Clinical and Molecular Portraits of Pediatric RASopathies: A Study of 118 Genotype‐Confirmed Cases

ABSTRACT

RASopathies are clinically and genetically heterogeneous disorders resulting from dysregulation of the RAS/MAPK pathway. We present a single‐center retrospective cohort of 118 genotype‐confirmed pediatric patients diagnosed with RASopathy. Noonan syndrome was the most common clinical diagnosis, followed by NF‐Noonan syndrome, Legius syndrome, LEOPARD syndrome, Costello syndrome, and cardiofaciocutaneous syndrome. The most frequent clinical features included short stature, pulmonary stenosis, pectus deformities, and neurocognitive delay. Through molecular analysis, we identified distinct variants across major RAS/MAPK pathway genes, with a considerable proportion of patients carrying rare or less frequently reported variants, including those in LZTR1, RIT1, RAF1, HRAS, BRAF, RASA2, KRAS, CBL, SHOC2, and MAP2K2. In addition to well‐established genotype–phenotype correlations, we report several uncommon or previously unrecognized features: microcephaly in multiple PTPN11‐positive patients and renal agenesis in a child with a novel HRAS variant diagnosed with Costello syndrome. Furthermore, we identified three patients harboring RASA2 variants, providing additional support for its emerging role in Noonan syndrome. This study expands the clinical and molecular spectrum of pediatric RASopathies. It emphasizes the need for comprehensive genetic evaluation and regular follow‐up in the context of overlapping clinical features, supporting accurate diagnosis and identifying candidates for emerging targeted therapies.

Genotype–phenotype correlations were evaluated in 118 genotype‐confirmed pediatric RASopathy patients. Microcephaly emerged as an under‐recognized feature in PTPN11‐related Noonan syndrome. Rare and noncoding variants across multiple genes expanded the molecular spectrum, refining genotype–phenotype correlations and highlighting phenotypic variability in pediatric RASopathies.

1. Introduction

The RASopathies are a group of developmental disorders caused by pathogenic variants in genes encoding components or regulators of the Ras/mitogen‐activated protein kinase (RAS/MAPK) pathway. This pathway is crucial in regulating the cell cycle, cellular growth, differentiation, and senescence. Dysregulation of this pathway leads to a variety of developmental defects [1]. The cumulative incidence of RASopathies is approximately 1 in 1000 live births [2]. While each type of RASopathies presents a unique phenotype, they also share different kinds of overlapping characteristics, such as dysmorphic facial features, congenital heart defects, cutaneous and musculoskeletal abnormalities, neurodevelopmental abnormalities, and an increased risk of malignancy [1, 3].

Noonan Syndrome (NS, MIM #163950), an autosomal dominant inherited condition, is the most prevalent and frequently observed form serving as a prototype for this group of disorders. The prevalence of NS is approximately 1 in 1000–2500 newborns. NS is presented with highly variable features, which include ptosis, posteriorly rotated and low‐set ears, short stature, pectus deformity, congenital heart defects such as pulmonary valve stenosis (PS) and hypertrophic cardiomyopathy (HCM), cryptorchidism, various coagulation defects, lymphatic dysplasias, and a variable degree of neurocognitive delay [1, 3, 4].

Cardiofaciocutaneous syndrome (CFCS, MIM #115150), Costello syndrome (CS, MIM #218040), Legius syndrome (LS, MIM #611431), Neurofibromatosis‐Noonan syndrome (NF‐NS, MIM #613113), Noonan‐like syndrome with loose anagen hair (NS/LAH, MIM #607721), and Noonan syndrome with multiple lentigines (LEOPARD, MIM #151100) also arise from defects in the RAS/MAPK pathway and give rise to other disorders within the RASopathy group and are classified as NS‐related disorders [1, 2, 3, 5].

To date, several genes associated with the RAS/MAPK pathway have been identified as causing NS and other related disorders in groups of RASopathies. Pathogenic variants in PTPN11, SOS1, SOS2, RAF1, RIT1, LZTR1, MAP2K1, MAP2K2, BRAF, KRAS, NRAS, HRAS, SHOC2, CBL, SPRED1, NF1, RRAS2, RASA2, PPP1CB, MAPK1, and MRAS, which function within the RAS/MAPK pathway, have been reported to cause disease [2, 6]. In recent studies, SYNGAP1, LZTR1, RIT1, PPP1CB, MAPK1, and SHOC2 have garnered increased attention due to their possible genotype–phenotype correlations and hold promise for advancements in treatment options [7, 8]. Although pathogenic variants in canonical genes like PTPN11, SOS1, and RAF1 account for most cases, rare variants in genes such as LZTR1, RIT1, SHOC2, and CBL are increasingly being recognized as important contributors to the broad clinical spectrum of RASopathies. These rare variants are often associated with distinctive or atypical phenotypes, expanding the known spectrum of these disorders and emphasizing the need for comprehensive genetic testing, even beyond the most commonly implicated genes [9, 10].

In this single‐center study, we report one of the largest fully genetically confirmed pediatric RASopathy cohorts to date, comprising 118 patients, and present comprehensive clinical and molecular findings, including the identification of several rare variants. In addition, we investigated genotype–phenotype correlations to enhance our understanding of the diverse clinical presentations characteristic of RASopathies.

