Feasibility of Ultra-Rapid Exome Sequencing in Critically Ill Infants and Children With Suspected Monogenic Conditions in the Australian Public Health Care System | Critical Care Medicine | JAMA | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Steps for Non-Rapid Genomic Testing Compared With Ultra-Rapid Genomic Testing
Steps for Non-Rapid Genomic Testing Compared With Ultra-Rapid Genomic Testing
Figure 2.  Recruitment and Summary of Ultra-Rapid Exome Sequencing Cohort
Recruitment and Summary of Ultra-Rapid Exome Sequencing Cohort

aOccurred 3 months after the ultra-rapid exome sequencing report and was considered by clinicians to have been influenced by the genomic report.

bSome patients had changes in management from more than 1 category.

Table 1.  Comparison of Multisite Ultra-Rapid Exome Sequencing Cohort vs Rapid Exome Sequencing Cohort From Prior Study8
Comparison of Multisite Ultra-Rapid Exome Sequencing Cohort vs Rapid Exome Sequencing Cohort From Prior Study
Table 2.  Presenting Clinical Features, Ultra-Rapid Exome Sequencing Results, and Subsequent Clinical Course in a Representative Selection of Patients
Presenting Clinical Features, Ultra-Rapid Exome Sequencing Results, and Subsequent Clinical Course in a Representative Selection of Patients
1.
Farnaes  L, Hildreth  A, Sweeney  NM,  et al.  Rapid whole-genome sequencing decreases infant morbidity and cost of hospitalization.   NPJ Genom Med. 2018;3:10. doi:10.1038/s41525-018-0049-4PubMedGoogle ScholarCrossref
2.
French  CE, Delon  I, Dolling  H,  et al; NIHR BioResource—Rare Disease; Next Generation Children Project.  Whole genome sequencing reveals that genetic conditions are frequent in intensively ill children.   Intensive Care Med. 2019;45(5):627-636. doi:10.1007/s00134-019-05552-xPubMedGoogle ScholarCrossref
3.
Gubbels  CS, VanNoy  GE, Madden  JA,  et al.  Prospective, phenotype-driven selection of critically ill neonates for rapid exome sequencing is associated with high diagnostic yield.   Genet Med. 2020;22(4):736-744. doi:10.1038/s41436-019-0708-6PubMedGoogle ScholarCrossref
4.
Kingsmore  SF, Cakici  JA, Clark  MM,  et al; RCIGM Investigators.  A randomized, controlled trial of the analytic and diagnostic performance of singleton and trio, rapid genome and exome sequencing in ill infants.   Am J Hum Genet. 2019;105(4):719-733. doi:10.1016/j.ajhg.2019.08.009PubMedGoogle ScholarCrossref
5.
Meng  L, Pammi  M, Saronwala  A,  et al.  Use of exome sequencing for infants in intensive care units: ascertainment of severe single-gene disorders and effect on medical management.   JAMA Pediatr. 2017;171(12):e173438. doi:10.1001/jamapediatrics.2017.3438PubMedGoogle Scholar
6.
Mestek-Boukhibar  L, Clement  E, Jones  WD,  et al.  Rapid paediatric sequencing (RaPS): comprehensive real-life workflow for rapid diagnosis of critically ill children.   J Med Genet. 2018;55(11):721-728. doi:10.1136/jmedgenet-2018-105396PubMedGoogle ScholarCrossref
7.
Sanford  EF, Clark  MM, Farnaes  L,  et al; RCIGM Investigators.  Rapid whole genome sequencing has clinical utility in children in the PICU.   Pediatr Crit Care Med. 2019;20(11):1007-1020. doi:10.1097/PCC.0000000000002056PubMedGoogle ScholarCrossref
8.
Stark  Z, Lunke  S, Brett  GR,  et al; Melbourne Genomics Health Alliance.  Meeting the challenges of implementing rapid genomic testing in acute pediatric care.   Genet Med. 2018;20(12):1554-1563. doi:10.1038/gim.2018.37PubMedGoogle ScholarCrossref
9.
Petrikin  JE, Cakici  JA, Clark  MM,  et al.  The NSIGHT1-randomized controlled trial: rapid whole-genome sequencing for accelerated etiologic diagnosis in critically ill infants.   NPJ Genom Med. 2018;3:6. doi:10.1038/s41525-018-0045-8PubMedGoogle ScholarCrossref
10.
Greenhalgh  T, Papoutsi  C.  Spreading and scaling up innovation and improvement.   BMJ. 2019;365:l2068. doi:10.1136/bmj.l2068PubMedGoogle ScholarCrossref
11.
Braithwaite  J, Churruca  K, Long  JC, Ellis  LA, Herkes  J.  When complexity science meets implementation science: a theoretical and empirical analysis of systems change.   BMC Med. 2018;16(1):63. doi:10.1186/s12916-018-1057-zPubMedGoogle ScholarCrossref
12.
Curran  GM, Bauer  M, Mittman  B, Pyne  JM, Stetler  C.  Effectiveness-implementation hybrid designs: combining elements of clinical effectiveness and implementation research to enhance public health impact.   Med Care. 2012;50(3):217-226. doi:10.1097/MLR.0b013e3182408812PubMedGoogle ScholarCrossref
13.
Damschroder  LJ, Aron  DC, Keith  RE, Kirsh  SR, Alexander  JA, Lowery  JC.  Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science.   Implement Sci. 2009;4:50. doi:10.1186/1748-5908-4-50PubMedGoogle ScholarCrossref
14.
Stark  Z, Nisselle  A, McClaren  B,  et al.  Attitudes of Australian health professionals towards rapid genomic testing in neonatal and paediatric intensive care.   Eur J Hum Genet. 2019;27(10):1493-1501. doi:10.1038/s41431-019-0429-yPubMedGoogle ScholarCrossref
15.
Harris  PA, Taylor  R, Thielke  R, Payne  J, Gonzalez  N, Conde  JG.  Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support.   J Biomed Inform. 2009;42(2):377-381. doi:10.1016/j.jbi.2008.08.010PubMedGoogle ScholarCrossref
16.
Richards  S, Aziz  N, Bale  S,  et al; ACMG Laboratory Quality Assurance Committee.  Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology.   Genet Med. 2015;17(5):405-424. doi:10.1038/gim.2015.30PubMedGoogle ScholarCrossref
17.
Akesson  LS, Eggers  S, Love  CJ,  et al.  Early diagnosis of Pearson syndrome in neonatal intensive care following rapid mitochondrial genome sequencing in tandem with exome sequencing.   Eur J Hum Genet. 2019;27(12):1821-1826. doi:10.1038/s41431-019-0477-3PubMedGoogle ScholarCrossref
18.
Sobreira  N, Schiettecatte  F, Valle  D, Hamosh  A.  GeneMatcher: a matching tool for connecting investigators with an interest in the same gene.   Hum Mutat. 2015;36(10):928-930. doi:10.1002/humu.22844PubMedGoogle ScholarCrossref
19.
Rehman  AU, Najafi  M, Kambouris  M,  et al.  Biallelic loss of function variants in PPP1R21 cause a neurodevelopmental syndrome with impaired endocytic function.   Hum Mutat. 2019;40(3):267-280.PubMedGoogle ScholarCrossref
20.
Cooper  MS, Stark  Z, Lunke  S, Zhao  T, Amor  DJ.  IREB2-associated neurodegeneration.   Brain. 2019;142(8):e40. doi:10.1093/brain/awz183PubMedGoogle Scholar
21.
Costain  G, Ghosh  MC, Maio  N,  et al.  Absence of iron-responsive element-binding protein 2 causes a novel neurodegenerative syndrome.   Brain. 2019;142(5):1195-1202. doi:10.1093/brain/awz072PubMedGoogle ScholarCrossref
22.
Richmond  CM, Campbell  S, Foo  HW,  et al.  Rapid identification of biallelic SPTB mutation in a neonate with severe congenital hemolytic anemia and liver failiure.   Mol Syndromol. 2020;11(1):50-55. doi:10.1159/000505886PubMedGoogle ScholarCrossref
23.
Clark  MM, Hildreth  A, Batalov  S,  et al.  Diagnosis of genetic diseases in seriously ill children by rapid whole-genome sequencing and automated phenotyping and interpretation.   Sci Transl Med. 2019;11(489):eaat6177. doi:10.1126/scitranslmed.aat6177PubMedGoogle Scholar
24.
Stark  Z, Boughtwood  T, Phillips  P,  et al.  Australian Genomics: a federated model for integrating genomics into healthcare.   Am J Hum Genet. 2019;105(1):7-14. doi:10.1016/j.ajhg.2019.06.003PubMedGoogle ScholarCrossref
25.
Friedman  JM, Bombard  Y, Cornel  MC,  et al; Paediatric Task Team of the Global Alliance for Genomics and Health Regulatory and Ethics Work Stream.  Genome-wide sequencing in acutely ill infants: genomic medicine’s critical application?   Genet Med. 2019;21(2):498-504. doi:10.1038/s41436-018-0055-zPubMedGoogle ScholarCrossref
26.
Grosse  SD, Farnaes  L.  Genomic sequencing in acutely ill infants: what will it take to demonstrate clinical value?   Genet Med. 2019;21(2):269-271. doi:10.1038/s41436-018-0124-3PubMedGoogle ScholarCrossref
Original Investigation
June 23/30, 2020

