Neurofibromatosis Type 1 (NF1)

Literature and regulatory information updated: 4 October 2026.

Von Recklinghausen disease: constitutional NF1, mosaic NF1 (segmental/localised and generalised), NF1 microdeletion syndrome (17q11.2), the NF1–Noonan phenotype and Watson syndrome, with differential diagnosis from Legius syndrome (SPRED1)

0. Initial Reference Table

1. Introduction

Neurofibromatosis type 1 is a multisystem genetic disease with autosomal dominant inheritance, caused by pathogenic loss-of-function variants in the NF1 tumour suppressor gene at 17q11.2, and characterised by multiple café-au-lait macules, freckling in the axillary and inguinal folds, cutaneous, subcutaneous and plexiform neurofibromas, Lisch nodules in the iris and a predisposition to benign and malignant tumours of the nervous system and other organs [1, 3]. It is one of the most frequent autosomal dominant diseases in humans, with a birth incidence of around 1 in 2,000 to 1 in 2,500 and one of the highest known mutation rates for a human gene [1, 3]. Expressivity is extremely variable, even among relatives carrying the same variant, and many manifestations have age-dependent penetrance, making genetic counselling and lifelong follow-up central aspects of clinical management [1, 3].

The disease's eponym derives from the monograph published in 1882 by the German pathologist Friedrich Daniel von Recklinghausen, Ueber die multiplen Fibrome der Haut und ihre Beziehung zu den multiplen Neuromen, in which he demonstrated that the multiple cutaneous tumours he described originated in peripheral nerves and coined the term “neurofibroma” [28, 29]. For almost a century, the literature did not consistently distinguish the “peripheral” form from the “central” form, characterised by bilateral vestibular schwannomas; the nosological distinction was established at the National Institutes of Health (NIH) Consensus Conference in 1987, published in 1988, which defined the seven clinical diagnostic criteria for NF1 and the criteria for neurofibromatosis type 2 (NF2) [4, 10]. Riccardi had laid the groundwork by proposing a numerical classification based on the presence of café-au-lait macules, freckling, ocular signs, neurofibromas and specific complications [4].

The molecular phase began with mapping the NF1 locus to the pericentromeric region of chromosome 17 and culminated in 1990 with cloning of the gene by two independent groups: Wallace and colleagues, in Science, identified a large transcript disrupted in three patients with NF1, and Viskochil and colleagues, in Cell, demonstrated deletions and a translocation disrupting the cloned gene at the NF1 locus [30, 31]. In the same year, Xu and colleagues showed that the gene product, neurofibromin, contains a domain related to GTPase-activating protein (GAP), establishing the link between the disease and the RAS pathway [32]. NF1 thus became the first human condition attributed to the RAS/MAPK pathway and the prototype of the RASopathies, a group including Noonan, Costello and cardiofaciocutaneous syndromes [4, 33].

In 2007, Brems and colleagues described in Nature Genetics families with café-au-lait macules and freckling without neurofibromas or Lisch nodules, carrying heterozygous loss-of-function variants in SPRED1, an entity subsequently named Legius syndrome [34]. This discovery, the characterisation of NF1 mosaicism, identification of choroidal abnormalities by optical coherence tomography and availability of molecular tests with sensitivity above 95% prompted the international revision of the diagnostic criteria, published in 2021 by Legius and colleagues in Genetics in Medicine, sponsored by the Children's Tumor Foundation [4]. Approval of selumetinib by the Food and Drug Administration (FDA) in April 2020 for children with symptomatic inoperable plexiform neurofibromas marked the first therapy targeting the disease's molecular pathway; in 2025, mirdametinib was approved and selumetinib extended to adults and children from 1 year [4, 18–23].

2. Molecular Genetics and Pathophysiology

2.1 The NF1 gene and neurofibromin

The NF1 gene is located at 17q11.2, spans around 350 kb of genomic DNA, contains 60 exons and produces an 11- to 13-kb messenger RNA; three exons undergo alternative splicing (9a, 23a and 48a) [3]. Three genes in reverse orientation (EVI2A, EVI2B and OMG) are inserted within intron 27b [1, 35]. Neurofibromin is a ubiquitously expressed cytoplasmic protein of 2818 amino acids, with higher levels in neurones, Schwann cells, oligodendrocytes, astrocytes and leucocytes [35]. Its central GAP-related domain (GRD, encoded mainly by exons 20 to 27a) accelerates hydrolysis of guanosine triphosphate (GTP) bound to RAS proteins, converting active RAS (RAS-GTP) to inactive RAS (RAS-GDP) and restraining downstream signalling through the RAF–MEK–ERK and PI3K–AKT–mTOR cascades [1, 4, 32]. Neurofibromin also has a cysteine- and serine-rich domain (CSRD), a Sec14-PH domain with lipid affinity and a SPRED1-binding domain, an interaction that recruits neurofibromin to the plasma membrane; it also regulates adenylyl cyclase and intracellular cyclic AMP production, a function relevant to astrocyte growth and cognition [1, 35].

2.2 Spectrum of pathogenic variants

More than 3,000 different pathogenic variants have been described, distributed throughout the entire coding region and non-coding regions, without hotspots encompassing more than a small proportion of families [1]. The spectrum includes nonsense, frameshift, splicing (including exonic and deep intronic variants that alter RNA processing) and missense variants, in-frame deletions of one to several codons, initiation codon variants, intragenic deletions and duplications, microdeletions encompassing the gene and flanking genes, balanced translocations and insertions of Alu/LINE elements [4]. Around 70% of germline variants truncate neurofibromin and 22–30% affect splicing, more than one third of which are not detected by genomic DNA sequencing restricted to coding regions; Messiaen and colleagues demonstrated in 2000 that comprehensive analysis combining RNA and DNA identifies 95% of variants and reveals a high frequency of unusual splicing defects [1, 3, 36]. De novo variants are predominantly paternal in origin, without a paternal age effect, a pattern reflecting the background rate of new mutations, which are mostly paternal, rather than a mechanism of spermatogonial selective advantage such as that observed in FGFR2, FGFR3 or PTPN11 [3]. If the Finnish incidence is correct and fertility in affected individuals is reduced by half, the NF1 mutation rate exceeds 1 in 8,000 gametes, one of the highest in the human genome [1].

2.3 Disease mechanism: haploinsufficiency and the “second hit”

The constitutional mechanism is loss of function of one allele [1]. ClinGen assigns NF1 a haploinsufficiency score of 3 (sufficient evidence for dosage pathogenicity) and a triplosensitivity score of 0 (no available evidence), supporting the pathogenicity of whole-gene deletions but not duplications [37]. In characteristic disease lesions — cutaneous and plexiform neurofibromas, café-au-lait macules, tibial pseudarthrosis, gliomas, MPNST, phaeochromocytoma and gastrointestinal stromal tumours (GIST) — a second somatic variant inactivating the remaining allele is identified, according to Knudson's model; in café-au-lait macules, biallelic loss occurs in melanocytes rather than keratinocytes or fibroblasts [4, 38]. In neurofibromas, the cell of origin is the Schwann cell (or its precursor) with biallelic NF1 loss, whose expansion depends on a microenvironment containing haploinsufficient cells, particularly mast cells, macrophages and fibroblasts, recruited by factors such as stem cell factor (KIT ligand) [35]. Progression from plexiform neurofibroma to atypical neurofibromatous neoplasm of uncertain biological potential (ANNUBP, formerly “atypical neurofibroma”) is associated with loss of CDKN2A/CDKN2B, and progression to MPNST additionally involves inactivation of components of Polycomb repressive complex 2 (SUZ12 or EED) and, frequently, TP53 [11, 12]. Somatic NF1 variants also occur in sporadic tumours in individuals without NF1 — MPNST, phaeochromocytomas, juvenile myelomonocytic leukaemia (JMML), gliomas, breast cancer, melanoma (23% of cutaneous, 26% of acral, 34% of mucosal and 6% of uveal melanomas), lung adenocarcinoma and ovarian cancer — with mutational spectra distinct from germline variants [1, 39].

2.4 NF1 microdeletion syndrome (17q11.2)

Whole-NF1-gene deletions occur in 5–11% of individuals with NF1 and constitute the most frequent recurrent changes [1, 3]. The 1.4-Mb type 1 deletion results from non-allelic homologous recombination between low-copy repeats NF1-REPa and NF1-REPc, occurs predominantly in maternal meiosis and encompasses 14 protein-coding genes (including SUZ12) and 4 microRNA genes [3, 12]. The 1.2-Mb type 2 deletion is mediated by recombination between SUZ12 (formerly JJAZ1) and its pseudogene SUZ12P1, frequently arises in postzygotic mitosis and is therefore associated with mosaicism; Kehrer-Sawatzki and colleagues found a significant frequency of mosaic type 2 deletions in sporadic patients without dysmorphism or intellectual impairment and recommended deletion-sensitive analysis (MLPA or FISH) when no intragenic variant is identified [3]. There are also type 3 deletions (1.0 Mb, between NF1-REPb and NF1-REPc) and atypical deletions with variable breakpoints [12].

Patients with a non-mosaic type 1 microdeletion have dysmorphic facies (coarse appearance, flat forehead, ocular hypertelorism, broad nasal tip, low-set ears and broad neck), overgrowth between 2 and 6 years with a phenotype that may resemble Weaver syndrome, large hands and feet, joint hyperlaxity, muscular hypotonia, more pronounced cognitive delay, more frequent cardiovascular malformations and more numerous and earlier onset cutaneous, subcutaneous, plexiform and spinal neurofibromas [1, 12, 28]. An internal neurofibroma burden above 3,000 ml is significantly more frequent in these patients, and lifetime MPNST risk reaches 16–26%, compared with 8–13% in carriers of intragenic variants; codeletion of SUZ12 contributes to the increased risk [12, 40]. The ERN GENTURIS guideline classifies an NF1 microdeletion including SUZ12 as a high-risk group for MPNST [11].

2.5 Genotype–phenotype correlations

Although intrafamilial variability limits individual prediction, correlations with an impact on counselling are established [1, 4]. The three-base-pair in-frame deletion c.2970_2972delAAT, removing methionine residue 992 (p.Met992del), described by Upadhyaya and colleagues in 2007, is associated with typical pigmentary manifestations without cutaneous, subcutaneous or superficial plexiform neurofibromas; around one quarter of carriers do not meet diagnostic criteria, but non-optic pathway gliomas, learning difficulties and Noonan features occur at frequencies similar to those in classic NF1 [1, 41]. Missense variants affecting arginine 1809 (p.Arg1809Cys, p.Arg1809Ser, p.Arg1809Pro, p.Arg1809Leu, p.Arg1809Gly) were studied in 136 individuals by Rojnueangnit and colleagues: café-au-lait macules with or without freckling and Lisch nodules, absence of externally visible plexiform neurofibromas and evident cutaneous neurofibromas, Noonan features in around 25%, significantly more prevalent pulmonary stenosis and short stature (p < 0.0001) and developmental delay or learning difficulties in more than 50%; these variants affect 1.23% of unrelated probands in the University of Alabama at Birmingham (UAB) cohort [42].

Conversely, missense variants in the five contiguous codons 844 to 848 (leucine 844, cysteine 845, alanine 846, leucine 847 and glycine 848), located in the cysteine- and serine-rich domain, were identified in 162 individuals (0.8% of UAB probands) and are associated with a severe phenotype: major superficial plexiform neurofibromas and symptomatic spinal neurofibromas more prevalent than in classic cohorts (p < 0.0001), optic pathway gliomas and skeletal abnormalities in almost half the cases, and a greater predisposition to malignancies (p = 0.0061) [43]. Recurrent missense variants at methionine 1149, arginine 1276 and lysine 1423, present in 1.8% of probands, are associated with a Noonan-like phenotype in around 25%; p.Arg1276 and p.Lys1423 are associated with cardiovascular abnormalities, particularly pulmonary stenosis, and p.Arg1276 with symptomatic spinal neurofibromas, whereas p.Met1149 produces a mild, essentially pigmentary phenotype [44]. The p.Arg1038Gly variant is associated with mild pigmentary manifestations, few neurofibromas and frequent NF1–Noonan features [1]. Non-truncating variants are also associated with a higher frequency of pulmonary stenosis [3]. Amino acid positions in these correlations follow the historical numbering of transcript NM_000267 (2818-amino-acid protein, without exon 23a); in MANE Select transcript NM_001042492.3, used by ClinVar, residues downstream of exon 23a are shifted by 21 positions, such that p.Arg1809Cys corresponds to NM_001042492.3:c.5488C>T, p.(Arg1830Cys) (VCV000208853) and p.Lys1423 to p.Lys1444, whereas codons 844–848, p.Met992del, p.Met1149 and p.Arg1276 retain the same numbering in both transcripts (for example, NM_001042492.3:c.3827G>A, p.(Arg1276Gln), VCV000068341) [6]. The literature should therefore be read with both numberings stated when comparing a laboratory report with published series.

2.6 NF1 mosaicism

Around 30–50% of patients with NF1 are sporadic cases (founders), and a proportion result from a variant acquired after fertilisation, present only in a subpopulation of cells [4]. Depending on the timing of the postzygotic mutation and the type of progenitor cells affected, the phenotype may be exclusively pigmentary (mutation in melanocyte precursors), exclusively tumoral (neurofibromas due to a mutation in Schwann cell precursors) or mixed, and distribution may range from a single body segment that does not cross the midline to several regions simulating a generalised phenotype [4, 38, 45]. Maertens and colleagues demonstrated in 2007, through molecular dissection of isolated lesions, that the same first-hit variant is present in melanocytes from café-au-lait macules and Schwann cells from neurofibromas of the same patient, and absent from unaffected tissues [38]. Pure gonadal mosaicism is rare but documented: Lázaro and colleagues described in 1994 a clinically normal father with two affected children, carrying the variant in around 10% of spermatozoa [3, 46]. The term “segmental NF1” (or “neurofibromatosis type V”) should be abandoned in favour of “mosaic NF1”, with “localised mosaic NF1” preferred for forms confined to one segment [1, 47].

2.7 NF1–Noonan phenotype and Watson syndrome

The NF1–Noonan phenotype occurs in around 12% of individuals with NF1 and includes ocular hypertelorism, downslanting palpebral fissures, ptosis, low-set ears, webbed neck, a pectus abnormality and pulmonary valve stenosis [1]. De Luca and colleagues identified heterozygous NF1 variants in 16 of 17 patients with well-characterised NFNS, with a high prevalence of in-frame defects in exons 24 and 25, which encode part of the GRD domain, and no PTPN11 variant; no lesion in NF1 exons 11 to 27 was found in 100 PTPN11-negative patients with Noonan syndrome, supporting NFNS as an allelic variant of NF1 rather than a distinct entity [48]. Pathogenic variants in NF1 and PTPN11 rarely coexist [1]. Watson syndrome, described as pulmonary valve stenosis, café-au-lait macules, short stature and mild intellectual impairment, was linked to NF1 by Allanson and colleagues in 1991 and is caused by NF1 variants, frequently producing mutant rather than truncated protein [3, 49]. Spinal neurofibromatosis (OMIM 162210), another allelic presentation, is characterised by bilateral neurofibromas of all spinal nerve roots, with few or no cutaneous manifestations of NF1, and despite the name also occurs in sporadic cases [1, 50]. In Ruggieri and colleagues' critical review, which applied strict criteria to the 98 published cases, the 49 patients meeting them (9 families, 1 mixed family with classic NF1 and 27 sporadic cases) had café-au-lait macules in 67% and other NF1 manifestations less frequently, and missense variants were associated with an increased risk of spinal neurofibromatosis compared with classic NF1 (OR 6.16; 95% CI 3.14–13.11; p = 0.0001) [50]. NF1 and SPRED1 are part of the RAS/MAPK pathway, explaining the facial and cutaneous overlap between NF1, NFNS, Legius syndrome and Noonan syndrome [33].

2.8 Legius syndrome: molecular genetics

Legius syndrome is caused by heterozygous loss-of-function variants in SPRED1 (15q14), encoding Spred1 (Sprouty-related, EVH1 domain-containing protein 1) [5, 34]. Spred1 contains an N-terminal EVH1 domain, a central KIT-binding domain and a cysteine-rich C-terminal SPR domain; it binds neurofibromin and recruits it to the membrane, where it inactivates RAS, so that Spred1 loss produces a functional effect convergent with that of neurofibromin loss [5]. As in NF1, café-au-lait macules result from a second hit: Brems and colleagues identified an acquired somatic variant in the wild-type SPRED1 allele in melanocytes from one macule, in addition to the germline variant, indicating that macule formation requires complete gene inactivation [34]. ClinGen assigns SPRED1 a haploinsufficiency score of 3 and a triplosensitivity score of 0 [37]. Sequencing detects 89% of pathogenic variants and deletion/duplication analysis 10% [5]. SPRED1 variants explain around 1% of sporadic cases and 19% of familial cases with exclusively pigmentary manifestations [4]. Somatic SPRED1 variants occur in sporadic mucosal and acral melanomas, without hereditary implications [5].

3. Clinical Spectrum and Manifestations by System

3.1 Frequency of manifestations and age-dependent expression

NF1 is a disease with extremely variable expressivity, whose progression and severity may differ over the lifetime of the same individual and between relatives with the same variant [1]. Table 1 brings together the cumulative lifetime frequencies described in GeneReviews, and Table 2 the age at onset of manifestations according to the Oxford Desk Reference.

Table 1. Frequency of the main manifestations of NF1 [1, 3, 11, 51]

Table 2. Manifestations with age-dependent expression [3]

The French PNDS (National Diagnosis and Care Protocol) describes the same usual sequence of lesion appearance: café-au-lait macules, plexiform neurofibroma, axillary or inguinal freckling, Lisch nodules, subcutaneous neurofibromas and, finally, cutaneous neurofibromas; the characteristic bone lesions are detected in the first year of life and symptomatic optic pathway glioma arises mainly between 3 and 6 years of age [17]. Penetrance is around 95% at 8 years and close to 100% at 20 years; at 8 years, 97% of children with NF1 meet at least two NIH criteria, whereas around half of children under 2 years with sporadic NF1 meet only one criterion, which explains the diagnostic delay in this age group [17]. Some PNDS frequencies differ from those in Table 1 because they derive from other series: long-bone dysplasia and sphenoid dysplasia in less than 10% each, scoliosis in 10–25%, osteopenia in 45% and osteoporosis in 25% of patients assessed by densitometry, epilepsy in 3–13%, autism spectrum disorder in 25–40%, migraine in 34–65%, hydrocephalus in around 2% and macrocephaly in 25–50% [17]. These differences mainly reflect the definition criteria (for example, radiological dysplasia versus pseudarthrosis) and the hospital-based recruitment of the series, rather than true variation between populations.

