Will a handheld ultrasound scanner be applicable for screening for heart abnormalities in newborns and children?

There is significant interest in opportunities to provide echocardiography services for detection of congenital heart disease with portable, or even handheld, devices in remote areas or third world countries where conventional ultrasound systems may not be available. We tested a handheld system (HHS) (SonoHeart, SonoSite Inc, Bothell, Wash) equipped with a broadband, 7- to 4-MHz, miniaturized, curved, linear-array transducer and implemented with an improved directional Doppler flow map. All echocardiography scanning was performed in the neonatal nursery, pediatric intensive care department, or pediatric echocardiography laboratory of our institution. We reviewed limited echocardiography view sequences sequentially obtained by the same expert examiner (D.J.S.) in 50 infants and children (age: 1 day to 6 years), with preoperative or postoperative forms of congenital heart disease. Each patient was studied twice, once with a conventional full-feature system (FFS) and then a limited scan with the HHS using similar frequency transducers. The cardiologist (D.J.S.) and blinded research laboratory reviewers (X.L., G.K.M., R.A.R.) read the FFS and HHS image sequences for diagnosis and for grading the quality of the anatomic and flow feature images. The studies were performed and reviewed with the examiner and reviewers blinded to patient diagnosis. The major diagnoses (eg, patent ductus arteriosus, atrio-ventricular (AV) canal, peripheral pulmonary valve stenosis, aortic coarctation, atrial septal defect, ventricular septal defect, preoperative or postoperative tetralogy of Fallot, and mitral regurgitation) were made by both readers, who were unaware of each other's diagnosis results. Furthermore, the average composite HHS cardiac anatomic feature score on a scale of 0 (not visualized) to 3 (visualized precisely) from the parasternal long-axis and 4- or 5-chamber view for cardiac anatomy were 2.67 ± 0.49 (SD) and 2.50 ± 0.55, respectively, versus 2.73 ± 0.45 and 2.55 ± 0.54 for the FFS. The mean flow feature score, comprising all views, was 2.67 ± 0.45 (HHS) versus 2.72 ± 0.48 (FFS). The P values for all above comparisons were >.05. Image quality of the FFS anatomic structures were, thus, not statistically different from the HHS. Although the color cosmetic was different for the HHS directional (nonvelocity) map, only 9% of 150 total findings (including structural abnormalities and flow features, none of which were critical) were missed, whereas the other 91% regur