Proteus Syndrome

AKT1-related Proteus syndrome and AKT1-related overgrowth spectrum

Literature and regulatory information updated: 29 October 2026.

0. Initial Reference Table

1. Introduction

1.1. Original description and name

Michael Cohen Jr. and Peter Hayden described a “newly recognised hamartomatous syndrome” in 1979 [19], and in 1983 Hans-Rudolf Wiedemann and colleagues brought together patients with partial gigantism of the hands and feet, naevi, hemihypertrophy, subcutaneous tumours, macrocephaly or other cranial anomalies and visceral involvement, which they named Proteus syndrome, after the polymorphous Greek god, because of the variability and the change in phenotype over time [20, 21]. One of the NIH imaging series attributes the first modern description to Temtamy and Rogers, in 1976 [7, 22]. Michael Cohen proposed that Joseph Carey Merrick, the nineteenth-century Englishman known as the “Elephant Man”, had Proteus syndrome rather than neurofibromatosis, a hypothesis now accepted by the GeneReviews authors and by the authors of the article that identified the molecular cause [1, 8, 21]. The expression “Elephant Man disease” should be restricted to the historical context [1].

1.2. Clinical criteria and separation from other forms of overgrowth

For two decades Proteus syndrome was a diagnosis of exclusion for almost all forms of asymmetric overgrowth, which generated diagnostic confusion and inconsistent natural history data. In March 1998, a workshop at the NIH established diagnostic criteria, a differential diagnosis and an evaluation protocol, published by Biesecker and colleagues in 1999 [23]. These criteria required mosaic distribution, a progressive course and sporadic occurrence, together with one category A criterion (cerebriform connective tissue naevus), two category B criteria or three category C criteria [23, 24]. They were revised in 2006 [2, 25]. Strict application of these criteria made it possible to separate from Proteus syndrome entities such as hemihyperplasia with multiple lipomatosis (1998), CLOVE syndrome (2007), fibroadipose overgrowth (2012) and isolated macrodactyly, later grouped in the PIK3CA-related overgrowth spectrum (PROS) [1, 2]. Some patients formerly described as having “Proteus syndrome” with germline PTEN variants are now considered to have PTEN hamartoma tumour syndrome (PHTS); Orphanet keeps the entity “Proteus-like syndrome” (ORPHA:2969) for them [9, 21, 26].

1.3. Molecular discovery

The absence of familial recurrence and the discordance between monozygotic twins supported the hypothesis formulated by Rudolf Happle in 1986: the syndrome would result from somatic mosaicism for a mutation that is lethal in the constitutional state [1, 8]. In 2011, Lindhurst and colleagues sequenced the exomes of paired affected and unaffected samples from patients followed at the NIH and found the activating variant c.49G>A, p.(Glu17Lys) of the AKT1 oncogene in 26 of 29 patients (90%), with mutant allele fractions between 1% and 47% in tissues and cell lines and increased activation of the signalling pathway [8]. In 2019, Sapp and colleagues replaced the phenotypic criteria with a weighted points system integrated with the molecular result, which defines AKT1-related Proteus syndrome and the AKT1-related overgrowth spectrum [2]. The first alternative variant, c.49_50delinsAG, p.(Glu17Arg), was described in 2020, and an intragenic duplication in 2021 [1, 7].

1.4. Entities within scope

This article covers Proteus syndrome in the strict sense (with or without molecular confirmation) and AROS, which share the mechanism and management. Conditions with an overlapping phenotype but a different cause (PROS, PHTS, Beckwith-Wiedemann syndrome, neurofibromatosis type 1) appear in the differential diagnosis. The general approach to overgrowth syndromes and cutaneous skeletal hypophosphataemia syndrome, caused by mosaic HRAS/NRAS variants, are mentioned in the differential diagnosis in section 6 [27].

2. Molecular Genetics and Pathophysiology

2.1. The AKT1 gene and the PI3K-AKT-mTOR pathway

AKT1 (14q32.33) encodes one of the three AKT serine/threonine kinases, downstream effectors of phosphatidylinositol 3-kinase (PI3K) that phosphorylate a large set of substrates and control glucose uptake, proliferation, cell survival, growth and angiogenesis [1, 4]. The pathway transmits signals from growth factor receptor tyrosine kinases and inhibits apoptosis [1]. The Oxford Desk Reference organises the overgrowth syndromes into two families, the epigenetic regulators and the PI3K/mTOR pathway; in the latter, overgrowth results from activating variants in AKT1, AKT3, PIK3CA or PIK3R2, or from loss of function of inhibitors such as PTEN [27]. In the same pathway, activating AKT2 variants cause adipose tissue dysregulation and hypoglycaemia, and AKT3 variants cause hemimegalencephaly [1].

2.2. Mechanism of the p.(Glu17Lys) variant

The p.(Glu17Lys) substitution lies in the lipid-binding pocket of the pleckstrin homology (PH) domain. It alters the electrostatic interactions of this pocket, creates new hydrogen bonds with the phosphoinositide ligand and activates the kinase through pathological localisation to the plasma membrane, independently of the ligand [1, 4]. The mechanism is gain of function [1]. In patient cells, mutant lines showed more AKT phosphorylation at serine 473 than control lines, and two single-cell clones derived from the same culture, which differed only in mutational status, also differed in phosphorylation [8]. The p.(Glu17Arg) variant is likewise constitutively activating: in fibroblasts from an overgrown digit, phosphorylation increased between 2- and 25-fold relative to fibroblasts without the variant, in proportion to the allele fraction [7].

2.3. Mosaicism and lethality of the constitutional state

All variants associated with the syndrome are postzygotic and are present in a variable fraction of cells [1]. In the 2011 series, the variant was detected in 75 of 97 affected samples, in 11 of 20 samples of uncertain classification and in 13 of 41 clinically unaffected samples, showing that the absence of a phenotype in a tissue does not exclude the presence of the variant [8]. The autopsy study of one patient confirmed an overall, but not absolute, correlation between allele fraction and overgrowth: some overgrown tissues had no detectable variant and others, apparently normal, contained it [1]. No known patient has a body burden close to 50%, which is attributed to the embryonic lethality of the non-mosaic state [2]. The 2020 ERN-Skin consensus redefined cutaneous mosaicism for the genomic era, because the classical definition (coexistence of cells with at least two genotypes in an individual derived from a single zygote) encompasses the somatic mutations that everyone accumulates before and after birth: it calls a mosaic skin anomaly the coexistence of genotypes already established at birth that produces a disease phenotype, and a mosaic disorder the situation in which the postzygotic variant explains the entire phenotype, without requiring it to be visible at birth [28]. Proteus syndrome meets this second definition and is the example chosen by the consensus for the class “autosomal dominant mosaic anomaly, lethal in the germline”: normal germline genotype and a single heterozygous postzygotic variant in utero that would be lethal in the constitutional state and is therefore not transmitted [28]. The classification by genetic mechanism and transmission potential was proposed because it is clinically useful, as it defines the sample to collect, the type of test and the counselling (sections 5.3 and 8.1) [28]. In mice, ubiquitous expression of Akt1 p.(Glu17Lys) causes vascular defects and early embryonic lethality [1]. A 2026 mouse model with endogenously regulated mosaic expression showed that miransertib crosses the placenta and that prenatal treatment of the dams with 60 or 90 mg/kg/day increased neonatal survival (15 of 63 versus 6 of 152; p=0.0003); this is a preclinical result [29].

2.4. Variant spectrum and nomenclature

The conversion between NM_005163.2 and the MANE Select transcript NM_001382430.1 was confirmed in VariantValidator, which gives the same c.49 position in both transcripts (GRCh37 coordinate: NC_000014.8:g.105246551C>T) [5]. ClinVar classifies c.49G>A as pathogenic (multiple submitters, no conflicts; record VCV000013983) and, in the somatic context, as oncogenic with tier I clinical impact [30]. There are no founder variants, because the disease is not transmitted.

2.5. Genotype–phenotype correlations, penetrance and expressivity

As almost all patients have the same variant, phenotypic variability results from the tissue distribution and fraction of the mutant clone, not from the allele [1, 2]. The NIH authors described a minimally affected man, with manifestations only in the legs and a low allele fraction, and acknowledge that people with even more limited mosaicism may exist [2]. Penetrance cannot in practice be assessed in a non-inherited mosaic disorder; in theory, a very low fraction might not cause symptoms [1]. The only patient with p.(Glu17Arg) studied in detail had severe disease and died at 7.5 years of bullous and restrictive lung disease, but a single case does not allow a correlation to be established [7]. In the series of children with plantar cerebriform naevus, the growth rate of the naevus showed a non-significant positive correlation with allele fraction (coefficient 0.62; p=0.10; n=8) [31].