2. Materials and Methods

This study retrospectively evaluates 118 pediatric patients who were clinically suspected of having RASopathies and subsequently received a confirmed molecular diagnosis at the Pediatric Genetics Clinic of Ankara Bilkent City Children's Hospital, a tertiary care center, over 10 years. Demographic, clinical, radiological data, and molecular analysis results were extracted from hospital records. The clinical parameters assessed included age, sex, anthropometric measurements, presenting symptoms, personal and family medical history, dysmorphic facial features, cardiac anomalies, pectus deformities, neurodevelopmental status, ectodermal findings, cryptorchidism, and findings from abdominal ultrasonography and echocardiography. Molecular analysis results were reviewed to confirm the clinical diagnosis and also to assess genotype–phenotype correlations. Patients without a confirmed molecular diagnosis were excluded from the study. Among those with NF1 variants, individuals who met the diagnostic criteria for classical Neurofibromatosis Type 1 (NF1) were excluded, whereas patients with an NF‐Noonan phenotype were retained in the cohort.

All procedures performed in studies involving human participants were conducted in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments. The Clinical Research Ethics Committee of Ankara Bilkent City Hospital approved the study with decision number TABED 1‐25‐970. Written consent for the use of clinical images was obtained from the patients' families.

DNA isolation was performed using the QIAGEN QIAamp DNA Mini Kit with the QIAsymphony AS 4‐channel automatic pipetting and isolation system, following the manufacturer's protocols. The obtained data were analyzed using the QIAGEN Clinical Insight (QCITM) analysis software (QIAGEN, Hilden, Germany). The reported variants were compared with the in‐house database or well‐known data in the literature. The variant review was performed using the HGMD, ClinVar, and VarSome databases and classified according to ACMG criteria. The QIAseqTM Targeted DNA RASopathy Panel was used according to the following protocol for sequencing all exons and exon‐intron junctions of the genes BRAF, CBL, HRAS, KAT6B, KRAS, LZTR1, MAP2K1, MAP2K2, NF1, NRAS, PTPN11, RAF1, RASA2, RIT1, SHOC2, SOS1, SOS2, and SPRED1.

3. Results

3.1. Clinical Features

Of the 118 children with RASopathy included in this study, 47 were female (39.8%), and 71 were male (60.2%), resulting in a male‐to‐female ratio of 1.5:1. The median age at diagnosis was 6.8 years. Among our cohort, 79 patients were diagnosed with NS, 21 with NF‐NS, 5 with LS, 5 with LEOPARD Syndrome, 4 with CS, 3 with CFCS, and one patient with NS/LAH.

According to Van der Burgt's diagnostic criteria [4] 87 of the 105 patients (82.8%) diagnosed with NS, LEOPARD, and NF‐Noonan phenotype fulfilled the criteria based on characteristic facial dysmorphic features and at least one major criterion. While 14 patients (13.3%) were diagnosed based on suggestive facial features accompanied by two major or three minor criteria, 4 patients (3.8%) did not fully meet the criteria but were still genetically diagnosed due to strong clinical suspicion.

In our cohort of 118 pediatric patients with RASopathies, the most common clinical findings are summarized in Figure 1. Short stature (n = 69, 58.5%) and pulmonary stenosis (n = 50, 42.4%) were the two most frequently observed features. Pectus deformities, neurocognitive delay, and ectodermal findings were also common. Notably, microcephaly was identified in 9.3% (n = 11) of patients. A total of three patients in our cohort developed malignancies. Juvenile myelomonocytic leukemia (JMML) was observed in a patient diagnosed with NS, while another NS patient developed a giant cell bone granuloma. Additionally, one patient with CS presented with both embryonal rhabdomyosarcoma and secondary neuroblastoma.

FIGURE 1.

The common clinical findings in our cohort of 118 patients. [Colour figure can be viewed at wileyonlinelibrary.com]

Among the 93 patients for whom consanguinity information was available, 21 (22.5%) had a positive history of consanguineous parentage. Notably, only two of the patients presenting with microcephaly had a history of consanguinity.

3.2. Molecular Findings

Among 118 pediatric patients with a clinical diagnosis of RASopathies, the identified genes and their frequencies are presented in Figure 2. In total, 119 variants were detected across the cohort, representing 88 distinct molecular variants. Of these, 96 (80.6%) were classified as pathogenic (P), 13 (10.9%) as likely pathogenic (LP), and 10 (8.4%) as variants of uncertain significance (VUS) according to ACMG 2015 criteria. One patient with compound heterozygous LZTR1 variants carried both a pathogenic and a VUS allele. Regarding variant type, the majority were missense changes (n = 97, 81.5%), followed by splice‐site variants (n = 10, 8.4%), nonsense (n = 5, 4.2%), frameshift (n = 4, 3.3%), two in‐frame, and a noncoding variant.

FIGURE 2.

Molecularly identified genes in our cohort of 118 patients. [Colour figure can be viewed at wileyonlinelibrary.com]

The clinical features of patients harboring PTPN11 variants are listed in Table 1. The vast majority of these variants were missense (100%). Five patients with PTPN11 variants were clinically followed with LEOPARD, while the others were diagnosed with NS. Among NS patients with PTPN11 variants (n = 44), PS was present in 29 patients (65.9%), and HCM in 4 (9.1%). Short stature was observed in 31 patients (70.4%). Thirteen patients had pectus deformities (29.5%), 8 had microcephaly (18%), and 13 had developmental delay (29.5%).

TABLE 1.

PTPN11 variants and clinical features in our cohort.