Feasibility of Ultra-Rapid Exome Sequencing in Critically Ill Infants and Children With Suspected Monogenic Conditions in the Australian Public Health Care System

Australian Genomics Health Alliance Acute Care Flagship
JAMA. 2020;323(24):2503-2511. doi:10.1001/jama.2020.7671
Key Points

Question  Can ultra-rapid genomic testing be performed for critically ill pediatric patients in a public health care system?

Findings  This multisite descriptive feasibility study included 12 Australian hospitals and 2 laboratories. Among the 108 critically ill infants and children with suspected monogenic conditions who had ultra-rapid genomic testing, the mean time to genomic test report was 3.3 days and the molecular diagnostic yield was 51%.

Meaning  This study suggests feasibility of ultra-rapid genomic testing in a public setting for critically ill pediatric patients. However, further research is needed to understand its clinical value.

Abstract

Importance  Widespread adoption of rapid genomic testing in pediatric critical care requires robust clinical and laboratory pathways that provide equitable and consistent service across health care systems.

Objective  To prospectively evaluate the performance of a multicenter network for ultra-rapid genomic diagnosis in a public health care system.

Design, Setting, and Participants  Descriptive feasibility study of critically ill pediatric patients with suspected monogenic conditions treated at 12 Australian hospitals between March 2018 and February 2019, with data collected to May 2019. A formal implementation strategy emphasizing communication and feedback, standardized processes, coordination, distributed leadership, and collective learning was used to facilitate adoption.

Exposures  Ultra-rapid exome sequencing.

Main Outcomes and Measures  The primary outcome was time from sample receipt to ultra-rapid exome sequencing report. The secondary outcomes were the molecular diagnostic yield, the change in clinical management after the ultra-rapid exome sequencing report, the time from hospital admission to the laboratory report, and the proportion of laboratory reports returned prior to death or hospital discharge.

Results  The study population included 108 patients with a median age of 28 days (range, 0 days to 17 years); 34% were female; and 57% were from neonatal intensive care units, 33% were from pediatric intensive care units, and 9% were from other hospital wards. The mean time from sample receipt to ultra-rapid exome sequencing report was 3.3 days (95% CI, 3.2-3.5 days) and the median time was 3 days (range, 2-7 days). The mean time from hospital admission to ultra-rapid exome sequencing report was 17.5 days (95% CI, 14.6-21.1 days) and 93 reports (86%) were issued prior to death or hospital discharge. A molecular diagnosis was established in 55 patients (51%). Eleven diagnoses (20%) resulted from using the following approaches to augment standard exome sequencing analysis: mitochondrial genome sequencing analysis, exome sequencing–based copy number analysis, use of international databases to identify novel gene–disease associations, and additional phenotyping and RNA analysis. In 42 of 55 patients (76%) with a molecular diagnosis and 6 of 53 patients (11%) without a molecular diagnosis, the ultra-rapid exome sequencing result was considered as having influenced clinical management. Targeted treatments were initiated in 12 patients (11%), treatment was redirected toward palliative care in 14 patients (13%), and surveillance for specific complications was initiated in 19 patients (18%).