2.6. Allelic and germline disorders

The same c.49G>A variant occurs as an isolated somatic event in several cancers, mainly breast cancer, and also colon, ovarian and endometrial cancer, and in 6–9% of sporadic non-syndromic meningiomas, mainly of the meningothelial and transitional subtypes [1, 2, 4, 32]. It has also been described in sclerosing pneumocytoma [1]. The NIH authors consider that a sporadic meningothelial meningioma with the variant could correspond to a minimal form of Proteus syndrome, but assign it only 1 point in the criteria, insufficient for either designation [2, 32]. In hormone receptor-positive, HER2-negative advanced breast cancer, PIK3CA, AKT1 or PTEN alterations are a biomarker for capivasertib (Truqap®) combined with fulvestrant, an oncological indication unrelated to the syndrome [4, 33]. OMIM also records an autosomal dominant germline association, Cowden syndrome type 6 (OMIM 615109), whose gene–disease validity was classified as limited by ClinGen in November 2024 [4, 34].

3. Clinical Spectrum and Manifestations by System

3.1. Data sources and overall natural history

Almost all quantitative data come from the NIH natural history cohort (protocol 94-HG-0132, NCT00001403, active since 1994), which brings together North American and other patients evaluated at a single centre; the following figures describe that cohort, with a probable bias towards more severely affected patients, and not the general population [11, 35]. Most patients have minimal or absent manifestations at birth, such as subtle asymmetry or faint linear naevi that go unnoticed [1]. Asymmetric overgrowth typically begins between 6 and 18 months, most often in the feet or hands, and onset has been described up to 12 years; it progresses rapidly and continuously during childhood and stabilises after adolescence [1, 21, 24]. A minority (probably fewer than 5%) present prenatally with hemimegalencephaly, often with neuronal migration defects and later intellectual disability, and there is a single report of generalised overgrowth diagnosed before birth [1]. In the cohort of 64 patients analysed by Sapp and colleagues, the median age at diagnosis was about 19 months, 75% had been diagnosed by 3.2 years and the latest diagnosis occurred at 24 years [11]. Final height is normal, as is somatic growth in adolescence [21].

3.2. Skeleton

Bony overgrowth is the manifestation that distinguishes the syndrome from other forms of overgrowth: it is postnatal, rapid, progressive and distorts bone architecture, especially in periarticular regions [1, 36]. In an affected area, the typical progression is a 15% increase at 1 year of age, 30% at 3 years and 100% at 6 years [1]. On radiographs, the tubular bones of the limbs, the vertebrae and the skull develop irregular calcified overgrowth that, over time, can make the bone unrecognisable; limb length discrepancies of 20 cm and scolioses greater than 90° have been described [1]. Any bone may be affected [1]. Smith's Recognizable Patterns of Human Malformation lists hemihypertrophy, scoliosis, kyphosis, hip dislocation, angular knee deformities, hallux and foot valgus, macrodactyly, clinodactyly, skull hyperostosis, vertebral bodies with asymmetric overgrowth and posterior scalloping (megaspondylodysplasia), segmentation defects, early degenerative changes and coarse ribs and scapulae [21]. Overgrowth of the patella, often with severe knee flexion contracture, and iliotibial band contractures that worsen the valgus are frequent; cervical scoliosis, rare in idiopathic scoliosis, is common and may compromise the spinal canal and the trachea [36]. Scoliosis in the syndrome has an age at onset similar to idiopathic scoliosis but can progress very rapidly; a girl with scoliosis at 8 years died at 15 of respiratory failure [36]. In the NIH cardiothoracic imaging series, scoliosis due to bony overgrowth was present in 94% of the 38 patients studied [22]. Cranial hyperostosis was present in 29 of 61 patients (48%; ages 2 to 66 years), predominantly frontal or frontoparietal (23 of 29), unilateral and parasagittal, due to expansion of the diploë [32]. Bone Dysplasias states that melorheostosis-like anomalies of the axial skeleton associated with fibrolipomatous lesions should raise suspicion of the syndrome [37].

3.3. Skin

The cerebriform connective tissue naevus (CCTN) is present in most patients and is almost pathognomonic: a firm lesion with deep grooves and gyri resembling those of the cerebral cortex, most frequent on the sole of the foot and also on the hands, the alae nasi, the ears and near the lacrimal puncta [1]. It is rare in infancy, appears during childhood and progresses until the end of adolescence, rarely thereafter [1]. In a series of 152 photographs of 22 patients (median age 6.9 years), the relative area of plantar CCTN increased by 5.6% per year in children (95% confidence interval, 95% CI, 3.7–7.6), so that a naevus covering 20% of the sole would approach 50% within five years; the confluent papules and nodules that precede it are a precursor lesion, and the centre of the lesions converges near the metatarsal heads, suggesting a contribution from mechanical stress [31]. By late adolescence, the grooves become deep enough to hinder hygiene and cause malodour, and there is a risk of pressure ulcers [1]. The naevus should not be confused with the marked plantar or palmar creasing of other forms of overgrowth [1]. The linear verrucous epidermal naevus, pigmented and rough, follows the lines of Blaschko, appears in the first months and remains stable [1, 24]. Hypertrichosis or acne along the lines of Blaschko are less frequent but relatively specific findings [1]. Smith also describes generalised skin thickening, flat non-organoid epidermal naevi and, occasionally, café-au-lait macules and gyriform hyperplasia on the nose [21]. The NIH cross-sectional hair study (patients evaluated between 1996 and 2019, all with AKT1 c.49G>A confirmed in tissue; 45 of 51 with sufficient records or photographs; 28 male; median 7 years, 2–40) found asymmetric body hypertrichosis in 29 of 45 (64%): unilateral blaschkoid patches over normal skin or over epidermal naevus in 16 (36%), unilateral non-blaschkoid patches in 11 (24%), some over CCTN, epidermal naevus or capillary-venous malformation, and hypertrichosis of most of a limb in 10 (22%), almost always (9 of 10) on an overgrown limb [38]. Blaschkoid hypertrichosis, which GeneReviews lists among the less frequent findings, thus affected one-third of this reference cohort, suggesting that its frequency was underestimated in earlier descriptions, based on isolated cases without molecular confirmation [1, 38]. Diffuse or patchy changes in hair density or colour occurred in 11 (24%): five with altered density, including alopecia over an exostosis and one aplasia cutis congenita, and four with depigmented locks or hairs [38]. The retrospective design and limited longitudinal follow-up are the stated limitations, and the frequency may be underestimated because it depends on the records [38]. The ERN-Skin table of mosaic skin disorders summarises the cutaneous phenotype of the syndrome as keratinocytic epidermal naevus, vascular malformations, lipomas and CCTN, with progressive postnatal overgrowth as the dominant extracutaneous manifestation [28].

3.4. Adipose tissue

Adipose tissue overgrowth arises mainly in infancy and may appear at new sites throughout childhood and young adulthood; many patients also have marked regional lipoatrophy, and both coexist in the same patient [1]. These masses are not encapsulated lipomas, despite the common use of the term [1]. Smith reports asymmetric subcutaneous adipose overgrowth mainly on the trunk and abdominal, pelvic, pancreatic and spinal lipomatosis [21]. Fatty infiltration of the myocardium, especially of the interventricular septum, was found by echocardiography in 7 of 20 patients (35%) and by cardiac magnetic resonance imaging (MRI) in 9 of 20 (45%), without changes in ventricular dimensions or function (mean ejection fraction 60%) [22, 39]. Non-gated chest computed tomography (CT) detected this fat in only 7 of 34 patients (21%) [22].

