Prevalence of Genetic Diagnoses in a Cohort with Valvar Pulmonary Stenosis

Abstract

Background.

Valvar pulmonary stenosis (vPS) accounts for 8–12% of congenital heart disease (CHD) cases. Multiple genetic syndromes are associated with vPS, most commonly Noonan syndrome, but the etiology is unknown in most cases. We analyzed genomic data from a large cohort with vPS to determine the prevalence of genetic diagnosis.

Methods.

The Pediatric Cardiac Genomics Consortium database was queried to identify probands with vPS without complex CHD or aneuploidy and with existing whole exome or genome sequencing. A custom analysis workflow was used to identify likely pathogenic or pathogenic variants in disease-associated genes. Demographic and phenotypic characteristics were compared between groups with and without molecular diagnoses.

Results.

Data from 119 probands (105 trios) were included. A molecular diagnosis was identified in 22 (18%); 17 (14%) had Noonan syndrome or a related disorder. Extracardiac and/or neurodevelopmental comorbidities were seen in 67/119 (56%) of probands. Molecular diagnosis was more common in those with extracardiac and/or neurodevelopmental phenotypes than those without (18/67 vs 4/52, p=0.0086).

Conclusions.

Clinicians should have high suspicion for a genetic diagnosis in individuals with vPS, particularly if additional phenotypes are present. Our results suggest that clinicians should consider offering sequencing of at least the known CHD and RASopathy genes to all individuals with vPS, regardless of whether that individual has extracardiac or neurodevelopmental phenotypes present.

Introduction

Congenital heart disease (CHD) affects 1 in 100 live births and is associated with a range of morbidity and mortality¹. CHD has both monogenic and complex etiologies: Approximately 13% is attributed to chromosome abnormalities (e.g., trisomy 21, 22q11 deletion syndrome)² and among those without a gross chromosome abnormality, approximately 8–10% have causative de novo pathogenic variants identified via exome sequencing^{3, 4}.The prevalence of causative pathogenic variants rises to 28% when considering individuals with CHD in combination with neurodevelopmental delay and/or extracardiac anomalies^{3, 5}.

While CHD categories are generally distinguished by anatomic and physiologic differences, these categories can also be distinguished by heritability differences. Studies have shown that heterotaxy and right ventricular outflow tract obstruction (RVOTO) are the most heritable types of CHD. For RVOTO, the relative risk for recurrence in first degree relatives is 48.6, versus 12.9 for left ventricular outflow tract obstructive defects and 11.7 for conotruncal defects^{6, 7}. The most common RVOTO, valvar pulmonary stenosis (vPS), is estimated to represent 8–12% of all CHD^8–10. However, there is a paucity of studies regarding the prevalence of genetic diagnoses among individuals with vPS. Prior studies of large CHD cohorts have either excluded vPS or grouped it into an “other” category, which includes such CHD as atrioventricular canal defects and atrial septal defects^3–5.

Our recently published retrospective study identified genetic diagnoses in 10% of 204 children with vPS, but also revealed that approximately 20% of the cohort had extracardiac manifestations and had not received any genetic services¹¹. Similar results were reported by Bell and colleagues (2021), with 16% of 686 patients with vPS having a genetic diagnosis but without the entire cohort having received genetic testing¹². Noonan syndrome (NS), which is the second most common syndromic cause of CHD, is highly associated with vPS^{13, 14} and was the most common genetic diagnosis in both cohorts (6–9%)^{11, 12}. Individuals with NS frequently have extracardiac manifestations such as short stature, poor feeding, genitourinary anomalies, developmental delay, thrombocytopenia, and ocular anomalies as well as characteristic facial features¹³. However, the majority of vPS is identified in infancy or early childhood¹⁵ when many of these extracardiac issues are not apparent or are still unknown.

Earlier diagnosis of genetic disorders and understanding how genetic variants are associated with outcomes can improve healthcare delivery by allowing anticipatory rather than reactionary guidance and management^{16, 17}, while simultaneously making possible more accurate familial risk counseling. From a cardiac standpoint, multiple studies have shown that individuals with vPS due to NS are more likely to need intervention and re-intervention^18–20. Further, a recent study identified worse outcomes in children with NS versus children without NS hospitalized for cardiac surgery: longer hospitalization, higher cost, higher mortality, and increased risk of chylothorax²¹. In addition to complications from structural heart defects, infants and children with NS are at risk for failure to thrive, feeding intolerance and reflux, thrombocytopenia, and progressive cardiomyopathy¹³. Therefore, improved understanding of the likelihood that vPS may be associated with a genetic diagnosis could impact genetic testing recommendations for infants and children with vPS, leading to earlier diagnosis and improved medical care. In order to determine the prevalence of genetic diagnoses and extracardiac comorbidities in a cohort of patients with vPS, we analyzed existing sequencing data from patients with vPS enrolled in the Pediatric Cardiac Genomics Consortium (PCGC). We hypothesized that the prevalence of genetic diagnoses would be greater among this sequenced cohort than among previously published, incompletely sequenced, retrospective cohorts.