Conclusions and Relevance  This study suggests feasibility of ultra-rapid genomic testing in critically ill pediatric patients with suspected monogenic conditions in the Australian public health care system. However, further research is needed to understand the clinical value of such testing, and the generalizability of the findings to other health care settings.

Introduction

Quiz Ref IDMore than 10 single-center studies including more than 600 critically ill neonatal and pediatric patients with suspected genetic conditions have demonstrated genomic testing has a high diagnostic yield and influences clinical management.1-9 In the critical care setting, genomic testing results need to be delivered rapidly to be useful and the mean turnaround times achieved by nascent rapid testing programs have ranged from 7 to 23 days. Although studies of rapid genomic testing have primarily included infants in neonatal intensive care units (NICUs),1,3-5,9 a minority of studies have included other critically ill children such as those in pediatric intensive care units (PICUs).2,6-8 These early results from multiple centers have created momentum for broader implementation of rapid genomic sequencing in critically ill neonatal and pediatric patients worldwide.

Translating the experience of academic single-center studies into implementation across health care systems requires the development of robust clinical and laboratory pathways to deliver rapid genomic testing reliably and consistently. Furthermore, the successful spread and scaling up of innovations in health care systems is complex, and necessitates both structure and adaptability to the local context, benefitting from a participatory culture, distributed leadership, transparent assessment of outcomes, collective learning, and adaptation.10 Developing purposeful implementation strategies that recognize the iterative, nonlinear, and multicomponent nature of change has the potential to streamline the process of scaling up rapid diagnostic programs across health care systems, preventing costly trial and error at multiple sites.11

The objective of this study was to examine the feasibility of providing ultra-rapid genomic testing for critically ill neonatal and pediatric patients with suspected monogenic conditions across multiple centers in Australia.

Methods

The Australian Genomics Acute Care study received human research ethics committee approval from Melbourne Health (HREC/16/MH251). Parents provided written informed consent for participation in the study.

Study Design and Participants

The study adopted a hybrid implementation-effectiveness study design, with simultaneous evaluation of processes and outcomes.12 An implementation strategy was developed based on prior experience with rapid exome sequencing (target results in <21 days) implementation at 2 Australian centers8 and using the Consolidated Framework for Implementation Research,13 which is a theory-informed set of constructs associated with effective implementation. Twelve participating hospitals and 2 laboratories formed a network aiming to deliver genomic results within 5 days. Professionals at the participating sites were surveyed to determine the preferred clinical service delivery model and implementation readiness.14

A multidisciplinary working group developed the patient selection criteria and the clinical and laboratory pathways. Clinical geneticists acted as formal implementation leaders at each site. Process communication was addressed through the development of standard operating procedures for patient recruitment, test ordering, sample labeling, tracking, and transportation. Structured electronic communication procedures were implemented to facilitate the timely exchange of information between the clinical and laboratory teams. The ultra-rapid exome sequencing report and the turnaround time for each case were distributed to the study leadership group, and the summary statistics were provided bimonthly to the entire working group. Components of the implementation strategy were mapped to the Consolidated Framework for Implementation Research13 (eTable 1 in the Supplement).

Recruitment occurred prospectively between March 2018 and February 2019, with data collected to May 2019. The 12 tertiary hospitals included 5 women’s hospitals, 3 women’s and children’s hospitals, and 4 children’s hospitals. Of the latter, 2 were quaternary hospitals caring for complex patients from throughout Australia and overseas. The total number of NICU beds was 451 and the total number of PICU beds was 135. Quiz Ref IDPatients were eligible if admitted to a participating NICU or PICU and had been referred to the clinical genetics service for a suspected monogenic condition. Other hospitalized patients were considered for inclusion if a rapid result was likely to alter clinical management (eg, organ transplant decisions). Patients were ineligible if (1) a monogenic etiology was considered unlikely (eg, isolated congenital diaphragmatic hernia); (2) a secure clinical diagnosis (such as Apert syndrome) was made; or (3) death or hospital discharge were imminent.

An expert panel of study investigators (S.L., M.W., C.P., C.P.B., J.P., M.F.B., T.R., J.C., and Z.S.) discussed referrals electronically to determine study eligibility. Other clinicians participated on an as-needed basis (S.A.S., D.M., E.P.K., K.B.H., and S.M.W.). When panel members disagreed about eligibility, recruitment proceeded with majority approval. A core clinical data set was collected at recruitment (including Human Phenotype Ontology terms) and was managed using REDCap, which is a secure, web-based application designed to support data capture for research studies.15 Chromosomal microarray was performed prior to enrollment if the likelihood of a chromosomal condition was considered high by the referring teams (eg, in patients with multiple congenital abnormalities) and concurrently with ultra-rapid exome sequencing when the pretest probability was considered low. If a mitochondrial condition was suspected, rapid mitochondrial genome sequencing was performed concurrently with ultra-rapid exome sequencing after obtaining specific approval from the panel. The steps for the diagnostic ultra-rapid exome sequencing pathway appear in Figure 1 and are compared with non-rapid testing.

Exome Sequencing, Data Analysis, and Interpretation

Ultra-rapid exome sequencing was performed in both the parents and the child (trio) when possible at 2 laboratories accredited by the National Association of Testing Authorities. The sequencing and bioinformatics analyses are described in eMethods 1 in the Supplement. Variants relevant to patient phenotype were classified based on the standards of the American College of Medical Genetics and Genomics.16 All pathogenic and likely pathogenic variants identified in this study have been deposited in ClinVar (submission IDs: SUB6379721, SCV000929966, and SCV000929967; SUB7027998 and SUB7195751 are still under review).