3.2 Cutaneous manifestations

The café-au-lait macules of NF1 are ovoid macules with well-defined borders, a uniform colour slightly darker than the surrounding skin and a usual size of 1 to 3 cm, although they may be smaller, larger, lighter, darker or irregular; they appear in infancy and early childhood and, once established, remain stable, potentially becoming difficult to identify in older people [1]. They do not occur on the palms or soles, and the Brazilian chapter in Síndromes Genéticas: Reconhecer para Tratar adds that they usually spare the scalp, eyebrows and face [1, 52]. A macule with a raised or soft texture suggests an underlying plexiform neurofibroma [1]. In very light or very dark skin, a Wood's lamp facilitates identification [1, 3]. Smith states that 99% of affected individuals have six or more macules larger than 5 mm by the first year of life and that axillary or inguinal freckling is present in 90% by 7 years [28], whereas the Oxford Desk Reference emphasises that six macules are often not present before 6 years [53]. Freckling appears in areas of skin-to-skin friction (axillae, groins and inframammary region) and diffusely on the trunk, proximal limbs and neck; the 2021 consensus notes that these are, more precisely, lentiginous macules, because they occur in areas not exposed to the sun [1, 4]. Some patients have subtle generalised increased pigmentation compared with unaffected relatives [3]. Juvenile xanthogranuloma (yellowish or orange papules, often on the forehead, which regress at the end of childhood) and naevus anaemicus are more frequent than in the general population and may support the diagnosis in young children, but were not included in the 2021 criteria [1, 3, 4].

The French PNDS adds details of natural history: café-au-lait macules are present in 99% of patients, often appear in the first 3 months, increase in number until 2 years of age and fade or disappear after 50 years; freckling, present in around 85%, is rare before 2 years and becomes evident between 4 and 6 years, sometimes spread across the trunk; neither pigmentary lesion has malignant potential [17]. Naevus anaemicus (a white macule with an irregular outline and normal texture, mainly on the trunk) occurs in around 50% of patients and is particularly specific in childhood. Violaceous macules are bluish macules, compressible on palpation, sometimes smaller than 1 cm, that histologically correspond to neurofibromatous tissue around capillaries and venules; they are a subtle, early marker of cutaneous neurofibroma, but do not always progress to one [17]. Juvenile xanthogranuloma is usually solitary, on the head or neck, appears before 2 years and regresses before 5; given the rarity and the controversial nature of the association with JMML, the PNDS advises careful clinical examination (including palpation of lymph nodes) and allows a full blood count, without systematic haematological follow-up. Around 20% of patients report pruritus, generalised or localised to neurofibromas, moderate, frequently daily and predominantly in the evening [17].

3.3 Neurofibromas

Neurofibromas are benign nerve sheath tumours composed of Schwann cells, fibroblasts, perineural cells and mast cells, which may involve virtually any nerve [1]. Cutaneous neurofibromas are discrete masses ranging from 1–2 mm to several centimetres, sessile or pedunculated, soft to firm in consistency, skin-coloured or pinkish, brownish or bluish, sometimes pruritic; they are rare in children, appear from late childhood, increase in number and size at puberty and during pregnancy, and are present in almost all adults, numbering from a few to thousands [1, 3, 28]. Smith records neurofibromas in 48% of 10-year-old children and 84% at 20 years, with an increase between 50 and 70 years [28]. Subcutaneous neurofibromas are mobile rubbery nodules beneath the skin, isolated or in a “string of beads” along a nerve, sometimes painful, present in around 15% of adults on clinical examination [1].

Plexiform neurofibromas involve multiple fascicles or branches of a nerve or plexus and are present in around half of patients, mostly internally and identifiable only by magnetic resonance imaging (MRI) [1, 11]. They grow chiefly in childhood and adolescence and stabilise in adulthood; they may cause pain, disfigurement, hypertrophy and erosion of adjacent structures and functional deficits [1]. The superficial diffuse form is associated with thickening, hypertrophy and hyperpigmentation of the skin, with a “bag of worms” feel on palpation, and the nodular form produces a “string of beads” sensation [1]. Most disfiguring diffuse plexiform neurofibromas become evident in the first 2 years of life, sometimes heralded by a macule of increased pigmentation [3]. Orbital and periorbital plexiform neurofibromas have an incidence below 10% in children with NF1 and cause deprivation or anisometropic amblyopia, glaucoma, ptosis, proptosis and disfigurement [11]. The ERN GENTURIS guideline defines as symptomatic plexiform neurofibromas those with persistent refractory pain, disfigurement, actual or potential functional impairment (neurological, bladder, bowel, respiratory or swallowing) or haemorrhage [11].

The French PNDS organises neurofibromas by clinicopathological correlation: cutaneous neurofibromas (present in 95% of adults, appearing mainly at puberty) are histologically diffuse neurofibromas; subcutaneous or peripheral nodular neurofibromas (around 20% of adults, from adolescence onwards), palpable rather than visible, isolated or in a string-of-beads pattern, frequently painful and sometimes with distant paraesthesiae, are nodular neurofibromas; plexiform neurofibromas (20–50% of patients) are diagnosed in early childhood, when they grow fastest, and the overlying skin may show brown or pink pigmentation, hypertrichosis and a soft, wrinkled texture; facial plexiform neurofibromas are associated with hemihypertrophy, bone dysplasia and a sphenoidal cleft [17]. Internal neurofibromas, not accessible on examination, are nodular or plexiform and strongly associated with the development of MPNST; a deep, mainly paraspinal, form has a worse prognosis [17]. Dysplastic neurofibromas (the PNDS term for the lesion that current consensus calls ANNUBP) appear after early childhood, grow faster than the plexiform neurofibroma in which they arise, have a nodular appearance on MRI and ¹⁸F-FDG PET hypermetabolism greater than that of the surrounding plexiform neurofibroma; CDKN2A/B deletion is the first step in this progression [17]. Plexiform neurofibromas grow slowly as part of their natural course before 25 years, the more so the younger the patient, which must be taken into account when interpreting an increase in volume [17].

Reference images of the condition are available in GeneReviews — café-au-lait macules (https://www.ncbi.nlm.nih.gov/books/NBK1109/figure/nf1.F1/), cutaneous neurofibromas (https://www.ncbi.nlm.nih.gov/books/NBK1109/figure/nf1.F2/) and a plexiform neurofibroma (https://www.ncbi.nlm.nih.gov/books/NBK1109/figure/nf1.F3/) — and in Figures 1 and 2 of Smith's chapter, showing café-au-lait macules, axillary freckling, pectus excavatum, Lisch nodules, plexiform neurofibromas of the upper trunk and orbit, and distal tibial pseudarthrosis [1, 28].

3.4 Central nervous system tumours

Low-grade gliomas are the most frequent central nervous system (CNS) tumours in NF1, described in around 20% of patients, and 66–75% of all CNS tumours in this population are optic pathway gliomas (OPGs), with an incidence of up to 15% in children, mean age at diagnosis of 4.5 years and the period of greatest risk in the first 6 years of life [3, 11]. Around 40–50% of children with OPG develop symptoms, but only 15–20% of patients with OPG require treatment for progression [11]. Symptoms include loss of visual acuity, visual field defect, strabismus, nystagmus, proptosis, optic disc pallor or oedema and, in chiasmatic gliomas, precocious puberty, present in 39% of children with chiasmatic OPG according to Smith [3, 11, 28]. Tubular enlargement of the optic nerves, frequently with elongation and kinking and extension to the chiasm, is characteristic of NF1-associated OPGs [3]. These tumours have a more benign course than sporadic OPGs, frequently remain stable and decline in prevalence from around 20% in children to below 5% in older adults, suggesting spontaneous regression [1]. Non-optic pathway low-grade gliomas (particularly pilocytic astrocytomas of the basal ganglia, thalamus, cerebellum and brainstem) occur in 4–5% of patients, with a mean age of 7 years and coexistence with OPG in 50–60% of children; second CNS gliomas occur in 17–20% of patients with OPG [1, 11]. High-grade gliomas, whose risk is 50 times that in the general population, arise chiefly in adults and in the cerebral hemispheres, contributing to mortality, although their prognosis is better than that of sporadic tumours [11].

T2-hyperintense lesions, termed unidentified bright objects (UBOs) or focal areas of abnormal signal intensity (FASIs), occur in more than half of children (60% according to Smith) in the basal ganglia, thalamus, brainstem, cerebellum and optic pathways, without mass effect or corresponding findings on T1 or computed tomography (CT); they represent spongiform myelinopathy, peak at around 7 years and subsequently regress, without a consistent correlation with seizures or cognitive profile [1, 3, 28]. They were not included in the 2021 criteria because of their lack of specificity and the need for anaesthesia in children [4]. Aqueductal stenosis is a rare but recognised complication in infancy, and a rapid increase in head circumference requires MRI to exclude it [3, 54].

The French PNDS adds natural history data on optic pathway glioma: a frequency of 15–20% in children in hospital series (probably lower in the overall NF1 population), a median age at detection of 4.2 years, 5-year survival of 90%, symptomatic progression in 30–50% and no progression in around half; around one third of affected children will require treatment [17]. The risk of symptomatic glioma is greatest before 7 years; female sex and familial forms are associated with a greater likelihood of progression, and retrochiasmatic involvement with a worse visual prognosis; the presence of optic pathway glioma is associated with a higher frequency of a second brain tumour outside the optic pathway [17]. Precocious puberty due to a chiasmatic glioma is not a sign of tumour progression and does not in itself justify treatment; hypotropia with limited elevation, mild exophthalmos and a relative afferent pupillary defect characterise anterior optic nerve gliomas, and often the only clue is optic disc pallor or an abnormal OCT [17]. For UBOs (unidentified bright objects), the PNDS gives prevalences above 93% before 15 years, of 57% between 16 and 30 years and of 29% after 30, with a sensitivity of 70–97% and a specificity of 79–100% for NF1; they are hyperintense on T2/FLAIR, isointense on T1, frequently nodular and well circumscribed, smaller than 1.5 cm, without mass effect or gadolinium enhancement and absent from the cortex and the periventricular region; failure to meet these criteria should raise suspicion of a tumour [17]. Extra-optic brainstem gliomas are often asymptomatic and do not require treatment, but symptomatic ones are more frequently of higher grade [17].

3.5 Neurological and neurocognitive manifestations

Macrocephaly (24% with a head circumference above 2 standard deviations) and short stature (13% below 2 standard deviations) are frequent [3]. Learning difficulties affect 50–60% of children (62% in McGaughran's series), with deficits in visuospatial skills, executive function, memory, language, attention and motor skills; mean intelligence quotient is around 1 standard deviation below the population mean (mean 88 according to Smith), and overt intellectual impairment (IQ < 70) occurs in 4–8%, around twice the general population, and is more frequent in microdeletions [1, 3, 28]. Attention deficit hyperactivity disorder (ADHD) affects 30–50% of children and adolescents and may persist into adulthood; 25% of children meet criteria for autism spectrum disorder (ASD); sleep disorders, anxiety, depression and mood disorders are more frequent than in the general population [1, 28]. Hypotonia and abnormalities of coordination, balance and fine motor skills are common [1]. Seizures occur in around 5–7%, are more frequently focal and associated with tumour, infarction or mesial temporal sclerosis [1]. Headaches, particularly migraine, affect at least half of patients [1]. Diffuse polyneuropathy, often associated with multiple nerve root tumours, affects a small percentage of patients and is associated with an increased risk of MPNST [1, 11, 52].

The French PNDS describes the neurocognitive profile by domain: visuospatial disorders (perception, spatial orientation and memory, visuoconstruction) are the most frequent and affect geometric and logico-mathematical learning; ADHD in NF1 is more frequently of the inattentive or combined type than of the hyperactive-impulsive type; articulation and oral language disorders occur, with repercussions on reading and spelling, together with executive function and working memory deficits, difficulties in verbal and visual memory encoding, social cognition disorders and ASD in 25–40% (compared with 1–2% in the general population), developmental coordination disorder, motor and execution slowness and fatigue linked to the attentional cost [17]. The IQ distribution is shifted to the left, with a mean of around 90, and intellectual disability (IQ < 70) occurs in 4–8% [17]. Epilepsy (3–13%) is mainly focal, begins in childhood or young adulthood and requires imaging to look for a structural lesion (tumour, aqueductal stenosis, vasculopathy); around 30% of epilepsies are drug-resistant, in association with severe intellectual disability [17]. Headaches are not associated with the presence of intracranial tumours or UBOs, but a change in pattern, a new-onset headache or an abnormal neurological examination warrant imaging and investigation for severe hypertension, intracranial hypertension or hydrocephalus [17]. Hydrocephalus (around 2%, mainly in adults, usually due to stenosis of the aqueduct of Sylvius from subependymal glial proliferation or, more rarely, to a mesencephalic glioma) requires urgent neurosurgery when symptomatic, because of the risk to life and vision, although some forms are well tolerated without intervention [17]. Spinal cord compression, mainly in adults, results from neurofibroma, MPNST, an intramedullary tumour or scoliosis [17]. Peripheral neuropathies (symmetrical sensory polyneuropathies, demyelinating or axonal, including small-fibre neuropathy) are associated with multiple and subcutaneous neurofibromas and with MPNST; in a new neuropathy, a compressive neurofibroma or MPNST should be sought, these being present in around 25% of cases [17]. The risk of multiple sclerosis in adults is around twice that in the general population, although the association is rare (< 1%), and its semiology may be confused with spinal cord compression by an intraspinal neurofibroma [17]. Chronic pain affects around one third of patients, may begin early, is neuropathic or nociceptive and originates from nodular or plexiform neurofibromas, MPNST, skeletal deformities, scoliosis, pseudarthrosis or fractures; pain of increasing intensity requires MPNST to be excluded first [17]. Depressive symptoms occur in 20–50% of patients, together with anxiety and personality disorders; cognitive difficulties persist into adulthood and may compromise employment, and the psychological impact, which cannot be predicted from the apparent severity, should be assessed systematically and repeatedly [17].

3.6 Ophthalmological manifestations

Lisch nodules are melanocytic hamartomas of the iris, asymptomatic rounded yellowish-brown elevations identified by slit-lamp biomicroscopy and distinguishable from iris freckles by their three-dimensional elevation; they are present in almost all adults but in fewer than half of children under 5 years [1, 52]. Choroidal abnormalities, proliferations of Schwann cells arranged in concentric rings around an axon, are not visible on conventional ophthalmoscopy, but are identified by near-infrared reflectance (NIR) imaging or optical coherence tomography (OCT) in 82–98% of patients, increase with age and distinguish NF1 from Legius syndrome [1, 4]. Retinal microvascular abnormalities and hyperpigmented fundus spots are also observed [1]. Glaucoma occurs almost exclusively in the form with an extensive ipsilateral orbital plexiform neurofibroma, and neovascular glaucoma and vasoproliferative tumours are rare [52, 55].

According to the French PNDS, Lisch nodules appear progressively from 3 years of age and reach a prevalence close to 100% by around 30 years, and hyperreflective choroidal spots (Yasunari spots) sometimes precede the nodules; one third of patients have, on fundoscopy, a corkscrew tortuosity, usually single and unilateral, of a second- or third-order retinal venule [17]. In association with a plexiform neurofibroma of the ipsilateral upper eyelid, congenital ectropion uveae and high unilateral axial myopia may coexist (with or without glaucoma; without glaucoma the term buphthalmos without glaucoma is used); glaucoma, congenital or juvenile and frequently severe, is present in 20–50% of children with orbitofacial involvement and results mainly from progressive endothelialisation of the iridocorneal angle, so it should be screened for early and treated in a specialist setting [17]. Papilloedema may result from optic pathway glioma, aqueductal stenosis or pseudotumour cerebri secondary to a spinal cord tumour and may progress to optic atrophy; retinal hamartomas with a risk of detachment, vasoproliferative tumours and ischaemia due to retinal vasculopathy have also been described [17].

3.7 Skeletal manifestations

Long-bone dysplasia, almost always unilateral and more frequent in the tibia and fibula, is congenital and manifests in infancy as anterolateral bowing of the leg, with narrowing of the medullary canal and cortical thickening at the apex of the curve; it may progress to fracture and pseudarthrosis (2% of patients), whose surgical treatment is difficult [1, 3, 4]. Sphenoid wing dysplasia presents with orbital asymmetry, may be progressive and produce pulsatile enophthalmos; it does not count as a separate criterion when coexisting with an ipsilateral orbital plexiform neurofibroma [1, 4]. Scoliosis affects around 10% of patients (11% according to the Oxford Desk Reference); the dystrophic form, with dysplastic vertebrae and acute angulation in a short segment, appears between 6 and 10 years and may progress within months, whereas the non-dystrophic form resembles idiopathic adolescent scoliosis [1, 3, 28]. Other abnormalities include pectus excavatum, vertebral body scalloping due to dural ectasia, tapering pedicles, enlarged intervertebral foramina, cystic and lytic bone lesions and short stature [28, 52]. Osteopenia is frequent at all ages and osteoporosis occurs earlier than in the general population (~20% in middle adulthood), associated with low 25-hydroxyvitamin D concentrations, elevated parathyroid hormone and increased bone resorption [1]. In intramedullary long-bone lesions, the differential diagnosis with fibrous dysplasia takes into account that, when the hip is involved, NF1 usually produces coxa valga, whereas fibrous dysplasia produces “shepherd's crook” coxa vara [56].