Methods

Our study is a secondary analysis of de-identified individual-level data from the PCGC’s Congenital Heart Disease Network Study (CHD GENES: Clinicaltrials.gov identifier NCT01196182).²² All participants or their parents provided informed consent (at individual sites) prior to enrollment in CHD GENES. Analysis of the de-identified dataset (sequencing and phenotype data from participants with vPS in CHD Genes) was performed under CCHMC IRB protocol 2020–0045. Full methods are available as supplementary material. The data used for this study are available upon request to the PCGC (https://benchtobassinet.com/?page_id=221) or through dbGaP (https://www.ncbi.nlm.nih.gov/gap/). The dbGaP IDs of the 119 probands included in our study are listed in supplemental material (Supplemental Table I).

Results

Cohort

We identified 119 unrelated probands meeting our inclusion criteria (Supplemental Table I). Cohort demographics and phenotypes are summarized in Table 1 and cohort breakdown by test type (trio or singleton, genome or exome) is summarized in Figure 1. Of note, the age of probands with NDD was significantly older than those without NDD (p=2.8e-4).

Table 1:

Cohort Characteristics by Phenotype Group

	Isolated	ECA	NDD	ECA+NDD
Total	52	20	21	26
Sex
Female	51% (27)	35% (7)	19% (4)	73% (19)
Age (years)	9.49	4.22	16.69	9.19
Ethnicity
Non-Hispanic	84% (44)	80% (16)	85% (18)	76% (20)
Hispanic	15% (8)	20% (4)	14% (3)	23% (6)
Race
Don’t know	0% (0)	5% (1)	0% (0)	0% (0)
White	80% (42)	65% (13)	80% (17)	84% (22)
Black or African American	9% (5)	0% (0)	9% (2)	0% (0)
Asian	5% (3)	15% (3)	4% (1)	0% (0)
More than one race	3% (2)	15% (3)	4% (1)	15% (4)
Molecular Dx
RAS	3% (2)	15% (3)	9% (2)	38% (10)
OTHER	3% (2)	10% (2)	0% (0)	3% (1)
Not identified	92% (48)	75% (15)	90% (19)	57% (15)
Pair-wise Comparisons	Sex (M/F)	Molecular Dx (Y/N)	Age
Isolated (52) vs ECA/NDD (67)	p=0.46	p=0.01	p=0.19
ECA (46) vs no ECA (73)	p=0.19	p<0.01	p=0.13
NDD (47) vs no NDD (72)	p=1.0	P=0.05	p=< 0.01

Figure 1:

Overview of analysis pipeline and results

Molecular analysis results

Among 105 proband-parent trios and 14 singleton probands, we identified 22 (18%) with pathogenic or likely pathogenic variants in genes known to be associated with CHD (gene list available in Supplemental Table II). Among those 22 probands, there were 17 with a RASopathy. Among the group with a RASopathy diagnosis, one proband inherited their PTPN11 variant from their father and one PTPN11 variant was identified in a singleton genome, making inheritance impossible to determine. The remaining 15 RASopathy variants occurred de novo. Other diagnoses identified in the cohort include CHARGE syndrome (CHD7), “CHD, dysmorphic facial features, and intellectual development disorder” (CDK13), and Mowat-Wilson syndrome (ZEB2). Table 2 lists genes and variants for the 22 probands with a molecular diagnosis. The majority of the variants identified have been reported in ClinVar as likely pathogenic or pathogenic. Of those not present in ClinVar, three are classified as pathogenic based on ACMG/AMP criteria (GATA4, NOTCH1, ZEB2),²³ and one is classified as likely pathogenic (NF1). Therefore, we considered these latter 4 variants to be diagnostic for the associated syndromes.