ID	Clinical diagnosis	Short stature (< 3 centile)	Typical facial dysmorphology	Cardiac defect	Neurocognitive delay	Other findings	PTPN11 variant (c.)	Exon number	Amino acid change (p.)	ACMG criteria	Variant type
23	NS	+	+	PS	−	−	174C>G	3	N58K	P	Missense
42	NS	+	+	PS	−	Microcephaly	174C>G	3	N58K	P	Missense
90	NS	+	+	−	−	−	178G>A	3	G60S	P	Missense
46	NS	+	−	PS	−	−	181G>A	3	D61N	P	Missense
41	NS	+	+	−	−	−	184T>G	3	Y62D	P	Missense
14	NS	−	+	AVSD	+	−	188A>G	3	Y63C	P	Missense
21	NS	+	+	PS	−	−	188A>G	3	Y63C	P	Missense
49	NS	+	+	−	+	Bifid pelvis	188A>G	3	Y63C	P	Missense
111	NS	−	+	−	−		188A>G	3	Y63C	P	Missense
112	NS	+	+	−	−	Microcephaly	188A>G	3	Y63C	P	Missense
48	NS	+	+	−	+	−	205G>C	3	E69Q	P	Missense
55	NS	+	+	PS	−	−	211T>G	3	F71V	P	Missense
39	NS	+	+	PS	−	−	215C>G	3	A72G	P	Missense
89	NS	−	−	PS	−	−	215C>G	3	A72G	P	Missense
8	NS	+	+	PS	−	JMML, exitus	218C>T	3	T73I	P	Missense
37	NS	−	+	PS	−	Iris coloboma, bilateral sensorineural hearing loss	218C>T	3	T73I	P	Missense
66	NS	−	+	PS	−	JMML	226G>A	3	E76K	LP	Missense
3	NS	−	+	PS	+	Microcephaly	236A>G	3	Q79R	P	Missense
44	NS	+	+	PS	+	−	417G>C	4	E139D	P	Missense
85	NS	+	+	ASD	+	Long QT	417G>C	4	E139D	P	Missense
5	NS	−	+	+	+	−	781C>T	7	L261F	P	Missense
99	NS	+	+	PS	−	−	794G>A	7	R265Q	P	Missense
73	LEOPARD	−	−	−	−	−	836A>G	7	Y279C	P	Missense
96	LEOPARD	+	+	HCM	+	−	836A>G	7	Y279C	P	Missense
19	NS	−	+	PS	+	−	844A>G	7	I282V	P	Missense
116	NS	+	+	PS	−	Microcephaly, obesity	844A>G	7	I282V	P	Missense
83	NS	+	+	−	−	−	853 T>C	8	F285L	P	Missense
52	NS	+	+	PS	−	−	854 T>C	8	F285S	P	Missense
43	NS	+	+	−	−	Ectopic kidney	855 T>G	8	F285L	P	Missense
27	NS	+	+	PS	+	Microcephaly	922A>G	8	N308D	P	Missense
47	NS	+	+	PS	+	Osteoporosis, Behçet's Disease	922A>G	8	N308D	P	Missense
51	NS	+	+	−	−	−	922A>G	8	N308D	P	Missense
81	NS	+	−	PS	−	−	922A>G	8	N308D	P	Missense
91	NS	−	−	PS	−	−	922A>G	8	N308D	P	Missense
106	NS	−	+	PS, HCM	−	−	923A>G	8	N308S	P	Missense
82	NS	−	−	PS, HCM	−	−	1391G>C	12	G464A	P	Missense
107	LEOPARD	−	−	−	−	Long QT, sensorineural hearing loss	1402A>C	12	T468P	P	Missense
13	NS	−	+	HCM	+	Autism spectrum disorder	1403C>T	12	T468M	P	Missense
40	NS	−	+	−	−	Myoclonic epilepsy	1403C>T	12	T468M	P	Missense
20	NS	+	+	PS	−	Benign external hydrocephaly	1472C>T	13	P491L	P	Missense
88	NS	+	+	ASD	−	Microcephaly	1472C>T	13	P491L	P	Missense
54	LEOPARD	−	−	−	−	Macrocephaly	1493G>T	13	R498L	P	Missense
114	NS	+	−	PS, HCM	+	Microcephaly	1507G>A	13	G503R	P	Missense
84	NS	+	+	PS	+	−	1510A>G	13	M504V	P	Missense
61	NS	+	+	PS		Iris coloboma, microcephaly	1510A>G	13	M504V	P	Missense
50	NS	+	+	PS	−	−	1510A>G	13	M504V	P	Missense
29	NS	+	+	PS	−	−	1510A>G	13	M504V	P	Missense
95	NS	+	+	PS	−		1510A>G	13	M504V	P	Missense
94	LEOPARD	+	+	−	+	LEOPARD confirmed mother	1529A>C	13	Q510P	P	Missense

Among the 49 patients carrying a PTPN11 variant, the three most recurrent variants were p.Tyr63Cys, p.Asn308Asp, and p.Met504Val, each detected in five patients. The most frequently affected exons were exons 3 (n = 19), 7 (n = 6), 8 (n = 11), 12 (n = 4), and 13 (n = 9). Exon 3 variants were the most common and were associated with short stature, typical Noonan facial dysmorphology, and pulmonary stenosis, occasionally accompanied by neurocognitive delay. Variants located in exons 7 and 8 showed a broader phenotypic spectrum, including hypertrophic cardiomyopathy, autism spectrum disorder, and microcephaly, while short stature remained frequent. Exon 12–13 variants were often associated with pulmonary stenosis and neurocognitive delay, and several corresponded to LEOPARD syndrome diagnoses. Within our cohort, LEOPARD syndrome cases were exclusively associated with variants located in exons 7, 12, and 13.