Outcome Measures

The primary outcome was time from sample receipt to ultra-rapid exome sequencing report. The secondary outcomes were the molecular diagnostic yield, the change in clinical management after the ultra-rapid exome sequencing report, the time from hospital admission to the laboratory report, and the proportion of laboratory reports returned prior to death or hospital discharge. The time from hospital admission to laboratory report was measured in calendar days and included the following time points: referral to clinical genetics service, clinical genetics assessment, patient proposal for ultra-rapid exome sequencing, approval by the expert panel, consent, transportation, and laboratory report issue.

The changes in clinical management were collected from the medical records by the clinical geneticist site leaders (M.W., C.P., C.P.B., J.P., S.A.S., E.I.K., M.E., A.V., and Z.S.) 3 months after the laboratory report using a structured data collection instrument (eMethods 2 in the Supplement). Quiz Ref IDThese changes in clinical management were grouped into 3 main categories: targeted treatments, redirection of treatment toward palliative care, and targeted surveillance (investigations and subspecialist referrals aimed at known complications). Referring clinical geneticists subjectively rated the usefulness of ultra-rapid exome sequencing using a 5-point scale ranging from “very useful” to “not useful at all.”

Statistical Analysis

Descriptive statistics were used for participant characteristics, referral source, the duration of the clinical and laboratory components of the diagnostic trajectory, the molecular diagnostic yield, and the changes in clinical management after the ultra-rapid exome sequencing report. The 95% CIs for the means of the nonnormally distributed quantitative results (the normality of the data was assessed using the Shapiro-Wilk test) were calculated using bias-corrected, accelerated bootstrapping from the boot package version 1.3-24 in R version 3.6.2 (R Foundation for Statistical Computing) with a 0.95 confidence limit and 10 000 iterations. No patients were lost to follow-up and all data collection was complete.

Results
Participant Demographics and Indications for Testing

The program considered 131 patients over 11 months. Of these 131 patients, the expert panel excluded 9 (7%) because they did not meet the study criteria, and 14 (11%) were approved and subsequently withdrawn by the referring teams (Figure 2). Of the 108 patients approved for ultra-rapid exome sequencing, panel members disagreed regarding the likelihood of a monogenic condition in 15 patients (14%).

The median age of participants was 28 days (range, 0 days-17 years); 34% were female; 62 patients (57%) were admitted to NICUs, 36 patients (33%) were admitted to PICUs, and 10 patients (9%) were from other hospital wards. The most common clinical indication for ultra-rapid exome sequencing was neurological signs and symptoms (33%) such as hypotonia or seizures (Table 1). Seventy-five patients (69%) originated from 3 pediatric hospitals. In 105 cases (97%), patients were analyzed together with both parents (trio); in 3 cases, patients were analyzed together with only 1 parent (duo) due to the unavailability of the other parent.

Mitochondrial genome testing occurred concurrently with ultra-rapid exome sequencing in 7 patients (6%). The mean number of clinical geneticist inpatient consultations was 2.8 (range, 1-7 consultations). Genetic counselors were involved with 95 families (88%), and the mean number of consultations was 2.0 (range, 1-6 consultations).

Patient characteristics, the indications for testing, the molecular diagnostic yield, the changes in clinical management after the ultra-rapid exome sequencing report, and the time from hospital admission to the ultra-rapid exome sequencing report appear in Table 1 and are compared against the previously published Australian rapid exome sequencing study,8 which aimed to deliver results within 21 days at 2 pediatric hospitals.

Primary End Point: Time From Sample Receipt to Ultra-Rapid Exome Sequencing Report

One hundred and two reports (94%) were issued within the performance target of 5 calendar days. The mean time from sample receipt to ultra-rapid exome sequencing report was 3.3 days (95% CI, 3.2-3.5 days) and the median time was 3 days (range, 2-7 days). The mean time from hospital admission to ultra-rapid exome sequencing report was 17.5 days (95% CI, 14.6-21.1 days). The longest component was time from hospital admission to clinical genetics referral, which had a mean time of 8.1 days (95% CI, 6.0-10.8 days). The mean sample transportation time was 1.6 days (95% CI, 1.4-1.9 days) despite the furthest hospital being 1600 km away from the laboratory site. Laboratory 1 performed 94 tests and laboratory 2 performed 14 tests. Data on time from hospital admission to ultra-rapid exome sequencing report, broken down per patient and per referral site, appear in eTable 2 in the Supplement.

Secondary End Points
Molecular Diagnostic Yield

A total of 56 molecular diagnoses were made for 55 patients (51%) (Table 1 and Table 2). Routine ultra-rapid exome sequencing trio analyses resulted in 45 molecular diagnoses, with an additional 11 molecular diagnoses made through extended approaches. Of these 11 molecular diagnoses, 7 were made as part of the original analysis, including 3 through hypothesis-driven identification of copy number variants affecting single genes from the exome sequencing data (10-kb homozygous deletion of NEU1 causing sialidosis type 1, 12-kb deletion in ABCC6 in trans with pathogenic variant causing arterial calcification of infancy, and 12-kb hemizygous deletion in ATP7A causing Menkes disease). Two molecular diagnoses of Pearson syndrome due to mitochondrial deletions were achieved through concurrent mitochondrial genome sequencing.17

Two further molecular diagnoses were made by submitting novel gene candidates to GeneMatcher18 (an international data sharing platform) and matching to multiple other cases in real time, enabling establishment of novel gene–disease associations and diagnostic reporting.19 Four diagnoses were initially reported as variants of uncertain significance but upgraded with additional data, including publication of another novel gene, IREB2, 6 months later.20,21 Rapid RNA studies confirmed the pathogenicity of a homozygous splice-site ASNS variant in 10 days, and 2 other variants of uncertain significance were reclassified as likely pathogenic after further phenotyping during a period of 6 months.22 One patient with pontocerebellar hypoplasia was found to have an unbalanced translocation on concurrent chromosomal microarray, and was not included in the ultra-rapid exome sequencing diagnostic yield.

Among 35 of the 62 patients enrolled from the NICU, the molecular diagnostic yield was 56.0% (95% CI, 41.9%-66.1%). Among 17 of the 36 patients enrolled from the PICU, the molecular diagnostic yield was 47.0% (95% CI, 27.8%-66.1%). Among 3 of the 10 patients enrolled from other hospital wards, the molecular diagnostic yield was 30% (95% CI, 0%-50%). Among 4 of the 15 patients about whom there was disagreement among the expert panel members, the molecular diagnostic yield was 27.0% (95% CI, 6.7%-46.7%). Among 51 of the 93 patients approved by all the expert panel members, the molecular diagnostic yield was 57.0% (95% CI, 43.0%-63.4%).