Table 2:

Molecular Diagnoses

dbGaP ID	Gene	cDNA variant	Amino Acid Change	WES			WGS			Inh	ClinVar
				F	M	P	F	M	P
1690	CDK13	c.2525A>G	N842S	29	33	63				DNM	Pathogenic/Likely pathogenic
4802	CHD7	c.7803C>G	Y2601Ter	30	25	45				DNM	Pathogenic
2036	GATA4	c.601C>T	R201Ter	48	43	45				DNM	N/A*
26233‡	MRAS	c.203C>T	T68I	14	14	14				DNM	Likely pathogenic
17207‡	NF1	c.3545T>G	V1182G	32	28	91				DNM	N/A†
6459	NOTCH1	c.4918del	A1640Hfs*9	7	8	7	25	29	20	DNM	N/A*
1182	PTPN11	c.922A>G	N308D	47	39	68				DNM	Pathogenic
2457‡	PTPN11	c.1505C>T	S502L	33	33	70				DNM	Pathogenic/Likely pathogenic
2724	PTPN11	c.1507G>A	G503R						31	unk	Pathogenic
8406	PTPN11	c.598A>T	N200Y	46	32	55				DNM	Pathogenic/Likely pathogenic
8830	PTPN11	c.802G>T	G268C	37	44	121				DNM	Pathogenic/Likely pathogenic
14763	PTPN11	c.922A>G	N308D	29	32	67				DNM	Pathogenic
18768	PTPN11	c.922A>G	N308D	18	16	16				PAT	Pathogenic
23646	PTPN11	c.922A>G	N308D				27	35	30	DNM	Pathogenic
23689‡	PTPN11	c.923A>C	N308T	18	32	33				DNM	Pathogenic
292‡	RAF1	c.788T>G	V263G	29	28	44				DNM	Likely pathogenic
2681‡	RAF1	c.770C>T	S257L	30	16	27				DNM	Pathogenic
2167‡	RIT1	c.170C>G	A57G	25	14	10				DNM	Pathogenic
15999‡	RIT1	c.170C>G	A57G	10	10	12				DNM	Pathogenic
953‡	SOS1	c.1654A>G	R552G	32	28	34				DNM	Pathogenic
5698‡	SOS1	c.797C>A	T266K	42	41	165				DNM	Pathogenic
7321‡	ZEB2	c.1859_1860del	I620Sfs*10	65	30	83				DNM	N/A*

WES, whole exome sequencing; WGS, whole genome sequencing; Inh, inheritance; DNM, de novo variant; PAT, paternally inherited variant; N/A, not available

^*Meets ACMG/AMP criteria for pathogenic: PVS1 (loss of function is known mechanism of disease), PM2 (not in gnomAD), PP3 (pathogenic predictions in silico)

^†Meets ACMG/AMP criteria for likely pathogenic: PM2 (not in gnomAD), PM5 (alternative variant classified as pathogenic, V1182L and V1182F), PP2 (non-VUS missense path rate), PP3 (pathogenic predictions in silico)

^‡Clinical diagnosis of Noonan syndrome at study enrollment (n=11). Subject 1–02085 also had clinical diagnosis of NS but did not have any de novo candidate causative variants identified.

“Genetic syndrome diagnosis” is one of the fields included in the CHD-GENES study case report form filled out at the time of enrollment. Twelve of the 119 patients had a clinical diagnosis of NS noted at enrollment (Study IDs are noted in Table 2). Interestingly, only 10 of these 12 had a molecular diagnosis of NS confirmed on our exome/genome analysis. Of the remaining two, one had a pathogenic variant in ZEB2 (causative for Mowat-Wilson syndrome but not currently associated with NS), and one did not have any detected RAS variants of pathogenic or uncertain significance. Thus overall there were 7 probands without a NS diagnosis reported at the time of enrollment, in whom we identified pathogenic variants in RAS/MAPK pathway genes. The age range of these individuals was 4 months to 21 years (mean 7.9 years, median 4.2 years). The age range of the 10 probands with clinical diagnosis of NS at entry, and concordant molecular results on our analysis, was newborn to 19.4 years (mean 7.9 years, median 6.8 years).

Probands with both exome and whole genome sequence data

Eleven of the 105 probands with trio exome data also had genome trio data available. In one such proband, a de novo, likely pathogenic truncating variant in NOTCH1 was identified. This variant was not identified in the exome data, likely due to poor coverage (7–8 reads for proband and both parents). In comparison, coverage on the genome was 20–29 reads. Of note, this particular locus (9:136504772_GC/G) had low coverage (<10 reads) in more than 1/3 of exome samples. The remaining 10 probands with both exome and whole genome sequence did not have a pathogenic or likely pathogenic variant identified in a gene with known association to human disease.