3.5. Vessels and thromboembolism

Cutaneous capillary malformations, a prominent venous pattern and varicosities are frequent; large and complex malformations affect some patients, and lymphatic malformations, which can progress, often accompany lipomatous overgrowth [1]. Arteriovenous malformations are uncommon [1]. The table of the genetics of lymphatic anomalies reproduced in the Oxford Desk Reference includes Proteus syndrome among the somatic AKT1 gain-of-function causes that may present with lymphoedema [40]. Unlike other forms of overgrowth, skeletal overgrowth also occurs in regions without vascular malformation [1]. In a prospective series of 22 patients (median 13.5 years), 18 (82%) had venous malformations, superficial in all and deep or visceral in half; 4 of these 18 had a history of DVT in an affected limb and 5 of the 22 (23%) of superficial thrombophlebitis [41]. Magnetic resonance angiography and venography showed, for example, a diffusely enlarged inferior vena cava, portal vein dilatation and gluteal varicosities [41]. In the series of 57 patients evaluated between 1997 and 2012, 6 of the 10 deaths were caused by DVT with bilateral PE, at a median age of 17 years, and 6 of the 47 living patients had had non-fatal thromboembolism, always postoperative and in an affected limb [15]. DVT presents with a palpable subcutaneous cord, oedema, erythema, pain and distal venous congestion; PE presents with dyspnoea, chest pain, cough and haemoptysis, and may be asymptomatic; the rarity of these thromboses in children delays diagnosis [1]. The hypercoagulability panel was abnormal in 11 of 21 patients (52%) in the first series and in 16 of 22 (73%) in the second, with a positive lupus anticoagulant in half, without a single pattern [15, 41]. The endothelial markers thrombomodulin, sVCAM and P-selectin were increased relative to controls, suggesting a contribution from AKT-dependent endothelial dysfunction [41].

3.6. Lung and chest

In a review of 39 patients, cystic lung disease affected 26 (67%), with a mean age of 17.1 years; 81% of those affected had heterogeneous emphysema-like regions, 12% mass effect and one had pneumothorax; 38% reported dyspnoea and 35% respiratory infections, and 96% had scoliosis and abnormal pulmonary function tests [42]. The disease progressed in 7 of 10 patients followed up, including all 5 aged under 20 years, and three improved after lung resection [42]. In the cardiothoracic imaging series, the most frequent findings were pulmonary venous dilatation (62%), band-like pulmonary scarring (56%) and hyperlucent parenchyma (50%) [22]. Bullous disease presents with exercise intolerance or as an imaging finding and can progress within a few months; the association with scoliosis produces restrictive lung disease [1, 36]. Smith reports emphysema, cysts and pulmonary scarring and associates cystic emphysematous disease with premature death [21].

3.7. Viscera

Overgrown tonsils and adenoids, often unilateral, splenomegaly and asymmetric enlargement of the kidneys and testes are frequent [1]. Smith also describes nephromegaly, hydronephrosis, renal calculi and haemangiomas, gastromegaly, colonic polyps, an enlarged thymus, hypertrophic cardiomyopathy and conduction disorders [21]. In the abdominal imaging series of 38 patients, 11 (29%) had focal hepatosplenic lesions, 9 splenomegaly, 12 portal vein dilatation and 4 hepatomegaly; splenomegaly was associated with lower platelet counts (179 ± 87 versus 253 ± 57 × 10³/µL), without signs of portal hypertension, and there were no visceral vascular malformations in patients without cutaneous malformations [43]. Severe and complex overgrowth of the uterus, cervix and ovaries, difficult to distinguish from neoplasia, has been described [1].

3.8. Tumours

The GeneReviews authors acknowledge a modest but significant increase in many tumour types, mostly benign, rather than a restricted set [1]. Meningiomas are the most frequent intracranial tumour: in the series of 61 patients they were found in four (at 9, 19, 24 and 34 years; two with multiple tumours), almost all adjacent to the frontal hyperostosis, and in one of them the variant was present in the tumour and in the overlying bone; in the syndrome, diagnosis usually occurs before 30 years, unlike the usual 40–70 years in the general population [21, 32]. Ovarian cystadenomas and monomorphic adenoma of the parotid at an early age are part of the criteria [2]. Among 64 men with a confirmed diagnosis, 12 (19%) underwent surgery for paratesticular masses, at a mean age of 9 years; the tumours were usually unilateral, small (median 1.6 cm) and slow-growing, but half recurred or had a metachronous tumour, sometimes with more atypical histology; histology ranged from Müllerian-type papillary cystadenoma to low-grade papillary adenocarcinoma, with expression of PAX8, WT1 and hormone receptors, and all seven sequenced tumours had c.49G>A without other oncogenic alterations [44]. Ovarian tumours may have borderline histology and should be assessed by pathologists with experience in gynaecological pathology and in the hyperplasia of the syndrome [1]. Low-grade serous ovarian carcinoma in an adolescent, ductal carcinoma in situ of the breast in a 28-year-old woman, mesothelioma, papillary thyroid carcinoma, yolk sac tumour and papillary adenoma of the epididymis have been described [21, 45, 46].

3.9. Nervous system, eye and development

Hemimegalencephaly, when present, is congenital and is associated with malformations of cortical development, epilepsy and intellectual disability [1]. Smith reports moderate intellectual disability in 20% of patients, but the Oxford Desk Reference considers the intelligence quotient usually normal; the discrepancy probably reflects the proportion of patients with central nervous system involvement in each series [21, 24]. Hydrocephalus, schizencephaly, cerebral arteriovenous malformations, spinal lipomatosis, perineural cysts, spinal canal stenosis due to vertebral anomalies or tumour infiltration, and ocular manifestations such as strabismus, epibulbar dermoid, microphthalmia, myopia, cataracts and nystagmus have been described [21]. The facial phenotype (dolichocephaly, long face, down-slanting palpebral fissures or mild ptosis, low nasal bridge, wide or anteverted nares and open mouth at rest) develops during childhood [1].

3.10. Psychosocial dimension

Overgrowth causes functional limitation and disfigurement, and parents describe stigma by association [1]. Uncertainty about the course and late diagnosis weigh on families: many patients diagnosed late suffered prolonged emotional and psychosocial consequences from having been misdiagnosed [1, 11].

3.11. AKT1-related overgrowth spectrum

Because it is a mosaic disorder, variation is continuous, and the threshold between Proteus syndrome and AROS is arbitrary [1, 2]. Patients above and below the threshold may have severe complications, so each needs an individual surveillance and treatment plan [1].

4. Clinical Recognition and Warning Signs

4.1. When to suspect

Suspicion arises in a young child, without relevant changes at birth, who between the first and second year of life develops asymmetric and progressive overgrowth of a foot, a hand or digits, with bony distortion and without a strict topographic relationship to vascular malformations [1, 24]. Targeted examination looks for the connective tissue naevus, especially on the soles, the linear epidermal naevus, adipose tissue dysregulation and hyperostosis of the skull or external auditory canal [1, 27]. The Oxford Desk Reference recommends examining the soles of the feet in any child evaluated for overgrowth, because connective tissue naevi point towards the syndrome [27]. In adults and adolescents, suspicion may arise from an early meningioma with hyperostosis, genital cystadenomas at an early age, cystic lung disease or venous thromboembolism in a patient with asymmetric overgrowth [2, 32]. The syndrome can be confused with neurofibromatosis type 1 (lipomas, pigmentary changes, overgrowth of limbs and digits, deep plantar creases) and with encephalocraniocutaneous lipomatosis in facial asymmetry [47–49]. The presence of substantial prenatal extracranial overgrowth or “ballooning” soft tissue overgrowth from birth argues against the diagnosis and towards PROS [2].

4.2. Warning and emergency signs

There is no newborn screening for the disease.

5. Diagnosis: Criteria, Tests and Algorithms

5.1. 2019 clinical criteria

The current criteria, published in 2019 by Sapp and colleagues and adopted by GeneReviews, require the three general criteria (mosaic distribution of lesions, sporadic occurrence and progressive course) and a score obtained from positive and negative criteria [1, 2]. They were built from 65 NIH patients with a mosaic AKT1 variant and 10 patients with severe PROS, those most likely to cause confusion; patients with an AKT1 variant scored between 0 and 26 (possible range −10 to 28) and patients with CLOVES between −3 and 6 [2]. Compared with the 1999 and 2006 versions, DVT and PE, testicular cystadenomas and cystadenocarcinomas, meningiomas and cardiac septal lipoma were added, as well as two negative criteria frequent in CLOVES [2]. The 2019 GeneReviews revision added pulmonary bullae [1].

Source: [1, 2].

Source: [1, 2].

The requirement for two overgrown organs prevents isolated hepatomegaly from leading to AROS, and the score of 1 for tumours prevents an isolated meningioma with the variant from being classified as the syndrome [2]. The authors recommend testing AKT1 even with 15 or more points, because the diagnosis has serious consequences and the result guides targeted therapies [2]. Among several hundred patients with somatic overgrowth who did not meet the criteria, none had c.49G>A, and almost all those who met them had the variant, which constitutes external validation of the criteria [2].

5.2. Historical 1999 criteria

The 1999 Biesecker criteria still appear in reference works and in publications prior to 2019 [21, 24]. They required mosaic distribution, a progressive course and sporadic occurrence, and one A criterion (CCTN), two B criteria (linear epidermal naevus; disproportionate overgrowth of limbs, skull, auditory canal, vertebrae or viscera; bilateral ovarian cystadenomas or monomorphic adenoma of the parotid before 20 years) or three C criteria (dysregulated adipose tissue; vascular malformation; facial phenotype) [23, 24]. They do not incorporate the molecular result, and have been replaced by the 2019 version [2].

