Neonatologists perform cardiac ultrasound (NPCU) is an important assessment tool to assess cardiac hemodynamics in premature and unwell infants. In premature infants < 30 weeks gestation routinely receiving NPCU studies, we identified 109 (5%) with a de novo diagnosis of congenital heart disease (CHD) and 11 with critical or major CHD conditions requiring early cardiac intervention, who would have otherwise been diagnosed later. All patients where there was suspected CHD on NPCU were appropriately referred to cardiology. The NPCU was able to facilitate prompter cardiology referral and assist cardiologists in earlier diagnosis in all cases. No major or critical CHD were missed on NPCU flagged cardiac scans. Our study suggests that NPCU, although not intended for this purpose, may detect de novo CHD despite the implementation of modern prenatal imaging and pulse oximetry saturation screening techniques. As such, NPCU-trained clinicians should be adept at recognizing abnormalities in cardiac structures during hemodynamic assessment.

In Australia, training for neonatologists in cardiac ultrasounds is guided by the Australasian Society for Ultrasound in Medicine, resulting in a Certificate of Clinician-Performed Ultrasound; this demonstrates significant hands-on training in functional echocardiography [18]. This training includes overlap with features from both North American and European curriculum [14, 16]. NPCU is not a screening tool for CHD and does not require trainees to exclude CHD [13], our training programs rigorously teaches functional hemodynamic assessments of the heart where anatomical review is required and significant CHD (such as left ventricular outflow obstruction) can be detected [14, 16]. This is partly because appropriate hemodynamic assessment, for example, assessment of a PDA or pulmonary hypertension, is strongly dependent upon anatomical variance and may be limited by the presence of CHD. Effective NPCU services should be able to identify lesions using structured training and appropriate supervision [6, 16]. They should then closely collaborate with cardiology services to ensure optimal follow-up and care [18, 19]. However, in various centers access to cardiology oversight is not possible [20, 21]. If a center is implementing NPCU, adequate attention to training and support is paramount to the reliable detection of CHD and subsequent appropriate hemodynamic assessment [12, 16]. We suggest that NPCU training must include teaching about recognition of early and significant CHD with demonstration of normal cardiac chambers, connections, and valves prior to hemodynamic assessment for clinical decision making [22]. Currently, the degree to which this is practiced varies widely across countries and neonatology centers with focus on standard curriculum and protocols required [20, 23].

CHD continues to be detected by the NPCU despite multiple advances in prenatal and postnatal screening for CHD. In Australia, obstetric ultrasonography has seen recent advances that have contributed to the detection of CHD [24]. NPCU may also be considered within the spectrum of tools available to detect CHD, along with pulse oximetry screening, and are being increasingly used in Australia [8, 10]. Each screening measure, although not designed for this purpose, may opportunistically detect critical CHD despite limitations in detecting certain types of CHD, such as PS [20, 25]. With appropriate training and oversight via pediatric cardiology services, NPCU may improve upon current screening for CHD, almost as an extension of the clinical examination [13, 26]. Other Australian studies have shown de novo CHD diagnosis by NPCU in 14% of CHD cases, but with a vast majority were still diagnosed prenatally [19].

CHD detection, within this study, occurs at similar rates to prior studies but continues to produce a high rate of false positives, due to the need for clinicians to over report CHD based on limitations in training and experience [13, 18]. In our cohort, CHD was diagnosed at a higher rate if flagged during NPCU with early diagnosis being a positive influence on clinical cardiology management. Our focus is less upon the exact CHD diagnosis but rather the prompt recognition of abnormal anatomy, leading to the need for referral. Decreased precision in recognizing more minor anomalies is expected in sonographers with varying experiences and in those with evolving or complex cardiac lesions [27]. There remain many CHD subtypes that may not be apparent on early NPCU, such as PS and ASD, and we expect the rate of detection of these lesions to be lower for NPCU than if the scan was completed later in life. For example, ASDs may continue to be difficult to ascertain from a PFO within early life and PS may progress with chronological age. Errors in detection continue to occur within other screening methods and may be related to sonographer experience, the timing of the scan within the context of chronological age and the weight/gestational age of the baby being scanned [28, 29]. The rate of detection of NPCU continues to be user dependent and we continue to expect that some of these lesions will not be routinely determined on NPCU, as they may not be obvious in the first few days of life when NPCU is routinely performed.

This data is from a tertiary-level NICU with close collaborative ties to pediatric cardiology services and we must acknowledge how this setup varies across the world [16, 20, 21, 23]. In other centers, NPCU may be performed with variable training and protocols without standards that are translatable elsewhere [12, 23, 30]. However, delays in accessing timely cardiology services or the requirement of a neonatal transfers warrants focus on improved CHD detection and management on whatever scan is completed whether it be POCUS, NPE or TnECHO. The accuracy of diagnosis by neonatologists needs to be further explored, and anatomic concordance data is not available for this study because the role of NPCU is not to diagnose and is left to our cardiologists to review scans suspicious for CHD. Without cardiology oversight the successful detection and management of CHD may not always be possible. For example, in many of our NPCU scans hemodynamic assessment was abandoned because of the significant CHD, making assessment inaccurate and requiring an urgent formal echocardiogram. The retrospective nature of the study design is a limitation. Earlier scans were completed by a range of neonatologists, at differing levels of experience and with indications for NPCU being historically driven as opposed to current Australasian training which is becoming more rigid, and guideline driven. We acknowledge that minor CHD may still exist but remain undiagnosed in the true negative group, and that younger members of the cohort may also receive cardiac intervention in the future.

NPCU can accurately identify CHD in infants without a prenatal diagnosis and may lead to earlier identification, referral, and treatment for CHD. The success of the NPCU involves clear and structured guidelines for its performance, appropriate training, accreditation, and supervision, as well as strong collaboration between neonatal and cardiology services. CHD continues to remain within the realm of cardiologists to diagnose and treat. Prospective studies and training developments will assist in appropriate detection of and referral of CHD during neonatologist performed cardiac ultrasound.