Copy number variant analysis

We identified a paternally inherited 8p23 microdeletion in one proband which affected a single exon of CSMD1. No other clinically significant CNVs were identified, although the analysis was significantly limited by exome data quality.

Phenotypic predictors of genetic diagnosis

The majority of probands with a molecular diagnosis had extracardiac anomalies and/or neurodevelopmental differences (ECANDD) in addition to vPS (Table 3). Overall, 27% (18/67) of those within the ECANDD group had a molecular diagnosis versus 8% (4/52) of the isolated group (p= 0.0086, Fisher’s exact test, OR 95% CI 1.39, 13.98). However, when analysis was restricted to the cohort without ECA (top row of Table 3), there was no difference in prevalence of molecular diagnoses between probands with or without NDD (8% vs 9%, 4/52 vs 2/21, p=1.0, Fisher’s exact test). The strongest predictor of a molecular diagnosis in our cohort was presence of ECA.

Table 3:

Genetic Diagnoses and Phenotype Breakdown

	Neurodevelopment Abnormal
Extracardiac	No	Yes
No	52 (48% M, 8% GNT)	21 (81% M, 10% GNT)
	1 GATA4 1 NOTCH1 2 PTPN11	2 PTPN11
Yes	20 (65% M, 25% GNT)	26 (27% M, 42% GNT)
	1 CDK13 1 CHD7 3 PTPN11	1 MRAS 1 NF1 2 PTPN11 2 RAF1	2 RIT1 2 SOS1 1 ZEB2

M, Male; GNT, genetic diagnosis; shaded cells indicate non-isolated

When considering the specific ECA reported in the cohort (Table 4), the most common system affected was the head/face, with craniofacial anomalies reported in 23 of the 46 probands with ECA manifestations. Dysmorphic facies was the only ECA that was significantly associated with presence of a genetic diagnosis. Co-occurrence of ECA in individual patients is illustrated in Supplemental Figure I. Interestingly, only 19 of the 46 had a clinical genetics exam documented on their CHD Genes case report forms, and 12/19 had dysmorphic facies documented versus 11 with dysmorphic facies out of 27 without clinical genetics exam documented.

Table 4:

Extracardiac Manifestations by Phenotype Category^* (n=46)

	Molecular Diagnosis		No molecular diagnosis
	With phenotype	Without phenotype	With phenotype	Without phenotype	Uncorrected P value†	FDR-corrected P value
Dysmorphic facies	14	2	9	21	0.000	0.010
Head abnormalities	6	10	1	29	0.005	0.052
Chest abnormalities	4	12	0	30	0.011	0.069
Ear abnormalities	8	8	4	26	0.013	0.069
Nose abnormalities	3	13	0	30	0.037	0.135
Neck abnormalities	3	13	0	30	0.037	0.135
Hand abnormalities	3	13	1	29	0.114	0.359
Neurological abnormalities	4	12	2	28	0.163	0.448
Eye abnormalities	7	9	8	22	0.325	0.743
Skeletal abnormalities	1	15	6	24	0.394	0.743
Genitourinary abnormalities	3	13	3	27	0.405	0.743
Feet abnormalities	3	13	3	27	0.405	0.743
Airway abnormalities	0	16	2	28	0.536	0.851
Mouth abnormalities	0	16	3	27	0.542	0.851
Abdominal abnormalities‡	1	15	5	25	0.649	0.952
Dermatologic abnormalities	2	14	6	24	0.694	0.954
Palate abnormalities	1	15	1	29	1.000	1.000
Pulmonary abnormalities	1	15	1	29	1.000	1.000
Renal abnormalities	2	14	3	27	1.000	1.000
Endocrinologic abnormalities	1	15	1	29	1.000	1.000
Hematologic abnormalities	0	16	1	29	1.000	1.000
Immunologic abnormalities	0	16	1	29	1.000	1.000

^*See form 105 for detailed list of phenotypes in each category

^†Fisher’s exact test

^‡No hepatic abnormalities were reported in any of the probands in this cohort

We next looked for association between additional cardiac phenotypes, presence of ECANDD, and genetic diagnoses (Supplemental Table III). Overall, 80/119 (67%) had an additional cardiac phenotype present, which was most commonly an atrial septal defect (present in 60 probands). Probands with additional cardiac phenotypes present were more likely to have ECA (37/80 with ECA vs 9/39 without ECA, p=0.01, Fisher’s exact test) but were not more likely to have a molecular diagnosis (16/80 with genetic diagnosis vs 6/39 without, p=0.5, Fisher’s exact test). We also tabulated available information on probands with “dysplastic pulmonary valve” listed as a diagnosis (n=13, Supplemental Table IV) since this diagnosis has previously been reported to be associated with Noonan syndrome. Five of 13 probands with dysplastic pulmonary valve have a RASopathy, versus 12/106 without dysplastic pulmonary valve listed (p=0.02, Fisher’s exact test).