In addition to the core Noonan features, several PTPN11‐positive patients exhibited remarkable associated findings. JMML and iris coloboma were each observed in two patients, and two others had sensorineural hearing loss, while long QT interval was also documented in two individuals. Additional isolated findings included ectopic kidney, osteoporosis with Behçet's disease, benign external hydrocephaly, obesity, autism spectrum disorder, and myoclonic epilepsy. Family history was notable in four cases: two diagnosed with NS and tested positive for the same PTPN11 variant, one with a father suspected of NS, and another with maternal LEOPARD syndrome.

The remaining pathogenic variants were distributed across several genes related to RASopathies and are summarized in Table 3. SOS1 variants were found in 10 patients; PS was present in 8 (80%), typical craniofacial features were observed in 9 (90%), and short stature was present in 7 of them (70%). The most common SOS1 variant was R552G, observed in three unrelated patients (30%). Two of these patients had short stature and typical facial features, while the third also had PS, an atrial septal defect (ASD), and a giant cell granuloma in the mandible (Figure 3). K107E was seen in two patients (20%) with variable congenital heart findings, and one of them also had microcephaly.

TABLE 3.

Rare variants and clinical features in our cohort.

ID	Clinical diagnosis	Short stature (< 3 centile)	Typical facial dysmorphology	Cardiac defect	Neurocognitive delay	Ectodermal findings	Other findings	Affected gene	Amino acid change	ACMG criteria	Variant type
1	NS	+	−	PS, HCM	−	−	−	RIT1	A77T	P	Missense
22	NS	+	−	PS	−	−	−	RIT1	A41P	P	Missense
60	NS	−	+	PS, HCM	−	−	−	RIT1	K23Q	P	Missense
77	NS	+	+	PS	−	−	Exitus	RIT1	A57G	P	Missense
2	NS	−	−	−	+	−	−	LZTR1	−	LP	Splice acceptor
12	NS	+	−	−	+	+	Microcephaly	LZTR1	R478W	VUS	Missense
56	NS	+	+	−	+	−	−	LZTR1	−	VUS//P	Splice donor//splice donor
64	NS	+	+	HCM	+	−		LZTR1	A116V	LP	Missense
93	NS	−	−	−	+	−	−	LZTR1	G301S	VUS	Missense
17	CS	−	−	HCM	+	+	Ectomesenchymoma	HRAS	G12S	P	Missense
31	CS	+	−	HCM	+	+	Microcephaly	HRAS	G12S	P	Missense
118	CS	+	−	PS, HCM	−	+	Severe malnutrition	HRAS	G12S	P	Missense
58	CS	+	−	PS, VSD, ASD	+	+	Right kidney agenesis, hearing loss	HRAS	−	VUS	Non‐coding (3′ UTR)
18	NS	+	+	+	−	−	Cleft palate, cleft lip	MAP2K2	M147V	P	Missense
65	CFC	−	−	+	+	+	Left hemiparesia, epilepsy, scoliosis	MAP2K2	N216D	LP	Missense
30	NS	+	+	HCM	+	−	Macrocephaly	RAF1	S257L	P	Missense
100	NS	+	−	HCM	−	−	Received MEK inhibitor, exitus	RAF1	S257L	P	Missense
53	NS	−	+	+	−	+	−	RAF1	H175R	P	Missense
70	NS	−	+	PS	+	−	−	RAF1	S259T	P	Missense
16	NS	−	+	PS	+	−	Macrocephaly	RAF1	P261S	P	Missense
45	NS	+	+	PS	−	−	−	RAF1	N140K	VUS	Missense
24	NS	+	+	HCM	−	−	Hearing loss	RAF1	P261R	P	Missense
10	NS	+	+	−	−	−	−	BRAF	−	LP	Splice acceptor
11	CFC	+	+	+	+	−	Macrocephaly, epilepsy	BRAF	Q257R	P	Missense
57	CFC	−	−	PS	−	+	Macrocephaly	BRAF	K499E	P	Missense
63	NS	−	+	+	−	−	Macrocephaly	BRAF	F595L	P	Missense
4	NS	−	−	PS	−	−	−	RASA2	−	VUS	Splice acceptor
28	NS	+	+	PS, ASD, VSD	+	−	−	RASA2	−	VUS	Splice donor
108	NS	−	−	VSD, noncompaction	−	−	−	RASA2	−	VUS	Splice acceptor
62	NS	+	+	+	+	−	Macrocephaly	KRAS	V141I	P	Missense
79	NS	−	+	+	−	−	Macrocephaly	KRAS	A146G	LP	Missense
38	NS‐LH	+	−	−	+	−	Macrocephaly	SHOC2	S2G	P	Missense
97	LS	−	−	−	−	−	−	SPRED1	I60Yfs*18	P	Frameshift
101	LS	−	−	−	−	−	−	SPRED1	R16	P	Nonsense
103	LS	−	−	−	−	−	−	SPRED1	R16	P	Nonsense
109	LS	−	−	−	−	−	−	SPRED1	D154Gfs*21	P	Frameshift
102	LS	+	−	−	−	−	Scoliosis, bifid pelvis	SPRED1	K77Nfs*45	P	Frameshift
6	NS	−	+	PS	−	−	Scoliosis	SOS1	K107E	P	Missense
7	NS	+	+	PS	+	−	Microcephaly	SOS1	K107E	P	Missense
26	NS	+	+	PS	−	−	−	SOS1	E891K	P	Missense
59	NS	−	−	PS	−	−	−	SOS1	W432R	P	Missense
68	NS	+	+	PS	−	−	−	SOS1	R552G	P	Missense
80	NS	+	+	PS	−	−	−	SOS1	R552G	P	Missense
92	NS	+	+	−	−	−	−	SOS1	C282R	LP	Missense
110	NS	−	+	PS, ASD	−	+	Giant cell granuloma	SOS1	R552G	P	Missense
113	NS	+	+	−	−	−	−	SOS1	M269T	LP	Missense
117	NS	−	+	PS	−	−	Macrocephaly	SOS1	S548R	P	Missense
104	NS	−	−	HCM	−	−	−	CBL	R246*	LP	Nonsense

FIGURE 3.

