-Capillary Malformation Syndrome. 2013 Dec 12 [updated 2021 Mar 18]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–. Available from: https://www.ncbi.nlm.nih.gov/books/NBK174452/ PMID: 24354023.
16. Sudarsanam A, Ardern-Holmes SL. Sturge-Weber syndrome: from the past to the present. Eur J Paediatr Neurol. 2014;18(3):257-66. doi:10.1016/j.ejpn.2013.10.003. PMID: 24275166.
17. Sabeti S, Ball KL, Burkhart C, Eichenfield L, Fernandez Faith E, Frieden IJ, et al. Consensus Statement for the Management and Treatment of Port-Wine Birthmarks in Sturge-Weber Syndrome. JAMA Dermatol. 2021;157(1):98-104. doi:10.1001/jamadermatol.2020.4226. PMID: 33175124. PMC8547264.
18. Waelchli R, Aylett SE, Robinson K, Chong WK, Martinez AE, Kinsler VA. New vascular classification of port-wine stains: improving prediction of Sturge-Weber risk. Br J Dermatol. 2014;171(4):861-7. doi:10.1111/bjd.13203. PMID: 24976116. PMC4284033.
19. Haute Autorité de Santé. Protocole National de Diagnostic et de Soins (PNDS): Syndrome de Sturge Weber. Centres de référence Epilepsies Rares (CRéER) et MAGEC; filières DéfiScience et FIMARAD. Saint-Denis La Plaine: HAS; 2024-02-09 (texte de juillet 2023). Available from: https://www.has-sante.fr/jcms/p_3493383
20. Sabeti S, Ball KL, Bhattacharya SK, Bitrian E, Blieden LS, Brandt JD, et al. Consensus Statement for the Management and Treatment of Sturge-Weber Syndrome: Neurology, Neuroimaging, and Ophthalmology Recommendations. Pediatr Neurol. 2021;121:59-66. doi:10.1016/j.pediatrneurol.2021.04.013. PMID: 34153815. PMC9107097.
21. De la Torre AJ, Luat AF, Juhász C, Ho ML, Argersinger DP, Cavuoto KM, et al. A Multidisciplinary Consensus for Clinical Care and Research Needs for Sturge-Weber Syndrome. Pediatr Neurol. 2018;84:11-20. doi:10.1016/j.pediatrneurol.2018.04.005. PMID: 29803545. PMC6317878.
22. El Hachem M, Diociaiuti A, Galeotti A, Grussu F, Gusson E, Ferretti A, et al. Multidisciplinary, multicenter consensus for the care of patients affected with Sturge-Weber syndrome. Orphanet J Rare Dis. 2025;20(1):28. doi:10.1186/s13023-024-03527-w. PMID: 39819452. PMC11740666.
23. Yeom S, Comi AM. Updates on Sturge-Weber Syndrome. Stroke. 2022;53(12):3769-3779. doi:10.1161/STROKEAHA.122.038585. PMID: 36263782. PMC11062639.
24. Jazz Pharmaceuticals. EPIDIOLEX (cannabidiol) oral solution. Prescribing information. NDA 210365. Label effective 2026-05-29. Available from: Drugs@FDA, NDA 210365.
25. European Medicines Agency. Epidyolex (cannabidiol): EPAR. Marketing authorisation 19/09/2019; EC decision 16/07/2026. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/epidyolex
26. VASCERN — Vascular Anomalies Working Group (VASCA-WG). Available from: https://vascern.eu/groupe/vascular-anomalies/
27. EpiCARE — European Reference Network for rare and complex epilepsies. Available from: https://epi-care.eu/
28. ERN-Skin. Available from: https://ern-skin.eu/
29. ERN-ITHACA. Available from: https://ern-ithaca.eu/
30. The Sturge-Weber Foundation. Available from: http://www.sturge-weber.org/
31. Associazione Sturge-Weber Italia. Available from: https://sturgeweberitalia.org/
32. Ribeiro EM, coord. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. Síndrome de Klippel-Trenaunay-Weber; p. 202–5.
33. Garzon MC, Epstein LG, Heyer GL, Frommelt PC, Orbach DB, Baylis AL, et al. PHACE Syndrome: Consensus-Derived Diagnosis and Care Recommendations. J Pediatr. 2016;178:24-33.e2. doi:10.1016/j.jpeds.2016.07.054. PMID: 27659028. PMC6599593.
