Rapid Genome Sequencing Shows Diagnostic Utility In Infants With Congenital Heart Defects

Congenital heart disease (CHD) is the most common birth defect and a leading cause of infant mortality. CHD often has a genetic etiology and recent studies demonstrate utility in genetic testing. In clinical practice, decisions around genetic testing choices continue to evolve, and the incorporation of rapid genome sequencing (rGS) in CHD has not been well studied. Though smaller studies demonstrate the value of rGS, they also highlight the burden of results interpretation. We analyze genetic testing in CHD at two time-points, in 2018 and 2022–2023, across a change in clinical testing guidelines from chromosome microarray (CMA) to rGS. Analysis of 421 hospitalized infants with CHD demonstrated consistent genetic testing across time. Overall, after incorporation of rGS in 2022–2023, the diagnostic yield was 6.8% higher compared to 2018, and this pattern was consistent across all patient subtypes analyzed. In 2018, CMA was the most common test performed, with diagnostic results for CHD in 14.3%, while in 2022–2023, rGS was the most frequent test performed, with results diagnostic for CHD in 16.9%. Additionally, rGS identified 44% more unique genetic diagnoses than CMA. This is the largest study to highlight the value of rGS in CHD and has important implications for management.


INTRODUCTION
2][3][4][5][6][7] The etiology of CHD is often genetic. 8Multiple studies have demonstrated diagnostic utility of genetic testing in patients with CHD, 9- 16 including a recent multicenter cohort analysis of newborns with critical CHD, which con rmed abnormal genetic testing in up to 44% of newborns with CHD. 17 In 2008, the American Heart Association and the American Academy of Pediatrics recommended genetic testing for patients with CHD and highlighted some of the reasons to obtain a genetic diagnosis. 18One of these reasons is that syndromic features are di cult to identify in newborns, therefore genetic testing is often necessary to identify syndromes. 18Moreover, genetic testing can identify comorbidity in other organ systems that would bene t from increased screening and identify risk for developmental delay that would bene t from early intervention services.Furthermore, genetic testing may identify risks in family members.Finally, genetic testing increasingly informs prognosis and can guide surgical care and medical management in CHD patients. 19thin the past decade, chromosome microarray (CMA) emerged as the preferred genetic testing approach in patients with CHD. 17 This shift in testing strategy to newer testing modalities, like CMA, can increase genetic diagnoses in newborns with CHD. 20ditionally, studies demonstrate CMA can be utilized to identify genotype-phenotype associations in CHD. 21However, recently the genetic testing landscape has seen a shift to newer modalities including genome sequencing.
Genome sequencing offers the bene t of identifying structural variation detected by CMA as well as changes at the nucleotide level in a single test. 22As costs have decreased and laboratory work ow capacity increased, genomic sequencing has become feasible as a rst-line clinical test. 22Over 20 recent studies have shown that genome sequencing is useful in critically ill newborns, with higher diagnostic yield and increased clinical utility. 23Genome sequencing generates more actionable results and can even reduce the cost of care. 23,24While there are limited studies speci c to CHD patients, two recent reports utilized rapid genome sequencing (rGS) in CHD patients and showed a high diagnostic yield of 27% and 46%. 25,26 is important to note that while genome sequencing has proven valuable, there are limitations.One of the biggest challenges involves the burden of interpreting results.In the recent study "Genomic Medicine in Ill Neonates and Infants," genomic sequencing led to a high diagnostic yield; however, when up to 67% of results were discordant between labs due to technical differences, variant interpretation differences or both. 27These discrepancies highlight the need for further evaluation of comprehensive genome sequencing in CHD.
In 2010, the International Standard Cytogenomic Array Consortium released a consensus statement, recommending CMA analysis as rst-tier genetic test for most patients with congenital anomalies, including CHD. 28 In 2021, the American College of Medical Genetics and Genomics updated genetic testing recommendations for newborns with congenital anomalies, incorporating consideration of exome and genome sequencing as rst-tier testing modalities. 29Consistent with these international guidelines, in 2014, our center implemented clinical guidelines for universal CMA in all hospitalized infants with CHD.In 2019, we changed the protocol to incorporate molecular-based testing, and in July 2022, we changed our guidelines to universal rGS for all infants with CHD.Here, we analyze genetic testing rate and yield in 2018 and again in 2022-2023, across a shift from CMA to rGS.We show that rGS has diagnostic utility across a range of patient subtypes, comparable to CMA.These ndings have important implications for patients with the most common birth defect, CHD.