5.3. Molecular testing

Identification of the variant requires DNA from affected tissue, usually a punch skin biopsy of a clinically evident lesion, such as the CCTN or the epidermal naevus, where the allele fraction tends to be higher; more than one sample may be needed [1]. GeneReviews is aware of only two patients with the variant identified in peripheral blood; in the original series, blood DNA was always negative by Sanger sequencing and only two samples were positive (8–9%) by a restriction enzyme assay; negative blood does not exclude the diagnosis, and the GeneReviews authors advise against it as a sample [1, 8]. The methodology must detect mosaicism (allele fraction below 50%), and some patients have fractions below 1%, difficult or impossible to detect by some methods [1].

The 50% proportion refers to the percentage of probands with a variant detectable by targeted analysis, not to the clinical sensitivity of a well-performed test: in the NIH cohort, 48 of 49 patients (98%) with adequate tissue had molecular confirmation, and in the 2011 series the variant was found in 26 of 29 patients [8, 11]. The p.(Glu17Arg) case was negative for c.49G>A and for the recurrent PIK3CA variants by restriction enzyme and was identified only by a 153-gene pathway panel, with a fraction of 11% in a fibroblast culture; the authors recommend looking for other AKT1 variants whenever c.49G>A is negative in a compatible patient [7]. For the class of germline-lethal mosaic anomalies, to which Proteus syndrome belongs, the 2020 ERN-Skin consensus recommends skin biopsy for diagnosis and blood only for comparison, DNA extraction directly from the biopsy and sequencing by a high-sensitivity method, unless the cell type carrying the variant is cultured; when a chromosomal anomaly is suspected, direct extraction followed by microarray or, preferably, copy number analysis by next-generation sequencing (NGS), or culture of the appropriate cell type with a karyotype targeted at detecting mosaicism [28]. The 2025 review by the same author, written for mosaic pigmentation disorders but with general sampling and sequencing principles, details the practical execution [50]:

Interpretation of a negative result depends on the detection limit declared by the laboratory relative to the expected allele fraction in the tissue: a test with a 5% limit does not exclude a variant that in that type of lesion usually appears at fractions of 1% or less, but it is reliable when the expected fraction is around 20%; the English National Health Service R327 panel declares a limit of 1% [50]. Confirmation by a second method, a common laboratory practice, may fail because it is less sensitive than the first, in which case a new biopsy and repetition with the initial method are justified [50]. When faced with a mosaic variant of uncertain significance, the options are discussion with experts in mosaicism, biopsy of a second independent lesion (the same variant in two lesions demonstrates clonality and, as a rule, causality) and, when available, functional studies [50]. With all tests negative, the review proposes reassessing the patient a few years later, or at 15 years of age in children, in the light of the tests then available [50]. The recommendation to avoid fibroblast culture applies mainly to pigmentary lesions; in Proteus syndrome the dermal fibroblasts of the CCTN carry the variant, and the p.(Glu17Arg) case was identified in a fibroblast culture, so culture of the right cell type is here an alternative accepted by the consensus itself [7, 28, 38]. The report should state the gene, the transcript with version (preferably MANE Select NM_001382430.1), the variant in HGVS, the tissue studied, the allele fraction and the mosaic status; c.49G>A is classified as pathogenic in ClinVar [5, 30]. Variants of uncertain significance in AKT1 can be discussed with the NIH group, which makes itself available for such review [1]. There is no biochemical biomarker; AKT phosphorylation has only been measured in research [4].

5.4. Complementary assessment of extent

After diagnosis, GeneReviews recommends a complete orthopaedic assessment, skeletal radiographs as a baseline, CT with three-dimensional reconstruction in significant scoliosis, a respiratory medicine consultation with pulmonary function tests, high-resolution chest CT if there are signs of bullous disease, developmental assessment and other imaging techniques as clinically indicated [1]. The 2007 orthopaedic workshop adds long-film radiographs of the spine and lower limbs at diagnosis, with the spine in the sitting position, and spinal MRI when scoliosis is diagnosed [36]. The NIH group proposes MRI of the chest, abdomen, pelvis and lower limbs at diagnosis, to define deep components (lymphatic and venous malformations, gastrointestinal and urogenital involvement), haematological assessment and baseline D-dimers in patients with vascular malformations [41]. Cardiac MRI detects the intramyocardial fat that non-gated CT may not show [22].

5.5. Practical algorithm

In the presence of progressive asymmetric overgrowth, the pathway proposed by the sources consulted is: confirm postnatal onset and the mosaic pattern and look for CCTN, epidermal naevus, hyperostosis and adipose dysregulation; calculate the 2019 score; obtain a biopsy of affected tissue, fresh, frozen or in a preservation medium and not formalin-fixed, and request AKT1 testing with a mosaic-sensitive method (deep-coverage NGS, with a known detection limit) [50], preferably by sequencing or a panel that includes PIK3CA and PTEN; if negative and suspicion persists, test a second tissue or perform deeper sequencing; apply the designation table and complete the assessment of extent [1, 2, 7, 36]. If the score is low and the clinical picture suggests PROS, the PROS diagnostic pathway is followed.

5.6. Prenatal diagnosis

Without a family history, mosaic AKT1 testing may be considered when hemimegalencephaly is observed incidentally on ultrasound; there has been one diagnosis in amniocytes, but most prenatal hemimegalencephalies have another diagnosis, mainly PROS [1]. With a family history of Proteus syndrome, prenatal diagnosis and preimplantation genetic testing are not indicated [1].

6. Differential Diagnosis

6.1. Principles

The distinction relies on two features: the natural history, with almost always postnatal onset and rapid progression, and the type of overgrowth, disproportionate, distorting and progressive, accompanied by CCTN [1]. The presence of substantial overgrowth already at birth and of “ballooning” soft tissue overgrowth is more typical of PROS and scores negatively in the criteria [2]. The differential includes disorders of the same pathway with a somatic cause (PROS), germline disorders with a second somatic variant (PHTS), imprinting disorders with hemihyperplasia and other mosaic disorders [24, 27].

6.2. Comparative table

6.3. Main differentials

PROS is the most frequent differential and the one that has caused the most confusion: CLOVES has asymmetric but proportionate prenatal overgrowth, with fat within the normal fascial planes, combined malformations and widely spaced feet, and Klippel-Trenaunay syndrome associates ipsilateral overgrowth with a capillary-lymphatic-venous malformation, without the progressive bony distortion of Proteus syndrome [1]. In the group of 10 patients with severe PROS used to build the criteria, the score of patients with CLOVES ranged from −3 to 6, below the threshold of 10 [2]. Hemihyperplasia with multiple lipomatosis, congenital and only slightly progressive, was often confused with the syndrome [1].

PHTS presents with asymmetric overgrowth, macrocephaly, cutaneous vascular malformations and tumour susceptibility, and its segmental forms (type 2 segmental Cowden, SOLAMEN) result from a germline PTEN variant with a second somatic variant [1, 21]. The distinction has practical consequences: PHTS is autosomal dominant, is transmitted to 50% of offspring and requires its own cancer surveillance, whereas Proteus syndrome is not transmitted [1, 21]. “Proteus-like” patients with germline PTEN belong to PHTS [9, 26].

In isolated asymmetry, about 20% of children have an 11p15 alteration of the Beckwith-Wiedemann spectrum, with a risk of embryonal tumours that justifies renal ultrasound surveillance in high-risk molecular forms; Proteus syndrome does not have this Wilms tumour risk described in the sources consulted [24]. Neurofibromatosis type 1 enters the differential because of pigmented lesions, localised overgrowth and facial asymmetry, and encephalocraniocutaneous lipomatosis shares lipomas, cranial hyperostosis, intellectual disability and seizures, but has adipose lesions present at birth and non-progressive and results from activating somatic FGFR1 variants [47–49, 55]. In the skeleton, melorheostosis-like anomalies of the axial skeleton associated with fibrolipomatous lesions suggest regional Proteus syndrome, and fibrous dysplasia is distinguished by well-demarcated ground-glass areas and frequent cranial involvement [37].