Finally, we looked at intervention status for the 33 probands for whom this information was available. The mean age of these 33 was markedly lower than the cohort as a whole (0.4 years versus 9 years). Given the very small numbers and that the data collection form does not distinguish between intervention for vPS versus for another CHD (such as a septal defect), we did not perform a statistical analysis but the data are presented in Supplemental Table V.

Discussion

We described a large cohort of sequenced individuals with vPS, thus overcoming one of the most significant limitations of previously described cohorts^{11, 12}. We found that nearly 20% of our cohort had an identifiable molecular genetic diagnosis on exome or genome sequencing. Of the 22 patients we identified with molecular diagnoses, the majority (17/22, 77%) had a pathogenic or likely pathogenic variant in a RASopathy gene. RASopathies are known to be highly associated with vPS. However, the prevalence of RASopathy diagnoses in our sequenced cohort (17/119, 14%) is nearly twice what was found by two retrospective cohort studies (6%−9%), in which the majority of probands had not received genetic testing, or information about percentage of probands with testing was not provided^{11, 12}. The comparison between our results and the prior studies suggests that RASopathies may be underdiagnosed among patients with vPS. Interestingly, we noted that 7 of the 17 probands with RASopathy molecular diagnosis did not enter the CHD Genes study with a syndrome diagnosis indicated on their case report forms. This included two probands without ECANDD (age 21 y and 0.3 y), 2 with ECA anomalies only (age 0.6 y and 4.7 y), 2 with ECA and NDD (age 3.8 y), and 2 with NDD only (age 14.7 y and 2.5 y). While data entry error and/or syndrome diagnosis after study entry cannot be ruled out, this suggests that lack of recognition of RASopathy diagnosis among individuals with vPS may occur across a range of ages and phenotypes.

Five of 22 probands (23%) with a molecular genetic diagnosis had a pathogenic/likely pathogenic variant in a non-RAS pathway-associated CHD gene. The reported ECA and NDD phenotypes of this group segregated with the corresponding molecular results as expected: two probands without ECA or NDD had variants in either GATA4 or NOTCH1 (both with known association to non-syndromic CHD^24–26), and three probands with ECA with or without NDD had pathogenic variants in established syndromic genes (CHD7, CDK13²⁷, ZEB2²⁸). Interestingly, the proband with a pathogenic ZEB2 variant (diagnostic of Mowat-Wilson syndrome) had a clinical diagnosis of Noonan syndrome reported on their case report form at study entry.

Coexistence of CHD with ECANDD is known to increase the likelihood of finding a genetic diagnosis⁵. This was true in our cohort as well; children with vPS and any ECANDD phenotype were significantly more likely to have a molecular diagnosis than those without ECANDD, and this relationship persists when NDD is not considered (e.g. comparison of children with vPS with or without NDD, to children with vPS and ECA phenotypes with or without NDD). The only ECA that was significantly associated with presence of molecular diagnosis was “dysmorphic facies,” which is consistent with what has been found in two retrospective studies of hospitalized infants with CHD^{29, 30}.

Among the those subjects without ECA, the presence of NDD did not correlate with increased likelihood of molecular diagnosis, as we observed similar diagnostic rates for those with vPS either with or without NDD. However, we used a broad definition of NDD and it is possible that limiting the NDD cohort to those with more severe disabilities might yield a different conclusion. There is a rapidly increasing number of genes in which pathogenic variants are associated with intellectual disability/developmental delay and many are also associated with various structural birth defects, including CHD (e.g. SIN3A³¹, AGO2³²). Therefore, we propose that absence of an NDD diagnosis in a child with vPS should not preclude offering genetic testing for CHD/RAS genes, but presence of NDD diagnosis should prompt consideration for broader clinical testing either as first test or through a tiered approach (e.g., gene panel followed by exome).