Study Population
We conducted a retrospective analysis of genetic testing practices and results at Riley Hospital for Children at Indiana University Health.To comprehensively identify patients with critical CHD, we utilized the Society of Thoracic Surgeons (STS) National Database.We included children who had surgical repair for CHD at 14 months old or younger between January 1, 2018 and December 31, 2018, as previously described, 20 and between July 1, 2022 and June 30, 2023.Indiana University Institutional Review Board approved the study (protocol #1408953015) and all research was performed in accordance with relevant ethical guidelines of the 1975 Declaration of Helsinki and all regulations as re ected in approval by the Institutional Review Board.A waiver of informed consent was approved by the Indiana University Institutional Review Board for this minimal risk study.The data was entered into the database deidenti ed, except for the date of birth, which was used for data integrity.Only summary results are presented within the research article and all data presented within the article is de-identi ed.

Data Collection
The patient's medical records were thoroughly reviewed, and all the study data were entered into a Research Electronic Data Capture (REDCap) database hosted at Indiana University.Additionally, data were obtained directly from the STS National Database.Variables included date of birth, surgery date, sex, race, family history of CHD, presence of other major congenital malformations, history of intrauterine growth restriction (IUGR) or being born small for gestational age (SGA), CHD group, infant of diabetic mother, maternal teratogens (alcohol, illegal drugs, tobacco, and prescription opioids), maternal infection, multiple gestation, gestational age, birth weight, number of cardiac surgeries, deceased status, and extracorporeal membrane oxygenation (ECMO).The medical record was reviewed to identify documentation of additional congenital anomalies, which were further classi ed by location, including the brain, ear, nose or throat, eye, chest, lung, diaphragm, congenital diaphragmatic hernia, gastrointestinal and abdominal wall, kidney, spleen, pancreas/annular pancreas, liver and gallbladder, anus, imperforate anus or anal atresia, genitourinary, ribs and vertebrae, limbs, lymphatic, lymphatic dysplasia, skin, arteriovenous malformation, and umbilical and categories with too few patients was grouped as "other" during statistical analysis.Participants with at least one congenital anomaly, in addition to CHD, were de ned as having multiple congenital anomalies (MCA) to distinguish them from those with isolated CHD.

Genetic Testing Review
All the available genetic testing results were collected from the medical record, including prenatal cell-free DNA screening (cfDNA), chromosome analysis (karyotype), CMA, uorescent in situ hybridization analysis (FISH), single-gene sequencing, gene panel sequencing, exome sequencing (ES) and rGS.The clinical testing laboratory report was used for result interpretation.Testing results from the clinical laboratory report were evaluated as "normal," "pathogenic," "likely pathogenic," "variant of unknown signi cance," "likely benign," or "not reported/unknown."All genetic testing results were independently reviewed by both a genetic counselor and medical geneticist with expertise in cardiovascular genetics and classi ed as "diagnostic for CHD" versus "not diagnostic for CHD." Discrepancies were reviewed and resolved by a second medical geneticist with expertise in cardiovascular genetics.

Statistical Analysis
Univariate comparisons of patients' characteristics, testing rate, and testing yield were analyzed using Pearson's Chi-square test (if all expected cell counts in contingency table were ≥ 5) or Fisher's exact test (if any expected cell counts were < 5) for categorical variables, and one-way analysis of variance for continuous variables.All analyses were conducted in R version 4.0.2(The R Foundation for Statistical Computing, Vienna, Austria).

Characteristics of the Cohort
The study included 421 CHD patients, among which 308 had isolated CHD and 113 had CHD + MCA (Table 1, Supplemental Table 1).More than half of the patients (57%) were male, and the majority (81.7%) were white.Among the total cohort, 78.4% were born full term and 13.5% were IUGR or SGA.Additionally, 13.1% of the patients had mothers with diabetes, 11.6% were exposed to maternal teratogens (such as alcohol, tobacco, or illicit drugs), and 24.0% were exposed to maternal infection.In the cohort, patients required an average of 1.6 cardiac surgeries, and 11.6% of the total cohort needed ECMO.At the time of data collection, 8.8% of the patients were deceased (Table 1).
In 2014, our institution implemented guidelines that recommended CMA for most infants hospitalized with CHD; in 2022, we updated guidelines to recommend rGS as a rst-tier testing modality.We analyzed a time period in 2018, where CMA was recommended, and 2022-2023, where rGS was the rst-tier testing modality.There were no signi cant differences between the two time periods by characteristics of: sex, race, maternal diabetes, maternal infection, fetal growth restriction, ECMO, or vitality status.However, there was a signi cantly higher rate of family CHD history, speci cally as ascertained by a genetics provider in 2022-2023, and a higher frequency of congenital anomalies identi ed in 2022-2023, possibly both re ecting the higher rate of medical genetics involvement in 2022-2023 (Tables 1 and 2).There was a smaller number of preterm births and a slightly larger number of late preterm births in 2022-2023, compared to 2018.Overall, the patient characteristics were similar between the 2018 and 2022-2023 cohort.2).

