Stroke Imaging Marker for detection of hyperacute ischemic changes in NCCT - advances
Ryszard Gomółka , R.M. Chrzan , Andrzej Urbanik , W.L. Nowiński , Antoni Grzanka
AbstractNCCT states a first-line diagnostic procedure for the emergency assessment of acute stroke. Although the NCCT is highly specific in differentiation of intracerebral hemorrhage from ischemia, it has a poor sensitivity in evaluation of acute ischemic stroke. Hence, computer-aided diagnosis is able to improve the performance. In 2013 we introduced a computational method for detection and localization of visible infarcts in NCCT. Herein, we aimed to evaluate and extend our method to localize a non-visible hyperacute ischemia by means of Stroke Imaging Marker (SIM). Based on the SIM and its components: ratio of percentile differences in subranges of HU distribution (P-ratio), ratio of voxels count in ranges of brain CT intensity (N-ratio), median HU attenuation value (MAV); the infarct localization was performed in 140 early and follow-up scans of 70 patients. The infarct was not visible to a radiologist or to an experienced stroke neuroradiologist in any of the early scans. The infarcted hemisphere detection rate and sensitivity of infarct localization were measured by comparing the detected region in the initial scan, with the gold standard set in respective follow-up examination. The best performance of the algorithm was found for the P-ratio including 7 percentile subranges within the range of 35th- 75th percentile. The modified SIM provided 76% ischemic hemisphere detection rate and 54% sensitivity in spatial localization of hyperacute ischemia (68% among properly detected sides). The improved SIM is a dedicated and potentially useful tool for hyperacute non-visible brain infarcts detection in NCCT and may contribute to decision to triage.
|Journal series||Journal of Neurology & Neurophysiology , ISSN 2155-9562|
|Publication size in sheets||3.15|
|Score|| = 0.0, 27-03-2017, ArticleFromJournal|
= 5.0, 27-03-2017, ArticleFromJournal
* presented citation count is obtained through Internet information analysis and it is close to the number calculated by the Publish or Perish system.