The French PNDS adds that long-bone dysplasia may also affect the fibula, radius and ulna and is diagnosed clinically in the presence of limb bowing or a pathological fracture, with a risk of secondary pseudarthrosis due to fractures with delayed union [17]. Sphenoid dysplasia, congenital, usually unilateral and involving the greater wing, may be primary, with a temporal meningeal anomaly and accumulation of cerebrospinal fluid, or secondary to an orbital plexiform neurofibroma (more than 50% of cases), and may cause pulsatile exophthalmos or herniation of the temporal lobe [17]. Calvarial defects in the region of the lambdoid suture, palpable and asymptomatic, should be sought before excision of scalp neurofibromas, because of the risk of underlying bone aplasia [17]. Vertebral dysplasias (scalloping, thinned pedicles, enlarged intervertebral foramina) are found in more than 70% of patients on spinal MRI; dysplastic scolioses, in young children, progress very rapidly in around 5% of cases, with a restrictive syndrome due to impaired spinal growth and a risk of paraplegia due to rotatory dislocation, and the associated vertebral dysplasias are the main aggravating factor [17]. Mineralisation abnormalities are associated with abnormalities of calcium–phosphate metabolism and with vitamin D deficiency in children and adults, with an increased fracture risk; the effect of vitamin D supplementation on bone density and fractures in NF1 has not been established and bisphosphonates have not been shown to be effective, so the PNDS recommends following the supplementation recommendations for at-risk populations, guided by measurement of 25-hydroxyvitamin D [17]. Macrocephaly, due to increased brain volume, rarely has consequences, but an acceleration in head circumference growth requires exclusion of hydrocephalus due to aqueductal stenosis, Chiari malformation or an intracranial tumour [17]. Non-ossifying fibromas, generally incidental, appear between 10 and 20 years in the distal metaphyses of the femur and tibia and in the humeral heads; the former “Jaffe–Campanacci syndrome” is now considered a form of NF1, and large lesions require surveillance because of the risk of fracture [17].

3.8 Vascular and cardiac manifestations

Hypertension occurs in at least 15–20% of patients, more frequently in adults, and may result from renal artery stenosis or middle aortic syndrome (particularly under 20 years), phaeochromocytoma or paraganglioma, or be essential [1, 3]. NF1 vasculopathy, due to intimal and medial hyperplasia, causes stenoses, ectasia and aneurysms of aortic branches, carotid, cerebral and coronary arteries, and moyamoya, whose risk triples after cranial irradiation; stroke is more frequent and occurs earlier than in the general population [1, 28]. Congenital circulatory system abnormalities are 3.35 times more frequent (Finland), with pulmonary valve stenosis (1–2% of patients, associated with non-truncating variants) and mitral valve abnormalities the most common; congenital heart defects and hypertrophic cardiomyopathy are more frequent in microdeletions [1, 3]. Pulmonary hypertension, rare but severe in older adults, is associated with diffuse pulmonary disease (upper-lobe cysts and bullae and basal interstitial disease) present in 10–20% of adults [1].

The French PNDS reports hypertension in 16% of children, with a frequency that increases with age and, in children, most often originating from renal artery stenosis (around 1% of patients with NF1); in hypertension it recommends CT or MR angiography of the renal arteries, because Doppler ultrasound, being less sensitive, cannot exclude stenosis, and it stresses the possibility of recurrence after revascularisation and the need for prolonged surveillance [17]. Congenital heart defects are around 3 times more frequent (0.4–27% depending on the series, generally small), predominantly pulmonary stenosis associated with in-frame missense variants and with the NF1–Noonan and Watson phenotypes, followed by aortic stenosis, coarctation, atrial septal defects and hypertrophic cardiomyopathy; abnormalities of diastolic function in children, early cardiac remodelling in young hypertensive patients and intracardiac neurofibromas have been described [17]. Vascular dysplasias (estimated incidence 0.4–6.4%) are a frequent cause of death before 30 years; in pooled series, renal artery stenosis accounts for 41%, cerebrovascular vasculopathy for 19% and abdominal aortic abnormalities for 12%, and small-vessel involvement by endothelial dysplasia explains haemorrhages that are spontaneous, follow minimal trauma or occur during surgery [17]. Cerebrovascular involvement (around 5%, children and young adults) consists mainly of occlusion of the large anterior cerebral arteries, with involvement of the internal carotid artery at its exit from the cavernous sinus, up to the moyamoya pattern; in adults, strokes complicate aneurysms, emboli or arteriovenous fistulae, and the risk of stroke and moyamoya increases after cranial irradiation [17]. Although studies have not demonstrated haemostatic abnormalities, the PNDS recommends haemostasis assessment before any surgery, dental procedure or obstetric intervention, because of the haemorrhagic risk of neurofibroma excision [17]. Pulmonary manifestations (emphysematous bullae, basal interstitial fibrosis) occur in around 5%; NF1-associated pulmonary arterial hypertension, attributed to vascular dysplasia, predominates in females, is diagnosed at around 64 years, with a DLCO of around 30% and NYHA class III–IV heart failure in 90% of cases, and has a poor prognosis despite specific treatment, lung transplantation being an option for eligible patients [17].

3.9 Endocrine, growth and pregnancy manifestations

Precocious puberty is usually associated with optic chiasm glioma, and growth hormone hypersecretion may occur in the same setting; delayed puberty is common, and pubertal growth velocity is reduced in both sexes [1, 3]. NF1-specific growth charts are available [28]. During pregnancy, preterm delivery, caesarean section, hypertension and placental abruption are more frequent; many women report an increase in the number and size of neurofibromas, but a whole-body MRI study did not demonstrate a persistent increase in tumour volume [1].

The French PNDS adds that birth weight is significantly higher than in the general population, that up to one third of patients have short stature (with normal growth until puberty and subsequent deceleration), associated with a normal somatotropic axis as well as with growth hormone deficiency or excess, and that precocious puberty occurs in around 3% of patients, one fifth of whom have an optic pathway glioma, which should therefore be sought; conversely, 10–35% of children with optic pathway glioma have precocious puberty [17]. Oestrogen, progesterone, androgen and growth hormone receptors have been described in neurofibromas, which would explain their development at puberty and in pregnancy, but no clinical or epidemiological data contraindicate steroid hormones or growth hormone; high-dose progestogen-only contraception could increase the size of neurofibromas, unlike combined oestrogen–progestogen contraception, but, given the scarcity of data, no contraceptive method is contraindicated, and there are no published data on ovarian stimulation, including for preimplantation genetic diagnosis [17]. Coexistence with multiple endocrine neoplasia is possible without increased incidence or a causal relationship [17].

3.10 Cancer spectrum

Lifetime cumulative cancer risk is 59.6% (general population 30.8%), with a median age at diagnosis of 39 years and particularly high standardised incidence ratios in children and women under 30 years; life expectancy is reduced by 8 to 15 years, chiefly because of malignancies [11]. In the Finnish registry of 1,404 patients (19,076 person-years), standardised incidence ratios (SIRs) were 2,056 for MPNST and 37.5 for CNS tumours, and breast cancer SIR was 11.1 in women under 40 years [14]. Table 3 summarises the tumour spectrum.

Table 3. Neoplasms associated with NF1 [1, 3, 11, 13, 14, 39, 57, 58]

MPNST is the most frequent malignancy and the main cause of reduced life expectancy [1, 3]. Evans and colleagues calculated, in a population registry of 4.1 million inhabitants in North West England, an annual incidence of 1.6 per 1,000 patients and a lifetime risk of 8–13%, with a median age at diagnosis of 26 years, compared with 62 years in sporadic MPNST [13]. Most, if not all, arise in pre-existing diffuse or nodular plexiform neurofibromas, particularly deep lesions (brachial plexus, sciatic nerve); almost all present with persistent pain, rapid growth or a change in consistency [1, 3]. According to ERN GENTURIS, high-risk groups comprise NF1 microdeletion including SUZ12, missense variants in codons 844–848, previous ANNUBP, high internal tumour burden on whole-body MRI (or large or multiple plexiform neurofibromas in its absence), neurofibromatous neuropathy, previous radiotherapy and a relative with NF1 and MPNST [11].

Breast cancer risk is 2 to 3 times that in the general population over a lifetime and 4 to 5 times before 50 years; the increase in incidence is essentially confined to women between 30 and 50 years, and tumours are more frequently oestrogen and progesterone receptor negative and HER2 positive, factors associated with poorer prognosis [3, 11]. NF1 is listed among moderate-risk breast cancer genes in the Oxford Desk Reference, and the ESMO handbook includes it in the group of highly actionable genes for which germline-directed analysis is recommended when detected in tumours of patients under 30 years [57, 59]. GISTs in NF1 are typically multiple, located in the small intestine and lack KIT and PDGFRA alterations, making them insensitive to imatinib [11, 39]. Familial GIST due to KIT and PDGFRA is part of the differential diagnosis. Phaeochromocytoma and paraganglioma occur in 1–5% of adults in cross-sectional studies, frequently as incidental findings, and NF1 accounts for around 1% of phaeochromocytomas in series of predisposing diseases [11, 58]. Although an association between juvenile xanthogranuloma and JMML has been described (more than 25 cases), a registry study did not demonstrate an increased risk, and leukaemia screening is not recommended [3, 11].

The French PNDS summarises the overall cancer risk as around 4 times that of the general population, with an estimated lifetime risk of 60% and around 20% of patients having had a cancer by 50 years; cancer is the main cause of death and soft-tissue and brain tumours are the most frequent [17]. MPNST accounts for 38–45% of cancer deaths, has a lifetime risk of 8–16%, arises mainly between 20 and 35 years, around 10 years earlier than in sporadic forms, and is high grade in 70% of cases; 5-year survival has risen from 21% in older series to 34–60% in series after 2000, and the PNDS concludes from these series that the prognosis of MPNST in NF1 is similar to that of sporadic MPNST [17]. The risk factors collated by the PNDS are numerous subcutaneous neurofibromas, internal neurofibromas, peripheral neuropathy, whole-gene deletion, previous radiotherapy and dysplastic neurofibromas; in children MPNST is exceptional (risk of around 0.5%), with survival and treatment response similar to those in adults [17]. Table 3A summarises the poor prognostic factors in adults.

Table 3A. Poor prognostic factors for MPNST in adults with NF1, according to the French PNDS [17]

For breast cancer, the PNDS gives a risk 2 to 5 times that of the population, mortality up to 5 times higher and a cumulative risk of contralateral cancer of 27% at 20 years after the first tumour; after 50 years the excess risk is lower, of the order of 2-fold [17]. In the five-country cohort underlying this last figure (142 women with NF1 and breast cancer, median age 46.9 years), the cumulative risk of contralateral cancer was 26.5% at 20 years and 10-year cancer-specific survival was 64.2%, compared with 91.2% in an age-matched increased-risk population without NF1 [60]. The PNDS attributes to the Finnish registry a cumulative risk of 23% at 80 years (around 10% in the general population) and of around 10% at 50 years (around 2%), and 5-year survival of 68% versus 88%; the values published in that registry (1,404 patients with NF1, 32 breast cancers) are a lifetime risk, up to 85 years, of 18.0% (95% CI 11.4–24.1%) versus 9.7% in the Finnish female population, a cumulative risk of 7.8% at 50 years versus 2.1% and of 4.7% at 40 years versus 0.45%, and 5-year overall survival of 68.1% in patients with NF1 versus 82.0% in controls matched for age, year of diagnosis and oestrogen receptor status (P = 0.053) [61]. The most common subtype is invasive ductal carcinoma, and treatment does not differ from that of patients without NF1 [17]. Coexistence of pathogenic NF1 and BRCA1 variants is documented in rare cases: in a woman with NF1 and breast cancer with a suggestive personal or family history, the PNDS recommends a breast–ovarian predisposition gene panel according to the French criteria (BRCA1, BRCA2, PALB2, CDH1, MLH1, MSH2, MSH6, PMS2, RAD51C, RAD51D, PTEN and TP53 panel; indications include, among others, breast cancer before 41 years, triple-negative before 51, bilateral with the first before 51, in a man, three breast cancers in the same lineage with one before 60, and non-borderline, non-germ-cell ovarian cancer before 70–76 years or high-grade serous ovarian cancer) [17]. Numerous cutaneous or intramammary neurofibromas may hinder surveillance and delay diagnosis.

For phaeochromocytoma and paraganglioma, the PNDS reports frequencies of 1–15% depending on the study, with 8–15% in prospective studies with consecutive inclusion (the classic figure of 1% is an underestimate owing to asymptomatic or non-secreting forms), a mean age at diagnosis close to 40 years, unilaterality in 60–80%, adrenal location in more than 90% and absence of metastases in more than 90%; around 80% are asymptomatic and more than half non-secreting, but secreting tumours may decompensate acutely during surgery, and metachronous contralateral phaeochromocytoma occurs in around 20% [17]. Warning signs are hypertensive crises with sweating, palpitations, headache, anxiety or agitation, fluctuating or resistant hypertension, malaise, weight loss, abdominal or lumbar pain and intermittent hyperglycaemia or diabetes [17]. Diagnosis is based on plasma or urinary metanephrines and adrenal CT or MRI; ¹⁸F-DOPA PET, when available, is preferred to MIBG scintigraphy, particularly for extra-abdominal paragangliomas [17]. GISTs occur in 6–7% of patients, present with abdominal pain, obstruction, haemorrhage or perforation and, because they are located in the small intestine, frequently escape upper endoscopy and colonoscopy, requiring CT, MRI or PET and sometimes guided fine-needle biopsy; most are indolent, of spindle-cell type with a low mitotic index, but there are aggressive forms with hepatic and peritoneal metastases [17]. Gastrointestinal neuroendocrine tumours (duodenal and pancreatic somatostatinomas, insulinomas, small-intestinal carcinoids) should prompt a search for MEN1, and rhabdomyosarcoma (around 1%) arises early, mainly in the genitourinary tract, as a palpable mass [17].

3.11 Manifestations of Legius syndrome

Legius syndrome is characterised by multiple café-au-lait macules, with or without intertriginous freckling, without neurofibromas, optic pathway gliomas, Lisch nodules or typical NF1 bone lesions [3, 5]. Around half of patients meet the NIH pigmentary criteria for NF1 [4, 5]. Other manifestations include macrocephaly (head circumference ≥ 97th percentile in around 40% in one cohort), subcutaneous lipomas (32% in the initial series, less frequent in subsequent series), Noonan facial features (14–40%), learning difficulties, ADHD and speech and motor delay, pectus, scoliosis, postaxial polydactyly and, rarely, pulmonary stenosis, seizures and hearing loss [5]. In the GeneReviews review, pectus or sternal abnormalities were described in 7 of 30 patients in one series, scoliosis in six patients (one congenital and one requiring surgery), postaxial polydactyly in four, seizures in seven, hearing loss in four (mild in two) and pulmonary valve stenosis in four, with isolated reports of mitral valve prolapse and paroxysmal atrial tachycardia [5]. Lisch nodules have been described only in two brothers from one family, and two patients had T2-hyperintense lesions on brain MRI, so these lesions alone do not distinguish the two conditions [5]. Cognitive performance is less affected than in NF1: in a study of 15 patients, performance IQ was lower than in unaffected relatives, without a difference in overall IQ [5]. In a comparative imaging study, 89% of patients with NF1 had FASIs, compared with 0% of the 16 patients with Legius syndrome [5]. No clinically relevant tumour predisposition has been demonstrated; a slightly increased risk of acute leukaemia has been suggested (one case of acute myeloblastic leukaemia and a SPRED1 frameshift variant identified in 230 screened paediatric leukaemias), and a 2022 review (Ney and colleagues) concluded that cancer incidence in Legius syndrome is similar to that in the general population and that tumour surveillance is not recommended [3, 5, 62]. The tumours described correspond to isolated reports — Wilms tumour and tubular colonic adenoma in the same patient, non-small-cell lung carcinoma, tenosynovial giant cell tumour, ovarian dermoid tumour, abdominal wall desmoid tumour and unilateral vestibular schwannoma — and do not define an established risk [5]. Estimated prevalence is 1:46,000 to 1:75,000 (around 4% of patients followed in neurofibromatosis clinics) [4, 5].

4. Clinical Recognition and Warning Signs

4.1 When to suspect NF1

Clinical suspicion of NF1 should arise in the presence of any manifestation included in the diagnostic criteria: six or more café-au-lait macules larger than 5 mm before puberty or larger than 15 mm after puberty, axillary or inguinal freckling, two or more neurofibromas or one plexiform neurofibroma, optic pathway glioma, two or more Lisch nodules or choroidal abnormalities, a characteristic bone lesion, an affected parent or a germline pathogenic NF1 variant [1, 4]. NF1 is the most frequent cause of multiple café-au-lait macules; 10–20% of adults have at least one macule, but more than one in a normal newborn is rare [53]. The Oxford Desk Reference systematises history-taking and examination: detailed family history with an active search for relatives with macules, cutaneous nodules or tumours; medical and surgical history that may reveal complications not attributed to NF1; development and schooling, with attention to language, reading, visuospatial skills, motor skills and attention; seizures and other neurological symptoms; and pain or rapid growth of soft-tissue lesions [3]. Examination should record height, weight, head circumference and blood pressure; examine the skin (with a Wood's lamp in light skin); carefully document soft-tissue lesions; assess for scoliosis, tibial bowing in children under 2 years, a pulmonary stenosis murmur, precocious puberty and visual acuity and visual fields in each eye; and examine parents without relying on self-assessment, because localised mosaic NF1 is easily overlooked [3]. The French PNDS (National Diagnosis and Care Protocol) sets a referral threshold: any child with at least one diagnostic criterion and any adult with two or more should be referred to an NF1 expert for initial assessment, confirmation of the diagnosis and definition of the surveillance strategy; at each visit to the attending physician, the factors associated with morbidity and mortality and the main complications should be sought [17].

4.2 Warning signs requiring urgent assessment

Certain signs require rapid access to specialist assessment and imaging. Table 4 brings together the warning signs defined by the Oxford Desk Reference, GeneReviews and ERN GENTURIS.