34. Shah AD, Alexieff P, Tatachar P. Sturge-Weber Syndrome: A Narrative Review of Clinical Presentation and Updates on Management. J Clin Med. 2025;14(7). doi:10.3390/jcm14072182. PMID: 40217631. PMC11989965.
35. Roach ES. Neurocutaneous syndromes. Pediatr Clin North Am. 1992;39(4):591-620. doi:10.1016/s0031-3955(16)38367-5. PMID: 1635798.
36. Lo W, Marchuk DA, Ball KL, Juhász C, Jordan LC, Ewen JB, et al. Updates and future horizons on the understanding, diagnosis, and treatment of Sturge-Weber syndrome brain involvement. Dev Med Child Neurol. 2012;54(3):214-23. doi:10.1111/j.1469-8749.2011.04169.x. PMID: 22191476. PMC3805257.
37. Brain Vascular Malformation Consortium (RDCRN). Available from: https://www.rarediseasesnetwork.org/bvmc
38. Wen T, Wang L, Luo H, Tang L. Sturge-Weber syndrome secondary glaucoma: From Pathogenesis to Treatment. Eye Vis (Lond). 2025;12(1):16. doi:10.1186/s40662-025-00432-6. PMID: 40241121. PMC12004580.
39. Sebold AJ, Day AM, Ewen J, Adamek J, Byars A, Cohen B, et al. Sirolimus Treatment in Sturge-Weber Syndrome. Pediatr Neurol. 2021;115:29-40. doi:10.1016/j.pediatrneurol.2020.10.013. PMID: 33316689. PMC8209677.
40. Xu T, Janssen V, Reinhard NR, Sobrevals-Alcaraz P, van Es RM, de Haan A, et al. Calcineurin-NFAT-DSCR1.4 signaling as druggable axis in Gαq-R183Q-driven capillary malformations. Angiogenesis. 2026;29(2):16. doi:10.1007/s10456-026-10029-9. PMID: 41636844. PMC12872783.
41. Nasim S, Baig M, Wylie-Sears J, Vivero MP, Smits P, Holm A, et al. MAPK signaling and angiopoietin-2 contribute to endothelial permeability in capillary malformations. Proc Natl Acad Sci U S A. 2025;122(47):e2509528122. doi:10.1073/pnas.2509528122. PMID: 41259140. PMC12664009.
42. Truver N, Solomon C, Comi A. Acute Seizure Susceptibility and Chronic Vascular Malformation in a Developmental Mouse Model of Sturge-Weber Syndrome. Int J Mol Sci. 2026;27(8). doi:10.3390/ijms27083519. PMID: 42074164. PMC13115601.
43. ClinVar. VCV000050853 — NM_002072.5(GNAQ):c.548G>A (p.Arg183Gln). Available from: https://www.ncbi.nlm.nih.gov/clinvar/variation/50853/
44. ClinVar. VCV000548668 — NM_002067.5(GNA11):c.547C>T (p.Arg183Cys). Available from: https://www.ncbi.nlm.nih.gov/clinvar/variation/548668/
45. Yeom S, Cohen B, Weiss CR, Montano C, Wohler E, Sobreira N, et al. Genetic testing in the evaluation of individuals with clinical diagnosis of atypical Sturge-Weber syndrome. Am J Med Genet A. 2023;191(4):983-994. doi:10.1002/ajmg.a.63106. PMID: 36710374.
46. ESMO Handbook — Cancer Genomics. Lugano: ESMO Press. Skin cancers: non-melanoma and melanoma; p. 96.
47. Zhou Q, Zheng JW, Yang XJ, Wang HJ, Ma D, Qin ZP. Detection of RASA1 mutations in patients with sporadic Sturge-Weber syndrome. Childs Nerv Syst. 2011;27(4):603-7. doi:10.1007/s00381-010-1258-z. PMID: 20821215.
48. Kossoff EH, Buck C, Freeman JM. Outcomes of 32 hemispherectomies for Sturge-Weber syndrome worldwide. Neurology. 2002;59(11):1735-8. doi:10.1212/01.wnl.0000035639.54567.5c. PMID: 12473761.