Changes in Clinical Management After the Ultra-Rapid Exome Sequencing Report and Perceived Utility

After the ultra-rapid exome sequencing report, changes in clinical management were identified by clinicians for 48 of 108 patients (44%). Among the 55 patients receiving a molecular diagnosis, 42 (76%) were considered as having had a subsequent change in clinical management, including targeted surveillance to identify complications of the condition in 19 patients (18%). The ultra-rapid exome sequencing result formed part of the palliative care discussions in 14 patients (13%). The molecular diagnosis was considered to have contributed to targeted treatment in 12 patients (11%).

Among the 53 patients who did not receive a molecular diagnosis through ultra-rapid exome sequencing, 6 (11%) were considered as having had a change in clinical management (eg, cancellation of planned tissue biopsies and cessation of medications) as a result of nongenetic diagnoses being thought to be more likely. Changes in clinical management were considered as having occurred in 35 of 62 patients (56%) from the NICU and in 17 of 36 patients (47%) from the PICU. Illustrative examples appear in Table 2 and the full data appear in eTable 3 in the Supplement.

Ultra-rapid sequencing reports establishing a molecular diagnosis were perceived as “very useful” or “useful” by referring clinical geneticists in 52 of 55 cases (95%) and were perceived as “neutral” in 3 of 55 cases (5%). Ultra-rapid exome sequencing reports that did not establish a molecular diagnosis were perceived as “very useful” or “useful” in 31 of 53 cases (58%), were perceived as “neutral” in 20 of 53 cases (38%), and were perceived as “not useful” in 2 of 53 cases (4%).

Discussion

Quiz Ref IDThis prospective study of 108 critically ill infants and children from multiple sites across Australia demonstrates the feasibility of a national, highly coordinated clinical and laboratory ultra-rapid genomic diagnosis program in a public health care system, with a 3-fold reduction in the mean time from hospital admission to molecular diagnosis compared with the prior Australian rapid exome sequencing study8 conducted in 2016-2017 at 2 centers. A greater proportion of reports were returned prior to hospital discharge or death and the reports formed part of the palliative care decisions in a greater proportion of the patients who died. The ultra-rapid exome sequencing report was perceived as contributing to the clinical management of the patients regardless of whether a molecular diagnosis was made.

Quiz Ref IDTwenty percent of the molecular diagnoses were made using approaches that extend usual clinical exome sequencing diagnostic practice, including mitochondrial genome sequencing, identification of single-gene copy number variants from exome sequencing data, real-time use of international databases to establish novel gene–disease associations, RNA studies, and additional phenotyping. There is a need to integrate these ancillary approaches into the diagnostic process, and to develop pathways for rapid turnaround times. Genome sequencing is likely to be the preferred modality for rapid diagnostics in the future4 because in addition to shortening processing times, it confers the capacity to simultaneously evaluate mitochondrial DNA, structural genomic variants, and repeat expansions. Consolidation of diagnostic approaches to multiple variant types into a single assay may shorten time to test initiation and thus overall time to molecular diagnosis.

A partially centralized model of laboratory service delivery was adopted during this feasibility study, with one laboratory processing the majority of samples and providing redundancy for a second laboratory. The 2 laboratories together were able to provide consistent service based on clinical demand, with 94% of tests reported before the target time of 5 calendar days. The logistical challenges in coordinating an ultra-rapid diagnosis program involving multiple clinical and laboratory sites were considerable, and specifically designed procedures for sample identification, shipping, and tracking were invaluable in preventing delays.

The lack of health care funding, the underpreparedness of the pediatric workforce, and the resource-intensive nature of rapid testing represent significant barriers to widespread adoption.8,14,23 The potential role for automated phenotyping and interpretation in the scaling up of whole genome sequencing was recently highlighted in a largely retrospective cohort23 with a mean time saving of 22 hours. However, rapid genomic testing challenges all the components of the clinical and laboratory pathway, necessitating timely patient identification, skilled assessment, appropriate counseling, and a high degree of communication and coordination in emotionally charged and rapidly evolving clinical situations.8

Health care is a complex adaptive system and the successful adoption of innovations such as rapid genomic testing relies not only on the scalability of the laboratory testing, but on the multidisciplinary teams of intensive care physicians, medical subspecialists, clinical geneticists, genetic counselors, laboratory scientists, and bioinformaticians simultaneously changing practices across multiple sites to provide consistent service. The creation of a collaborative learning community was used to accelerate implementation and promote a unified approach across different professional groups and multiple sites in Australia through shared decision-making, regular feedback regarding process and molecular diagnostic outcomes, opportunities for reflection, and the sharing of expertise. The creation of the network was greatly facilitated by the Australian Genomics Health Alliance, a government-funded national genomic medicine initiative, which provided governance, research infrastructure, and operational support.24

In addition to the substantial reduction in the mean time to molecular diagnosis achieved by the program, there was little variability in the duration of each step of the diagnostic pathway, indicative of a high degree of consistency. However, there was marked variation in the rates of recruitment between sites, which may be partly explained by the uneven distribution of the Australian population and the transfer of complex patients to specialist pediatric hospitals. It may also partly represent different levels of clinician engagement, prior experience, and perceptions of utility at different sites. The longest component of the diagnostic pathway remained time from hospital admission to test initiation, suggesting that further efforts to shorten the diagnostic trajectory need to focus on changing clinician practice as much as they do on shortening laboratory turnaround times.

Limitations

This study has several limitations. First, all clinical and laboratory services participated on a voluntary basis, and expressed high levels of implementation readiness prior to study commencement.14

Second, all participating hospitals had clinical genetics services on site, and the service delivery model implemented was led by the clinical genetics service as per participating center preferences.14 The program leveraged interdisciplinary expertise in patient selection and data interpretation and therefore outcomes such as molecular diagnostic yield may not be generalizable to other health care systems or service delivery models.

