H-Index
12
Scimago Lab
powered by Scopus
eISSN: 1941-5923

Logo



The ethics are in the numbers: A bayesian approach to the management of incidental findings in pediatric magnetic resonance imaging research

Tonya White

Am J Case Rep 2009; 10:22-26

ID: 869591

Published: 2009-02-25


Background: The last 15 years there has seen an exponential increase in the number of studies utilizing magnetic resonance imaging to study brain development and pathology in pediatric populations. During this period, scanners have undergone considerable technological advances and are able to provide striking, high-resolution details of the gross anatomy of the developing brain. Reviewing these high resolution images occasionally results in the detection of potentially abnormal findings, even in the brains of asymptomatic and healthy research subjects.
Case Report: Incidental findings have created an ethical dilemma as to the best approach to manage these artifacts, which may range from true pathology to developmental differences. The case presented involves an incidental subcortical finding on MRI in a healthy 10 year old boy. The clinical management of the finding, the interaction with the family, and the longitudinal outcome is presented.
Conclusions: Few of the studies to date have presented the longitudinal course and outcome of individuals with such artifacts. The high base rate of incidental findings (13–37%) coupled with the relatively low base rate of pediatric neurological disorders will result in high false positive rates. This translates to many families being concerned that their child has brain pathology, when this is not true. Bayesian statistics can be used to determine the approximate rates of false to true positives. However, studies that determine the sensitivity and specificity of specific incidental findings, coupled with longitudinal studies of subjects with incidental findings are required to provide more accurate rates of the true and false positives.

Keywords: Pediatric Imaging, Incidental Findings, Bioethics, Bayesian, Magnetic Resonance Imaging



Back