Personal smartphones for neonatal diagnostic imaging: A prospective crossover study

Abstract

Aim: Neonatal clinicians may be asked to review X-ray images when unable to directly access the original image. Transmitting an X-ray image to a smartphone is a technique increasingly being used by clinicians in a number of settings. Minimal data exist on its use in the neonatal setting. Our aim was to compare the ability of neonatal clinicians to correctly diagnose a pneumothorax from a chest radiograph (CXR) viewed on a smartphone, paralleled with the same image viewed on a computer screen.

Methods: We investigated the accuracy, sensitivity, specificity, positive predictive value and negative predicative value of two methods of viewing CXR images to diagnose pneumothorax, compared with a gold standard radiologist report. Clinicians were presented with 40 CXR images on two occasions, at least 1 week apart. Images were viewed once on a smartphone and once on a computer screen. Both the viewing method and viewing order of the images were randomised. The clinical details of the infant at the time the CXR was taken were provided, and participants were asked if a pneumothorax was present.

Results: Twenty-one clinicians viewed all CXR images using both viewing methods (840 paired observations). There was no difference in accuracy of detecting pneumothorax between viewing methods, 81% correctly identified a pneumothorax using the smartphone, versus 80% using the computer screen (P = 0.40 (95% CI), difference -4 to 1.5%).

Conclusion: Diagnosis of neonatal pneumothorax was as accurate using a smartphone as viewing CXR images displayed on a computer screen when clinicians were presented with X-rays with diagnosis of pneumothorax.

Keywords: diagnostic accuracy; infant; newborn; pneumothorax; smartphone; teleradiology.