Diagnostic errors from digital stitching of scoliosis images – the importance of evaluating the source images prior to making a final diagnosis

Abstract

Digital methods are used for scoliosis imaging. Separate images of the thoracic and lumbar/sacral spine are acquired. These are then stitched into a single image. This process has some potential disadvantages. To assess anatomical errors in digitally stitched scoliosis films. Eighty-six scoliosis studies were evaluated for stitching errors. The incidence of a stitching error resulting in an abnormality on the stitched image that could not be verified on the source images was calculated. Subgroups were analyzed based on the presence/absence of spinal hardware and standing/supine technique. Fourteen exams (16%) had stitching errors that could result in a false diagnosis if not correlated with the source images. The majority were errors of vertebral alignment. There was no significant difference in error rate with the presence or absence of hardware (P = 0.73) or patient positioning (P = 0.34). 16% of digital scoliosis exams had stitching errors that could result in a false diagnosis. The error rate was not influenced by spinal hardware or patient positioning. These results stress the importance of correlating any abnormality on the stitched image with the exam's source images; if not, a false diagnosis of abnormality will be made.