Automated infarction core delineation using cerebral and perfused blood volume maps
Language English Country Germany Media print-electronic
Document type Journal Article, Research Support, Non-U.S. Gov't
- MeSH
- Blood Volume * MeSH
- Humans MeSH
- Brain diagnostic imaging physiopathology MeSH
- Brain Infarction diagnosis physiopathology MeSH
- Cerebrovascular Circulation physiology MeSH
- Tomography, X-Ray Computed methods MeSH
- Regional Blood Flow physiology MeSH
- Software * MeSH
- Check Tag
- Humans MeSH
- Publication type
- Journal Article MeSH
- Research Support, Non-U.S. Gov't MeSH
PURPOSE: Thrombolytic therapy in patients with acute ischemic stroke is contraindicated when the infarction core exceeds a given threshold. To date, there are no standardized guidelines for computed tomography infarction core assessment. Current practice involves use of thresholding methods, where the results are further adjusted by an experienced physician. An automated method for infarction core delineation and volume measurement was developed and tested. MATERIALS AND METHODS: CT postprocessing software was developed for analysis of whole brain perfused blood volume (PBV) and cerebral blood volume (CBV) maps. The program was designed for potential use with mean transit time (MTT) or cerebral blood flow (CBF) maps. The proposed method was tested on a set of 12 patients on both PBV and CBV maps with whole brain coverage by comparison with the results of a simple thresholding method and with manually marked findings provided by two independent physicians. RESULTS: The proposed method produced a marked infarct core volume corresponding to 53 % of the manually delineated volumes. The simple thresholding method with the optimal threshold, using the same dataset, marked 15[Formula: see text] larger volume compared to the volume delineated by physicians. CONCLUSION: An automated infarction core segmentation method based on local neighborhood features was developed and tested, demonstrating its utility in distinguishing between infarcted and non-infarcted areas, as well as reduction in the number of false positives and volume error.
See more in PubMed
AJNR Am J Neuroradiol. 1996 Nov-Dec;17(10):1861-9 PubMed
Stroke. 1991 Mar;22(3):324-30 PubMed
Stroke. 2011 Jul;42(7):1923-8 PubMed
Stroke. 2006 Apr;37(4):979-85 PubMed
Cerebrovasc Dis. 2011;31(3):238-45 PubMed
Lancet. 2006 May 27;367(9524):1747-57 PubMed
Radiographics. 2006 Oct;26 Suppl 1:S75-95 PubMed
Stroke. 2011 Dec;42(12):3435-40 PubMed
AJNR Am J Neuroradiol. 2006 Jan;27(1):20-5 PubMed
Stroke. 2012 Oct;43(10):2648-53 PubMed
Radiographics. 1992 Mar;12(2):345-55 PubMed
