PURPOSE: Analysis of patterns and causes of artifacts found in endographic visualization of spiral CT data. MATERIALS AND METHODS: A straight perspex tube with a diameter of 20 mm was scanned in three different positions with ten scan protocols of varying table feed, increment, kernel, and signal to noise ratio (Tomoscan AVE, Philips). The resulting 30 CT datasets were visualized as a virtual endoscopy (VE) with ten different visualization protocols (Easy Vision 4.2, Philips) of varying threshold, resolution, and perspective angle. 300 VE datasets were analyzed by two radiologists and compared with the visualization of a software-generated tube in order to differentiate scanning and software artifacts. RESULTS: Five different classes of artifacts have been identified. Two of them result from the scanning process and two from the specific visualization method. Spiral patterns and the unevenness of the tube wall vary with the scanning parameter. Moiré-like patterns are caused by the VE software and depend on the visualization matrix. A high perspective angle distorts the size and form of the tube and makes it difficult for the observer to locate his position within the tube. The appearance of pseudoforamina depends on both the scanning and the visualization parameters. CONCLUSION: The knowledge of the patterns and potential causes of artifacts in endographic visualization of spiral CT scans are the basis for interpretation and optimization of this visualization method.
PURPOSE: Analysis of patterns and causes of artifacts found in endographic visualization of spiral CT data. MATERIALS AND METHODS: A straight perspex tube with a diameter of 20 mm was scanned in three different positions with ten scan protocols of varying table feed, increment, kernel, and signal to noise ratio (Tomoscan AVE, Philips). The resulting 30 CT datasets were visualized as a virtual endoscopy (VE) with ten different visualization protocols (Easy Vision 4.2, Philips) of varying threshold, resolution, and perspective angle. 300 VE datasets were analyzed by two radiologists and compared with the visualization of a software-generated tube in order to differentiate scanning and software artifacts. RESULTS: Five different classes of artifacts have been identified. Two of them result from the scanning process and two from the specific visualization method. Spiral patterns and the unevenness of the tube wall vary with the scanning parameter. Moiré-like patterns are caused by the VE software and depend on the visualization matrix. A high perspective angle distorts the size and form of the tube and makes it difficult for the observer to locate his position within the tube. The appearance of pseudoforamina depends on both the scanning and the visualization parameters. CONCLUSION: The knowledge of the patterns and potential causes of artifacts in endographic visualization of spiral CT scans are the basis for interpretation and optimization of this visualization method.
Authors: Dagmar Honnef; Joachim E Wildberger; Marco Das; Christian Hohl; Andreas H Mahnken; Michael Barker; Rolf W Günther; Gundula Staatz Journal: Eur Radiol Date: 2006-04-19 Impact factor: 5.315