Sharbell Yousef Hashoul MD,MBA, Presenter: Nothing to Disclose

Tamar Gaspar MD, Abstract Co-Author: Nothing to Disclose

Nathan Peled MD, Abstract Co-Author: Nothing to Disclose

There is an increasing interest in a reliable method for the prediction of stroke risk factors. It is a well known fact that there is a correlation between stroke incidence and carotid calcium burden. Hence, it would be useful to have an automatic tool for detection of these plaques. In particular, such a tool could be applied to automatic analysis of incidental findings on CT scans of the neck performed without contrast. The purpose of this paper is to evaluate effectiveness and precision of such an automatic CAD tool for evaluation of carotid calcium burden.

An experienced radiologist and CAD independently assessed 18 carotid, non-contrast, CT scans (Philips Brilliance, spiral non-gated 64x0.625mm, 1mm reconstructed slice thickness, 1mm increments). For each image cross-section, carotid calcium areas were tagged. A single point was used for each connected component. Then, for each patient, the number of matching scores, false negatives and false positives were computed. We analyzed 13 patients with significant calcium burden (which subsequently developed strokes) and 5 patients with no evidence of calcium (and no stroke).

CAD successfully detected all 13 patients with significant calcium burden. For all of 5 healthy patients no calcium was detected (i.e. per patient level both sensitivity and specificity have been 100%) For patients with detected calcium, on average, 86 % of the plaque formations were detected. For calcium-free areas, false positive rates were 1.88%.

Rapid, automatic carotid calcium plaque detection and volume assessment is feasible and simple. This provides an easy tool for assessing stroke risk using CT scans performed in search of various other pathologies.

The correlation between stroke incidence and carotid calcium burden is well known. Our aim was to evaluate effectiveness and precision of an automatic CAD tool for evaluation of carotid calcium burden

Hashoul, S, Gaspar, T, Peled, N, Automatic Carotid Calcium Burden Estimation as a Tool for Stroke Risk Prediction  .  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11034563.html