Genetic Testing Results
Next, we analyzed genetic testing results and initially segregated the results into two categories, normal versus abnormal.We referred all abnormal results to an expert in cardiovascular genetics for further categorization as diagnostic for CHD (Table 2).Out of the 327 total patients who underwent genetic testing, 88 patients (26.9%) had diagnostic results (Table 2).Among the 226 tested patients with isolated CHD, 51 patients (22.6%) had diagnostic results while a higher number of patients with CHD+MCA had diagnostic results, including 37 out of 101 (36.6%) tested patients (Table 2).
The diagnostic yield was 6.8% Overall, the diagnostic yield was 6.8% higher in 2022-2023 compared to 2018, after the incorporation of rGS, and this pattern was consistent across patient subtypes analyzed.

Medical Genetic Involvement
In 2014, new guidelines were implemented at our site that recommended medical genetics evaluation for all patients with CHD to assist in genetic testing.Out of 421 patients, medical genetics was involved with 288 patients (68.4%) (Table 2).The frequency of medical genetics involvement increased between the two study periods, with 121 of 190 patients (63.7%) being evaluated in 2018 compared to 167 of 231 (72.3%) in 2022-2023.However, at 5% level this was not statistically signi cant (p=0.074)(Table 2).
Additionally, the reported family history of CHD also increased from 2018 to 2022-2023.This could be due to an increase in medical history taking, which may be a result of the rise in medical genetics involvement.4).
We observed a signi cant change in the testing rate for multiple testing modalities between 2018 and 2022-2023; however, only chromosome analysis and cfDNA demonstrated a signi cant change in testing yield.
In 2018, 10 of 23 chromosome analysis results were diagnostic (43.5%), while in 2022-2023, there were 25 of 32 diagnostic results (78.1%), representing a signi cant increase (p=0.019)(Table 4).Targeted chromosome analysis was used in suspected aneuploidies, and indeed, all abnormal chromosome analysis results in the cohort represent a diagnosis of Trisomy 21, aside from two Turner Syndrome diagnoses and one larger deletion.The signi cant increase in yield in 2022-2023 likely re ects a bettertargeted use of chromosome analysis for suspected aneuploidies in the rGS era.
ES showed a trend toward increased use, with a consistent yield, although the numbers of patients were small.In 2018, 6 patients were tested, and diagnostic ndings were observed in 2 (33%).In 2022-2023,18 patients were tested, and diagnostic ndings were observed in 7 (38.9%).
Overall, the results demonstrate that the incorporation of rGS increased the diagnostic yield of genetic testing (6.8% increase in diagnostic yield in 2022-2023 )(Table 2), and rGS had a higher diagnostic yield when used as a rst-line modality in CHD (yield of CMA 14.3% in 2018 and yield of rGS 16.9% in 2022-2023)(Table 4).*The term "yield" for cfDNA refers to a positive screening result, which generally requires some con rmatory genetic test via prenatal amniocentesis or postnatally; therefore, it only represents abnormal results, whereas other modalities were reviewed and determined by an expert and determined diagnostic for CHD.

CMA and rGS Detailed Results
to the ndings, the inclusion of rGS improved the yield of genetic testing, resulting in a 6.8% increase in diagnostic yield between 2018 and 2022-2023 (Table 2).Additionally, when used as the rst line of testing, rGS exhibited a higher diagnostic yield at the patient level, (16.9% yield of rGS in 2022-2023 vs 14.3% for CMA in 2018 (Table 4)).However, each patient often had multiple abnormal ndings reported.Therefore, next, we analyzed CMA and rGS at the level of testing results breakdown when both time period results were grouped (Supplemental Table 2) (in contrast to the results per patient in Table 4.)Among the CMA results, 62.6% were negative or normal, while in the case of rGS, this gure was 74.1%.15.9% of CMA results were pathogenic variants compared to 12.0% of rGS results.Likely pathogenic results represented 2.7% of CMA results and 1.3% of rGS results.
Variants of uncertain signi cance made up 14.8% of CMA results and 12% of rGS results.Overall, the results breakdown between CMA and rGS were mostly consistent.However, there was a slightly increased proportion of abnormal results in CMA (Supplemental Table 2); in the setting of increased diagnostic results of rGS (Tables 2-4), this indicates CMA may have a higher number of nondiagnostic ndings compared to rGS.
All diagnostic genetic testing results from the two most common modalities, CMA and rGS are provided (Table 5).In total, more than 20 unique genetic ndings were identi ed through both CMA and rGS, with rGS detecting 44% more unique genetic diagnoses compared to CMA (13 versus 9) (Table 5).