49. Valery CB, Iannotti I, Kossoff EH, Zabel A, Cohen B, Ou Y, et al. Retrospective Analysis of Presymptomatic Treatment In Sturge-Weber Syndrome. Ann Child Neurol Soc. 2024;2(1):60-72. doi:10.1002/cns3.20058. PMID: 38745912. PMC11090403.
50. Day AM, Hammill AM, Juhász C, Pinto AL, Roach ES, McCulloch CE, et al. Hypothesis: Presymptomatic treatment of Sturge-Weber Syndrome With Aspirin and Antiepileptic Drugs May Delay Seizure Onset. Pediatr Neurol. 2019;90:8-12. doi:10.1016/j.pediatrneurol.2018.04.009. PMID: 30482419. PMC7288478.
51. Tillmann RP, Ray K, Aylett SE. Transient episodes of hemiparesis in Sturge Weber Syndrome – causes, incidence and recovery. Eur J Paediatr Neurol. 2020;25:90–96. doi:10.1016/j.ejpn.2019.11.001. PMID: 31894017.
52. Bosnyák E, Behen ME, Guy WC, Asano E, Chugani HT, Juhász C. Predictors of Cognitive Functions in Children With Sturge-Weber Syndrome: A Longitudinal Study. Pediatr Neurol. 2016;61:38-45. doi:10.1016/j.pediatrneurol.2016.05.012. PMID: 27353695. PMC4983234.
53. Turin E, Grados MA, Tierney E, Ferenc LM, Zabel A, Comi AM. Behavioral and psychiatric features of Sturge-Weber syndrome. J Nerv Ment Dis. 2010;198(12):905-13. doi:10.1097/NMD.0b013e3181fe75ee. PMID: 21135644.
54. Sujansky E, Conradi S. Outcome of Sturge-Weber syndrome in 52 adults. Am J Med Genet. 1995;57(1):35–45. doi:10.1002/ajmg.1320570110. PMID: 7645596.
55. Kavanaugh B, Sreenivasan A, Bachur C, Papazoglou A, Comi A, Zabel TA. Intellectual and adaptive functioning in Sturge-Weber Syndrome. Child Neuropsychol. 2016;22(6):635–648. doi:10.1080/09297049.2015.1028349. PMID: 25952468.
56. Aggarwal NK, Gandham SB, Weinstein R, Saltzmann R, Walton DS. Heterochromia iridis and pertinent clinical findings in patients with glaucoma associated with Sturge-Weber syndrome. J Pediatr Ophthalmol Strabismus. 2010;47(6):361-5. doi:10.3928/01913913-20100218-01. PMID: 20210282.
57. Choudhary T, Senthil S. Visual field loss in Sturge-Weber syndrome and phakomatosis pigmentovascularis: Implications for glaucoma assessment. Indian J Ophthalmol. 2026;74(10):1525-1530. doi:10.4103/IJO.IJO_1518_26. PMID: 42809258.
58. Longo R, Gusson E, Lorenzetto E, Polinelli L, Malvasi M, Panozzo G, et al. Retinal Vascular Density and Vessel Diameter in Sturge-Weber Syndrome Assessed by OCT-Angiography. J Clin Med. 2025;14(19). doi:10.3390/jcm14197061. PMID: 41096140. PMC12524786.
59. Miller RS, Ball KL, Comi AM, Germain-Lee EL. Growth hormone deficiency in Sturge-Weber syndrome. Arch Dis Child. 2006;91(4):340-1. doi:10.1136/adc.2005.082578. PMID: 16551788. PMC2065976.
60. Comi AM, Bellamkonda S, Ferenc LM, Cohen BA, Germain-Lee EL. Central hypothyroidism and Sturge-Weber syndrome. Pediatr Neurol. 2008;39(1):58-62. doi:10.1016/j.pediatrneurol.2008.03.018. PMID: 18555176.
61. Hadid SA, Noor L, Baer T, Jacobson RI, Brutsaert E. Hypothalamic-pituitary dysfunction in Sturge-Weber syndrome: case report and review of the literature. J Pediatr Endocrinol Metab. 2024;37(1):80-83. doi:10.1515/jpem-2023-0408. PMID: 38006605.