Table 4. Warning signs in NF1 [1, 3, 11]

Facial dysmorphism (particularly hypertelorism), heart defect, intellectual impairment, overgrowth, large hands and feet and early numerous neurofibromas suggest a 17q11.2 microdeletion, whose identification has implications because of the increased MPNST risk [3, 12]. Noonan features, pulmonary stenosis and short stature suggest a missense variant (p.Arg1809, p.Arg1276, p.Lys1423) or an in-frame defect in the GRD domain [42, 44, 48]. Affected siblings with clinically normal parents, consanguinity, café-au-lait macules with irregular borders and hypopigmented areas, or a family history of Lynch syndrome raise the possibility of constitutional mismatch repair deficiency (CMMRD) [1, 4, 63].

4.3 Severity stratification: “complicated” NF1 and the “at-risk” phenotype (French PNDS)

The French PNDS (National Diagnosis and Care Protocol) distinguishes uncomplicated NF1, the most frequent form, from “complicated” NF1, defined by the presence of at least one of the manifestations in Table 4A; the classification is dynamic, because an uncomplicated patient may develop a complication and a complication may resolve, and patients with complicated NF1 should be followed in an NF1 centre with a multidisciplinary team [17].

Table 4A. Manifestations defining “complicated” NF1 according to the French PNDS [17]

The “at-risk” MPNST phenotype is defined by the presence of 10 or more subcutaneous neurofibromas or at least one internal neurofibroma, frequently associated with peripheral neuropathy; in these patients the PNDS justifies, by expert recommendation, early screening for MPNST by whole-body MRI (to assess the internal neurofibroma burden) combined with ¹⁸F-FDG PET (PET-MRI or PET-CT), whereas systematic screening for internal neurofibromas by MRI is not recommended in all patients and is performed in the presence of a clinical warning sign [17]. When the diagnosis is disclosed, the patient and family should be informed of the signs requiring prompt consultation: pain, increase in size or change in texture of a neurofibroma, and neurological deficit [17].

For assessment of the physical and psychological impact, the PNDS lists non-specific and specific instruments: the Riccardi severity index (grades 1 to 4, from forme fruste to a severe form with involvement that is difficult to control or reduced life expectancy) and the Ablon visibility index (grades 1 to 3), neither of them recently updated; the Skindex-France, DLQI and SF-36 quality-of-life scales; the specific INF1-QOL questionnaire, which correlates with severity; in children, the CDLQI, DISABKIDS and PedSQL, bearing in mind that parents overestimate the burden felt by the child; and the Burden of NF1 (BoN) questionnaire, which assesses daily burden beyond quality of life [17].

5. Diagnosis: Criteria, Tests and Algorithms

5.1 Revised 2021 criteria (Legius et al.)

The 1988 NIH criteria required two or more of seven manifestations: six or more café-au-lait macules (>5 mm before puberty, >15 mm afterwards), two or more neurofibromas or one plexiform neurofibroma, axillary or inguinal freckling, optic glioma, two or more Lisch nodules, a characteristic bone lesion (sphenoid dysplasia, long-bone cortical thinning with or without pseudarthrosis) and an affected first-degree relative [3, 10, 28]. In the North West England registry, the frequencies of the criteria were 86.7% for macules, 59.4% for cutaneous neurofibromas, 45.5% for subcutaneous neurofibromas, 15.3% for plexiform neurofibromas, 83.8% for axillary freckling, 42.3% for inguinal freckling, 63% for Lisch nodules and 71.2% for a first-degree relative [3, 51]. In the cohort of 2,432 individuals referred for molecular testing analysed by Messiaen and colleagues, the NIH criteria had, below 7 years, sensitivity of 58.2%, specificity of 88.6%, positive predictive value of 80.5% and negative predictive value of 71.5% for detection of a variant in NF1 or SPRED1; from 7 years, sensitivity rose to 85.3% and specificity was 74.3% [4, 64].

The I-NF-DC group's modified Delphi process, initiated in 2017 with 76 experts (74 participated) and completed in January 2020, produced the criteria presented in Table 5 [4].

Table 5. Revised diagnostic criteria for NF1 (2021) [4]

ᵃ If only café-au-lait macules and freckling are present, the most likely diagnosis is NF1, but exceptionally this may be another condition, such as Legius syndrome; at least one of the two pigmentary findings should be bilateral. ᵇ Sphenoid wing dysplasia does not constitute a separate criterion in the presence of an ipsilateral orbital plexiform neurofibroma.

Changes from 1988 were inclusion of choroidal abnormalities as an alternative to Lisch nodules because of their high sensitivity and specificity and ability to distinguish NF1 from Legius syndrome; introduction of genetic diagnosis without implying that testing is mandatory; replacement of “long-bone cortical thinning” with “anterolateral tibial bowing”, which is the primary lesion; the requirement for bilaterality of at least one pigmentary finding; and restriction of the family criterion to “one parent”, excluding siblings (if only siblings are affected, CMMRD should be considered) and offspring (an adult with only one criterion and an affected child should raise suspicion of mosaicism) [4]. Only variants classified as pathogenic count as a criterion; likely pathogenic variants, variants of uncertain significance (VUS) or variants identified only in tumour tissue do not meet the criterion, and isolated identification of an NF1 variant by panel, exome or genome sequencing without a clinical manifestation does not establish the diagnosis: it requires confirmation by an orthogonal method and verification that it is not a somatic variant due to clonal haematopoiesis, therapy, haematological malignancy or circulating tumour cells [1, 4]. MRI FASIs, naevus anaemicus and juvenile xanthogranuloma were not included [4]. The consensus emphasises that systematic investigation of every manifestation is unnecessary, that biopsies or examinations under general anaesthesia should not be performed solely to confirm criteria, and that, in the absence of molecular testing, serial observation confirms the diagnosis in most children; caution is needed when diagnosing NF1 in the absence of macules and neurofibromas [4]. Genetic testing was recommended with consensus above 80% for patients with segmental findings, families with two affected siblings and clinically normal parents, and children who meet criteria only through pigmentary findings [4].

The French PNDS (National Diagnosis and Care Protocol) adopts the revised 2021 criteria and justifies the revision by the low sensitivity of the 1988 criteria in early childhood and by the pigmentary overlap with Legius syndrome: in a study of 71 patients under 20 years with six or more café-au-lait macules and no other criterion, 66% had NF1, 9% Legius syndrome and 25% no associated disease [17]. It also reproduces an algorithm published by Ben-Shachar and colleagues that classifies children with isolated café-au-lait macules and no affected parent as being at low or high risk of NF1 before molecular testing, and lists clinical, imaging and molecular signs that may complement the criteria: molecular analysis of NF1, UBOs on MRI, juvenile xanthogranuloma, naevus anaemicus, hypochromic macules, diffuse hyperpigmentation, vascular malformations (violaceous macules and ruby spots), hyperreflective choroidal spots, increased head circumference and hypertelorism, neurodevelopmental disorders, headache and epilepsy, endocrine abnormalities (precocious puberty, faltering height and weight, bone mineralisation abnormalities, hypovitaminosis D) and tumours [17].

5.2 Criteria for mosaic NF1

Table 6. Diagnostic criteria for mosaic NF1 (2021): any one of the following is sufficient [4]

If only macules and freckling with a segmental distribution are present, the most likely diagnosis is mosaic NF1, but this may be mosaic Legius syndrome or CMMRD [4]. Detection of the variant in patients with a localised phenotype requires techniques with appropriate sensitivity for low allele fractions and, if absent in blood, analysis of the relevant affected tissue: cultured melanocytes from café-au-lait macules (not keratinocytes or fibroblasts) or Schwann cells from neurofibromas [4]. García-Romero and colleagues' systematic review brought together 320 individuals from 111 articles (1977–2012): most had only pigmentary changes or only neurofibromas, those with an exclusively pigmentary phenotype were identified earlier, 76% had mosaicism localised to one segment and the remainder generalised, 29% of the 157 individual cases had NF1-associated complications and, in one large series, 6.5% had offspring with full NF1; the authors recommend complete physical examination, blood and skin genetic testing, counselling and surveillance in all suspected cases [47]. Distribution ranges from a narrow band to a quadrant and occasionally half the body, usually unilateral but sometimes bilateral, symmetrical or asymmetrical [3, 45].

5.3 Criteria for Legius syndrome

Table 7. Diagnostic criteria for Legius syndrome and mosaic Legius syndrome (2021) [4]

The Legius syndrome GeneReviews, updated in December 2025, reproduces these criteria with a threshold of five or more macules and accepts pathogenic or likely pathogenic SPRED1 variants [5]. A diagnosis of Legius syndrome can be established with certainty only by identifying a SPRED1 variant [3, 5].

5.4 Molecular testing strategy

Table 8. Molecular testing methods in NF1 and Legius syndrome [1, 4, 5]

The combined approach of DNA sequencing and dosage analysis detects a pathogenic variant in around 90% of classic non-founder patients (second generation), rising to 95–97% with RNA sequencing; rates are lower in individuals with few symptoms or an atypical presentation [4]. In the South African cohort of 26 patients investigated by an exon panel and MLPA, without RNA analysis, the detection rate was 58% [65]. A negative test does not exclude NF1 [1]. The French PNDS stresses the usefulness of molecular testing mainly in unusual or incomplete phenotypes and, in children with multiple macules and no family history, to distinguish NF1 from Legius syndrome, Noonan syndrome and CMMRD; it reports a sensitivity above 95% for next-generation sequencing in routine practice, more than 3,500 NF1 variants recorded in the LOVD database, large deletions in 5–10% of patients and few genotype–phenotype correlations beyond the variants already described, the variability being attributed in part to modifier genes [17]. Identification of a pathogenic or likely pathogenic variant allows targeted testing of relatives and is an essential prerequisite for prenatal or preimplantation diagnosis (section 8.4) [17]. The proposed algorithm, integrating recommendations from GeneReviews, the 2021 consensus and the Oxford Desk Reference, follows the steps in Table 9.

Table 9. Diagnostic algorithm [1, 3–5]

5.5 Initial assessment after diagnosis

Initial assessment includes a dermatological examination for cutaneous and plexiform neurofibromas; ophthalmological assessment with fundoscopy, slit lamp, NIR imaging or OCT and visual acuity; neurological examination assessing for seizures, headaches and pain; developmental and neuropsychological assessment; assessment for asymmetry, long-bone, sphenoid and vertebral dysplasia, scoliosis and recurrent fractures; blood pressure, with investigation for renovascular disease, middle aortic syndrome, structural renal abnormalities and phaeochromocytoma in hypertensive patients; clinical assessment for congenital heart defects and cardiomyopathy; recording height, weight and head circumference on appropriate charts; and genetic counselling [1]. Routine brain MRI in asymptomatic individuals is controversial: proponents cite early detection of structural abnormalities, tumours and vasculopathy, and opponents the clinical uncertainty of FASIs, cost, sedation in young children and repeated examinations for reassurance [1, 3]. The Oxford Desk Reference indicates MRI for rapid head circumference growth, focal neurological signs, epilepsy, visual abnormalities, precocious puberty or growth deceleration [3]. The Brazilian chapter also includes MRI for nodular and plexiform neurofibromas for precise localisation and knee radiographs in adolescence [52]. In the initial assessment, the French PNDS adds screening of all first-degree relatives, with full clinical examination, slit-lamp biomicroscopy and genetic testing in case of doubt, identification of high-risk patients (10 or more subcutaneous neurofibromas), the offer of consultations with a psychologist and a genetic counsellor, and provision of the contact details of the reference or competence centre and of patient associations [17].

5.6 Limitations of genetic tests and report content

Test performance depends on the method. Sequencing restricted to coding genomic DNA does not detect some splicing variants (including deep intronic variants), deletions and duplications of one or more exons require dosage analysis (MLPA or validated coverage analysis), whole-gene deletions and 17q11.2 microdeletions require MLPA or microarray, balanced rearrangements and mobile element insertions may require karyotyping, genomic sequencing or RNA analysis, and mosaicism requires a technique sensitive to low allele fractions and, when blood testing is negative, investigation of affected tissue (cultured melanocytes from macules or Schwann cells from neurofibromas) [1, 4]. A negative result does not exclude NF1 in a phenotype meeting the clinical criteria [1, 4]. The report should identify the gene, versioned transcript (MANE Select NM_001042492.3 for NF1; NM_152594.3 for SPRED1), normalised HGVS nomenclature, zygosity and allele fraction (relevant to the mosaicism criterion), classification according to ACMG/AMP criteria and method limitations; when the literature uses the historical numbering NM_000267, both should be stated (section 2.5) [4, 6]. For example, the variant in case 10.1, NM_001042492.3:c.1381C>T, p.(Arg461Ter), is classified as pathogenic in ClinVar (VCV000237514), and the deletion p.Met992del as pathogenic/likely pathogenic (VCV000000363) [6]. Only pathogenic variants count as a 2021 diagnostic criterion; a variant of uncertain significance does not become pathogenic through coexistence with café-au-lait macules, segregation in a small family or response to a MEK inhibitor, and RNA, segregation or tissue studies should be described as additional investigation until formal reclassification [4].

6. Differential Diagnosis

More than 100 genetic conditions with café-au-lait macules or other isolated manifestations of NF1 have been described, but few are confused with it [1, 52]. In most adults, the diagnosis is not in doubt; in most children with only macules and freckling, the phenotype is indistinguishable from Legius syndrome [3]. Table 10 presents the entities to consider.

Table 10. Differential diagnosis of NF1 [1, 3, 5, 33, 52, 53, 56, 63, 66–68]

Distinguishing NF1 from Legius syndrome has direct implications for surveillance, because NF1 cancer and ophthalmological surveillance do not apply to Legius syndrome [5]. Clinically, the distinction may be impossible in a young child, since neurofibromas and Lisch nodules arise only in later childhood or adolescence; examination of the parents may distinguish the conditions, and in simplex cases reassessment after adolescence or molecular testing is necessary [1, 5]. Among 253 individuals with familial café-au-lait macules without other manifestations, 86.6% had an NF1 variant, 7.1% a SPRED1 variant and 6.3% neither [1, 64]. Phenotypes associated with p.Met992del and p.Arg1809 variants resemble Legius syndrome because of the paucity of neurofibromas [5, 41, 42]. CMMRD has an NF1-like cutaneous phenotype, with a high risk of childhood malignancies and a 25% recurrence risk among siblings because it is autosomal recessive; it should be actively considered in consanguineous families with café-au-lait macules [3, 63, 69]. The French PNDS (National Diagnosis and Care Protocol) also notes two overlap situations: coexistence of NF1 with multiple endocrine neoplasia, without a causal relationship, in which phaeochromocytoma and skeletal abnormalities may be attributed to either condition, and multiple sclerosis in adults, whose semiology may be confused with spinal cord compression by an intraspinal neurofibroma [17].

7. Treatments and Risk Management Strategies

7.1 General principles

There is no curative treatment for NF1; management is multidisciplinary, ideally in a neurofibromatosis clinic with experienced specialists, and combines structured clinical surveillance, symptomatic treatment of complications and, for symptomatic inoperable plexiform neurofibromas, targeted therapy with MEK inhibitors [1, 11]. Radiotherapy should be avoided whenever possible because of the risk of MPNST and other gliomas in the irradiated field, malignant transformation of pilocytic astrocytoma and moyamoya [1, 3, 11]. In Europe, ERN GENTURIS published tumour surveillance guidelines in 2023 (67 recommendations after three Delphi rounds with 32 experts from 11 countries, 6 patient representatives and 59 external experts); in the USA, the American Academy of Pediatrics (AAP), with the American College of Medical Genetics and Genomics (ACMG), published paediatric supervision recommendations in 2019, and ACMG recommendations for adults in 2018 [11, 15, 16]. In France, the PNDS (French National Diagnosis and Care Protocol) of the Haute Autorité de Santé, drawn up by the Neurofibromatoses reference centre (first version 2016, updated with a literature review from August 2013 to December 2020 and text of August 2021), covers diagnosis, treatment of complications, pregnancy, mosaic NF1 and follow-up by age; its recommendations rest mainly on expert agreement, and the differences from the other guidelines are brought together in section 7.11 [17].

7.2 Targeted therapy: MEK inhibitors

Loss of neurofibromin maintains RAS in the active state and the RAF–MEK–ERK cascade constitutively signalling; selective allosteric MEK1/2 inhibitors block ERK phosphorylation and reduce proliferation of NF1⁻/⁻ Schwann cells and plexiform neurofibroma volume [1, 4]. Table 11 presents the approved drugs.

Table 11. MEK inhibitors approved for NF1-associated plexiform neurofibromas [18–25, 70–72]

In the SPRINT trial, initiated in August 2015, the most frequent manifestations of plexiform neurofibromas were disfigurement (44 patients), motor dysfunction (33) and pain (26); after one year of treatment, pain intensity decreased by a mean of 2 points, with clinically significant improvement in pain interference with everyday life (38% according to children and 50% according to parents), quality of life (48% and 58%), strength (56%) and range of movement (38%); five patients discontinued because of toxicity and six had progression [70]. Median time to response was 7.2 months [18]. In KOMET, the first randomised placebo-controlled trial in adults, selumetinib (25 mg/m² 2 times/day) produced a rapid response (median 3.7 months), with ORR 20% (14/71) compared with 5% (4/74) at cycle 16, reduction in chronic pain and analgesic use, without a significant difference in PlexiQoL; 86% of responders maintained their response for at least 6 months [19, 71]. In ReNeu, the largest multicentre plexiform neurofibroma trial reported by 2025, the most frequent treatment-related adverse effects were acneiform dermatitis, diarrhoea and nausea in adults and acneiform dermatitis, diarrhoea and paronychia in children, with early sustained improvements in pain and quality of life [72].

Management of adverse effects follows the official labels [18, 22]. Before starting, an echocardiogram should be obtained (repeated every 3 months in the first year and subsequently every 6 months with selumetinib), a complete ophthalmological assessment (repeated periodically and for visual symptoms), CPK and, in women of childbearing potential, a pregnancy test and effective contraception [18, 22]. A grade 2 reduction in ejection fraction (40–50%, or a 10–19% fall from baseline) occurred in 17% of children and 14% of adults treated with selumetinib; diarrhoea, occurring in 59% of children (grade 3 in 10%), is treated with loperamide from the first episode of loose stools and increased fluid intake; acneiform rash (47% in children) and paronychia are controlled with skin care and targeted dermatological treatment; increased CPK, observed in 70–73% of patients, is usually asymptomatic but requires exclusion of rhabdomyolysis; retinal vein occlusion requires permanent discontinuation, and retinal pigment epithelial detachment temporary interruption with OCT every 3 weeks until resolution and restarting at a reduced dose [18]. For grade 3 or intolerable toxicities, selumetinib is withheld and reintroduced with one or two dose reductions, with permanent discontinuation if two reductions are not tolerated [18]. Complications of bone healing during limb lengthening on selumetinib have been described, warranting coordination with orthopaedics [74].

