Esther Lim Yuh MD, PhD, Presenter: Nothing to Disclose

Alisa D. Gean MD, Abstract Co-Author: Consultant, Advanced Quantum Sensors Consultant, Cambridge Neuroscience, Inc Speakers Bureau, Schering AG (Berlex Inc) Speakers Bureau, Bracco Group Speakers Bureau, General Electric Company

Geoffrey T. Manley MD, PhD, Abstract Co-Author: Nothing to Disclose

Max Wintermark MD, Abstract Co-Author: Consultant, PAION AG

To determine the accuracy of an automatic computer assessment of noncontrast CT studies of the brain obtained in patients with suspected traumatic brain injury, employing a board-certified neuroradiologist as the gold standard.

The noncontrast CT studies of the brain obtained in 33 consecutive patients with traumatic brain injury were retrospectively identified and independently reviewed by a board-certified neuroradiologist and by a computer algorithm for each of the following features: 1) presence or absence of subarachnoid hemorrhage, 2) presence or absence of clinically significant midline shift (≥ 5 mm), 3) normal, partly effaced, or completely effaced basal cisterns, and 4) presence or absence of a large (≥ 25 milliliters) intracranial hemorrhagic collection. Using the neuroradiologist interpretation as gold standard, sensitivity and specificity of the computer algorithm for each of the above-mentioned features were calculated.

Sensitivities and specificities of the computer algorithm were as follows: 1) sensitivity 100% and specificity 58.3% for presence of subarachnoid hemorrhage, 2) sensitivity 92.3% and specificity 85% for clinically significant midline shift (≥ 5 mm), 3) sensitivity 93.8% and specificity 100% for effacement of basal cisterns, and 4) sensitivity 50% and specificity 90.3% for intracranial ≥ 25 milliliters hemorrhagic collections.

The tested automatic computer assessment of noncontrast CT studies of the brain accurately characterizes the typical features seen in case of traumatic brain injury.

Computer assessment of noncontrast CT studies of the brain might increase reproducibility and improve workflow in the setting of emergency neuroradiological evaluation of patients with suspected traum

Yuh, E, Gean, A, Manley, G, Wintermark, M, Automatic Computer Assessment of Computed Tomography (CT) Studies of the Brain in Patients with Suspected Traumatic Brain Injury.  Radiological Society of North America 2007 Scientific Assembly and Annual Meeting, November 25 - November 30, 2007 ,Chicago IL. http://archive.rsna.org/2007/5003459.html