7. Treatments and Risk Management Strategies

7.1. Framework and quality of evidence

There are no published clinical practice guidelines for Proteus syndrome [1]. The available recommendations are expert opinion and retrospective single-centre series: GeneReviews (updated in 2023), the 2007 NIH orthopaedic workshop published in 2011, the 2019 thrombosis surveillance recommendations and the 2022 lung disease data [1, 36, 41, 42]. No disease-modifying treatment is authorised, no therapy has been tested in a randomised trial and AKT1 has no ClinGen actionability assertion [4]. Management is multidisciplinary and coordinated, tailored to each patient's manifestations, with the geneticist, orthopaedics, physical and rehabilitation medicine, dermatology, respiratory medicine, haematology, surgery and psychology [1, 25].

7.2. Management by manifestation

7.3. Orthopaedics

The distinction between the normal limb and the overgrown limb is the central rule: the shorter limb is the normal one, and lengthening it risks exceeding the bone's capacity for expansion and compromising the functional limb [1, 36]. Epiphysiostasis makes it possible to reduce the predicted discrepancy in children still far from puberty, and epiphysiodesis is preferred from adolescence onwards; both allow the child to be mobilised rapidly, which reduces thrombotic risk [36]. In the series of 8 patients operated on at the NIH between 2005 and 2015 for lower limb discrepancy, the first surgery was performed at a mean age of 9.4 years (6.1–13.6), with a discrepancy of 5.0 cm (1.8–10.0); after 23 procedures and a mean follow-up of 4.6 years, the mean final discrepancy in the 7 patients without overcorrection was 2.6 cm (0.6–7.2), and the only complications were two mild knee valgus deformities; this is level IV evidence [57]. For the long bones, measurement of inseam length helps decide the timing of epiphysiodesis [36]. All procedures should be considered high risk and be performed by surgeons experienced in the syndrome, but excessive delay may make the patient inoperable because of progression of other manifestations [36]. CT with three-dimensional reconstruction helps to plan spinal surgery, because the vertebrae become progressively deformed [1]. Scoliosis should be corrected when indicated, because its progression can cause fatal restrictive lung disease [1]. There is no experience with joint prostheses, which may increase the risk of DVT and require revisions [36].

7.4. Prevention and treatment of venous thromboembolism

Thrombotic risk is concentrated in the perioperative period and in the affected limbs, but spontaneous events also occur [15]. The NIH group's recommendations are: haematological assessment and baseline D-dimers in patients with vascular malformations; before surgery, Doppler ultrasound of the venous anatomy and thrombophilia testing; a haematology consultation for prophylaxis before any procedure predisposing to DVT; and anticoagulation of confirmed events according to the American College of Chest Physicians guidelines [1, 41]. D-dimers were useful as a screening test: in the 2017 series, all patients with values above 1 µg/mL had thrombosis, and a threshold of 0.5 µg/mL (STA Liatest assay) had a sensitivity of 100% and a positive predictive value of 37% (4 of 11); in the 2019 series, however, 10 of 22 patients had D-dimers above 0.50 µg/mL without thrombosis on imaging, which limits specificity [15, 41].

The enoxaparin dose of 0.5 mg/kg every 12 hours is that of the NIH surgical protocol, which the same authors also suggest for patients with PROS and vascular malformations; GeneReviews leaves the decision on prophylaxis for each procedure to the haematology consultation [1, 41]. The efficacy of prophylaxis has not been formally demonstrated, and the 2007 workshop considered a trial impracticable [36]. The authors stress that, even with prophylaxis, deaths from PE have occurred after spinal surgery [1].

7.5. Targeted therapy: miransertib

Miransertib (MK-7075, formerly ARQ 092) is an oral, allosteric inhibitor selective for the three AKT isoforms, with a 50% inhibitory concentration of 5.0 nM for AKT1, developed for oncology [58]. In fibroblasts with p.(Glu17Lys), concentrations of 31–500 nM reduced AKT phosphorylation, and concentrations 10 to 20 times higher were needed to reduce cell viability [58]. The therapeutic rationale differs from that in oncology: the aim is to attenuate signalling for years, possibly for life, with lower doses, without interfering with normal growth and development [59]. Miransertib has no marketing authorisation in any jurisdiction and no trade name; it can only be used in clinical trials or expanded access programmes [4, 12].

The adverse effects described at these doses were mainly grade 1–2: dry mouth, mucositis and stomatitis, pharyngitis, nausea, headache, maculopapular rash, sinus tachycardia, transient elevations of aminotransferases and bilirubin, hyperglycaemia, hypertriglyceridaemia and neutropenia [58, 59, 61]. The only dose-limiting toxicity in phase 0/1 was an elevation of liver tests suggestive of steatohepatitis, grades 2 and 3, in cycle 8, which resolved after discontinuation [58]. In oncology, the dose-limiting toxicities of miransertib were hyperglycaemia and rash at 80 mg/day [59]. In the patient followed for five years, safety monitoring included examination, laboratory tests and electrocardiogram every 12 weeks, telephone contact every 4 weeks, and there was no delay in healing after epiphysiodesis or after cruciate ligament reconstruction [59]. In a child with the p.(Glu17Arg) variant and advanced lung disease, miransertib at 5 mg/m² was discontinued after 3 months because of clinical deterioration [7].

7.6. Clinical trials

None of these studies has a control group, and the phase 0/1 study and individual reports do not allow the effect of the drug to be distinguished from the natural course; the authors of the five-year case note that the slowing of the CCTN close to adulthood may partly reflect the natural history itself [59]. The CCTN growth rate of 5.6% per year in untreated children serves as a historical comparison and underpins the primary outcome of the phase 2 trial [14, 31]. In MOSAIC, the Proteus syndrome-specific cohort included a single patient, so the safety data mainly reflect PROS [61, 62].

7.7. Regulatory status

7.8. Developments in the last twelve months

Between November 2025 and October 2026, the NIH phase 2 trial reached its primary completion date (25 June 2026) with 38 patients enrolled, but there are no published results or press releases (registry, ongoing trial level) [14]. Publications included the Akt1 p.(Glu17Lys) mosaicism mouse model with rescue of lethality by prenatal miransertib (preclinical) [29], a validated liquid chromatography–mass spectrometry method for measuring miransertib in patient plasma (pharmacokinetics) [64] and the largest series of paratesticular tumours in the syndrome (observational study) [44]. The 2025 MOSAIC safety publication is the most recent trial result [61]. There was no documented regulatory change at the FDA or the European Medicines Agency (EMA) for the syndrome in the sources consulted.

7.9. Agents and circumstances to avoid

Drugs that increase the risk of DVT or are procoagulant and drugs that stimulate growth, such as androgenic steroids and growth hormone, should be avoided [1]. Distraction osteogenesis of the normal limb, liposuction of lipomatous overgrowth and prolonged immobilisation without prophylaxis are circumstances to avoid [1, 36, 41]. Dehydration, surgery and immobilisation increase thrombotic risk in patients with venous malformations [41].

7.10. Perioperative care and anaesthesia

Before any surgery, haematological assessment with a prophylaxis plan, venous Doppler ultrasound and thrombophilia testing are recommended [1, 41]. Bullous disease and restrictive disease due to scoliosis should be known before anaesthesia, and the 2007 workshop includes pulmonary embolism and pulmonary bullae among the non-orthopaedic complications that alter surgical risk [36]. MRI at early ages requires the anaesthetic risk to be weighed against the information obtained [41].

7.11. Pregnancy, lactation and transition

There are no data on the management of pregnancy; the risks are theoretical, mainly pelvic vein thrombosis [1]. The miransertib trials required effective contraception during treatment and for the following 90 days, and excluded pregnant and breastfeeding women, so there are no data on exposure in pregnancy [62, 63]. As the disease has paediatric onset and almost all mortality occurs before 22 years, the transition to adult care coincides with the phase in which overgrowth slows; adult follow-up maintains surveillance for tumours, the lungs and thrombosis [1, 11].

8. Reproductive Options and Family Counselling

8.1. Mechanism and recurrence risks

Proteus syndrome and AROS are not inherited: all people with a molecular diagnosis are mosaic for c.49G>A, c.49_50delinsAG or an AKT1 duplication, indicating a post-fertilisation mutation in a cell of the multicellular embryo [1]. There are no confirmed cases of vertical transmission or sibling recurrence, and discordant monozygotic twins have been described [1, 8]. Orphanet classifies the type of inheritance as “not applicable” [3].