(a) An 8‐year‐old girl with a RASA2 variant, ptosis, pectus deformity, and a scar from prior cardiac surgery was remarkable. (b) 11‐year‐old boy diagnosed with NS, harboring a KRAS variant with a gestalt Noonan face. (c, d) An infant carrying a RIT1 variant, with severe cardiac involvement and a typical Noonan gestalt face, died from complications of cardiac disease. (e, f) The patient diagnosed with CFC, caused by a MEK2 variant, from different ages. (g) 14‐year‐old girl harboring a novel HRAS variant, diagnosed with CS. (h) An infant with a typical Noonan gestalt face, carrying one of the common PTPN11 variants. (i) 17‐year‐old patient with a giant cell mandibular granuloma, harboring a SOS1 variant. (j) A 13‐year‐old boy with a compound heterozygous LZTR1 variant, a typical gestalt Noonan face, and a pectus deformity is shown. [Colour figure can be viewed at wileyonlinelibrary.com]

RAF1 variants (n = 7) were uniformly associated with cardiac involvement (100%). Additionally, 4 RAF1 patients presented with short stature (57.1%), 2 of whom had macrocephaly (28%), and one patient presented with hearing loss (14%). One of the RAF1 patients with severe HCM received MEK inhibitor (trametinib) treatment, but died due to cardiac complications.

RIT1 variants (n = 4) were consistently associated with PS (100%); 3 of them presented with short stature (75%). Two patients had HCM, and one of them was given trametinib as a treatment option for severe cardiomyopathy. One patient harboring a RIT1 variant died from progressive cardiac disease (Figure 3).

Among patients harboring LZTR1 variants (n = 5), neurocognitive delay was the prominent feature, observed in four individuals (80%), whereas cardiac involvement was detected in only one patient. One patient carried compound heterozygous LZTR1 variants, while another was homozygous for an LZTR1 variant (Figure 3).

Variants in HRAS, BRAF, and KRAS were identified in a limited number of patients and were associated with the expected syndromic phenotypes within the RASopathy spectrum. HRAS variants (n = 4), consistent with Costello syndrome, were uniformly accompanied by ectodermal anomalies and intellectual disability. One patient developed both embryonal rhabdomyosarcoma and secondary neuroblastoma, while another presented with renal agenesis and hearing loss (Figure 3). BRAF variants were detected in four patients, two of whom were diagnosed with cardiofaciocutaneous syndrome. Macrocephaly and epilepsy were present in two patients. KRAS variants (n = 2) were associated with macrocephaly, ectodermal findings, growth retardation, and neurocognitive impairment in both individuals. SPRED1 variants were detected in five patients with clinical findings fully consistent with Legius syndrome (100%), all of whom had presented with café‐au‐lait macules (CALMs).

Variants in MAP2K2 were identified in two patients. They both presented with dysmorphic facial features and cardiac defects, but the patient with variant M147V also had neurocognitive delay, ectodermal findings, along with cleft lip and palate, further expanding the phenotypic spectrum associated with MAP2K2‐related disease. In addition, a pathogenic SHOC2 variant was identified in one patient fulfilling the clinical features of NS‐LH, accompanied by short stature and macrocephaly. Finally, RASA2 variants were detected in three patients with Noonan features, while a CBL variant, found in one patient with isolated HCM, was included as a rare RASopathy‐related pathogenic variant (Table 3).

Among the 21 patients with NF‐NS carrying a heterozygous NF1 (NM_001042492.3) variant (Table 2), the majority were missense changes (81%), with splice‐site (10%) and in‐frame deletion variants (10%) also represented. Typical facial dysmorphology was observed in all patients (100%), whereas short stature (< 3rd centile) was present in 10/21 cases (47.6%), most frequently in those carrying K191T, M992del, and R1276Q variants. Cardiac defects were identified in 4/21 patients (19%), including PS, ASD, and aortic dilatation, primarily in individuals with K191T and R1276Q. Neurocognitive delay was present in 38% of patients, while epilepsy was reported in 14%. Neurological findings were found particularly in carriers of F894S, T780K, and R1276Q. Notably, in‐frame deletions such as M992del and the missense variant R1276Q were consistently associated with Noonan‐like features and additional findings like macrocephaly and Chiari malformation.

TABLE 2.

Clinical features of NF‐NS phenotyped patients harboring NF1 variants.