The ERN GENTURIS guideline recommends considering MEK inhibitors, when incorporated into national care, for symptomatic inoperable plexiform neurofibromas, retaining surgery as the only potentially curative option and decision-making by a multidisciplinary team [11].

7.3 Other systemic therapies and ongoing trials

Cabozantinib, a multiple tyrosine kinase inhibitor, produced a partial response in 8 of 19 evaluable patients (42%) aged ≥16 years in a phase II Neurofibromatosis Clinical Trials Consortium trial, without progression during treatment, at the cost of gastrointestinal toxicity, hypothyroidism, fatigue and palmar–plantar erythrodysaesthesia requiring dose reduction or interruption in 9 patients [75]. For NF1-associated low-grade gliomas, selumetinib (25 mg/m² 2 times/day, the labelled dose for plexiform neurofibromas; glioma is not included in the label, so this is off-label use) achieved a sustained partial response in 10 of 25 patients aged 3 to 21 years (40%) with recurrent, refractory or progressive disease in stratum 3 of the Pediatric Brain Tumor Consortium phase II trial, and the phase III ACNS1831 trial (NCT03871257) compares selumetinib with carboplatin and vincristine in children with previously untreated NF1-associated low-grade glioma, with event-free survival and visual acuity as outcomes [76, 77]. For cutaneous neurofibromas, topical NFX-179 gel, a rapidly metabolised MEK inhibitor, reduced p-ERK levels by 47% at a concentration of 0.5% in a phase IIa trial (48 participants) and, in a phase IIb trial with 199 participants, achieved response rates of 44.2% (1.5% gel) and 34.8% (0.5%), compared with 24.1% with vehicle, with a phase III programme planned [78, 79]; these phase IIb results were disclosed in a press release and, when ClinicalTrials.gov was consulted in October 2026, no phase III NFX-179 trial was registered [80]. Also ongoing are the phase III ACNS1831 trial (active, not recruiting, primary completion expected in December 2027), a selumetinib trial for prevention of plexiform neurofibroma growth (NCT06188741), a mirdametinib trial in adult cutaneous neurofibromas (NCT06159166) and a mirdametinib trial in infants and young children with plexiform neurofibromas (NCT07850635, not yet recruiting) [80]. Approved indications for selumetinib and mirdametinib are restricted to NF1-associated plexiform neurofibromas and do not include optic pathway glioma, other gliomas or MPNST [18, 22]. MEK inhibitors as monotherapy or in combination, other targeted therapies, immunotherapies and gene therapy for MPNST and gliomas are under investigation, as are therapies for scoliosis, leukaemia, pain, constipation, hypertension and cognitive, behavioural and motor deficits [1].

7.4 Treatment of manifestations

Table 12. Treatment of NF1 manifestations [1, 3, 11, 52]

7.5 Surveillance: ERN GENTURIS and AAP/ACMG recommendations

Table 13. ERN GENTURIS 2023 tumour surveillance protocol [11]

Table 14. Supervision according to AAP/ACMG (children) and ACMG (adults), as summarised in GeneReviews [1, 15, 16]

Patients with a whole-gene microdeletion, large or growing plexiform neurofibromas, intracranial tumours, symptomatic vascular disease, progressive bone lesions or other severe manifestations require more frequent targeted follow-up [1]. The Oxford Desk Reference proposes annual follow-up of children without active problems by a paediatrician (symptoms, development and schooling, growth, blood pressure, skin, spine and vision), annual follow-up of adults by their general practitioner with access to specialist consultation (symptoms, blood pressure, skin and reinforcement of counselling) and specialist neurofibromatosis consultation for patients with complex surgical, medical or diagnostic problems; Smith proposes six-monthly complete physical assessment for structural and functional complications [3, 28]. With MPNST diagnosed at 20–40 years and reduced life expectancy, transition from paediatrics to adult care deserves specific attention, with whole-body MRI and brain MRI at this stage [11]. In ClinGen, adult-context NF1 actionability curation (report AC124, version 1.2.2, February 2022) assigned the semiquantitative score 9CB to breast cancer surveillance and 7CD to consultation with a professional experienced in NF1 as an intervention targeting MPNST, without a final actionability assertion (the initial stage of the process recorded as incomplete); there is no curation in the paediatric context [81].

7.6 Agents and circumstances to avoid

Radiotherapy is associated with a high risk of MPNST in the treated field and should be used only when absolutely necessary [1, 3]. CT should be avoided in children where possible, and breast MRI is preferred to mammography [11]. Activity restrictions may be warranted in tibial dysplasia and dystrophic scoliosis [1]. In Legius syndrome, routine cancer or ophthalmological surveillance is not indicated; surveillance focuses on development, behaviour, schooling, lipomas, scoliosis and seizures [5]. After diagnosis, GeneReviews recommends developmental assessment (motor, adaptive, cognitive and speech and language), ADHD screening from 12 months, skin examination for macules, freckling and lipomas, clinical assessment for scoliosis using the Adams forward bend test, enquiry about seizures with referral to neurology if suspected, and genetic counselling; these elements are repeated at each consultation, with hearing assessment when required and no imaging component [5].

7.7 Comparative regulatory status (United States and European Union)

The table distinguishes dates of initial authorisation, indication and formulation extensions, and authorisation type. Levels of evidence and decision-making should be separated: a peer-reviewed article (SPRINT, KOMET, ReNeu), a sponsor or foundation press release with preliminary results (NFX-179), an opinion of the EMA Committee for Medicinal Products for Human Use (CHMP), which is not yet an authorisation, and an FDA or European Commission decision. In the European Union, both MEK inhibitors have conditional authorisation, renewed annually while specific obligations are fulfilled. Authorisation also does not guarantee marketing, reimbursement or access in each country: in Portugal, access depends on INFARMED assessment (the public funding decision was not located in the sources consulted, although Portuguese series of patients treated with selumetinib since 2018 exist, section 9.2); in Brazil, selumetinib has been registered with ANVISA since 2021 for children from 2 years, but CONITEC recommended against incorporation into SUS (decision no. 1037/2025; Ordinance SECTICS/MS no. 74, dated 24/09/2025), despite recognising efficacy, because of uncertainty in the magnitude of effect and an incremental cost–utility ratio of around R$ 309.5 thousand per quality-adjusted life-year, even with the proposed 61.8% discount on the maximum price [27].

7.8 Pivotal trials and limitations of their interpretation

These trials do not allow ranking selumetinib and mirdametinib through indirect comparison. Only KOMET is randomised and placebo-controlled; SPRINT and ReNeu are single-arm trials, with different populations, durations, reading methods and temporal definitions of response. The difference between the published 70% and the 66% in the SPRINT regulatory analysis illustrates the effect of the criterion and timing of assessment. Response rate corresponds to a ≥20% volumetric reduction, a surrogate outcome whose translation into pain, function and quality of life was assessed in secondary outcomes; in KOMET, a 20% response means that around four in five adults did not achieve a response within the trial period, and the absence of a PlexiQoL difference should be read alongside the reduction in chronic pain. Adult plexiform neurofibromas grow more slowly than children's, limiting comparison between paediatric and adult trials. Medians of individual volumetric reduction are not equivalent to population means, and efficacy in infants, in asymptomatic plexiform neurofibromas or in growth prevention has not yet been demonstrated.

7.9 Extrapolation of evidence and off-label use

Extrapolation and off-label use are distinct concepts. Even within the indication, expected efficacy in an individual patient is sometimes extrapolated: for example, adult mirdametinib data come from a single-arm trial, and extension of selumetinib to children aged 1 to 2 years rests chiefly on formulation and pharmacokinetic data. Off-label use depends on the label text, age and jurisdiction: selumetinib in a 2-year-old child is within the indication in the USA with capsules or granules, but in the EU requires the granule formulation; below 1 year (selumetinib) or 2 years (mirdametinib), use is off-label in both jurisdictions; use of MEK inhibitors in optic pathway glioma, other gliomas, cutaneous neurofibromas or MPNST is off-label, and the glioma dose of 25 mg/m² twice daily is that of the PBTC trial (patients aged 3–21 years; no maximum dose defined in the source) rather than a label [18, 22, 76]. During pregnancy, MEK inhibitors are discouraged because of embryo-fetal toxicity, with pregnancy testing before starting and effective contraception [18, 22]. In Legius syndrome, without plexiform neurofibromas, these drugs have no indication. In Brazil, non-incorporation into SUS does not change the ANVISA-approved indication, but limits public access [27].

7.10 Therapeutic recommendations of the French PNDS (HAS, 2021)

The PNDS (French National Diagnosis and Care Protocol) starts from the premise that there is no curative treatment for NF1 from a genetic standpoint and that management of complications should ideally be guided by the opinion of an expert centre; unless otherwise stated, its recommendations represent professional consensus [17]. Several of the drugs cited are used outside the terms of their marketing authorisation: in France, such prescribing is permitted in the absence of a suitable alternative, when a temporary recommendation for use (RTU) exists or when the prescriber considers it indispensable in the light of scientific data, provided the patient is informed of the off-label nature, the risks and the reimbursement conditions, this is stated explicitly on the prescription and the rationale is recorded in the medical notes; the medicines agency (ANSM) may draw on the PNDS to establish an RTU, and exceptional reimbursement requires a prior RTU and an opinion from the HAS [17]. Table 12A brings together the recommendations that complement Table 12.

Table 12A. Therapeutic recommendations of the French PNDS not covered in Table 12 [17, 84–89]

MEK inhibitors in France. At the time of the 2021 text, two molecules were available in France for NF1: trametinib, as an oral solution, and selumetinib, whose European marketing authorisation (inoperable plexiform neurofibromas in children) replaced the temporary authorisations for use [17]. The PNDS discusses their indication in two situations: optic pathway glioma (second line) and inoperable, symptomatic plexiform or internal neurofibromas, where prescription is driven by the impact (pain uncontrolled despite properly managed analgesia or with adverse effects of analgesic treatment, compression, cosmetic harm) rather than by volume; the indication must be validated at the national multidisciplinary meeting of the FIMARAD rare disease network (filière) or at a specialist local meeting, which confirms inoperability and considers effective alternative treatments, particularly for low-grade gliomas [17]. Monitoring includes a cardiological assessment with transthoracic echocardiography and an ophthalmological examination before starting, after 1 month and then every 3 months, regular measurement of weight and height in children, given the lack of data on growth, and imaging reassessment of the target lesion together with review of symptoms [17]. For a glioma treated with a MEK inhibitor there are no data on prolonged treatment and, by analogy with other treatments for optic pathway glioma in NF1, a duration of 2 years is usually cited; for plexiform or internal neurofibromas the maximum duration is unknown, and it is advisable to wait for 6 to 12 months of effective treatment before declaring it ineffective [17]. Cutaneous toxicities (acneiform rash, alopecia, paronychia) are very common, sometimes severe, and may require interruption; they frequently call for dermatological follow-up, and an oral tetracycline may be considered for acneiform rash in the absence of contraindications [17]. As the PNDS gives no dose, the reference is the MASCC guideline for the cutaneous toxicity of EGFR inhibitors, an analogous condition (drug-induced papulopustular rash): doxycycline 100 mg twice daily (200 mg/day, the usual maximum label dose) or minocycline 100 mg/day, orally, in adults [90]; the doxycycline label advises against tetracyclines before 8 years of age, except in severe situations, because of the risk of permanent tooth discolouration, and specifies, from 8 years of age and below 45 kg, 4.4 mg/kg on the first day in two divided doses followed by 2.2 mg/kg/day, with the adult dose from 45 kg [91]. For trametinib, with no published dose for NF1, the reference is the oral solution label (paediatric low-grade gliomas with the NM_004333.6(BRAF):c.1799T>A, p.(Val600Glu) variant, in combination with dabrafenib): a single daily dose by weight band, from 0.3 mg at 8 kg to 2 mg from 51 kg — extrapolation from another indication [92].

MPNST: care pathway and treatment according to the PNDS. Before any biopsy or surgery, the strategy must be validated at a multidisciplinary meeting of the French sarcoma networks (NetSarc-ResOs, the sarcoma pathology reference network RRePS and the AP-HP sarcoma network RESAP), and the patient should be referred urgently to a centre within these networks [17]. When malignancy is clinically suspected, MRI and ¹⁸F-FDG PET-CT or PET-MRI are recommended: an SUVmax above 4.5 is the threshold suggestive of malignant transformation, and a tumour SUVmax/liver SUVmax ratio below 1.5 has a 99% predictive value for benignity, with histology required above that value [17]. Percutaneous biopsy targeted at the area of highest uptake had a positive predictive value of 100% and a negative predictive value of 99%, whereas surgical biopsy is not recommended as first line because of the risk of dissemination; histology should be reviewed by a pathologist of the RRePS network and the surgical specimen graded using the FNCLCC system [17]. Staging includes thin-section chest CT and whole-body imaging. Treatment is that of soft-tissue sarcomas: surgery with wide margins, without tumour exposure, aiming at R0, with consideration of re-excision and adjuvant treatment after R1; postoperative radiotherapy to the tumour bed for deep tumours of grade 2 or 3 and larger than 5 cm; adjuvant chemotherapy debatable, preferably within a trial; neoadjuvant chemotherapy if inoperable; radiotherapy alone or after incomplete surgery only with palliative intent and after a sarcoma multidisciplinary meeting has ruled out curative surgery [17]. The agents cited are doxorubicin, ifosfamide and dacarbazine (with marketing authorisation) and the etoposide–ifosfamide combination and pazopanib (off-label); in metastatic disease the objective response is around 20% and median progression-free survival about 4 months in adults, and in children the European EpSSG NRSTS 2005 protocol applies, which classifies MPNST among the “adult-type sarcomas” [17]. The PNDS gives no doses: in the SARC006 trial, specific to stage III or IV high-grade MPNST without prior chemotherapy (34 patients with NF1, median age 33 years, and 14 with sporadic tumours), the regimen consisted of 21-day cycles, 2 of ifosfamide–doxorubicin followed by 2 of ifosfamide–etoposide, with local control and 2 further cycles of each, for a total of 8, giving cumulative doses of 300 mg/m² of doxorubicin, 72 g/m² of ifosfamide and 2 g/m² of etoposide (equivalent, per cycle, to 75 mg/m² of doxorubicin, 9 g/m² of ifosfamide and 500 mg/m² of etoposide), with mesna and growth factors according to institutional protocol and no doses above the planned cumulative doses; the partial response rate was 17.9% (5/28) in patients with NF1 and 44.4% (4/9) in sporadic cases [93]. For pazopanib, the label for adult soft-tissue sarcoma specifies 800 mg orally once daily on an empty stomach (200 mg/day in moderate hepatic impairment) [94]; for dacarbazine, with no guideline dose for MPNST, the label (melanoma) specifies 250 mg/m²/day IV for 5 days every 3 weeks, or 2–4.5 mg/kg/day for 10 days every 4 weeks, with no maximum cumulative dose defined in the label — extrapolation [95]. Post-treatment surveillance is mainly clinical, with imaging for deep sites, at least every 3–6 months for 3 years, every 6 months for 2 years and annually thereafter; targeted therapies and immunotherapy remain investigational [17].

GIST. Treatment of NF1-associated GIST is complete surgical resection; these tumours respond poorly to imatinib, and the PNDS accepts sunitinib for metastatic disease [17], at the GIST label dose: 50 mg orally once daily for 4 weeks, followed by a 2-week break (6-week cycles) [96]. Phaeochromocytoma is treated surgically in the same way as outside NF1, with prior alpha and beta blockade; the indication depends on whether partial adrenalectomy is feasible (relevant because of the contralateral risk), on size, progression and secretion, and, if surgery is not indicated initially, six-monthly measurement of metanephrines and annual adrenal MRI are recommended [17]. Treatment of breast cancer is the same as for patients without NF1 [17].

7.11 Divergences between the French PNDS and other guidelines

The guidelines agree on the essentials (structured clinical surveillance, no routine brain imaging in asymptomatic children, immediate investigation of signs of MPNST, breast MRI from 30 years of age, MEK inhibitors for symptomatic inoperable plexiform neurofibromas), but differ in intervals, ages and indications. Table 12B sets out the divergences; for doses, the reference is the guideline or trial cited, with the label used for comparison.

Table 12B. Divergences between the French PNDS (2021) and the other guidelines and sources used in this article [1, 11, 15–18]

8. Reproductive Options and Family Counselling

8.1 Mode of transmission and recurrence risks

NF1 is inherited in an autosomal dominant manner: each child of an individual with constitutional NF1 has a 50% probability of inheriting the variant and, given penetrance approaching 100%, will develop disease manifestations, which may be considerably more or less severe than those of the parent [1, 3]. Around half of probands have an affected parent and the other half a de novo variant; in the North West England registry study, the proportion of new mutations was 28.8% [1, 3]. Establishing that parents are unaffected requires history-taking, a complete skin examination and ophthalmological examination with a slit lamp and, ideally, NIR imaging or OCT for Lisch nodules and choroidal abnormalities [1]. If parents are unaffected after this assessment, or the proband has a confirmed de novo variant, recurrence risk for siblings is below 1%, although higher than in the general population because of possible germline mosaicism, molecularly documented in several families, the first described by Lázaro and colleagues in 1994 [1, 3, 46]. Parental somatic and germline mosaicism may exist even without clinical signs and without detection of the variant in parental leucocyte DNA, so once the variant is identified in the proband, targeted parental testing with a mosaicism-sensitive technique is recommended [1]. Families with two or three different NF1 variants have been described, warranting caution before assuming that all affected relatives share the same variant [1].