No pregnancy has zero risk, but the evidence indicates that recurrence is not increased relative to the population [1]. The ERN-Skin consensus classifies the syndrome among the germline-lethal mosaic anomalies, and the 2025 review sets out the two conditions necessary for a mosaic variant to pass to the offspring as a constitutional disorder: being present in at least some of the gametes and being, in its exact form, compatible with life in the germline state; gonadal involvement cannot be inferred from the cutaneous pattern [28, 50]. In Proteus syndrome the second condition fails, but the same review acknowledges, without studies for most mosaic skin disorders, that carriers of germline-lethal variants may have an increased risk of miscarriage; knowing the parent's exact variant is what allows preimplantation genetic diagnosis to be discussed if that scenario arises [50]. This hypothesis has not been studied in Proteus syndrome and does not change the GeneReviews position, according to which prenatal diagnosis and preimplantation genetic testing are not indicated [1, 50]. Relatives do not need to be assessed for the disease [1].

8.2. Testing of relatives, prenatal and preimplantation diagnosis

There is no cascade testing, because there is no carrier state [1]. With a family history of Proteus syndrome, prenatal diagnosis and preimplantation genetic testing are not indicated [1]. Without a family history, mosaic AKT1 testing may be considered when fetal hemimegalencephaly is observed incidentally, bearing in mind that most of these cases have another diagnosis, mainly PROS [1].

8.3. Content of counselling

Counselling explains the somatic nature of the disease, the variability arising from mosaicism, the progressive course in childhood with stabilisation after adolescence and the absence of increased risk for relatives [1, 11]. The ideal time to discuss risks is before pregnancy, and affected young adults should receive counselling about risks to offspring and reproductive options [1]. The distinction from PHTS is decisive, because in PHTS transmission is autosomal dominant with a 50% risk to children [1]. Many families benefit from patient associations and psychosocial support [1].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Prevalence and what each figure measures

There is no population-based study of prevalence or incidence. The available values are estimates: Orphanet places the syndrome in the prevalence class below 1/1,000,000 in Europe, at birth and point prevalence, without a numerical value, based on the 2011 article and on experts [3]; GeneReviews reports about 100 patients known to the author and a “very rough” estimate of 1:1,000,000 to 1:10,000,000 [1]; the 2017 survival article considers the incidence probably of the order of 1/10,000,000 births [11]; and the 2018 European Union orphan designation estimated fewer than 0.01 per 10,000 people, that is, fewer than 500 people in the Union [12]. The NIH cohort, with 64 patients in the survival analysis and 65 with an AKT1 variant in the criteria revision, is the largest confirmed series and mainly includes patients from the United States, but it does not measure prevalence [2, 11]. The cardiothoracic imaging series reports a male:female ratio of 2:1, and the survival cohort included 36 men and 28 women [11, 22].

9.2. Europe, Portugal and Brazil

There are European series and cases, including the Italian and Spanish patients of the MOSAIC trial and European reports of severe gynaecological involvement published in the European Journal of Medical Genetics and the American Journal of Medical Genetics [1, 61]. The sources consulted contain no epidemiological data from Portugal or Brazil; the Brazilian book Síndromes Genéticas: Reconhecer para Tratar includes the syndrome in the differential diagnosis of Klippel-Trenaunay-Weber syndrome, with more marked bone and soft tissue hypertrophy, and cites a 2013 Brazilian case report [65]. There are no founder populations, given the somatic mechanism. The lack of national data in the sources consulted does not demonstrate that such data do not exist.

9.3. Reference networks and patient associations

In Europe, the syndrome falls within the scope of ERN-Skin, which has a thematic group dedicated to mosaic skin disorders and complex vascular malformations and has published a review with guidance on cutaneous mosaicism, as well as VASCERN (rare multisystemic vascular diseases) and ERN ITHACA (rare malformation syndromes and intellectual disability) [16, 28, 66, 67]. The Proteus Syndrome Foundation (USA) and the Proteus Syndrome Foundation UK support families, and GeneReviews also lists MedlinePlus and the Parent to Parent network [1, 17, 18]. The NIH natural history study continues to recruit patients of all ages [35].

10. Illustrative Clinical Case

Ours and colleagues published in 2021 the five-year course of a boy with Proteus syndrome treated with miransertib [59]. The patient had taken part in the NIH natural history study since the age of 6 years and entered the phase 0/1 pharmacodynamic trial at 13 years. He had a lower limb length discrepancy operated on several times, bilateral plantar CCTN, musculoskeletal pain in the lower back and lower limbs, eczema, seasonal rhinitis, iron-deficiency anaemia and irritable bowel syndrome; the variant NM_005163.2:c.49G>A, p.(Glu17Lys) (rs121434592; NM_001382430.1:c.49G>A in the MANE Select transcript) was identified in mosaic state in cultured fibroblasts from a biopsy of affected skin, by Sanger sequencing and by restriction fragment length polymorphism assay [5, 59]. Before treatment, the CCTN occupied 39.2% of the left sole and 9.5% of the right, and maximum pain on the 0 to 10 numerical scale was 7 [59]. He received 5 mg/m²/day for the 48 weeks of the trial; he asked to continue because of the improvement in pain and the slowing of the naevus; after a 9-month break he resumed 5 mg/m²/day, increased after 16 weeks to 10 mg/m²/day (20 mg) [59]. Over five years he had 33 adverse effects, 94% of grade 1–2; those attributed to the drug (headache, transient elevation of AST up to 38 U/L, ALT up to 32 U/L and total bilirubin up to 1.3 mg/dL, blood glucose up to 117 mg/dL and triglycerides up to 164 mg/dL) were grade 1 and resolved without dose adjustment [59]. The two grade 3 events, a tibial plateau fracture with anterior cruciate ligament rupture and a pharyngeal phlegmon, were considered unrelated; he also underwent percutaneous epiphysiodesis with transphyseal screws of the left distal femur and proximal tibia, with perioperative anticoagulant prophylaxis and without healing complications [59]. After five years, the CCTN occupied 44.6% of the left sole and 15.2% of the right, with annual growth of 1.2% in each foot versus 5.3% and 4.8% in the previous seven years; maximum pain fell to 0, mobility assessed by the mother improved (T-score from 39 to 56) and the paediatric quality of life index rose slightly; height remained between the 96th and 99th centiles and puberty was normal [59]. The figures of the article show serial photographs of the feet (https://pmc.ncbi.nlm.nih.gov/articles/PMC8751418/figure/MCS006134OURF1/), the change in the percentage of the sole occupied by the CCTN (https://pmc.ncbi.nlm.nih.gov/articles/PMC8751418/figure/MCS006134OURF2/) and the patient-reported outcomes (https://pmc.ncbi.nlm.nih.gov/articles/PMC8751418/figure/MCS006134OURF3/) [59]. The authors themselves stress that this is a single uncontrolled patient and that part of the slowing may reflect the natural deceleration close to adulthood; miransertib remains an investigational drug, with no place in routine practice outside trials [14, 59].

11. Prognosis and Long-Term Follow-Up

11.1. Prognosis

In the NIH cohort of 64 patients, 11 (17.2%) had died at the time of the review, 10 of them before 21.8 years, with a mean age at death of 14.4 years (standard deviation 11.3) and a median of 15.1 years; by Kaplan-Meier analysis, the probability of death by 22 years was 25% (75% of patients were alive at 21.8 years), and only one death occurred after that age [11]. GeneReviews summarises this as a projected mortality of 25% before 20 years [1]. Deaths are distributed relatively evenly from infancy to the beginning of the third decade, and the authors conclude that interventions to reduce mortality should target the paediatric age group [11]. The limitations of these figures are the size of the cohort, the absence of systematic recontact and ascertainment bias [11]. DVT with PE is the most common cause of death (6 of 10 deaths in the 2017 series), followed by cystic and restrictive lung disease, central nervous system tumours and abscesses [7, 15, 21]. The individual prognosis depends on the location and extent of overgrowth and on the presence of bullous disease, hemimegalencephaly or PE; mildly affected patients with good follow-up have a very favourable prognosis [1].

11.2. Surveillance plan

GeneReviews indicates respiratory tests “as needed”, the 2007 workshop proposes them annually when abnormal and the 2022 series concludes that screening for cystic disease in asymptomatic patients is indicated, especially before 20 years, the age at which all affected patients progressed [1, 36, 42]. Regarding tumours, GeneReviews does not recommend periodic imaging and bases surveillance on targeted history and examination every 6–12 months; the NIH series document paratesticular masses operated on in 19% of men, at a mean age of 9 years, and meningiomas before 30 years adjacent to cranial hyperostosis, but none of the sources consulted proposes an imaging protocol for these tumours [1, 32, 44].