Third, there is no consensus on how to evaluate rapid testing programs or what evidentiary threshold should be used for health care system funding.25,26 The optimal study design to assess the utility of rapid (and non-rapid) genomic testing for rare diseases remains an unresolved question, and a blinded randomized clinical trial9 of rapid genome sequencing vs standard care was terminated prematurely due to loss of equipoise. This current study, and many others,1-3,5-8 have opted for a descriptive approach, acknowledging this has the potential for biased reporting of outcomes, in particular in drawing a causal link between ultra-rapid exome sequencing results and subsequent changes in clinical management.

Fourth, this study was not designed or powered to measure differences in major clinical outcomes like morbidity and mortality against those that result from standard care of critically ill pediatric patients. There is a need to develop robust frameworks to measure the long-term clinical outcomes and performance of rapid diagnosis programs to assist with international benchmarking, service planning, funding, and iterative program development. Best practices in pretest and posttest counseling need to be informed by exploration of family experiences and preferences, and further analysis is required to establish cost-effectiveness.

Conclusions

This study suggests feasibility of ultra-rapid genomic testing in critically ill pediatric patients with suspected monogenic conditions in the Australian public health care system. However, further research is needed to understand the clinical value of such testing, and the generalizability of the findings to other health care settings.

Back to top
Article Information

Corresponding Author: Zornitza Stark, DM, Victorian Clinical Genetics Services, Murdoch Children’s Research Institute, 50 Flemington Rd, Parkville, VIC 3052, Australia (zornitza.stark@vcgs.org.au).

Accepted for Publication: April 23, 2020.

Authors/Members of the Australian Genomics Health Alliance Acute Care Flagship: Sebastian Lunke, PhD; Stefanie Eggers, PhD; Meredith Wilson, MBBS; Chirag Patel, MD; Christopher P. Barnett, MBBS; Jason Pinner, MBBS; Sarah A. Sandaradura, PhD; Michael F. Buckley, PhD; Emma I. Krzesinski, MBChB; Michelle G. de Silva, PhD; Gemma R. Brett, MGenCouns; Kirsten Boggs, MGenCouns; David Mowat, MBBS; Edwin P. Kirk, PhD; Lesley C. Adès, MBBS; Lauren S. Akesson, PhD; David J. Amor, PhD; Samantha Ayres, MGenCouns; Anne Baxendale, GradDipGenetCounsel; Sarah Borrie, MGenCouns; Alessandra Bray, PhD; Natasha J. Brown, PhD; Cheng Yee Chan, MBChB; Belinda Chong, PhD; Corrina Cliffe, BSc, Biomed; Martin B. Delatycki, PhD; Matthew Edwards, MD; George Elakis, BSc; Michael C. Fahey, PhD; Andrew Fennell, MBBChBAO; Lindsay Fowles, PhD; Lyndon Gallacher, MGenCouns; Megan Higgins, MBBS; Katherine B. Howell, PhD; Lauren Hunt, MSc; Matthew F. Hunter, MBChB; Kristi J. Jones, PhD; Sarah King, PhD; Smitha Kumble, MBBS; Sarah Lang, PhD; Maelle Le Moing, MSc; Alan Ma, MBBS; Dean Phelan, PhD; Michael C. J. Quinn, PhD; Anna Richards, MSc; Christopher M. Richmond, MBBS; Jessica Riseley, BSc; Jonathan Rodgers, MBBS; Rani Sachdev, MBBS; Simon Sadedin, PhD; Luregn J. Schlapbach, MD; Janine Smith, MBBS; Amanda Springer, MGenCouns; Natalie B. Tan, MBBS; Tiong Y. Tan, PhD; Suzanna L. Temple, PhD; Christiane Theda, MD; Anand Vasudevan, MBBS; Susan M. White, MBBS; Alison Yeung, MBBS; Ying Zhu, PhD; Melissa Martyn, PhD; Stephanie Best, PhD; Tony Roscioli, PhD; John Christodoulou, PhD; Zornitza Stark, DM.

Affiliations of Authors/Members of the Australian Genomics Health Alliance Acute Care Flagship: Australian Genomics Health Alliance, Parkville, Australia (Lunke, de Silva, Boggs, Ayres, Bray, King, Best, Christodoulou, Stark); Victorian Clinical Genetics Services, Murdoch Children’s Research Institute, Melbourne, Australia (Lunke, Eggers, de Silva, Brett, Akesson, Ayres, Brown, Chong, Delatycki, Gallacher, Kumble, Le Moing, Phelan, Richmond, Riseley, Sadedin, N. B. Tan, T. Y. Tan, White, Yeung, Christodoulou, Stark); University of Melbourne, Melbourne, Australia (Lunke, de Silva, Brett, Akesson, Amor, Brown, Delatycki, Gallacher, Howell, T. Y. Tan, Theda, White, Christodoulou, Stark); Sydney Children’s Hospitals Network-Westmead, Sydney, Australia (Wilson, Sandaradura, Boggs, Adès, Bray, Jones, Ma, Smith); University of Sydney, Sydney, Australia (Wilson, Sandaradura, Adès, Jones, Ma, Smith, Christodoulou); Genetic Health Queensland, Royal Brisbane and Women’s Hospital, Brisbane, Australia (Patel, Fowles, Higgins, Hunt, Quinn, Rodgers); Women’s and Children’s Hospital, North Adelaide, Australia (Barnett, Baxendale, Borrie); Sydney Children’s Hospitals Network-Randwick, Sydney, Australia (Pinner, Boggs, Mowat, Kirk, Bray, Sachdev); University of New South Wales, Sydney, Australia (Pinner, Mowat, Kirk, Roscioli); NSW Health Pathology Randwick Genomics Laboratory, Sydney, Australia (Buckley, Kirk, Chan, Cliffe, Elakis, Lang, Richards, Temple, Zhu, Roscioli); Monash Genetics, Monash Health, Melbourne, Australia (Krzesinski, Akesson, Fahey, Fennell, Hunter, Springer, Yeung); Department of Paediatrics, Monash University, Melbourne, Australia (Krzesinski, Fahey, Fennell, Hunter, Springer); Royal Children’s Hospital, Melbourne, Australia (Amor, Howell); Murdoch Children’s Research Institute, Melbourne, Australia (Amor, Howell, Theda, Martyn, Best); Neuroscience Research Australia, University of New South Wales, Sydney, Australia (Chan, Roscioli); Hunter Genetics, Newcastle, Australia (Edwards); Department of Paediatrics, School of Medicine, University of Western Sydney, Sydney, Australia (Edwards); University of Queensland, Brisbane, Australia (Higgins, Hunt); Centre for Cancer Biology, SA Pathology and University of South Australia, Adelaide (King); Paediatric Critical Care Research Group, Child Health Research Centre, the University of Queensland and Queensland Children’s Hospital, Brisbane, Australia (Schlapbach); Royal Women’s Hospital, Melbourne, Australia (Theda, Vasudevan); Melbourne Genomics Health Alliance, Melbourne, Australia (Martyn); Australian Institute of Health Innovation, Macquarie University, Sydney (Best).