In a parent with mosaic NF1, risk to offspring is below 50%, depends on the proportion of germ cells carrying the variant (which cannot be predicted from the fraction in blood or other tissues) and ranges from below 1% to 50%; the Oxford Desk Reference considers an overall value of 5% reasonable, and in one series included in García-Romero's review, 6.5% of individuals with mosaicism had children with full NF1 [3, 47]. If transmitted, the variant will be present in all the child's cells, and the phenotype is usually more severe than the parent's [1, 4]. The French PNDS (National Diagnosis and Care Protocol) stresses that, in mosaic NF1, the possibility of associated germline mosaicism results in a transmission risk that is significantly lower than in the generalised form but not zero, which always justifies genetic counselling [17]. In Legius syndrome, each child of a germline carrier has a 50% risk and many probands have an affected parent [5]. Type 1 microdeletions are predominantly de novo, maternal in origin, with low recurrence among siblings; a carrier transmits the deletion to 50% of children [12].

8.2 Information on variability

Counselling parents of a newly diagnosed healthy child requires a balance between the possibility of complications and disfigurement and the equal possibility of a life with few difficulties; support groups, with specialist professionals, are a resource at this stage [3]. Since expressivity cannot be predicted from relatives, predicting manifestations based on the presence or absence of complications in other family members is unwise, except for established genotype–phenotype correlations [1, 3]. The French PNDS describes the disclosure of the diagnosis as a planned, dedicated consultation, never improvised, held after the expert opinion or investigations have been obtained, adapted to whether the form is familial (guilt) or sporadic (anxiety about an unfamiliar disease) and to the patient's age, and explaining the natural history, variability, risks, the personalised follow-up plan and warning signs; psychological support, a consultation with a genetic counsellor, the possibility that other relatives are affected and the contact details of the specialist centre and of patient associations, which patients remain free to contact or not, should be offered systematically, together with guidance towards reliable information websites [17]. Table 15 reproduces the counselling values of Huson and McGaughran adapted in the Oxford Desk Reference.

Table 15. Frequency of NF1 complications for counselling purposes [3, 51, 97]

ᵃ McGaughran and colleagues observed learning difficulties of variable severity in 62%. ᵇ Median age at MPNST diagnosis is 26 years.

8.3 Predictive testing and assessment of relatives

Children of affected individuals are assessed annually until 2 years if there are no signs and again at 5 years; if the familial variant is known, genetic testing allows early diagnosis and avoids unnecessary surveillance of non-carriers [3]. Clinical assessment should be offered to other at-risk relatives, with molecular testing when the variant is known [3]. Targeted genetic testing is the only way to distinguish the exact diagnosis in a child with macules whose parent has Legius syndrome or mild NF1 [4, 5].

8.4 Reproductive options

Once the familial pathogenic variant is identified, prenatal diagnosis by chorionic villus sampling or amniocentesis and preimplantation genetic testing (PGT-M) are possible [1, 3]. Prenatal ultrasound is rarely informative, although exceptionally severe cases detected by ultrasound have been described [1]. Decisions about prenatal or preimplantation diagnosis are personal, should be discussed before pregnancy and are influenced by the inability to predict severity; the Oxford Desk Reference records that prenatal diagnosis is infrequently requested in UK practice [1, 3]. Other options include gamete donation and adoption. In pregnancy in a woman with NF1, biochemical screening for phaeochromocytoma (metanephrines) before or early in pregnancy, blood pressure surveillance and obstetric care attentive to the increased risks of preterm delivery, caesarean section, hypertension and placental abruption are recommended [1, 11]. MEK inhibitors have embryo-fetal toxicity and require effective contraception [18, 22]. In France, the PNDS places genetic counselling in the hands of a trained physician or a genetic counsellor; it can be repeated whenever necessary and is particularly indicated after diagnosis, before the test is ordered (information on its purpose, the personal and family consequences and the right not to undergo it, with the result given in an individual consultation, according to the 2013 good practice rules) and when parenthood is being planned [17]. Once the variant is identified, prenatal diagnosis and preimplantation diagnosis may be offered to couples at risk, whether the form is familial or sporadic, and non-invasive prenatal diagnosis on maternal plasma when the variant to be sought in the fetus is of paternal origin and is a point variant recognised as pathogenic; the 50% transmission risk is the same whichever parent is affected, and severity in the child cannot be predicted [17]. There are no published data on ovarian stimulation in women with NF1 in the context of preimplantation diagnosis, and no contraceptive method is contraindicated (section 3.9) [17].

9. Special Considerations: Population Diversity and Epidemiology

9.1 Global epidemiology

NF1 affects both sexes and all population groups, without a known founder effect: no pathogenic variant has been found in more than a small proportion of the families studied, and most variants are private [1, 52]. Lee and colleagues' meta-analysis (2023), including 12 studies, estimated a pooled prevalence of 1 in 3,164 (95% CI 1 in 2,132 to 1 in 4,712) and a birth incidence of 1 in 2,662 (95% CI 1 in 1,968 to 1 in 3,601); prevalence was higher in active screening studies (1 in 2,020) than in studies based on clinical records (1 in 4,329), suggesting underdiagnosis [8]. The authors emphasise that the estimates do not reflect worldwide prevalence, because 8 of the 10 NF1 studies were conducted in Europe (Finland 3, United Kingdom 2, Germany, Italy and Sweden) and the remainder in Cuba and Israel [8]. Table 16 brings together available values by region.

Table 16. NF1 prevalence and incidence by region [1–3, 7–9]

Differences between estimates chiefly reflect the ascertainment method (active screening of young populations versus registries), cohort age (prevalence is higher in younger age groups because of increased mortality) and inclusion of mild cases, rather than true differences in incidence between populations [1, 8]. There are no population-based prevalence studies for Africa, most of Asia, the Middle East outside Israel, Latin America outside Cuba, or Portugal or Brazil; in these settings, frequency should be considered unknown and should not be extrapolated from European data [8]. The French PNDS (National Diagnosis and Care Protocol) adopts a prevalence range of 1 in 3,000 to 1 in 6,000 and a birth incidence of 1 in 2,558 to 1 in 3,333, compiled from the international literature, with no French population-based estimate, and estimates the prevalence of mosaic NF1 (formerly “NF5”) at between 0.0014% and 0.002%, with bilateral distribution in about 6% of cases [17]. In mosaic NF1 the prognosis is better than in the generalised form, but transformation of plexiform neurofibromas into MPNST is possible and some studies have suggested an excess of cancers; digestive and neuro-ophthalmological forms have been described, and the PNDS recommends regular clinical surveillance [17].

9.2 Phenotypic and access differences between populations

Classic phenotypic data derive predominantly from European populations [98]. In a multicentre cross-sectional study of 1,313 Chinese patients compared with American patients, Chinese patients were diagnosed later and had a lower prevalence of spinal neurofibromas and fractures and a higher MPNST rate; four phenotypic subtypes were identified (severe neural, cutaneous, hereditary and mild), and plexiform neurofibroma risk was inversely proportional to the number of cutaneous neurofibromas and freckles [98]. In the first clinical and molecular characterisation of a Southern African cohort (26 patients), the phenotype was similar to that described, with a lower frequency of plexiform neurofibromas and a higher frequency of developmental delay and intellectual impairment, and the molecular detection rate without RNA analysis was 58%, highlighting limitations of laboratory resources in that setting [65]. In highly pigmented skin, café-au-lait macule contrast is lower and a Wood's lamp assists examination [1]. A pilot study compared patients followed at public hospitals in Campinas (Brazil) and Coimbra (Portugal): difficulties in accessing care were similar, but mean age at the first genetics consultation was 9.1 years in Brazilian patients and 15.5 years in Portuguese patients, with differences in access to geneticists and health literacy [99]. In Brazil, a retrospective analysis of 491 patients followed at two referral centres found plexiform neurofibromas in around 38%, of which 52.5% were symptomatic and inoperable, and MPNST in 2.9% [100]; an analysis of SUS administrative data (2008–2024) identified 3,748 patients under 18 years with code Q85.0 and a highly uneven regional distribution of records (4.22 per 10,000 paediatric inhabitants in the Federal District compared with 0.10 in Sergipe), chiefly reflecting access and coding rather than prevalence [101]; another combined analysis of a referral hospital and DataSUS described regional concentration of consultations and a low number of newborn diagnoses [102]. The Sociedade Brasileira de Oncologia Pediátrica (SOBOPE) published a Delphi consensus in 2025 with 24 recommendations on benign and malignant tumours in paediatric NF1 [103]. In Portugal, the Instituto Português de Oncologia in Lisbon described 54 patients with plexiform neurofibromas treated with selumetinib between 2018 and 2023 (median age 16.4 years), with stability in around 54–59% and partial response in around 29% [104], and, in a paediatric cohort of 181 patients, endocrinopathies in 23.2%, particularly pubertal abnormalities (12.2%) and growth hormone deficiency (9.9%), more frequent in the presence of optic pathway glioma [105]. These series describe patients from referral centres and do not replace population-based prevalence studies. Data on differences in penetrance, severity or cancer risk by ancestry are scarce and derive from limited-size cohorts; this gap should be acknowledged in counselling.

9.3 Sex and special populations

Excess mortality is greater in females in the Finnish registry, and breast cancer risk is concentrated in women between 30 and 50 years [1, 11]. Children and women under 30 years have particularly high standardised cancer incidence ratios [11]. Patients with a type 1 microdeletion, variants in codons 844–848 or previous radiotherapy are increased-risk populations warranting intensified surveillance [11, 12, 43]. Pregnant women with NF1 and patients undergoing elective surgery under general anaesthesia are special situations because of the risk of undiagnosed phaeochromocytoma [11].

The French PNDS considers pregnancy permissible in NF1, but with closer and more attentive obstetric follow-up [17]. Although neurofibromas appear to arise or grow under hormonal influence, a recent study found no difference in growth compared with non-pregnant women. On expert recommendation, lumbosacral MRI should be performed in the 2nd or 3rd trimester, or in the year before a planned conception, so that epidural analgesia can be authorised in the absence of a neurofibroma at the puncture site; hypertension is more frequent and may progress to pre-eclampsia; vascular dysplasias (aneurysms, wall thinning) predispose to arterial rupture, haemorrhage and postpartum haemorrhage, so the bleeding risk and any haemostatic abnormalities must be known to the anaesthetic and surgical teams, particularly for caesarean sections, which are more frequent; placental abruption and oligohydramnios are more frequent, in-hospital maternal mortality does not appear to be increased and, in the fetus, prematurity and fetal growth restriction are more frequent [17]. Table 16A summarises these complications.

Table 16A. Pregnancy and NF1 according to the French PNDS [17]

10. Illustrative Clinical Case

10.1 Illustrative case (composite, for teaching purposes)

A 3-year-old girl is referred to clinical genetics because of brownish skin macules present since the first months of life. Her non-consanguineous parents report that she started walking at 17 months and that her nursery teacher noted language delay. On examination, height is at the 10th percentile and head circumference at the 97th percentile; nine café-au-lait macules measuring 8 to 25 mm in greatest diameter are identified, distributed over the trunk and all four limbs, with bilateral axillary freckling. On the posterior aspect of the right thigh is an area of hyperpigmented, thickened, soft-textured skin measuring around 6 × 4 cm. Blood pressure is normal, and there is no cardiac murmur or tibial bowing. Slit-lamp ophthalmological examination shows no Lisch nodules; near-infrared reflectance imaging, performed without sedation, shows four choroidal abnormalities in the right eye and three in the left; visual acuity and visual fields estimated for age are normal. Her 34-year-old father denies symptoms; examination reveals seven café-au-lait macules, inguinal freckling, three cutaneous neurofibromas on the trunk and bilateral Lisch nodules.

The child meets the revised 2021 criteria through three independent routes: six or more macules larger than 5 mm before puberty, axillary freckling and two or more choroidal abnormalities; additionally, her father meets criteria A, so she also meets criterion B [4]. Choroidal abnormalities allow NF1 to be distinguished immediately from Legius syndrome, which would otherwise be indistinguishable at this age [4, 5]. Combined DNA and RNA molecular testing identifies in the proband and her father the heterozygous variant NM_001042492.3:c.1381C>T, p.(Arg461Ter) (classified as pathogenic in ClinVar, VCV000237514), introducing a premature termination codon in place of arginine 461; this is neither a microdeletion nor a variant with an established genotype–phenotype correlation, and the family is informed of variability of expression. Thigh MRI shows a diffuse plexiform neurofibroma involving branches of the sciatic nerve, measuring 68 ml, without distinct nodules.

The follow-up plan includes annual assessment by the NF1 team until 10 years, at least annual ophthalmological assessment (six-monthly if feasible) until 8 years with OCT, developmental screening and early language intervention, annual blood pressure measurement and clinical surveillance of the plexiform neurofibroma at every consultation, with parental education on warning signs (new pain, rapid growth, change in consistency, motor deficit) [11, 15]. At 7 years, the child develops limping and nocturnal thigh pain, with enlargement of the plexiform neurofibroma to 155 ml and sciatic nerve compression; ¹⁸F-FDG PET-MRI shows no uptake suggestive of malignant transformation, and the multidisciplinary team considers the tumour inoperable. After a baseline echocardiogram (ejection fraction 64%), normal ophthalmological assessment and CPK, she starts selumetinib (Koselugo®) 25 mg/m² 2 times daily in continuous 28-day cycles [18]. After 8 cycles, volumetric MRI shows a 31% reduction (partial response by REiNS criteria), with resolution of nocturnal pain. Adverse effects comprise grade 1 acneiform dermatitis, grade 2 paronychia treated by dermatology and grade 1 asymptomatic CPK elevation, without a need for dose reduction; three-monthly echocardiography remains normal [18]. Her father, who had never received follow-up, enters an adult clinic with whole-body MRI and brain MRI, and her 1-year-old sister is tested for the familial variant, with a negative result. At 30 years, the proband should begin annual breast MRI until 50 years [11].

10.2 Published case: retroperitoneal MPNST in an adult with NF1

Qian, Ye and Sun reported in 2026 in BMC Surgery a 30-year-old woman with numerous café-au-lait macules densely distributed over the head, neck, trunk and limbs, consistent with NF1, who presented with a pelvic mass initially attributed to the adnexa on gynaecological ultrasound [106]. Surgical exploration revealed retroperitoneal MPNST with metastases in the right pelvic wall and left rectus femoris muscle; exploratory laparotomy, partial colectomy, partial rectal resection, resection of the retroperitoneal masses and colostomy were performed, but only palliative resection without clear margins was possible because of extensive local invasion [106]. MPNST diagnosis was confirmed by histology and immunohistochemistry, and prognosis remained guarded [106]. The patient declined adjuvant therapy for financial reasons, and at 145 days postoperatively the residual lesions had not progressed on CT [106]. CT and MRI images, histology and operative findings are available in the original article (https://pmc.ncbi.nlm.nih.gov/articles/PMC12888607/figure/Fig1/ to https://pmc.ncbi.nlm.nih.gov/articles/PMC12888607/figure/Fig7/) [106]. The authors themselves note the absence of genetic testing; under the 2021 criteria, café-au-lait macules alone constitute only one criterion, so NF1 diagnosis in this report is not demonstrated according to current criteria, and MPNST is not a diagnostic criterion [4, 106]. The case illustrates the diagnostic difficulty of deep MPNSTs, presentation in the third decade of life and the absence of prior structured surveillance, and reinforces recommendations for whole-body MRI at transition to adulthood to quantify internal plexiform neurofibroma burden and immediate investigation by regional MRI and ¹⁸F-FDG PET for pain or tumour growth [11]. A Japanese real-world series of 11 children treated with selumetinib showed improvement in pain or motor dysfunction in all patients with these symptoms and tumour reduction in 64%, with 98% of adverse effects grade 1–2 and one permanent discontinuation because of dermatitis, but improvement in disfigurement in only one of the six patients with that problem [107].

11. Prognosis and Long-Term Follow-Up

Median life expectancy in NF1 is at least 8 years below that of the general population, with a reduction of 8 to 15 years across the various series; in the United Kingdom, Wilding and colleagues estimated median life expectancy of 70.4 years in men and 74 years in women, and Smith reports a mean age at death of 61.1 years [1, 3, 11, 28, 108]. The 2021 Brazilian chapter states that life expectancy is normal and survival depends on manifestations and tumours; British, Finnish and American population data contradict the first part of this statement, and the text adopts values from population-based studies [1, 52, 108]. The reduction is due almost entirely to malignancies, particularly MPNST, and vasculopathy; high-grade gliomas and breast cancer contribute to mortality, and excess mortality is greater in females [1, 3, 11]. Most patients have a benign course, and almost all have sufficient manifestations for diagnosis by 6 years [28]. Quality of life is lower than in comparison groups in children and adults because of cosmetic, medical, social and behavioural factors, and a Danish registry showed that people with NF1 have more than twice the likelihood of hospitalisation, with more frequent and longer admissions at all ages [1].

Long-term follow-up is organised by age and individual risk (Tables 13 and 14). In childhood, the focus is detection of optic pathway glioma (particularly until 8 years), symptomatic plexiform neurofibromas, bone dysplasia and scoliosis, hypertension and neurodevelopmental disorders, with early school support [11, 15]. In adolescence, transition to adult care includes whole-body MRI to estimate internal tumour burden, brain MRI, review of genetic and reproductive counselling and psychosocial support [11]. In adulthood, regular clinical assessment, education on MPNST warning signs, breast cancer screening by MRI between 30 and 50 years, blood pressure surveillance and biochemical phaeochromocytoma screening when indicated continue [11, 16]. Patients with an NF1 microdeletion, variants in codons 844–848, previous ANNUBP, high internal tumour burden, neuropathy, previous radiotherapy or a relative with MPNST require intensified individualised surveillance, sometimes with repeated whole-body MRI [11]. Patients on MEK inhibitors maintain cardiac, ophthalmological, skin and laboratory monitoring throughout treatment, whose optimal duration and long-term effects on growth and the skeleton remain under assessment [1, 18, 74]. In Legius syndrome, prognosis is favourable, without cancer surveillance, with follow-up focused on development and learning [5].

The French PNDS (National Diagnosis and Care Protocol) recommends lifelong surveillance from the time NF1 is suspected, with a full clinical examination and blood pressure measurement at every visit; complementary investigations are not routine and should be guided by clinical findings, and at-risk patients and children are followed primarily by a specialist team, with the remaining visits handled by the general practitioner, dermatologist or paediatrician [17]. Table 17 sets out the PNDS surveillance schedule, with the differences from ERN GENTURIS highlighted in section 7.11.