12. Summary and Final Considerations

Proteus syndrome is a non-inherited mosaic disorder, caused in almost all patients by the activating AKT1 variant c.49G>A, p.(Glu17Lys), present in only part of the tissues. It has almost no expression at birth, presents between 6 and 18 months with asymmetric overgrowth that distorts bone, progresses rapidly during childhood and stabilises after adolescence. The cerebriform connective tissue naevus is the most specific sign; the linear epidermal naevus, adipose tissue dysregulation, low-flow vascular malformations, visceral overgrowth, meningiomas, cystadenomas, cystic lung disease and venous thromboembolism complete the picture. Diagnosis uses the 2019 weighted criteria, which integrate the clinical score and molecular testing in affected tissue, and defines the AKT1-related overgrowth spectrum for forms with a low score. Blood cannot be used to exclude the disease, and methods must detect low allele fractions. The main differential diagnoses are PROS, congenital and more proportionate, and PTEN hamartoma tumour syndrome, which is inherited. In the NIH cohort, a quarter of patients died before 22 years, mainly from PE. Management, without formal guidelines, relies on experienced teams: epiphysiostasis and epiphysiodesis on the overgrown limb, never lengthening of the normal limb; timely treatment of scoliosis; respiratory surveillance; perioperative antithrombotic prophylaxis with a haematology plan; clinical tumour surveillance; and avoidance of procoagulant and growth-promoting drugs. Miransertib, an AKT inhibitor, reduced tissue AKT phosphorylation at low doses and was well tolerated in small studies, with signs of benefit on pain and on growth of the cerebriform naevus, but it is not authorised in any jurisdiction; the results of the NIH phase 2 trial, with primary completion in June 2026, are the next milestone. The disease is not transmitted, and the risk for relatives is that of the general population.