62. Lance EI, Sreenivasan AK, Zabel TA, Kossoff EH, Comi AM. Aspirin use in Sturge-Weber syndrome: side effects and clinical outcomes. J Child Neurol. 2013;28(2):213-8. doi:10.1177/0883073812463607. PMID: 23112247. PMC4373084.
63. Abbas K, Harrison B, Peter Chang TC, Edmunds B, Hammond B, Lueder GT, et al. Glaucoma Management in Sturge-Weber Syndrome Using the Delphi Process. Ophthalmol Glaucoma. 2026;9(3):336-342. doi:10.1016/j.ogla.2025.11.003. PMID: 41314339.
64. DéfiScience — Filière de santé maladies rares du neurodéveloppement. Carte d'urgence: Syndrome de Sturge Weber (ORPHA:3205). 2020 (2022 print version). Available from: https://defiscience.fr/app/uploads/2022/07/06-carte-durgence-syndrome-de-sturge-weber-print.pdf
65. Peret A, Arkin LM, Miller A, Drolet BA, Rebsamen SL, Brucker JL, et al. Imaging Sturge-Weber Syndrome with the Feed-and-Swaddle FAST Protocol. AJNR Am J Neuroradiol. 2026;. doi:10.3174/ajnr.A9608. PMID: 42648909.
66. Triana Junco PE, Sánchez-Carpintero I, López-Gutiérrez JC. Preventive treatment with oral sirolimus and aspirin in a newborn with severe Sturge-Weber syndrome. Pediatr Dermatol. 2019;36(4):524-527. doi:10.1111/pde.13841. PMID: 31134637.
67. Metry D, Heyer G, Hess C, Garzon M, Haggstrom A, Frommelt P, et al. Consensus Statement on Diagnostic Criteria for PHACE Syndrome. Pediatrics. 2009;124(5):1447-56. doi:10.1542/peds.2009-0082. PMID: 19858157.
68. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. Parkes Weber Syndrome (Capillary Malformation-Arteriovenous Malformation Syndrome, RASA1 Related Disorders); p. 728–31.
69. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. Klippel-Trenaunay Syndrome; p. 724–7.
70. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Hemihypertrophy and limb asymmetry; p. 164–7.
71. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Lissencephaly, polymicrogyria and neuronal migration; p. 216–21.
72. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Epilepsy in adults; p. 416–9.
73. Freedman D, Babatunde I, Morgan RL, Abu-Sawwa R, Albert D, Behbahani M, et al. American Epilepsy Society Clinical Practice Guideline: Infantile Epilepsy. Epilepsy Curr. 2026;26(3):174-202. doi:10.1177/15357597261433266. PMID: 41969626. PMC13061853.
74. European Medicines Agency. Buccolam (midazolam) oromucosal solution: summary of product characteristics. EU/1/11/709. Date of first authorisation 05/09/2011. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/buccolam
75. Diazepam rectal gel (reference Diastat, NDA 020648). Prescribing information, ANDA 217468 (Zydus Pharmaceuticals USA), label effective 2026-04-30. Available from: openFDA / DailyMed, set id 652714c7-25c4-4d05-979d-8eee3b6f2f4d.
76. European Medicines Agency. Diacomit (stiripentol): summary of product characteristics. Date of first authorisation 04/01/2007. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/diacomit
77. European Medicines Agency. Valproate and related substances — referral (Article 31): new measures to avoid valproate exposure in pregnancy endorsed. First published 07/06/2018. Available from: https://www.ema.europa.eu/en/medicines/human/referrals/valproate-related-substances-0
78. Novadoz Pharmaceuticals. Oxcarbazepine oral suspension (ANDA 209020; reference Trileptal). Prescribing information, label effective 2024-11-18. Available from: Drugs@FDA, ANDA 209020.
79. UCB. KEPPRA (levetiracetam) tablets and oral solution. Prescribing information. NDA 021035/021505. Label effective 2026-01-28. Available from: Drugs@FDA, NDA 021035.
80. Smegal LF, Vedmurthy P, Ryan M, Eagen M, Andrejow NW, Sweeney K, et al. Cannabidiol Treatment for Neurological, Cognitive, and Psychiatric Symptoms in Sturge-Weber Syndrome. Pediatr Neurol. 2023;139:24-34. doi:10.1016/j.pediatrneurol.2022.10.014. PMID: 36508880.
