Limits...
Semi-automatic measurement of the airway dimension by computed tomography using the full-width-half-maximum method: a study on the measurement accuracy according to the CT parameters and size of the airway.

Kim N, Seo JB, Song KS, Chae EJ, Kang SH - Korean J Radiol (2008 May-Jun)

Bottom Line: The measured values as determined by CT and the actual dimensions of the tubes were compared.There was no significant difference in accuracy among images with the use of variable slice thicknesses or a variable FOV.Below a 1-mm threshold, the measurement failed to represent the change of the real dimensions.

View Article: PubMed Central - PubMed

Affiliation: Department of Radiology and Research Institute of Radiology, University of Ulsan College of Medicine, Asan Medical Center, Seoul, Korea.

ABSTRACT

Objective: To assess the influence of variable factors such as the size of the airway and the CT imaging parameters such as the reconstruction kernel, field-of-view (FOV), and slice thickness on the automatic measurement of airway dimension.

Materials and methods: An airway phantom was fabricated that contained eleven poly-acryl tubes of various lumen diameters and wall thicknesses. The measured density of the poly-acryl wall was 150 HU, and the measured density of the airspace filled with polyurethane foam was -900 HU. CT images were obtained using a 16-MDCT (multidetector CT) scanner and were reconstructed with various reconstruction kernels, thicknesses and FOV. The luminal radius and wall thickness were measured using in-house software based on the full-width-half-maximum method. The measured values as determined by CT and the actual dimensions of the tubes were compared.

Results: Measurements were most accurate on images reconstructed with use of a standard kernel (mean error: -0.03 +/- 0.21 mm for wall thickness and -0.12 +/- 0.11 mm for the luminal radius). There was no significant difference in accuracy among images with the use of variable slice thicknesses or a variable FOV. Below a 1-mm threshold, the measurement failed to represent the change of the real dimensions.

Conclusion: Measurement accuracy was strongly influenced by the specific reconstruction kernel utilized. For accurate measurement, standardization of the imaging protocol and selection of the appropriate anatomic level are essential.

Show MeSH
Software design and image display.A. Schematic workflow diagram for full width at half maximum measurement approach.B. Image overlay of measured results on CT images of physical airway phantom (airway wall, dark gray; lumen, gray; average wall in case of wall outside 2SD, bright gray).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2627250&req=5

Figure 2: Software design and image display.A. Schematic workflow diagram for full width at half maximum measurement approach.B. Image overlay of measured results on CT images of physical airway phantom (airway wall, dark gray; lumen, gray; average wall in case of wall outside 2SD, bright gray).

Mentions: In-house software was developed at the Asan Medical Center for airway measurement. Figure 2 shows a schematic diagram of the data processing and analysis. The software was developed for interactively analyzing pulmonary images and for providing measurement tools for the quantification of the airways. For each of the eleven phantom tubes, the software provided a graphical user interactive interface to identify the approximate airway center by pointing with the use of a computer mouse. For pre-processing, the software magnified the image ten times using a bi-cubic interpolation algorithm, segmented the airway lumen using the lumen threshold parameter (HU), and calculated the center of inertia of the airway lumen. In general, the FWHM algorithm is regarded as a robust interpolation algorithm. The software was then used to make the half-maximum measurements and to collect the gray-level profiles for airway dimension measurements. By analyzing the 120 rays cast around 360 degrees from the center point, the luminal radius (inner boundary), outer boundary, wall thickness, and wall thickness inside two standard deviations (SD) were measured using the FWHM method which is one of the most typical airway wall measurement algorithms (6-8, 13). On a 10 × magnified image, the pixel values were interpolated along the ray using a bi-linear algorithm. Final estimations included the area of the airway lumen, the lumen radius, mean wall thickness, and mean wall area.


Semi-automatic measurement of the airway dimension by computed tomography using the full-width-half-maximum method: a study on the measurement accuracy according to the CT parameters and size of the airway.

Kim N, Seo JB, Song KS, Chae EJ, Kang SH - Korean J Radiol (2008 May-Jun)

Software design and image display.A. Schematic workflow diagram for full width at half maximum measurement approach.B. Image overlay of measured results on CT images of physical airway phantom (airway wall, dark gray; lumen, gray; average wall in case of wall outside 2SD, bright gray).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2627250&req=5

Figure 2: Software design and image display.A. Schematic workflow diagram for full width at half maximum measurement approach.B. Image overlay of measured results on CT images of physical airway phantom (airway wall, dark gray; lumen, gray; average wall in case of wall outside 2SD, bright gray).
Mentions: In-house software was developed at the Asan Medical Center for airway measurement. Figure 2 shows a schematic diagram of the data processing and analysis. The software was developed for interactively analyzing pulmonary images and for providing measurement tools for the quantification of the airways. For each of the eleven phantom tubes, the software provided a graphical user interactive interface to identify the approximate airway center by pointing with the use of a computer mouse. For pre-processing, the software magnified the image ten times using a bi-cubic interpolation algorithm, segmented the airway lumen using the lumen threshold parameter (HU), and calculated the center of inertia of the airway lumen. In general, the FWHM algorithm is regarded as a robust interpolation algorithm. The software was then used to make the half-maximum measurements and to collect the gray-level profiles for airway dimension measurements. By analyzing the 120 rays cast around 360 degrees from the center point, the luminal radius (inner boundary), outer boundary, wall thickness, and wall thickness inside two standard deviations (SD) were measured using the FWHM method which is one of the most typical airway wall measurement algorithms (6-8, 13). On a 10 × magnified image, the pixel values were interpolated along the ray using a bi-linear algorithm. Final estimations included the area of the airway lumen, the lumen radius, mean wall thickness, and mean wall area.

Bottom Line: The measured values as determined by CT and the actual dimensions of the tubes were compared.There was no significant difference in accuracy among images with the use of variable slice thicknesses or a variable FOV.Below a 1-mm threshold, the measurement failed to represent the change of the real dimensions.

View Article: PubMed Central - PubMed

Affiliation: Department of Radiology and Research Institute of Radiology, University of Ulsan College of Medicine, Asan Medical Center, Seoul, Korea.

ABSTRACT

Objective: To assess the influence of variable factors such as the size of the airway and the CT imaging parameters such as the reconstruction kernel, field-of-view (FOV), and slice thickness on the automatic measurement of airway dimension.

Materials and methods: An airway phantom was fabricated that contained eleven poly-acryl tubes of various lumen diameters and wall thicknesses. The measured density of the poly-acryl wall was 150 HU, and the measured density of the airspace filled with polyurethane foam was -900 HU. CT images were obtained using a 16-MDCT (multidetector CT) scanner and were reconstructed with various reconstruction kernels, thicknesses and FOV. The luminal radius and wall thickness were measured using in-house software based on the full-width-half-maximum method. The measured values as determined by CT and the actual dimensions of the tubes were compared.

Results: Measurements were most accurate on images reconstructed with use of a standard kernel (mean error: -0.03 +/- 0.21 mm for wall thickness and -0.12 +/- 0.11 mm for the luminal radius). There was no significant difference in accuracy among images with the use of variable slice thicknesses or a variable FOV. Below a 1-mm threshold, the measurement failed to represent the change of the real dimensions.

Conclusion: Measurement accuracy was strongly influenced by the specific reconstruction kernel utilized. For accurate measurement, standardization of the imaging protocol and selection of the appropriate anatomic level are essential.

Show MeSH