Patient ID	Short stature (< 3rd centile)	Typical facial dysmorphology	Cardiac defect	Neurocognitive delay	Other findings	NF1 (NM_001042492.3) variant (c.)	Exon number	Amino acid change (p.)	ACMG criteria	Variant type
15	+	+	Aortic dilatation	+	−	572A>C	5	K191T	VUS	Missense
25	+	+	−	−	Sensorineural hearing loss	572A>C	5	K191T	VUS	Missense
87	−	+	−	−	Low‐grade glial tumor, macrocephaly	669G>A	7	W223*	P	Nonsense
78	−	+	−	−	−	908T>C	9	L303P	LP	Missense
76	−	+	−	+	Optic glioma, macrocephaly	1885G>A	17	G629R	P	Missense
98	−	+	−	+	−	2339C>A	20	T780K	P	missense
35	−	+	−	−	−	2409 + 1G>C	Intron 20	−	P	Splice donor
36	−	+	−	−	−	2409 + 1G>C	Intron 20	−	P	Splice donor
32	+	+	−	+	−	3709‐2A>G	Intron 27	−	P	Splice acceptor
115	−	+	−	−	NF1‐diagnosed mother with plexiform neurofibromas	1756_1759del	16	T586Vfs*18	P	Frameshift
71	+	+	−	−	−	2540 T>C	21	L847P	P	Missense
69	−	+	−	+	Epilepsy, hypotonia, prematurity	2681T>C	21	F894S	LP	Missense
74	+	+	−	−	−	2970_2972delAAT	22	M992del	P	In frame
75	+	+	ASD	−	Macrocephaly, Chiari malformation	2970_2972delAAT	22	M992del	P	In frame
105	−	+	−	−	−	2999G>C	26	R1000P	P	Missense
67	+	+	−	−	−	3295A>C	25	K1009Q	LP	Missense
72	−	+	−	−	Bilateral optic atrophy	3706T>C	27	W1236R	LP	Missense
33	+	+	PS	+	Epilepsy	3827G>A	28	R1276Q	P	Missense
86	+	+	−	−	−	4084C>T	30	R1362*	P	Nonsense
9	−	+	−	+	−	5498T>C	37	L1833P	P	Missense
34	+	+	ASD	+	Vesicoureteral reflux	5609G>A	38	R1870Q	P	Missense

4. Discussion

Genotype–phenotype correlations have become increasingly significant in the clinical management of RASopathies, as they offer insight into the heterogeneity of presentation and allow for early identification of high‐risk features, particularly cardiac and neurodevelopmental complications [9, 10]. With the progress of precision medicine, recognizing rare or unusual variants in RAS/MAPK pathway genes is now crucial. These findings help clinicians make more accurate diagnoses and also guide patient‐specific treatment options, including the use of MEK inhibitors in appropriate cases [2, 11].

In our cohort, consistent with the previous data, PTPN11 is the most frequently mutated gene in NS, and the vast majority of pathogenic variants occur as missense mutations [3, 12]. The exon distribution observed in our PTPN11‐positive cohort aligns well with previously reported mutational hotspots. Exons 3, 7, 8, 12, and 13 were the most frequently affected, consistent with earlier descriptions of the molecular spectrum of PTPN11‐related NS [12, 13]. Variants in these regions were commonly associated with short stature and the typical Noonan facial gestalt, while cardiac defects appeared particularly enriched among changes in exons 3 and 12–13. Similar patterns have also been noted in more recent RASopathy cohorts, further supporting the role of exon location in shaping phenotypic expression [10]. PS emerged as the predominant cardiac abnormality in patients harboring PTPN11 variants, whereas HCM remained relatively uncommon [12, 13]. In particular, the association reported between exon 8 variants and PS [13] is supported by our findings, as nearly all individuals harboring exon 8 mutations presented with valvular obstruction. The high familial occurrence rate of PTPN11‐related NS described previously [12] was not observed in our cohort, as most cases were de novo. Neurodevelopmental delay has been reported in about one‐third of individuals with PTPN11 variants, and we observed a similar rate in our cohort [13].

Several genotype–phenotype associations described in earlier studies were also evident in our cohort. The variants T73I and E76K, both known to carry a high risk for myeloproliferative disease, including JMML [3, 13], were identified in children who developed JMML. Notably, our cohort included two patients with the T73I variant: one died from JMML‐related complications, while the other had bilateral sensorineural hearing loss and iris coloboma (Table 1).

While macrocephaly has been described in a substantial proportion of patients with PTPN11‐variants, an unexpected finding in our cohort was the occurrence of microcephaly in several individuals, particularly those carrying variants in exons 3 and 13 [13]. Microcephaly is not considered a classic feature of NS, yet it has been described as an uncommon or atypical cranial finding in recent RASopathy cohorts [10]. In our study, microcephaly was present in 9.3% of all patients, and notably, it occurred more frequently among those carrying PTPN11 variants (8/49) compared with other genotypes. Given SHP2's role in coordinating developmental signaling pathways that support early brain growth, alterations in its regulatory domains may influence neurodevelopment more broadly than previously appreciated [14]. Among the eight PTPN11 patients with microcephaly, age‐appropriate developmental assessments showed that five had motor and cognitive development consistent with their peers, whereas three of them exhibited developmental delay predominantly affecting speech and were receiving special education support.

LEOPARD‐associated variants such as Y279C and T468P were also present in our series and followed their expected phenotype, including a higher frequency of HCM, cutaneous findings, hypotonia, and hearing impairment [3, 13]. Additional variant‐specific features, including long QT interval, autism spectrum disorder, and ocular anomalies such as iris coloboma, further highlight the broad phenotypic variability observed in PTPN11‐related RASopathies (Table 1).

Hearing impairment has been reported in a variable proportion of patients with RASopathies and appears to be particularly enriched among PTPN11‐positive patients [3, 15, 16, 17]. In our cohort, however, hearing loss was identified in only 4.2% of patients, which is likely influenced by incomplete routine audiological screening in clinical follow‐up. Unlike prior observations emphasizing a predominantly PTPN11‐associated pattern, hearing loss in our cohort was also detected in patients with RAF1, NF1, and HRAS variants, suggesting broader genetic heterogeneity. These findings highlight the need for systematic hearing assessments to better capture the true spectrum of auditory involvement in pediatric RASopathies [16, 17].