81. ClinicalTrials.gov. NCT00639730 — Use of the Atkins Diet for Children With Sturge Weber Syndrome. Johns Hopkins University. Available from: https://clinicaltrials.gov/study/NCT00639730
82. ClinicalTrials.gov. NCT07778810 — Epidiolex Pilot Trial for Presymptomatic Treatment of Sturge-Weber Syndrome. Johns Hopkins University. Available from: https://clinicaltrials.gov/study/NCT07778810
83. Bay MJ, Kossoff EH, Lehmann CU, Zabel TA, Comi AM. Survey of aspirin use in Sturge-Weber syndrome. J Child Neurol. 2011;26(6):692-702. doi:10.1177/0883073810388646. PMID: 21427442.
84. Frank NA, Greuter L, Dill PE, Guzman R, Soleman J. Focal lesionectomy as surgical treatment of epilepsy in patients with Sturge-Weber syndrome: a case-based systematic review and meta-analysis. Neurosurg Focus. 2022;52(5):E4. doi:10.3171/2022.2.FOCUS21788. PMID: 35535828.
85. Mozaffari K, Krishnakumar A, Chen JS, Goel K, Wang A, Shlobin NA, et al. Seizure outcomes in children with Sturge-Weber syndrome undergoing epilepsy surgery: An individual participant data meta-analysis. Seizure. 2023;107:43-51. doi:10.1016/j.seizure.2023.03.008. PMID: 36958063.
86. Wang S, Liu QZ, Zhao R, Zhai X, Zhang K, Cai L, et al. Seizure, Motor, and Cognitive Outcomes After Epilepsy Surgery for Patients With Sturge-Weber Syndrome: Results From a Multicenter Study. Neurology. 2024;103(1):e209525. doi:10.1212/WNL.0000000000209525. PMID: 38875518. PMC11244739.
87. ClinicalTrials.gov. NCT07655492 — Developing a Tailored Neuropsychological Rehabilitation for SWS. Sapienza University of Rome. Available from: https://clinicaltrials.gov/study/NCT07655492
88. Allergan. BOTOX (onabotulinumtoxinA) for injection. Prescribing information. BLA 103000. Label effective 2023-11-18. Available from: openFDA / Drugs@FDA, BLA 103000.
89. Zhao Y, Tu P, Zhou G, Zhou Z, Lin X, Yang H, et al. Hemoporfin Photodynamic Therapy for Port-Wine Stain: A Randomized Controlled Trial. PLoS One. 2016;11(5):e0156219. doi:10.1371/journal.pone.0156219. PMID: 27227544. PMC4881994.
90. Nguyen L, Sorbe C, Schoen G, Schneider SW, Herberger K. Laser and light-based treatments for port-wine birthmarks - a systematic review and network meta-analysis. J Dtsch Dermatol Ges. 2025;23(3):293-301. doi:10.1111/ddg.15612. PMID: 39668417. PMC11887003.
91. Marqués L, Núñez-Córdoba JM, Aguado L, Pretel M, Boixeda P, Nagore E, et al. Topical rapamycin combined with pulsed dye laser in the treatment of capillary vascular malformations in Sturge-Weber syndrome: phase II, randomized, double-blind, intraindividual placebo-controlled clinical trial. J Am Acad Dermatol. 2015;72(1):151-8.e1. doi:10.1016/j.jaad.2014.10.011. PMID: 25455610.
92. Lidocaine and prilocaine cream 2.5%/2.5% (reference EMLA). Prescribing information, ANDA 219120 (Viona Pharmaceuticals), label effective 2025-12-01. Available from: openFDA / DailyMed, ANDA 219120.
93. Timolol maleate ophthalmic solution, preservative-free (Timoptic in Ocudose). Prescribing information, ANDA 212592 (Nordic Pharma), label effective 2026-08-06. Available from: openFDA / DailyMed, ANDA 212592.
94. Dorzolamide hydrochloride ophthalmic solution 2% (reference Trusopt). Prescribing information, ANDA 215660 (Sagent Pharmaceuticals), label effective 2026-01-09; dorzolamide-timolol ophthalmic solution, ANDA 202054, label effective 2026-06-12. Available from: openFDA / DailyMed.