13. References

1. Biesecker LG, Sapp JC. Proteus Syndrome. 2012 Aug 9 [updated 2023 May 25]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1993–2026. NBK99495. https://www.ncbi.nlm.nih.gov/books/NBK99495/
2. Sapp JC, Buser A, Burton-Akright J, Keppler-Noreuil KM, Biesecker LG. A dyadic genotype-phenotype approach to diagnostic criteria for Proteus syndrome. Am J Med Genet C Semin Med Genet. 2019;181(4):565-70. doi:10.1002/ajmg.c.31744. PMID: 31692258. PMCID: PMC7520016.
3. Orphanet. Proteus syndrome (ORPHA:744). https://www.orpha.net/en/disease/detail/744
4. GeneGuidance. AKT1 — gene and condition records.
5. VariantValidator. Validation of NM_005163.2:c.49G>A e c.49_50delinsAG (GRCh38, MANE Select). https://variantvalidator.org (accessed 2026-10-29).
6. Online Mendelian Inheritance in Man (OMIM). *164730 AKT serine/threonine kinase 1; AKT1. Baltimore: Johns Hopkins University. https://www.omim.org/entry/164730 (accessed 2026-10-29).
7. Buser A, Lindhurst MJ, Kondolf HC, Yourick MR, Keppler-Noreuil KM, Sapp JC, Biesecker LG. Allelic heterogeneity of Proteus syndrome. Cold Spring Harb Mol Case Stud. 2020;6(3):a005181. doi:10.1101/mcs.a005181. PMID: 32327430. PMCID: PMC7304363.
8. Lindhurst MJ, Sapp JC, Teer JK, Johnston JJ, Finn EM, Peters K, et al. A mosaic activating mutation in AKT1 associated with the Proteus syndrome. N Engl J Med. 2011;365(7):611-9. doi:10.1056/NEJMoa1104017. PMID: 21793738. PMCID: PMC3170413.
9. Orphanet. Proteus-like syndrome (ORPHA:2969). https://www.orpha.net/en/disease/detail/2969
10. Online Mendelian Inheritance in Man (OMIM). #176920 Proteus syndrome. Baltimore: Johns Hopkins University. https://www.omim.org/entry/176920 (accessed 2026-10-29).
11. Sapp JC, Hu L, Zhao J, Gruber A, Schwartz B, Ferrari D, Biesecker LG. Quantifying survival in patients with Proteus syndrome. Genet Med. 2017;19(12):1376-9. doi:10.1038/gim.2017.65. PMID: 28661492. PMCID: PMC5720922.
12. European Medicines Agency. EU/3/18/1997 – orphan designation for treatment of Proteus syndrome (miransertib). https://www.ema.europa.eu/en/medicines/human/orphan-designations/eu3181997 (accessed 2026-10-29).
13. ArQule Inc. FDA grants rare pediatric disease designation to ArQule's miransertib (ARQ 092) for the treatment of Proteus syndrome [press release]. 2017 Nov 1. https://www.biospace.com/fda-grants-rare-pediatric-disease-designation-to-arqule-s-miransertib-arq-092-for-the-treatment-of-proteus-syndrome
14. National Human Genome Research Institute. A multi-cohort phase 2 dose-escalation study of MK-7075 (miransertib) in Proteus syndrome. ClinicalTrials.gov NCT04316546. https://clinicaltrials.gov/study/NCT04316546 (updated 2026-07-17; accessed 2026-10-29).
15. Keppler-Noreuil KM, Lozier JN, Sapp JC, Biesecker LG. Characterization of thrombosis in patients with Proteus syndrome. Am J Med Genet A. 2017;173(9):2359-65. doi:10.1002/ajmg.a.38311. PMID: 28627093. PMCID: PMC5592090.
16. ERN-Skin. Thematic group MOSAIC – cutaneous mosaic disorders. https://ern-skin.eu/mosaic/ (accessed 2026-10-29).
17. Proteus Syndrome Foundation. https://www.proteus-syndrome.org (accessed 2026-10-29).
18. Proteus Syndrome Foundation UK. https://www.proteus-syndrome.org.uk (accessed 2026-10-29).
19. Cohen MM Jr, Hayden PW. A newly recognized hamartomatous syndrome. Birth Defects Orig Artic Ser. 1979;15(5B):291-6. PMID: 118782.
20. Wiedemann HR, Burgio GR, Aldenhoff P, Kunze J, Kaufmann HJ, Schirg E. The proteus syndrome. Partial gigantism of the hands and/or feet, nevi, hemihypertrophy, subcutaneous tumors, macrocephaly or other skull anomalies and possible accelerated growth and visceral affections. Eur J Pediatr. 1983;140(1):5-12. doi:10.1007/BF00661895. PMID: 6873112.
21. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 736–739.
22. Mirmomen SM, Arai AE, Turkbey EB, Bradley AJ, Sapp JC, Biesecker LG, Sirajuddin A. Cardiothoracic imaging findings of Proteus syndrome. Sci Rep. 2021;11(1):6577. doi:10.1038/s41598-021-86029-0. PMID: 33753828. PMCID: PMC7985501.
23. Biesecker LG, Happle R, Mulliken JB, Weksberg R, Graham JM Jr, Viljoen DL, Cohen MM Jr. Proteus syndrome: diagnostic criteria, differential diagnosis, and patient evaluation. Am J Med Genet. 1999;84(5):389-95. PMID: 10360391.
24. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 164–166.
25. Biesecker L. The challenges of Proteus syndrome: diagnosis and management. Eur J Hum Genet. 2006;14(11):1151-7. doi:10.1038/sj.ejhg.5201638. PMID: 16883308.
26. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 564–567.
27. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 272–275.
28. Kinsler VA, Boccara O, Fraitag S, Torrelo A, Vabres P, Diociaiuti A. Mosaic abnormalities of the skin: review and guidelines from the European Reference Network for rare skin diseases. Br J Dermatol. 2020;182(3):552-63. doi:10.1111/bjd.17924. PMID: 30920652.
29. Abdul Rahiman Sirajuddeen SR, Lindhurst MJ, Detels M, Laughner N, Johnston J, Brinster LR, Biesecker LG. Rescue of Proteus syndrome lethality in mice with prenatal miransertib treatment. Genet Med. 2026:102732. doi:10.1016/j.gim.2026.102732. PMID: 42788321. PMCID: PMC13695987.
30. ClinVar. VCV000013983 – NM_001382430.1(AKT1):c.49G>A (p.Glu17Lys). https://www.ncbi.nlm.nih.gov/clinvar/variation/13983/ (accessed 2026-10-29).
31. Nathan NR, Patel R, Crenshaw MM, Lindhurst MJ, Olsen C, Biesecker LG, Keppler-Noreuil KM, Darling TN. Pathogenetic insights from quantification of the cerebriform connective tissue nevus in Proteus syndrome. J Am Acad Dermatol. 2018;78(4):725-32. doi:10.1016/j.jaad.2017.10.018. PMID: 29042227. PMCID: PMC5857242.
32. Keppler-Noreuil KM, Baker EH, Sapp JC, Lindhurst MJ, Biesecker LG. Somatic AKT1 mutations cause meningiomas colocalizing with a characteristic pattern of cranial hyperostosis. Am J Med Genet A. 2016;170(10):2605-10. doi:10.1002/ajmg.a.37737. PMID: 27550858. PMCID: PMC5580816.
33. Collet L, et al. Breast cancer — genomics. In: ESMO Handbook — Cancer Genomics. Lugano: ESMO Press. p. 82–91.
34. Online Mendelian Inheritance in Man (OMIM). #615109 Cowden syndrome 6; CWS6. Baltimore: Johns Hopkins University. https://www.omim.org/entry/615109 (accessed 2026-10-29).
35. National Human Genome Research Institute. Study of Proteus syndrome and related congenital disorders. ClinicalTrials.gov NCT00001403. https://clinicaltrials.gov/study/NCT00001403 (accessed 2026-10-29).
36. Tosi LL, Sapp JC, Allen ES, O'Keefe RJ, Biesecker LG. Assessment and management of the orthopedic and other complications of Proteus syndrome. J Child Orthop. 2011;5(5):319-27. doi:10.1007/s11832-011-0350-6. PMID: 23024722. PMCID: PMC3179535.
37. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford University Press; 2018. p. 613–615.
38. Pithadia DJ, Roman JW, Sapp JC, Biesecker LG, Darling TN. Hypertrichotic patches as a mosaic manifestation of Proteus syndrome. J Am Acad Dermatol. 2021;84(2):415-24. doi:10.1016/j.jaad.2020.01.078. PMID: 32035943. PMCID: PMC7415737.
39. Hannoush H, Sachdev V, Brofferio A, Arai AE, LaRocca G, Sapp J, et al. Myocardial fat overgrowth in Proteus syndrome. Am J Med Genet A. 2015;167A(1):103-10. doi:10.1002/ajmg.a.36773. PMID: 25377688. PMCID: PMC4275354.
40. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 260–265.
41. Keppler-Noreuil KM, Lozier J, Oden N, Taneja A, Burton-Akright J, Sapp JC, Biesecker LG. Thrombosis risk factors in PIK3CA-related overgrowth spectrum and Proteus syndrome. Am J Med Genet C Semin Med Genet. 2019;181(4):571-81. doi:10.1002/ajmg.c.31735. PMID: 31490637. PMCID: PMC8513083.
42. Keppler-Noreuil KM, Burton-Akright J, Kleiner DE, Sapp JC, Lindhurst MJ, Han CG, Biesecker LG, Gochuico BR. Phenotypic features of cystic lung disease in Proteus syndrome: a clinical trial. Ann Am Thorac Soc. 2022;19(11):1871-80. doi:10.1513/AnnalsATS.202111-1214OC. PMID: 35839129. PMCID: PMC9667804.
43. Takyar V, Khattar D, Ling A, Patel R, Sapp JC, Kim SA, et al. Characterization of the hepatosplenic and portal venous findings in patients with Proteus syndrome. Am J Med Genet A. 2018;176(12):2677-84. doi:10.1002/ajmg.a.40636. PMID: 30346092. PMCID: PMC8020299.
44. Matoso A, Vang R, Xing D, Killian JK, Meltzer PS, Keppler-Noreuil KM, et al. Characterizing paratesticular neoplasms in Proteus syndrome. Am J Surg Pathol. 2026;50(3):326-37. doi:10.1097/PAS.0000000000002498. PMID: 41368924. PMCID: PMC13596702.
45. Leoni C, Gullo G, Resta N, Fagotti A, Onesimo R, Schwartz B, et al. First evidence of a therapeutic effect of miransertib in a teenager with Proteus syndrome and ovarian carcinoma. Am J Med Genet A. 2019;179(7):1319-24. doi:10.1002/ajmg.a.61160. PMID: 31058421.
46. Iqbal J, He G, Biesecker LG, Rosen P, Duray PH, Schwartzentruber D, et al. Morphological characterization of the breast in Proteus syndrome complicated by ductal carcinoma in situ. Ann Clin Lab Sci. 2006;36(4):469-74. PMID: 17127737.
47. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 222–223.
48. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 506–511.
49. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 146–149.
50. Kinsler VA. Mosaic disorders affecting pigmentation – part 2: how to make a genetic diagnosis. Br J Dermatol. 2025;193(6):1047-55. doi:10.1093/bjd/ljaf224.
51. Online Mendelian Inheritance in Man (OMIM). #612918 Congenital lipomatous overgrowth, vascular malformations, and epidermal nevi (CLOVES). Baltimore: Johns Hopkins University. https://www.omim.org/entry/612918 (accessed 2026-10-17).
52. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 224–227.
53. Online Mendelian Inheritance in Man (OMIM). #158350 Cowden syndrome 1; CWS1 (includes “Proteus-like syndrome”). Baltimore: Johns Hopkins University. https://www.omim.org/entry/158350 (accessed 2026-10-29).
54. Ribeiro EM, coord. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 216–218.
55. Online Mendelian Inheritance in Man (OMIM). #613001 Encephalocraniocutaneous lipomatosis; ECCL. Baltimore: Johns Hopkins University. https://www.omim.org/entry/613001 (accessed 2026-10-29).
56. Online Mendelian Inheritance in Man (OMIM). #163200 Schimmelpenning-Feuerstein-Mims syndrome; SFM. Baltimore: Johns Hopkins University. https://www.omim.org/entry/163200 (accessed 2026-10-29).
57. Crenshaw MM, Goerlich CG, Ivey LE, Sapp JC, Keppler-Noreuil KM, Scott AC, Biesecker LG, Tosi LL. Orthopaedic management of leg-length discrepancy in Proteus syndrome: a case series. J Pediatr Orthop. 2018;38(3):e138-44. doi:10.1097/BPO.0000000000001121. PMID: 29329145.
58. Keppler-Noreuil KM, Sapp JC, Lindhurst MJ, Darling TN, Burton-Akright J, Bagheri M, et al. Pharmacodynamic study of miransertib in individuals with Proteus syndrome. Am J Hum Genet. 2019;104(3):484-91. doi:10.1016/j.ajhg.2019.01.015. PMID: 30803705. PMCID: PMC6407523.
59. Ours CA, Sapp JC, Hodges MB, de Moya AJ, Biesecker LG. Case report: five-year experience of AKT inhibition with miransertib (MK-7075) in an individual with Proteus syndrome. Cold Spring Harb Mol Case Stud. 2021;7(6):a006134. doi:10.1101/mcs.a006134. PMID: 34649967. PMCID: PMC8751418.
60. Biesecker LG, Edwards M, O'Donnell S, Doherty P, MacDougall T, Tith K, et al. Clinical report: one year of treatment of Proteus syndrome with miransertib (ARQ 092). Cold Spring Harb Mol Case Stud. 2020;6(1):a004549. doi:10.1101/mcs.a004549. PMID: 32014856. PMCID: PMC6996520.
61. Eng W, Iacobas I, Perkins J, Zampino G, Leoni C, Buonuomo PS, et al. Safety findings from the phase 1/2 MOSAIC study of miransertib for patients with PIK3CA-related overgrowth spectrum or Proteus syndrome. Orphanet J Rare Dis. 2025;20:375. doi:10.1186/s13023-025-03831-z. PMID: 40713644. PMCID: PMC12296598.
62. ArQule/Merck Sharp & Dohme. Study of miransertib (MK-7075) in participants with PIK3CA-related overgrowth spectrum and Proteus syndrome (MOSAIC). ClinicalTrials.gov NCT03094832. https://clinicaltrials.gov/study/NCT03094832 (accessed 2026-10-29).
63. Merck Sharp & Dohme. Long-term safety extension of miransertib in PROS or Proteus syndrome (MK-7075-006). ClinicalTrials.gov NCT04980872. https://clinicaltrials.gov/study/NCT04980872 (accessed 2026-10-29).
64. Khan RMN, Arisa OT, Zeng Y, Rostagni O, Redick C, Paris KM, et al. Development and validation of an LC-MS/MS assay for the quantification of miransertib in human plasma and clinical application. J Chromatogr B Analyt Technol Biomed Life Sci. 2026;1281:125161. doi:10.1016/j.jchromb.2026.125161. PMID: 42224880.
65. Ribeiro EM, coord. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 202–205.
66. VASCERN – European Reference Network on Rare Multisystemic Vascular Diseases. https://vascern.eu (accessed 2026-10-29).
67. ERN ITHACA – European Reference Network on Rare Malformation Syndromes, Intellectual and Other Neurodevelopmental Disorders. https://ern-ithaca.eu (accessed 2026-10-29).