Author Contributions: Drs Lunke and Stark had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Lunke, Eggers, Barnett, Brett, Hunter, Theda, Martyn, Best, Roscioli, Christodoulou, Stark.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Lunke, Barnett, Brett, Amor, Ayres, Fahey, Schlapbach, White, Martyn, Best, Stark.

Critical revision of the manuscript for important intellectual content: Lunke, Eggers, Wilson, Patel, Barnett, Pinner, Sandaradura, Buckley, Krzesinski, de Silva, Brett, Boggs, Mowat, Kirk, Ades, Akesson, Amor, Baxendale, Schenscher, Bray, Brown, Chan, Chong, Cliffe, Delatycki, Edwards, Elakis, Fahey, Fennell, Fowles, Gallacher, Higgins, Howell, Hunt, Hunter, Jones, King, Kumble, Lang, Le Moing, Ma, Phelan, Quinn, Richards, Richmond, Riseley, Rodgers, Sachdev, Sadedin, Schlapbach, Smith, Springer, N. Tan, T. Tan, Temple, Theda, Vasudevan, White, Yeung, Zhu, Best, Roscioli, Christodoulou, Stark.

Statistical analysis: Lunke, Martyn.

Obtained funding: Lunke, Patel, Pinner, Amor, Hunter, Roscioli, Christodoulou, Stark.

Administrative, technical, or material support: Lunke, Eggers, Wilson, Patel, Barnett, Pinner, Buckley, de Silva, Brett, Boggs, Kirk, Akesson, Ayres, Schenscher, Bray, Brown, Chong, Cliffe, Elakis, Fahey, Fowles, Gallacher, Higgins, Hunt, Hunter, Jones, King, Lang, Le Moing, Phelan, Quinn, Richards, Richmond, Riseley, Sadedin, Springer, N. Tan, Temple, Vasudevan, White, Yeung, Roscioli.

Supervision: Lunke, Wilson, Barnett, Pinner, Buckley, Boggs, Brown, Hunter, Sachdev, T. Tan, Martyn, Best, Roscioli, Christodoulou, Stark.

Conflict of Interest Disclosures: Dr Fahey reported receiving personal fees for providing expert testimony in medicolegal cases; and receiving grants from the Fulbright Foundation. Dr Howell reported receiving personal fees from RogCon Biosciences Inc. Dr Theda reported receiving personal fees from Navi Medical Technologies Pty Ltd, Ventora Pty Ltd, the Royal Children’s Hospital, and Johns Hopkins University. No other disclosures were reported.

Funding/Support: The Australian Genomics Health Alliance (Australian Genomics) project was funded by grant GNT1113531 from the National Health and Medical Research Council Targeted Call for Research, grant 2017-906 from the Royal Children’s Hospital Foundation, and unspecified grants from the Sydney Children’s Hospital Network and the Channel 7 Children’s Research Foundation. The research conducted at the Murdoch Children’s Research Institute was supported by the Victorian government’s operational infrastructure support program.

Role of the Funder/Sponsor: The funders/sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We thank Andrew Sinclair, PhD (Murdoch Children’s Research Institute), for critically reading the manuscript without receiving compensation.