SOS1 was one of the most affected genes in our cohort (n = 10, 8.5%), in line with previously reported frequencies [13, 18]. All variants were missense, and affected individuals typically exhibited classic Noonan facial gestalt with a lower prevalence of neurodevelopmental delay compared to PTPN11‐positive patients. Cardiac involvement, particularly pulmonary stenosis, was prominent [9]. Notably, one patient with the p.Arg552Gly variant presented with a mandibular giant cell granuloma (Figure 3) consistent with earlier reports linking this variant to giant cell lesions [19].

In the literature, RAF1 variants have been reported in patients with NS, CFCS, CS, and LEOPARD syndromes [20]. All RAF1‐positive patients in our cohort were consistent with NS. Because these phenotypes can overlap, the final syndromic classification may change over time with clinical follow‐up; therefore, long‐term clinical surveillance is important [20]. Similar to previous reports, most of our patients showed a typical Noonan facial gestalt, and all had cardiac involvement. Different from the literature data [20, 21] HCM and PS were observed at similar frequencies in our cohort. In previous reports, RASopathy‐associated RAF1 variants are almost exclusively missense, and HCM is particularly prominent in patients carrying variants in hotspot regions [21]. In our cohort, four patients carried hotspot variants, and three of them had HCM as the main cardiac finding (Table 3). No patient with a nonhotspot variant developed HCM. A patient with the S257L variant received trametinib treatment but died due to severe cardiac disease. It is well recognized that patients with RAF1 variants may have a severe course due to progressive cardiac involvement [21], and trametinib treatment appears to be promising [11].

Previous studies have reported HCM in approximately 50%–70% of individuals with RIT1 mutations, higher than the broader NS population [22, 23]. Consistent with these findings, HCM was more frequent among our RIT1‐positive patients than among those carrying PTPN11 or SOS1 variants and occurred at levels similar to those observed in patients with RAF1 variants. Although short stature has generally been reported less frequently in RIT1‐related NS compared with PTPN11 or RAF1, 75% of our RIT1‐positive patients exhibited short stature, a rate that was even higher than that observed in our RAF1 cohort (57%) and PTPN11 cohort (70.4%) [22].

Although it remains debated whether NF‐NS represents a separate entity or a phenotypic subgroup of NF1, it is clinically recognized by a prominent Noonan‐like facial gestalt, pectus deformities, and cardiac findings. In classic NF1, pectus deformities are underexplored, and cardiac involvement is less frequent [24, 25, 26]. Consistent with the literature, all NF‐NS patients in our cohort displayed a typical Noonan facial gestalt, and the frequencies of neurocognitive impairment and macrocephaly were comparable to previous reports. Cardiac involvement was less common than expected in our cohort, and none of our NF‐NS patients had HCM [24, 25]. Although previous studies have described an increased rate of ADHD in individuals with the NF‐NS phenotype [25], no such features were observed in our group. Two patients in our cohort had epilepsy, and one patient had bilateral sensorineural hearing loss (Table 2).

Demonstrating an NF1 variant at the molecular level is essential for distinguishing NF‐NS from overlapping RASopathies. In line with the molecular patterns reported in the literature, the NF‐NS phenotype in our cohort was most often associated with NF1 missense and in‐frame variants, although nonsense variants may also occur [25, 27]. While NF‐NS has been linked to specific hotspot regions within the NF1 gene, it is not restricted to these sites, supporting a broader mutational spectrum [25].

Previous studies have shown that both monoallelic and biallelic variants in LZTR1 can cause NS, typically presenting with the characteristic Noonan facial gestalt and congenital cardiac anomalies, especially HCM [7, 28, 29, 30]. In dominantly inherited cases, pathogenic variants are most often missense, although nonsense and splice‐site alterations have also been reported, and affected individuals commonly exhibit learning difficulties in addition to the core Noonan features. In contrast, biallelic LZTR1 variants are associated with a more severe phenotype, frequently including early‐onset and severe HCM, autism spectrum disorder, epilepsy, and global developmental delay [29, 30], as well as reports of hematologic malignancies and tumor predisposition [29]. Recessive LZTR1‐related NS has been attributed to both missense variants and canonical or noncanonical splice defects [29, 30]. In contrast to previously published data, only one of our five patients harboring LZTR1 variants showed cardiac involvement (Table 3). Instead, all exhibited neurocognitive difficulties, such as learning problems, poor academic performance, and delayed speech acquisition. This pattern suggests that, within our cohort, LZTR1‐associated phenotypes may predominantly manifest with neurodevelopmental features.

In line with recent data on LS [31], pigmentary phenotype with café‐au‐lait macules, freckling, and no NF1‐suggestive ophthalmologic or cranial MRI findings, systemic involvement in our series was limited to skeletal anomalies (scoliosis). We did not observe neurofibromas or other tumor manifestations, supporting the concept of LS as an NF1‐like but nontumor‐predisposing RASopathy [31].

In our cohort, HRAS‐related CS, with the missense G12S variant as the most frequent genotype (3/4 patients), is consistent with previous reports describing it as the predominant pathogenic variant in approximately 80% of CS cases [32, 33]. Besides, these three patients displayed the classic CS phenotype; one patient carried a novel noncoding VUS HRAS variant (c.*9A>T); the clinical presentation was fully compatible with CS (Figure 3), and the negative parental segregation analysis supported the pathogenicity of this de novo variant. Notably, this patient also presented with a unilateral atrophic kidney, a finding not previously reported in association with CS. This highlights that CS may occasionally be associated with noncoding HRAS alterations and underscores the importance of considering such variants in the diagnostic evaluation. As expected for CS, which carries the highest malignancy risk among the RASopathies [34], one of our patients developed both rhabdomyosarcoma and neuroblastoma (ectomesenchymoma), aligning with the established oncogenic activity of HRAS [32].