95. Kalantary A, Halbert A, Franchina M, Heath Jeffery RC, O'Day R, Chen FK. Oral Sirolimus for Diffuse Choroidal Hemangioma in Sturge-Weber Syndrome. JAMA Ophthalmol. 2024;142(6):585-587. doi:10.1001/jamaophthalmol.2024.0854. PMID: 38635233.
96. Kaplan EH, Offermann EA, Sievers JW, Comi AM. Cannabidiol Treatment for Refractory Seizures in Sturge-Weber Syndrome. Pediatr Neurol. 2017;71:18-23.e2. doi:10.1016/j.pediatrneurol.2017.02.009. PMID: 28454984.
97. ClinicalTrials.gov. NCT02332655 — Cannabidiol Expanded Access Study in Medically Refractory SWS. Available from: https://clinicaltrials.gov/study/NCT02332655
98. ClinicalTrials.gov. NCT04447846 — Novel Cognitive Treatment Targets for Epidiolex in SWS (Phase II). Available from: https://clinicaltrials.gov/study/NCT04447846
99. Gupta SS, Joslyn KE, McKenney KD, Comi AM. Biomarker development in Sturge-Weber syndrome. J Neurodev Disord. 2025;17(1):50. doi:10.1186/s11689-025-09640-6. PMID: 40851064. PMC12376483.
100. McKenney KD, Kalb LG, Hammill AM, Comi AM. Development and Retrospective Application of Novel Outcome Measure: Sturge-Weber Syndrome Acute Crisis (SWAC) Index. Ann Child Neurol Soc. 2026;4(2):132-139. doi:10.1002/cns3.70059. PMID: 42563839. PMC13359142.
101. Pervan-Steel CL, Roy Chowdhury U, Sookdeo HK, Casale RA, Dosa PI, Htoo TM, et al. Ocular Hypotensive Properties and Biochemical Profile of QLS-101, a Novel ATP-Sensitive Potassium (KATP) Channel Opening Prodrug. Invest Ophthalmol Vis Sci. 2022;63(4):26. doi:10.1167/iovs.63.4.26. PMID: 35486069. PMC9055548.
102. ClinicalTrials.gov. NCT05495269 — QLS-101 in Adolescents With SWS-Related Glaucoma Due to Elevated EVP. Qlaris Bio. Available from: https://clinicaltrials.gov/study/NCT05495269
103. ClinicalTrials.gov. NCT04947124 — QLS-101 1% vs 2% in Adults With SWS-Related Glaucoma. Qlaris Bio. Available from: https://clinicaltrials.gov/study/NCT04947124
104. Sirolimus tablets/oral solution (ANDA201676; reference Rapamune). Prescribing information, label effective 2025-06-30. Available from: Drugs@FDA, ANDA 201676.
105. European Medicines Agency. Rapamune (sirolimus): EPAR. Marketing authorisation 13/03/2001; EC decision 03/06/2026. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/rapamune
106. ClinicalTrials.gov. NCT03047980 — Trial of Sirolimus for Cognitive Impairment in SWS. Available from: https://clinicaltrials.gov/study/NCT03047980
107. ClinicalTrials.gov. NCT01997255 — Adjunctive Everolimus (RAD 001) Therapy for Epilepsy in Children With Sturge-Weber Syndrome (SWS). Baylor College of Medicine. Available from: https://clinicaltrials.gov/study/NCT01997255
108. ClinicalTrials.gov. NCT01533376 — Treatment of Port-wine Mark in Sturge-Weber Syndrome Using Topical Timolol. Wills Eye. Available from: https://clinicaltrials.gov/study/NCT01533376
109. ClinicalTrials.gov. NCT01425944 — Innovative Approaches to Gauge Progression of Sturge-Weber Syndrome. Hugo W. Moser Research Institute at Kennedy Krieger. Available from: https://clinicaltrials.gov/study/NCT01425944
110. Olson J, McKee H. Pregnancy and refractory epilepsy in Sturge-Weber syndrome: Challenges and outcomes. Epileptic Disord. 2023;25(1):87-93. doi:10.1002/epd2.20058. PMID: 37032541.
111. Byrne B, O'Carroll RF, Doherty C. Sturge-Weber syndrome: Case report, literature review and proposed pregnancy care plan. Obstet Med. 2022;15(4):276-279. doi:10.1177/1753495X211002696. PMID: 36523885. PMC9745603.