References
1.
Farnaes  L, Hildreth  A, Sweeney  NM,  et al.  Rapid whole-genome sequencing decreases infant morbidity and cost of hospitalization.   NPJ Genom Med. 2018;3:10. doi:10.1038/s41525-018-0049-4PubMedGoogle ScholarCrossref
2.
French  CE, Delon  I, Dolling  H,  et al; NIHR BioResource—Rare Disease; Next Generation Children Project.  Whole genome sequencing reveals that genetic conditions are frequent in intensively ill children.   Intensive Care Med. 2019;45(5):627-636. doi:10.1007/s00134-019-05552-xPubMedGoogle ScholarCrossref
3.
Gubbels  CS, VanNoy  GE, Madden  JA,  et al.  Prospective, phenotype-driven selection of critically ill neonates for rapid exome sequencing is associated with high diagnostic yield.   Genet Med. 2020;22(4):736-744. doi:10.1038/s41436-019-0708-6PubMedGoogle ScholarCrossref
4.
Kingsmore  SF, Cakici  JA, Clark  MM,  et al; RCIGM Investigators.  A randomized, controlled trial of the analytic and diagnostic performance of singleton and trio, rapid genome and exome sequencing in ill infants.   Am J Hum Genet. 2019;105(4):719-733. doi:10.1016/j.ajhg.2019.08.009PubMedGoogle ScholarCrossref
5.
Meng  L, Pammi  M, Saronwala  A,  et al.  Use of exome sequencing for infants in intensive care units: ascertainment of severe single-gene disorders and effect on medical management.   JAMA Pediatr. 2017;171(12):e173438. doi:10.1001/jamapediatrics.2017.3438PubMedGoogle Scholar
6.
Mestek-Boukhibar  L, Clement  E, Jones  WD,  et al.  Rapid paediatric sequencing (RaPS): comprehensive real-life workflow for rapid diagnosis of critically ill children.   J Med Genet. 2018;55(11):721-728. doi:10.1136/jmedgenet-2018-105396PubMedGoogle ScholarCrossref
7.
Sanford  EF, Clark  MM, Farnaes  L,  et al; RCIGM Investigators.  Rapid whole genome sequencing has clinical utility in children in the PICU.   Pediatr Crit Care Med. 2019;20(11):1007-1020. doi:10.1097/PCC.0000000000002056PubMedGoogle ScholarCrossref
8.
Stark  Z, Lunke  S, Brett  GR,  et al; Melbourne Genomics Health Alliance.  Meeting the challenges of implementing rapid genomic testing in acute pediatric care.   Genet Med. 2018;20(12):1554-1563. doi:10.1038/gim.2018.37PubMedGoogle ScholarCrossref
9.
Petrikin  JE, Cakici  JA, Clark  MM,  et al.  The NSIGHT1-randomized controlled trial: rapid whole-genome sequencing for accelerated etiologic diagnosis in critically ill infants.   NPJ Genom Med. 2018;3:6. doi:10.1038/s41525-018-0045-8PubMedGoogle ScholarCrossref
10.
Greenhalgh  T, Papoutsi  C.  Spreading and scaling up innovation and improvement.   BMJ. 2019;365:l2068. doi:10.1136/bmj.l2068PubMedGoogle ScholarCrossref
11.
Braithwaite  J, Churruca  K, Long  JC, Ellis  LA, Herkes  J.  When complexity science meets implementation science: a theoretical and empirical analysis of systems change.   BMC Med. 2018;16(1):63. doi:10.1186/s12916-018-1057-zPubMedGoogle ScholarCrossref
12.
Curran  GM, Bauer  M, Mittman  B, Pyne  JM, Stetler  C.  Effectiveness-implementation hybrid designs: combining elements of clinical effectiveness and implementation research to enhance public health impact.   Med Care. 2012;50(3):217-226. doi:10.1097/MLR.0b013e3182408812PubMedGoogle ScholarCrossref
13.
Damschroder  LJ, Aron  DC, Keith  RE, Kirsh  SR, Alexander  JA, Lowery  JC.  Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science.   Implement Sci. 2009;4:50. doi:10.1186/1748-5908-4-50PubMedGoogle ScholarCrossref
14.
Stark  Z, Nisselle  A, McClaren  B,  et al.  Attitudes of Australian health professionals towards rapid genomic testing in neonatal and paediatric intensive care.   Eur J Hum Genet. 2019;27(10):1493-1501. doi:10.1038/s41431-019-0429-yPubMedGoogle ScholarCrossref
15.
Harris  PA, Taylor  R, Thielke  R, Payne  J, Gonzalez  N, Conde  JG.  Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support.   J Biomed Inform. 2009;42(2):377-381. doi:10.1016/j.jbi.2008.08.010PubMedGoogle ScholarCrossref
16.
Richards  S, Aziz  N, Bale  S,  et al; ACMG Laboratory Quality Assurance Committee.  Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology.   Genet Med. 2015;17(5):405-424. doi:10.1038/gim.2015.30PubMedGoogle ScholarCrossref
17.
Akesson  LS, Eggers  S, Love  CJ,  et al.  Early diagnosis of Pearson syndrome in neonatal intensive care following rapid mitochondrial genome sequencing in tandem with exome sequencing.   Eur J Hum Genet. 2019;27(12):1821-1826. doi:10.1038/s41431-019-0477-3PubMedGoogle ScholarCrossref
18.
Sobreira  N, Schiettecatte  F, Valle  D, Hamosh  A.  GeneMatcher: a matching tool for connecting investigators with an interest in the same gene.   Hum Mutat. 2015;36(10):928-930. doi:10.1002/humu.22844PubMedGoogle ScholarCrossref
19.
Rehman  AU, Najafi  M, Kambouris  M,  et al.  Biallelic loss of function variants in PPP1R21 cause a neurodevelopmental syndrome with impaired endocytic function.   Hum Mutat. 2019;40(3):267-280.PubMedGoogle ScholarCrossref
20.
Cooper  MS, Stark  Z, Lunke  S, Zhao  T, Amor  DJ.  IREB2-associated neurodegeneration.   Brain. 2019;142(8):e40. doi:10.1093/brain/awz183PubMedGoogle Scholar
21.
Costain  G, Ghosh  MC, Maio  N,  et al.  Absence of iron-responsive element-binding protein 2 causes a novel neurodegenerative syndrome.   Brain. 2019;142(5):1195-1202. doi:10.1093/brain/awz072PubMedGoogle ScholarCrossref
22.
Richmond  CM, Campbell  S, Foo  HW,  et al.  Rapid identification of biallelic SPTB mutation in a neonate with severe congenital hemolytic anemia and liver failiure.   Mol Syndromol. 2020;11(1):50-55. doi:10.1159/000505886PubMedGoogle ScholarCrossref
23.
Clark  MM, Hildreth  A, Batalov  S,  et al.  Diagnosis of genetic diseases in seriously ill children by rapid whole-genome sequencing and automated phenotyping and interpretation.   Sci Transl Med. 2019;11(489):eaat6177. doi:10.1126/scitranslmed.aat6177PubMedGoogle Scholar
24.
Stark  Z, Boughtwood  T, Phillips  P,  et al.  Australian Genomics: a federated model for integrating genomics into healthcare.   Am J Hum Genet. 2019;105(1):7-14. doi:10.1016/j.ajhg.2019.06.003PubMedGoogle ScholarCrossref
25.
Friedman  JM, Bombard  Y, Cornel  MC,  et al; Paediatric Task Team of the Global Alliance for Genomics and Health Regulatory and Ethics Work Stream.  Genome-wide sequencing in acutely ill infants: genomic medicine’s critical application?   Genet Med. 2019;21(2):498-504. doi:10.1038/s41436-018-0055-zPubMedGoogle ScholarCrossref
26.
Grosse  SD, Farnaes  L.  Genomic sequencing in acutely ill infants: what will it take to demonstrate clinical value?   Genet Med. 2019;21(2):269-271. doi:10.1038/s41436-018-0124-3PubMedGoogle ScholarCrossref
×