Discussion
This study is the largest to date evaluating a clinical program for rGS in infants with CHD.Our ndings demonstrate that rGS is a useful tool in infants with CHD, comparable or slightly improved when compared to modalities such as CMA, with a higher diagnostic yield and a more diverse spectrum of genetic diagnoses identi ed.The results of this study have signi cant implications for the management of infants with CHD, the most common birth defect.
Our study demonstrates the bene t of genetic testing in patients with CHD, across multiple patient subtypes and genetic testing modalities.The diagnostic yield of genetic testing across both time periods analyzed was found to be 26.9%(Table 2), which is consistent with the genetic contribution to CHD reported in epidemiological studies and similar to other studies of genetic testing in infants with CHD, which report yields 25%-50%, depending on ascertainment protocols, often re ecting clinical program and referral sources at each institution. 11,13,15Our yield is on the lower end because it represents only results con rmed clearly diagnostic for CHD after a secondary review.It is possible some of the excluded diagnoses will ultimately be found to be causative of CHD as well.Our study period represents an early phase of adoption of rGS testing, and it is possible that with longer rGS usage, we may see the diagnostic yield increase.Nonetheless, genetic testing showed high yields across almost every patient subtype, including both isolated CHD (22.6%) and patients with CHD+MCA (36.6%).
Our the potential to identify increased genetic diagnoses and impact patient care. 11,20 this study, we found that rGS is as useful as or slightly better than CMA in detecting genetic abnormalities in infants with CHD.
After the incorporation of rGS in 2022-2023, the overall diagnostic yield of genetic testing increased by 6.8% (Table 2).In directly comparing modalities, CMA had a diagnostic rate of 14.3% for CHD when used as a rst-tier diagnostic tool in 2018, whereas rGS had a diagnostic rate of 16.9% for CHD in 2022-2023 (Table 4).Furthermore, rGS identi ed 44% more unique genetic diagnoses when compared to CMA (Table 5).
These results are consistent with recent studies of genome sequencing in critically ill infants.One study found that in patients referred for genetics evaluation, genome sequencing had a yield of 41% in CHD, compared to 24% for CMA. 30Another study found that when compared directly to CMA or CMA with a gene panel, genome sequencing had a higher yield (8% and 13% versus 34%). 31There have also been recent reports of rGS speci cally in CHD.One study of a small, selected cohort of 24 CHD patients found that rGS improved patient care and reduced costs, with a diagnostic yield of up to 46%. 32Another study of a slightly larger population of 48 patients had a diagnostic yield of 27%. 25 Although diagnostic yield can be a awed metric since it is dependent on patient ascertainment and can be elevated in highly selected cohorts, there is clear value in rGS in CHD patients.Our approach to hospitalized infants with CHD is to test comprehensively and 77.7% underwent genetic testing.Our data from 2022-2023 captures our change in protocol to rGS and our cohort of 90 patients tested con rms that rGS is a useful diagnostic tool for infants with CHD.
Genome sequencing has immense potential as it provides a broader capture of genetic anomalies.Studies have shown that the laboratory work ow for genome sequencing can be simpler, with improving cost and diagnostic rates, making it a rst-tier genetic test. 24However, there are limitations to rGS, as previous studies have noted major discrepancies in variant interpretation that can impact its utilization. 27Nevertheless, our data showed that rGS may have had fewer nondiagnostic ndings, including fewer VUS compared to CMA (12% versus 14.8%, Supplemental Table 2.) Our data also revealed an increased rate of medical genetics involvement in 2022-2023, accompanying the shift to rGS, which may have helped overcome this limitation.
The increased utilization of genetics providers in the 2022-2023 cohort likely aided the interpretation of results.There were other differences between our 2018 and 2022-2023 cohorts, attributable to increased medical genetic involvement in 2022-2023, including a signi cantly higher rate of family CHD history ascertained by a genetics provider and a higher frequency of congenital anomalies identi ed in 2022-2023 (Table 1).Our institutional guidelines recommend consultation with medical genetics in all CHD patients; however, medical genetics involvement was not universal, occurring in 68.4% (Table 2).With the increasing complexity of genetic testing results and the changing testing recommendations, the involvement of medical geneticists is becoming increasingly important.Medical genetics can help with testing strategies and interpreting results, and multiple studies have shown that medical genetics can increase testing utilization and yield, as well as improve care. 11Although there is a shortage of genetics expertise in the workforce, we expect telehealth to provide an opportunity for increased involvement in the future.Additionally, we expect that the identi cation of emerging genes attributable to CHD and advanced technology in variant interpretation will likely continue to improve the diagnostic yield of genetic testing.
Our protocol recommended universal genetic testing in patients with CHD; however, in our cohort, the genetic testing rate was 77.7%, indicating screening was not universal and there was a lack of strict compliance with the recommended protocol.Single center studies indicate genetic testing is underutilized and highly variable [9][10][11][12][13][14][15][16] and our recent multicenter evaluation of newborns with CHD indicated genetic testing rates of 42-78% across 5 centers. 17Identifying the genetic cause of CHD has been shown to improve patient care by providing information on prognosis, identifying other organ system involvement, and allowing for the implementation of health supervision and early intervention.Moreover, identifying a genetic origin enables the assessment of recurrence risk for patients as well as the risk to other family members. 10,18,33,34Therefore, quality improvement efforts are necessary to ensure that guidelines for genetic testing are being followed.
We should note that our study has some limitations.First, we only used genetic testing results that were available within the electronic medical record, which means that we did not include any testing that may have been carried out at outside hospitals or outside the period of data collection.Moreover, our study only includes patients who underwent surgery soon after birth and, therefore, excludes less critically ill infants.On the other hand, it may not be representative of the most critically ill infants who died before surgery and who are at high risk of genetic diagnoses.To improve our understanding of this issue, future studies should aim to increase sample sizes and involve multiple institutions.
Although our study reports the diagnostic yield, it does not measure clinical utility.Clinical utility refers to the ability of genome sequencing to change medical or surgical management based on its results, and this level of detail is not captured in our data.
Studies have shown the clinical usefulness of genome sequencing in critically ill infants from small populations. 35Future studies should expand these efforts.