We did not see an association between presence of additional cardiac phenotypes (Supplemental Table III) and molecular diagnosis, although probands with additional cardiac phenotypes were more likely to have ECA. We did note that a diagnosis of pulmonary valve dysplasia was associated with molecular diagnosis, which has previously been observed in some studies¹¹(Supplemental Table IV). We interpret this result with caution, however, because it is possible that a child with both pulmonary valve dysplasia and stenosis may only have been labeled as having pulmonary stenosis, and there are no formal criteria for diagnosing pulmonary valve dysplasia. Therefore, this descriptor may not be uniformly applied across different clinicians and institutions.

A total of 97 probands in the cohort did not have a genetic molecular diagnosis identified, including 49 with ECA and/or NDD (15 ECA, 19 NDD, and 15 ECA+NDD). This could be due to a number of factors, including presence of a causative variant in a Mendelian gene that has not yet been associated with CHD, insufficient sequence coverage within a known gene(s), presence of a causative non-coding variant not sufficiently surveyed by sequencing, and/or multifactorial etiology. Of the 11 WES-negative probands that subsequently underwent WGS, there was 1 proband for whom we identified a causative variant that was missed on exome due to poor coverage. Since genome sequencing offers more even coverage of coding regions and can accurately detect copy number variation³³, it will certainly become an attractive one-test-covers-all option as cost continues to decrease.

Our study has a number of limitations. While the phenotyping was detailed at enrollment, a cross sectional study design does not allow for the evolution of clinical signs/symptoms and associated diagnoses over time. This is particularly important when considering the presence or absence of NDD diagnoses, which cannot be determined in a newborn and can be diagnosed over a range of ages. Consistent with this, the subcohort with just NDD was significantly older than those with ECA, ECA+NDD, or neither ECA or NDD. Based on the high overall percentage of patients with ECA and/or NDD in addition to vPS (58%) compared with other published CHD cohorts (typically quoted around 28%), our cohort may be biased toward patients with ECA and/or NDD manifestations and thus higher pre-test suspicion for a syndromic diagnosis. The majority of our subjects were surveyed using exome sequencing, which cannot determine presence or absence of copy number variations as effectively as whole genome sequence or microarray. Our CNV analysis did identify one proband with an inherited 8p23 deletion which was also noted on the proband’s case report form. However, our analysis indicated that the deletion was inherited and did not contain GATA4, which is typically deleted in the 8p23 microdeletion syndrome associated with CHD³⁴, therefore we did not consider this CNV to be of clear clinical significance. Interestingly, there were no diagnostic microdeletions found in our cohort (either reported at study entry, or detected in our CNV analysis). Our data, therefore, do not support a high yield of copy number analysis in this population, but this conclusion is tempered by limited ability to robustly assess copy number variation using exome data and the possibility that children with microdeletion syndromes (e.g. 22q11 deletion) might have not been prioritized for enrollment or sequencing in the CHD-GENES study. Microarray is often recommended as a first-tier test for infants with isolated CHD³⁰ but our results suggest that an alternative approach which includes both sequence analysis and copy number variant detection might be more efficient in patients with vPS. Finally, we did not have sufficient information about interventions that the subjects in our cohort required to determine if there might be an association between severity of vPS and likelihood of a genetic diagnosis.

In conclusion, our study demonstrates that a substantial proportion of patients with vPS harbor a pathogenic or likely pathogenic variant in a gene with known association to CHD, and likelihood of molecular genetic diagnosis is even higher when ECA manifestations are present. The true prevalence of genetic diagnoses in individuals with vPS is probably 10–20% or more, and the prevalence of RASopathies is at least 9–14%. Given the broad phenotypic heterogeneity of the most common syndrome identified in our cohort (NS), as well as our detection of “non-syndromic” CHD genetic diagnoses (NOTCH1, GATA4), our findings suggest that all children with vPS may benefit from comprehensive genetic testing-- including at a minimum sequencing of known RASopathy and CHD genes—regardless of whether they have ECA or NDD

Nonstandard abbreviations and acronyms:

ACMG

American College of Medical Genetics

AMP

Association for Molecular Pathology

CCHMC

Cincinnati Children’s Hospital Medical Center

CHD

congenital heart disease

CNV

copy number variant

ECA

extracardiac anomalies

NDD

neurodevelopmental delay

NS

Noonan syndrome

PCGC

pediatric cardiac genomic consortium

RVOTO

right ventricular outflow tract obstruction

vPS

valvar pulmonary stenosis

WES

whole exome sequencing

WGS

whole genome sequencing