Helen Branson MD, Abstract Co-Author: Nothing to Disclose

Andrea Schwarz Doria MD, Presenter: Nothing to Disclose

Rahim Moineddin, Abstract Co-Author: Nothing to Disclose

Manohar Meghraj Shroff MD, Abstract Co-Author: Nothing to Disclose

The use of intravenous (IV) contrast in CT scanning after an unenhanced scan is associated with increased radiation which is a key limitation for children, increased costs and morbidity risks. Our purpose was to determine the test performance of unenhanced CT scans for diagnosis of brain pathology compared with corresponding enhanced CT scans.

We reviewed 398 CT scans of brain in children before and after IV contrast for clinical indications other than trauma. Diagnosis of shunts, tumors and incomplete scans were excluded (n=22). Unenhanced and contrast-enhanced scans were randomly read, independently, by two pediatric radiologists, blinded to clinical information. The diagnosis of each scan (presence vs absence of abnormalities) was recorded according to anatomic sections (supratentorial, infratentorial, ventricles, skull and others), resulting in one out of 3 decisions: 1. No further information gained. 2. Contrast confirmed the diagnosis made pre-contrast. 3. Contrast changed the original diagnosis.

The interreader agreement rate for different anatomic regions varied between good (kappa coefficient, 0.63) and excellent (0.88). Overall, the sensitivity, specificity, positive and negative predictive values for unenhanced scans were 97% (95% confidence intervals [CI], 93-99%), 94% (89-97%), 92% (87-97%) and 98% (95-99%), respectively. Specifically, the sensitivity of unenhanced scans was slightly higher for diagnosis of supratentorial abnormal findings (90% [83-95%]) than infratentorial (84% [64-95%], ventricular (87% [79-93%]) or skull (76% [61-88%]) findings. On the other hand, the specificity of unenhanced scans for diagnosis of infratentorial (95% [93-97%]) and skull (95% [92-97%]) abnormalities was marginally higher compared with supratentorial (91% [87-94%]) and ventricular (83% [78-88%]) regions. The use of contrast material changed the original normal diagnosis in only 9/200 (4.5%) of the scans.

The use of IV contrast after an unenhanced scan of the brain in children does not significantly change the diagnosis in most cases. Unless there is a suspected abnormality on the pre-contrast scan, contrast is not indicated.

Branson, H, Doria, A, Moineddin, R, Shroff, M, Is Contrast Enhancement Really Needed after a Normal Unenhanced CT Scan of the Brain in Children? A Retrospective Review.  Radiological Society of North America 2005 Scientific Assembly and Annual Meeting, November 27 - December 2, 2005 ,Chicago IL. http://archive.rsna.org/2005/4411379.html