Table 17. Surveillance according to the French PNDS (HAS, 2021), by examination [17]

In the European Union, the relevant European reference network is ERN GENTURIS (genetic tumour risk syndromes), author of the 2023 tumour surveillance guideline, which allows discussion of complex cases between centres [11, 109]. Patient associations include the Children's Tumor Foundation and NF Network in the USA, the Associação Portuguesa de Neurofibromatose (APNF) in Portugal, and the Associação Mineira de Apoio às Pessoas com Neurofibromatoses (AMANF) in Brazil [79, 110–112]. In France, care is organised around the Neurofibromatoses reference centre, whose coordinating site is the dermatology department of the Hôpital Henri-Mondor (Créteil), with two constituent sites (Pitié-Salpêtrière, in Paris, for NF2 and the schwannomatoses, and the Centre Léon Bérard, in Lyon, for NF1 and related syndromes) and a network of regional centres of competence, including several paediatric and surgical sites in Île-de-France and one site in Fort-de-France (Martinique); at the time of the PNDS there was no NF1 centre of competence in Nouvelle-Aquitaine [17]. Indications for MEK inhibitors go through the national multidisciplinary meeting of the FIMARAD rare dermatological disease network (filière), MPNST through the NetSarc-ResOs, RRePS and RESAP sarcoma networks, and long-term condition status (off-list ALD for severe forms, ALD 30 for malignant tumours) ensures coverage of care costs [17]. The reference centre has agreements with the Association Neurofibromatoses et Recklinghausen (ANR), which took part in drafting the PNDS and funds research, and with the Cap NF foundation, created in 2017 on the initiative of the ANR to raise funds for research [17].

12. Summary and Final Considerations

Neurofibromatosis type 1 is an autosomal dominant RASopathy with an approximate incidence of 1 in 2,000 to 1 in 3,000 births, caused by loss of function of neurofibromin, a negative regulator of RAS, with around half of cases due to a de novo variant, unpredictable expressivity and age-dependent manifestations [1, 4, 8]. The two-hit model explains focal lesions, from café-au-lait macules to neurofibromas, gliomas and MPNST, and progression from plexiform neurofibroma to ANNUBP and MPNST involves successive losses of CDKN2A/B and the PRC2 complex [4, 11, 12]. The 2021 diagnostic criteria added choroidal abnormalities and molecular diagnosis, reformulated the tibial bone lesion, restricted the family criterion to one parent, require bilateral pigmentary findings and created criteria for NF1 mosaicism, Legius syndrome and mosaic Legius syndrome [4]. In a child with isolated macules and freckling, distinguishing NF1 from Legius syndrome depends on parental examination, choroidal imaging and NF1 and SPRED1 testing, and has direct consequences for surveillance [4, 5].

Genotype–phenotype correlations — 17q11.2 microdeletion and variants in codons 844–848 with a severe phenotype and high MPNST risk; p.Met992del, p.Arg1809 and p.Met1149 with a predominantly pigmentary phenotype; p.Arg1276 and p.Lys1423 with pulmonary stenosis and an NF1–Noonan phenotype — have become tools for counselling and surveillance stratification [12, 41–44]. Surveillance, systematised by ERN GENTURIS in 2023 and AAP/ACMG, is based on regular clinical assessment by experienced teams, ophthalmological surveillance until 8 years, whole-body MRI at transition to adulthood, annual breast MRI between 30 and 50 years, immediate investigation by MRI and ¹⁸F-FDG PET for MPNST signs and biochemical phaeochromocytoma screening in hypertension and pregnancy, avoiding unnecessary ionising radiation [11, 15, 16]. The MEK inhibitors selumetinib (Koselugo®) and mirdametinib (Gomekli®/Ezmekly®) are approved for symptomatic inoperable plexiform neurofibromas in children and adults, with response rates of 66–70% in children and 20–52% in adult and paediatric trials, and require cardiac, ophthalmological, skin and CPK monitoring [18–23, 70–72]. The French PNDS (National Diagnosis and Care Protocol) of the HAS adds a more intensive ophthalmological surveillance schedule before 6 years of age, periodic phaeochromocytoma screening from 35 years and selection for whole-body MRI with PET based on the at-risk phenotype, divergences set out in section 7.11 [17]. In the European Union, both have conditional authorisation, with minimum-age differences between USA and EU selumetinib labels, and in Brazil selumetinib is registered but has not been incorporated into SUS [27, 83]. Effective treatment of MPNST, cutaneous neurofibromas and cognitive deficits, the optimal duration of MEK inhibition and epidemiological and phenotypic characterisation of non-European populations, for which population-based data are absent, remain unresolved [1, 8, 11].

