Symbolic description of intracerebral vessels segmented from magnetic resonance angiograms and evaluation by comparison with X-ray angiograms

摘要: Abstract We describe and evaluate methods that create detailed vessel trees by linking vessels have been segmented from magnetic resonance angiograms (MRA). The tree-definition process can automatically exclude erroneous segmentations. parent–child connectivity information provided our is important to both surgical planning guidance of endovascular procedures. evaluated the branch connection accuracy 3D asking two neuroradiologists 140 connections comprising seven vascular against 17 digital subtraction angiography (DSA) views. Each reviewer rated each as (1) Correct, (2) Incorrect, (3) Partially correct (a minor error without clinical significance), or (4) Indeterminate. Analysis was summarized for evaluator calculating 95% confidence intervals proportion completely clinically acceptable (completely partially correct). In order protect overall Type I rate, alpha-splitting done using a top down strategy. additionally segmentation completeness examining slice in 11 MRA datasets determine unlabeled identifiable cross-section following segmentation. Results indicate only one judged incorrect reviewers. segmentations appeared complete within resolution limits. conclude permit creation image data. review literature compare other approaches own. provide examples useful visualizations enabled methodology taken visualization program now use.