Conclusions
Overall, this is the largest study highlighting the use of rGS in infants with CHD.We also demonstrate that overall, genetic testing is useful in infants with CHD, across a range of patient subtypes.We show that rGS has utility comparable or slightly improved to older modalities like CMA in infants with CHD.The ndings have important implications for the evaluation of CHD patients. Declarations

Table 2
Genetic testing rate, diagnostic yield and medical genetics involvement was a signi cant shift in use from CMA in 2018 to rGS in 2022-2023.CMA was the most common modality in 2018, used in 126 patients, and 18 patients (14.3%) had diagnostic results.The most frequent test performed in 2023-2023 was rGS, performed in 89 patients, with diagnostic results in 15 (16.9%) (Table As described previously, a medical geneticist with expertise in cardiovascular genetics analyzed the results of all genetic testing and determined whether each nding was diagnostic for CHD, and the overall diagnostic yield was 26.9% (Table2), whereas the yield of the individual testing modalities ranged from 6.4% to 63.6% (Table4).The most commonly performed tests were CMA and rGS.CMA had diagnostic results in 28 of 167 patients (16.8%), and rGS had diagnostic results in 15 of 90 (16.7%), overall.As described, there

Table 4
Testing yield by genetic testing modality in 2018 and 2022-2023
study examines two distinct time periods, in 2018 and 2022-2023.At our institution, we updated testing guidelines to recommend rGS in mid-2022.As a result, there was a signi cant shift from CMA being the most common test in 2018 to rGS being the most common test in 2022-2023 (Table3).Our institution follows guidelines for testing patients with congenital anomalies, which have evolved over time.Our guideline for CMA was implemented based on ISCA guidelines in 2014 and was in place from 2014 until 2018.It was updated in December of 2019 to an exome-based gene panel plus CMA.Based on the American College of Medical Genetics' 2021 guidelines for exome and genome sequencing in newborns with birth defects, we updated our guidelines in 2021 to Exome Sequencing (ES), and nally, in 2022 to rGS.The time period of 2022-2023 re ects a relatively new recommendation and a diverse patient population, which may represent protocol compliance at various stages.Nevertheless, rGS is the most commonly used test during this period, indicating a good uptake of the new protocol.Genetic testing protocols have shown value in the CHD population, including reducing costs and increasing genetic diagnoses.In a separate study, we demonstrated that genetic testing guidelines increased testing rates and shifted testing towards newer modalities, with