13. References

1. Friedman JM. Neurofibromatosis 1. 1998 Oct 2 [updated 2022 Apr 21; revised 2025 Apr 3]. In: Adam MP, Bick S, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1109/
2. Orphanet. Neurofibromatosis type 1 (ORPHA:636); Neurofibromatosis type 1 due to NF1 mutation or intragenic deletion (ORPHA:363700); 17q11 microdeletion syndrome (ORPHA:97685); Mosaic neurofibromatosis type 1 (ORPHA:634461); Neurofibromatosis-Noonan syndrome (ORPHA:638); Legius syndrome (ORPHA:137605); Mosaic Legius syndrome (ORPHA:634511) [Internet]. Paris: INSERM; 2026 [accessed 3 Oct 2026]. Available from: https://www.orpha.net/en/disease/detail/636. Prevalence data: Orphadata, Epidemiology of rare diseases (en_product9_prev.xml), https://www.orphadata.com/epidemiology/
3. Firth HV, Hurst JA. Neurofibromatosis type 1 (NF1). In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 506–10.
4. Legius E, Messiaen L, Wolkenstein P, Pancza P, Avery RA, Berman Y, et al. Revised diagnostic criteria for neurofibromatosis type 1 and Legius syndrome: an international consensus recommendation. Genet Med. 2021;23(8):1506–13. doi:10.1038/s41436-021-01170-5
5. Stevenson D, Legius E. Legius syndrome. 2010 Oct 14 [updated 2025 Dec 4]. In: Adam MP, Bick S, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK47312/
6. National Center for Biotechnology Information. ClinVar: NM_001042492.3(NF1):c.5488C>T (p.Arg1830Cys), VCV000208853; c.3827G>A (p.Arg1276Gln), VCV000068341; c.1381C>T (p.Arg461Ter), VCV000237514; c.2970_2972del (p.Met992del), VCV000000363; MANE transcripts for NF1 (Select NM_001042492.3; Plus Clinical NM_000267.4) and SPRED1 (NM_152594.3) according to Ensembl [Internet]. [accessed 4 Oct 2026]. Available from: https://www.ncbi.nlm.nih.gov/clinvar/variation/208853/
7. Kallionpää RA, Uusitalo E, Leppävirta J, Pöyhönen M, Peltonen S, Peltonen J. Prevalence of neurofibromatosis type 1 in the Finnish population. Genet Med. 2018;20(9):1082–6. doi:10.1038/gim.2017.215
8. Lee TJ, Chopra M, Kim RH, Parkin PC, Barnett-Tapia C. Incidence and prevalence of neurofibromatosis type 1 and 2: a systematic review and meta-analysis. Orphanet J Rare Dis. 2023;18(1):292. doi:10.1186/s13023-023-02911-2
9. Evans DG, Howard E, Giblin C, Clancy T, Spencer H, Huson SM, et al. Birth incidence and prevalence of tumor-prone syndromes: estimates from a UK family genetic register service. Am J Med Genet A. 2010;152A(2):327–32. doi:10.1002/ajmg.a.33139
10. National Institutes of Health Consensus Development Conference. Neurofibromatosis. Conference statement. Arch Neurol. 1988;45(5):575–8. PMID: 3128965
11. Carton C, Evans DG, Blanco I, Friedrich RE, Ferner RE, Farschtschi S, et al. ERN GENTURIS tumour surveillance guidelines for individuals with neurofibromatosis type 1. EClinicalMedicine. 2023;56:101818. doi:10.1016/j.eclinm.2022.101818
12. Kehrer-Sawatzki H, Mautner VF, Cooper DN. Emerging genotype-phenotype relationships in patients with large NF1 deletions. Hum Genet. 2017;136(4):349–76. doi:10.1007/s00439-017-1766-y
13. Evans DG, Baser ME, McGaughran J, Sharif S, Howard E, Moran A. Malignant peripheral nerve sheath tumours in neurofibromatosis 1. J Med Genet. 2002;39(5):311–4. doi:10.1136/jmg.39.5.311
14. Uusitalo E, Rantanen M, Kallionpää RA, Pöyhönen M, Leppävirta J, Ylä-Outinen H, et al. Distinctive cancer associations in patients with neurofibromatosis type 1. J Clin Oncol. 2016;34(17):1978–86. doi:10.1200/JCO.2015.65.3576
15. Miller DT, Freedenberg D, Schorry E, Ullrich NJ, Viskochil D, Korf BR; Council on Genetics; American College of Medical Genetics and Genomics. Health supervision for children with neurofibromatosis type 1. Pediatrics. 2019;143(5):e20190660. doi:10.1542/peds.2019-0660
16. Stewart DR, Korf BR, Nathanson KL, Stevenson DA, Yohay K. Care of adults with neurofibromatosis type 1: a clinical practice resource of the American College of Medical Genetics and Genomics (ACMG). Genet Med. 2018;20(7):671–82. doi:10.1038/gim.2018.28
17. Haute Autorité de Santé. Protocole National de Diagnostic et de Soins (PNDS): Neurofibromatose 1. Centre de référence labellisé Neurofibromatoses; texte d'août 2021. Saint-Denis La Plaine: HAS; 2021 Sep 2 [accessed 2026 Nov 2]. Available at: https://www.has-sante.fr/jcms/p_3283952
18. AstraZeneca Pharmaceuticals LP. KOSELUGO® (selumetinib) capsules and oral granules: prescribing information. Revised 05/2026. Silver Spring (MD): U.S. Food and Drug Administration; 2026. Available from: https://www.accessdata.fda.gov/drugsatfda_docs/label/2026/219943s003lbl.pdf
19. U.S. Food and Drug Administration. FDA approves selumetinib for pediatric patients 1 year of age and older with neurofibromatosis type 1 (2025 Sep 10); FDA approves selumetinib for adults with neurofibromatosis type 1 with symptomatic, inoperable plexiform neurofibromas (2025 Nov 19) [Internet]. Available from: https://www.fda.gov/drugs/resources-information-approved-drugs/fda-approves-selumetinib-adults-neurofibromatosis-type-1-symptomatic-inoperable-plexiform
20. European Medicines Agency. Koselugo (selumetinib): EPAR – medicine overview and product information [Internet]. Amsterdam: EMA; conditional marketing authorisation of 2021 Jun 17, with updates through 2026 [accessed 3 Oct 2026]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/koselugo
21. AstraZeneca. Koselugo approved in the EU for plexiform neurofibromas in adults with neurofibromatosis type 1 [Internet]. 2025 Oct 28. Available from: https://www.astrazeneca.com/media-centre/press-releases/2025/koselugo-approved-in-the-eu-for-plexiform-neurofibromas-in-adults-with-neurofibromatosis-type-1.html
22. U.S. Food and Drug Administration. FDA approves mirdametinib for adult and pediatric patients with neurofibromatosis type 1 who have symptomatic plexiform neurofibromas not amenable to complete resection [Internet]. 2025 Feb 11. Available from: https://www.fda.gov/drugs/resources-information-approved-drugs/fda-approves-mirdametinib-adult-and-pediatric-patients-neurofibromatosis-type-1-who-have-symptomatic
23. Merck KGaA (EMD Group). European Commission grants conditional approval of EZMEKLY® (mirdametinib) for NF1-associated plexiform neurofibromas in adults and children [Internet]. Darmstadt; 2025 Jul 18. Available from: https://www.emdgroup.com/en/news/ezmekly-european-commission-ok-18-07-2025
24. U.S. Food and Drug Administration. Drugs@FDA: GOMEKLI (mirdametinib), NDA 219389 (capsules) and NDA 219379 (tablets for oral suspension); original approval 2025 Feb 11, priority review [Internet]. [accessed 3 Oct 2026]. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=219389
25. European Medicines Agency. Ezmekly (mirdametinib): EPAR – medicine overview, EMEA/H/C/006460; conditional marketing authorisation of 2025 Jul 17 [Internet]. Amsterdam: EMA; [accessed 3 Oct 2026]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/ezmekly
26. European Medicines Agency. Ezmekly (mirdametinib): EPAR — procedural steps taken and scientific information after authorisation (renewal of conditional authorisation EMA/R/0000323237, decision of 29/07/2026; transfer EMA/T/0000334329) and summary of product characteristics [Internet]. Amsterdam: EMA; 2026 [accessed 4 Oct 2026]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/ezmekly
27. Comissão Nacional de Incorporação de Tecnologias no SUS (Conitec). Selumetinibe para o tratamento de pacientes pediátricos com neurofibromatose tipo 1 e neurofibroma plexiforme inoperável sintomático. Recommendation Report no. 1037. Brasília: Ministério da Saúde; 2025; Ordinance SECTICS/MS no. 74, dated 24 September 2025. Available from: https://www.gov.br/conitec/pt-br/midias/relatorios/2025/relatorio-final-1037-selumetinibe
28. Jones KL, Jones MC, del Campo M. Neurofibromatosis syndrome. In: Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 716–9.
29. von Recklinghausen F. Ueber die multiplen Fibrome der Haut und ihre Beziehung zu den multiplen Neuromen. Berlin: August Hirschwald; 1882.
30. Wallace MR, Marchuk DA, Andersen LB, Letcher R, Odeh HM, Saulino AM, et al. Type 1 neurofibromatosis gene: identification of a large transcript disrupted in three NF1 patients. Science. 1990;249(4965):181–6. doi:10.1126/science.2134734
31. Viskochil D, Buchberg AM, Xu G, Cawthon RM, Stevens J, Wolff RK, et al. Deletions and a translocation interrupt a cloned gene at the neurofibromatosis type 1 locus. Cell. 1990;62(1):187–92. doi:10.1016/0092-8674(90)90252-a
32. Xu GF, O'Connell P, Viskochil D, Cawthon R, Robertson M, Culver M, et al. The neurofibromatosis type 1 gene encodes a protein related to GAP. Cell. 1990;62(3):599–608. doi:10.1016/0092-8674(90)90024-9
33. Firth HV, Hurst JA. Noonan syndrome and the Ras/MAPK pathway syndromes: neuro-cardio-facial-cutaneous syndromes. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 512–4.
34. Brems H, Chmara M, Sahbatou M, Denayer E, Taniguchi K, Kato R, et al. Germline loss-of-function mutations in SPRED1 cause a neurofibromatosis 1-like phenotype. Nat Genet. 2007;39(9):1120–6. doi:10.1038/ng2113
35. Gutmann DH, Ferner RE, Listernick RH, Korf BR, Wolters PL, Johnson KJ. Neurofibromatosis type 1. Nat Rev Dis Primers. 2017;3:17004. doi:10.1038/nrdp.2017.4
36. Messiaen LM, Callens T, Mortier G, Beysen D, Vandenbroucke I, Van Roy N, et al. Exhaustive mutation analysis of the NF1 gene allows identification of 95% of mutations and reveals a high frequency of unusual splicing defects. Hum Mutat. 2000;15(6):541–55. doi:10.1002/1098-1004(200006)15:6<541::AID-HUMU6>3.0.CO;2-N
37. ClinGen. Dosage Sensitivity Curations: NF1 (HGNC:7765) and SPRED1 (HGNC:20249) [Internet]. Clinical Genome Resource; [accessed 3 Oct 2026]. Available from: https://search.clinicalgenome.org/kb/gene-dosage/HGNC:7765 and https://search.clinicalgenome.org/kb/gene-dosage/HGNC:20249
38. Maertens O, De Schepper S, Vandesompele J, Brems H, Heyns I, Janssens S, et al. Molecular dissection of isolated disease features in mosaic neurofibromatosis type 1. Am J Hum Genet. 2007;81(2):243–51. doi:10.1086/519562
39. Genomics in skin, neurological and rare cancers. In: ESMO Handbook — Cancer Genomics. Lugano: ESMO Press. p. 95–102.
40. De Raedt T, Brems H, Wolkenstein P, Vidaud D, Pilotti S, Perrone F, et al. Elevated risk for MPNST in NF1 microdeletion patients. Am J Hum Genet. 2003;72(5):1288–92. doi:10.1086/374821
41. Upadhyaya M, Huson SM, Davies M, Thomas N, Chuzhanova N, Giovannini S, et al. An absence of cutaneous neurofibromas associated with a 3-bp inframe deletion in exon 17 of the NF1 gene (c.2970-2972 delAAT): evidence of a clinically significant NF1 genotype-phenotype correlation. Am J Hum Genet. 2007;80(1):140–51. doi:10.1086/510781
42. Rojnueangnit K, Xie J, Gomes A, Sharp A, Callens T, Chen Y, et al. High incidence of Noonan syndrome features including short stature and pulmonic stenosis in patients carrying NF1 missense mutations affecting p.Arg1809: genotype-phenotype correlation. Hum Mutat. 2015;36(11):1052–63. doi:10.1002/humu.22832
43. Koczkowska M, Chen Y, Callens T, Gomes A, Sharp A, Johnson S, et al. Genotype-phenotype correlation in NF1: evidence for a more severe phenotype associated with missense mutations affecting NF1 codons 844–848. Am J Hum Genet. 2018;102(1):69–87. doi:10.1016/j.ajhg.2017.12.001
44. Koczkowska M, Callens T, Chen Y, Gomes A, Hicks AD, Sharp A, et al. Clinical spectrum of individuals with pathogenic NF1 missense variants affecting p.Met1149, p.Arg1276, and p.Lys1423: genotype-phenotype study in neurofibromatosis type 1. Hum Mutat. 2020;41(1):299–315. doi:10.1002/humu.23929
45. Ruggieri M, Huson SM. The clinical and diagnostic implications of mosaicism in the neurofibromatoses. Neurology. 2001;56(11):1433–43. doi:10.1212/wnl.56.11.1433
46. Lázaro C, Ravella A, Gaona A, Volpini V, Estivill X. Neurofibromatosis type 1 due to germ-line mosaicism in a clinically normal father. N Engl J Med. 1994;331(21):1403–7. doi:10.1056/NEJM199411243312102
47. García-Romero MT, Parkin P, Lara-Corrales I. Mosaic neurofibromatosis type 1: a systematic review. Pediatr Dermatol. 2016;33(1):9–17. doi:10.1111/pde.12673
48. De Luca A, Bottillo I, Sarkozy A, Carta C, Neri C, Bellacchio E, et al. NF1 gene mutations represent the major molecular event underlying neurofibromatosis-Noonan syndrome. Am J Hum Genet. 2005;77(6):1092–101. doi:10.1086/498454
49. Allanson JE, Upadhyaya M, Watson GH, Partington M, MacKenzie A, Lahey D, et al. Watson syndrome: is it a subtype of type 1 neurofibromatosis? J Med Genet. 1991;28(11):752–6. doi:10.1136/jmg.28.11.752
50. Ruggieri M, Polizzi A, Spalice A, Salpietro V, Caltabiano R, D'Orazi V, et al. The natural history of spinal neurofibromatosis: a critical review of clinical and genetic features. Clin Genet. 2015;87(5):401–10. doi:10.1111/cge.12498
51. McGaughran JM, Harris DI, Donnai D, Teare D, MacLeod R, Westerbeek R, et al. A clinical study of type 1 neurofibromatosis in north west England. J Med Genet. 1999;36(3):197–203. PMID: 10204844
52. Vasconcelos MC, Vasconcelos PHC, Vasconcelos JPC, Ribeiro EM. Neurofibromatose tipo 1 (NF1) ou doença de von Recklinghausen. In: Ribeiro EM, coordinator. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 216–8.
53. Firth HV, Hurst JA. Patchy pigmented skin lesions (including café-au-lait spots). In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 278–9.
54. Firth HV, Hurst JA. Hydrocephalus. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 173–4.
55. Firth HV, Hurst JA. Glaucoma. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 428.
56. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Fibrous dysplasia. In: Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford: Oxford University Press; 2018. p. 718.
57. Cancer predisposition — identification from somatic panels and conclusion. In: ESMO Handbook — Cancer Genomics. Lugano: ESMO Press. p. 29.
58. Firth HV, Hurst JA. Phaeochromocytoma and paraganglioma. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 608.
59. Firth HV, Hurst JA. Breast cancer. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 546.
60. Evans DGR, Kallionpää RA, Clementi M, Trevisson E, Mautner VF, Howell SJ, et al. Breast cancer in neurofibromatosis 1: survival and risk of contralateral breast cancer in a five country cohort study. Genet Med. 2020;22(2):398–406. doi:10.1038/s41436-019-0651-6
61. Uusitalo E, Kallionpää RA, Kurki S, Rantanen M, Pitkäniemi J, Kronqvist P, et al. Breast cancer in neurofibromatosis type 1: overrepresentation of unfavourable prognostic factors. Br J Cancer. 2017;116(2):211–7. doi:10.1038/bjc.2016.403. PMID: 27931045
62. Ney G, Gross A, Livinski A, Kratz CP, Stewart DR. Cancer incidence and surveillance strategies in individuals with RASopathies. Am J Med Genet C Semin Med Genet. 2022;190(4):530–40. doi:10.1002/ajmg.c.32018
63. Wimmer K, Kratz CP, Vasen HF, Caron O, Colas C, Entz-Werle N, et al. Diagnostic criteria for constitutional mismatch repair deficiency syndrome: suggestions of the European consortium 'care for CMMRD' (C4CMMRD). J Med Genet. 2014;51(6):355–65. doi:10.1136/jmedgenet-2014-102284
64. Messiaen L, Yao S, Brems H, Callens T, Sathienkijkanchai A, Denayer E, et al. Clinical and mutational spectrum of neurofibromatosis type 1-like syndrome. JAMA. 2009;302(19):2111–8. doi:10.1001/jama.2009.1663
65. Mudau MM, Dillon B, Smal C, Feben C, Honey E, Carstens N, et al. Mutation analysis and clinical profile of South African patients with neurofibromatosis type 1 (NF1) phenotype. Front Genet. 2024;15:1331278. doi:10.3389/fgene.2024.1331278
66. Firth HV, Hurst JA. Neurofibromatosis type 2 (NF2). In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 596.
67. Firth HV, Hurst JA. Lumps and bumps. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 222–3.
68. Firth HV, Hurst JA. Macrocephaly. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 224–5.
69. Firth HV, Hurst JA. Colorectal cancer (CRC). In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 556.
70. Gross AM, Wolters PL, Dombi E, Baldwin A, Whitcomb P, Fisher MJ, et al. Selumetinib in children with inoperable plexiform neurofibromas. N Engl J Med. 2020;382(15):1430–42. doi:10.1056/NEJMoa1912735
71. Chen AP, Coyne GO, Wolters PL, Martin S, Farschtschi S, Blanco I, et al. Efficacy and safety of selumetinib in adults with neurofibromatosis type 1 and symptomatic, inoperable plexiform neurofibromas (KOMET): a multicentre, international, randomised, placebo-controlled, parallel, double-blind, phase 3 study. Lancet. 2025;405(10496):2217–30. doi:10.1016/S0140-6736(25)00986-9
72. Moertel CL, Hirbe AC, Shuhaiber HH, Bielamowicz K, Sidhu A, Viskochil D, et al. ReNeu: a pivotal, phase IIb trial of mirdametinib in adults and children with symptomatic neurofibromatosis type 1-associated plexiform neurofibroma. J Clin Oncol. 2025;43(6):716–29. doi:10.1200/JCO.24.01034
73. European Medicines Agency. Koselugo (selumetinib): EPAR — product information (Annex I, Summary of Product Characteristics) [Internet]. Amsterdam: EMA [accessed 2026 Oct 26]. https://www.ema.europa.eu/en/documents/product-information/koselugo-epar-product-information_en.pdf
74. Safari Z, Tirta M, Rahbek O, Kold S. Impaired bone healing and fracture complications during limb lengthening in a neurofibromatosis type 1 patient receiving selumetinib: case report and literature review. J Orthop Case Rep. 2026;16(2):221–7. doi:10.13107/jocr.2026.v16.i02.6810
75. Fisher MJ, Shih CS, Rhodes SD, Armstrong AE, Wolters PL, Dombi E, et al. Cabozantinib for neurofibromatosis type 1-related plexiform neurofibromas: a phase 2 trial. Nat Med. 2021;27(1):165–73. doi:10.1038/s41591-020-01193-6
76. Fangusaro J, Onar-Thomas A, Young Poussaint T, Wu S, Ligon AH, Lindeman N, et al. Selumetinib in paediatric patients with BRAF-aberrant or neurofibromatosis type 1-associated recurrent, refractory, or progressive low-grade glioma: a multicentre, phase 2 trial. Lancet Oncol. 2019;20(7):1011–22. doi:10.1016/S1470-2045(19)30277-3
77. Children's Oncology Group. A phase 3 randomized study of selumetinib versus carboplatin/vincristine in newly diagnosed or previously untreated neurofibromatosis type 1 (NF1) associated low-grade glioma (ACNS1831). ClinicalTrials.gov identifier: NCT03871257 [Internet]. Available from: https://clinicaltrials.gov/study/NCT03871257
78. Sarin KY, Bradshaw M, O'Mara C, Shahryari J, Kincaid J, Kempers S, et al. Effect of NFX-179 MEK inhibitor on cutaneous neurofibromas in persons with neurofibromatosis type 1. Sci Adv. 2024;10(18):eadk4946. doi:10.1126/sciadv.adk4946
79. Children's Tumor Foundation. NFlection Therapeutics announces positive results from phase 2b clinical trial evaluating NFX-179 gel [Internet]. 2023 Nov. Available from: https://www.ctf.org/news/nflection-therapeutics-announces-positive-results-from-phase-2b-clinical-trial/
80. ClinicalTrials.gov. NCT03871257 (ACNS1831), NCT06188741, NCT06159166, NCT07850635; search for “NFX-179” [Internet]. Bethesda (MD): National Library of Medicine; [accessed 4 Oct 2026]. Available from: https://clinicaltrials.gov/study/NCT03871257
81. ClinGen Actionability Working Group. Neurofibromatosis type 1 (NF1) – Adult context, report AC124, release 1.2.2 [Internet]. Clinical Genome Resource; 2022 Feb 9 [accessed 3 Oct 2026]. Available from: https://actionability.clinicalgenome.org/ac/Adult/ui/stg2SummaryRpt?doc=AC124
82. AstraZeneca. KOSELUGO (selumetinib) capsules and oral granules: prescribing information (NDA 213756 and 219943; sections revised in 11/2025 and 05/2026) [Internet]. openFDA/DailyMed; 2026 [accessed 4 Oct 2026]. Available from: https://api.fda.gov/drug/label.json?search=openfda.brand_name:%22KOSELUGO%22
83. European Medicines Agency. Koselugo (selumetinib): EPAR — procedural steps taken and scientific information after authorisation (extension to adults EMA/VR/0000245231, opinion 18/09/2025, decision 24/10/2025; X/0018/G, granules in capsules, opinion 13/11/2025, decision 09/01/2026) and summary of product characteristics [Internet]. Amsterdam: EMA; 2026 [accessed 4 Oct 2026]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/koselugo
84. Lassaletta A, Scheinemann K, Zelcer SM, Hukin J, Wilson BA, Jabado N, et al. Phase II weekly vinblastine for chemotherapy-naïve children with progressive low-grade glioma: a Canadian Pediatric Brain Tumor Consortium study. J Clin Oncol. 2016;34(29):3537–43. doi:10.1200/JCO.2016.68.1585. PMID: 27573663
85. Gururangan S, Fangusaro J, Poussaint TY, McLendon RE, Onar-Thomas A, Wu S, et al. Efficacy of bevacizumab plus irinotecan in children with recurrent low-grade gliomas — a Pediatric Brain Tumor Consortium study. Neuro Oncol. 2014;16(2):310–7. doi:10.1093/neuonc/not154. PMID: 24311632
86. Averitas Pharma Inc. QUTENZA (capsaicin) 8% topical system: prescribing information (NDA 022395) [Internet]. DailyMed/openFDA; version of 2026 Jul 7 [accessed 2026 Nov 2]. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=3ffbbcb0-ad93-4f15-bb38-5da76a71c735
87. Major Pharmaceuticals. Sirolimus tablets: prescribing information (ANDA 214753) [Internet]. DailyMed/openFDA; version of 2025 Nov 10 [accessed 2026 Nov 2]. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=1bef7d8f-524b-4218-8fef-6c3f3628d33e
88. Zehou O, Ferkal S, Brugieres P, Barbarot S, Bastuji-Garin S, Combemale P, et al. Absence of efficacy of everolimus in neurofibromatosis 1-related plexiform neurofibromas: results from a phase 2a trial. J Invest Dermatol. 2019;139(3):718–20. doi:10.1016/j.jid.2018.09.016. PMID: 30339775
89. Sandoz Inc. RITALIN (methylphenidate hydrochloride) tablets: prescribing information (NDA 010187) [Internet]. DailyMed/openFDA; version of 2025 Feb 5 [accessed 2026 Nov 2]. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=d6fb2750-cdab-4749-ba0d-7534840a5892
90. Lacouture ME, Anadkat MJ, Bensadoun RJ, Bryce J, Chan A, Epstein JB, et al.; MASCC Skin Toxicity Study Group. Clinical practice guidelines for the prevention and treatment of EGFR inhibitor-associated dermatologic toxicities. Support Care Cancer. 2011;19(8):1079–95. doi:10.1007/s00520-011-1197-6. PMID: 21630130
91. St. Mary's Medical Park Pharmacy. Doxycycline: prescribing information (ANDA 209402) [Internet]. DailyMed/openFDA; version of 2026 May 1 [accessed 2026 Nov 2]. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=00b2aa98-dea2-90c3-e063-6294a90a35d6
92. Novartis Pharmaceuticals Corporation. MEKINIST (trametinib) tablets and for oral solution: prescribing information (NDA 204114 and 217513) [Internet]. DailyMed/openFDA; version of 2026 May 7 [accessed 2026 Nov 2]. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0002ad27-779d-42ab-83b5-bc65453412a1
93. Higham CS, Steinberg SM, Dombi E, Perry A, Helman LJ, Schuetze SM, et al. SARC006: phase II trial of chemotherapy in sporadic and neurofibromatosis type 1 associated chemotherapy-naive malignant peripheral nerve sheath tumors. Sarcoma. 2017;2017:8685638. doi:10.1155/2017/8685638. PMID: 29138631
94. Novartis Pharmaceuticals Corporation. VOTRIENT (pazopanib) tablets: prescribing information (NDA 022465) [Internet]. DailyMed/openFDA; version of 2025 Dec 16 [accessed 2026 Nov 2]. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=eeaaaf38-fb86-4d9f-a19d-0f61daac2fd7
95. Hikma Pharmaceuticals USA Inc. Dacarbazine for injection: prescribing information (ANDA 075812) [Internet]. DailyMed/openFDA; version of 2025 Sep 22 [accessed 2026 Nov 2]. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0d81315c-7021-4091-b703-ad135390c936
96. Pfizer Laboratories Div Pfizer Inc. SUTENT (sunitinib malate) capsules: prescribing information (NDA 021938) [Internet]. DailyMed/openFDA; version of 2026 Feb 20 [accessed 2026 Nov 2]. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=43a4d7f8-48ae-4a63-9108-2fa8e3ea9d9c
97. Huson SM, Compston DA, Harper PS. A genetic study of von Recklinghausen neurofibromatosis in south east Wales. II. Guidelines for genetic counselling. J Med Genet. 1989;26(11):712–21. doi:10.1136/jmg.26.11.712
98. Wang Z, Hu X, Mo H, Shen M, Romo CG, Vera K, et al. Phenotypic characterization of neurofibromatosis type 1 in a large Chinese cohort: a cross-sectional study. JAAD Int. 2026;24:181–90. doi:10.1016/j.jdin.2025.09.020
99. de Francisco DD, Silva IMW, Saraiva JM, Steiner CE, Gil-da-Silva-Lopes VL. From diagnosis to daily life: a comparative pilot study on healthcare access and challenges for neurofibromatosis type 1 in public systems of Brazil and Portugal. Clinics (Sao Paulo). 2026;81:100875. doi:10.1016/j.clinsp.2026.100875
100. Darrigo Junior LG, Ferraz VEF, Cormedi MCV, Araujo LHH, Magalhães MPS, Carneiro RC, et al. Epidemiological profile and clinical characteristics of 491 Brazilian patients with neurofibromatosis type 1. Brain Behav. 2022;12(6):e2599. doi:10.1002/brb3.2599. PMID: 35506373
101. Sonaglio V, Oigman Bellas G, Caran EMM, Decario CBL, Bunho APF, Longato M, et al. Neurofibromatosis type 1 in Brazil: pediatric care in the Unified Health System. Eur J Med Genet. 2026;83:105100. doi:10.1016/j.ejmg.2026.105100. PMID: 42551785
102. Lins MGDS, Morais PJ, Martinho JO, Nascimento MCAD, Vila APS, Castanhole-Nunes MMU, et al. Clinical and epidemiological characterisation of neurofibromatosis type 1: combined analysis of a reference hospital in Brazil and DataSUS. Genet Mol Biol. 2025;48(1):e20240144. doi:10.1590/1678-4685-GMB-2024-0144. PMID: 40402550
103. Darrigo Junior LG, Sonaglio V, Ferman SE, Caran E, Filho NPC, Epelman S, et al. Consensus on malignant and benign tumors in pediatric patients with neurofibromatosis type 1: on behalf of the Brazilian Society of Pediatric Oncology (SOBOPE). Curr Oncol. 2025;32(12):664. doi:10.3390/curroncol32120664. PMID: 41440192
104. Passos J, Soares MP, Salgado D, Nunes S, Cavaco D, Garrido PM, et al. A single-center case study series assessing the effect of selumetinib use in patients with neurofibromatosis-related plexiform neurofibromas. Neurooncol Adv. 2024;6(1):vdae177. doi:10.1093/noajnl/vdae177. PMID: 39605314
105. Regala C, Cavaco D, Maciel J, Figueiredo A, Damásio I, Pinheiro S, et al. Endocrine manifestations in a paediatric cohort of 181 patients with neurofibromatosis type 1. Eur J Endocrinol. 2025;193(2):216–22. doi:10.1093/ejendo/lvaf147. PMID: 40709404
106. Qian L, Ye X, Sun J. Retroperitoneal malignant peripheral nerve sheath tumor in a patient with neurofibromatosis type 1: a case report. BMC Surg. 2026;26(1):115. doi:10.1186/s12893-026-03495-x
107. Nishida Y, Nonobe N, Kidokoro H, Kato T, Takeichi T, Ikuta K, et al. Selumetinib for symptomatic, inoperable plexiform neurofibromas in pediatric patients with neurofibromatosis type 1: the first single-center real-world case series in Japan. Jpn J Clin Oncol. 2025;55(4):372–6. doi:10.1093/jjco/hyae184
108. Wilding A, Ingham SL, Lalloo F, Clancy T, Huson SM, Moran A, et al. Life expectancy in hereditary cancer predisposing diseases: an observational study. J Med Genet. 2012;49(4):264–9. doi:10.1136/jmedgenet-2011-100562
109. European Reference Network on Genetic Tumour Risk Syndromes (ERN GENTURIS) [Internet]. [accessed 4 Oct 2026]. Available from: https://www.genturis.eu
110. NF Network [Internet]. [accessed 4 Oct 2026]. Available from: https://www.nfnetwork.org
111. Associação Portuguesa de Neurofibromatose (APNF) [Internet]. [accessed 4 Oct 2026]. Available from: https://apnf.pt
112. Associação Mineira de Apoio às Pessoas com Neurofibromatoses (AMANF) [Internet]. [accessed 4 Oct 2026]. Available from: https://amanf.org.br