In the literature, more than half of individuals with BRAF variants are reported to exhibit a clinical picture consistent with CFCS, and a small proportion is related to NS [35]. Alongside cardiac involvement (75%), macrocephaly was also prominent in our BRAF‐positive patients (75%), whereas previous studies have emphasized neurodevelopmental delay and seizures as one of the predominant findings [35, 36], with macrocephaly described only rarely. Given the well‐recognized phenotypic overlap across the RASopathy spectrum, longitudinal clinical follow‐up and the monitoring of newly emerging features are essential. As expected, missense variants accounted for the majority of BRAF alterations in our series; however, in contrast to the existing literature [36] we identified a likely pathogenic noncoding variant in one patient with a phenotype fitting BRAF‐related NS, expanding the mutational spectrum and supporting the notion that BRAF‐associated phenotypes may be broader and more allele‐specific than previously recognized.

RASA2 is a rare contributor to the genetic spectrum of NS, with pathogenicity so far demonstrated mainly for heterozygous missense variants that impair its RAS‐GAP function and lead to increased RAS–ERK signaling [37]. In our cohort, we identified three patients who fulfilled the clinical diagnostic criteria for NS, all of whom had congenital cardiac involvement and carried a RASA2 VUS variant (Table 3 and Table S1). In one patient, the variant was considered clinically meaningful based on a phenotype highly consistent with NS. In the remaining two patients, although the variants were intronic, they were located within splicing‐relevant regions, and computational prediction tools supported a deleterious impact on gene function (Figure 3). However, these variants remain classified as VUS and should be interpreted with caution. The observed genotype–phenotype correlation in our cohort suggests a possible contribution of RASA2 variants to the NS phenotype; however, additional functional studies and larger cohorts are required to clarify their clinical significance. Accordingly, these findings should be considered observational and supportive of the emerging role of RASA2, rather than definitive evidence of pathogenicity.

Although classified as VUS according to ACMG criteria, these variants identified in our cohort were further evaluated through comprehensive phenotype–genotype correlations. The majority of these patients exhibited typical Noonan facial gestalt, while the others fulfilled supportive secondary criteria consistent with a RASopathy phenotype (Table S1). Importantly, all non‐NF1 cases received their molecular diagnosis through a targeted RASopathy gene panel based on clinical suspicion, reducing the likelihood of incidental findings. Patients harboring the same NF1 VUS variants presented with CALMs in addition to short stature and characteristic Noonan‐like facial features, compatible with the RASopathy spectrum.

Taken together, the combination of a clinical phenotype consistent with RASopathy features, gene relevance within the RAS/MAPK pathway, rarity in population databases, and predicted functional impact supports the likelihood that these VUS variants contribute to disease manifestation in our cohort. Nevertheless, there is a need for larger cohorts and functional studies to further clarify genotype–phenotype associations for rare and noncoding variants in RASopathies.

One of the limitations of this study is its retrospective design, which may restrict data completeness and consistency. The small numbers within certain genotype groups, such as RASA2, MEK2, CBL, and MAP2K2, limit the robustness of gene‐specific phenotype interpretations. Additionally, functional assays could not be performed for VUS variants, and clinical assessments were not standardized across all patients. Another limitation is that MRAS and RRAS2 were not included in the gene panel used in this cohort, which may have resulted in missed diagnoses in patients with Noonan‐like phenotypes. Despite these limitations, our cohort offers valuable insights into pediatric RASopathies and broadens the phenotypic spectrum associated with several rare genotypes.

This study presents one of the largest fully genotype‐confirmed RASopathy cohorts to date, focusing solely on pediatric patients, enabling a comprehensive evaluation of both common and rare variants and their phenotypic spectrum. Our extended cohort size allowed the identification of rare variants and provided meaningful contributions to the expanding genotypic and phenotypic spectrum of RASopathies. In particular, the observation of microcephaly in PTPN11‐related NS and the distinct gene‐specific clinical patterns across RAF1, RIT1, SOS1, and LZTR1 further refine genotype–phenotype correlations in pediatric RASopathies. Our findings underscore the possible overlap among RASopathy features and highlight the critical importance of longitudinal follow‐up. With the emergence of targeted therapies such as MEK inhibitors, molecular diagnosis has become increasingly central not only for confirming the phenotype but also for prognosis, genetic counseling, and personalized treatment strategies. Future multicenter studies involving larger, diverse pediatric populations are necessary to expand these genotype–phenotype associations, clarify the clinical significance of rare variants, and further advance precision medicine approaches for children with RASopathies.

Author Contributions

Aslı Genç: conceptualization, data curation, investigation, resources, visualization, writing – original draft. Elif Sarıkaya: data curation, investigation. Ahmet Cevdet Ceylan: formal analysis, methodology, data interpretation. Büşranur Çavdarlı: formal analysis, methodology, data interpretation. İbrahim İlker Çetin: resources, investigation. İlkay Erdoğan: resources, investigation. Esra Kılıç: conceptualization, supervision, data interpretation, resources, visualization, writing – review and editing.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Clinical characteristics of patients carrying VUS variants in our cohort of 118 patients.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.