Program Information
Automated Uniformity Measurements Using a Comprehensive Tomosynthesis QA Phantom
H Olafsdottir1*, D Goodenough2 , G Oskarsdottir3 , I Olafsson4 , J Fredriksson5 , R Mallozzi6 , A Healy7 , A Kristbjornsson8 , S Kristinsson9 , J Levy10 , (1) Raforninn Ehf., Reykjavik, Gullbringusysla, (2) George Washington University, Washington Dc, Washington, (3) Raforninn Ehf., Reykjavik, Gullbringusysla, (4) Raforninn Ehf., Reykjavik, Gullbringusysla, (5) Raforninn Ehf., Reykjavik, Gullbringusysla, (6) The Phantom Laboratory, Salem, New York, (7) The Phantom Laboratory, Salem, New York, (8) Raforninn Ehf., Reykjavik, Gullbringusysla, (9) Raforninn Ehf., Reykjavik, Gullbringusysla, (10) The Phantom Laboratory, Salem, New York
Presentations
SU-E-P-50 (Sunday, July 12, 2015) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose:
The purpose of this study is to look at an automated approach to measure uniformity in DBT using a comprehensive QA phantom versus using a separate uniformity phantom. Our question is whether the uniformity can be reasonably measured within the backdrop of the other test objects?
Methods:
Using the Tomophan ™ from The Phantom Laboratory, both the regional uniformity and the global uniformity was studied.
The uniform ROIs are selected carefully to minimize the effect from other test targets.To measure the regional uniformity two large (10mm radius) regions of interest (ROI) are placed in an anterior and lateral position within the phantom image. The locations of the ROIs are shown in figures 1 and 2. Both the mean value and standard deviation of the pixel value within those regions is computed and observed in the slice (z) direction.
The results of the global uniformity test indicate how well the imaging system performs in preserving the uniformity within a given slice. The global uniformity looks at the mean value and standard deviation of the pixel value within five ROIs all with a radius of 5 mm placed at various locations within the phantom. The five smaller ROIs indicate how well the imaging device performs in preserving the uniformity within the slice. Because tomosynthesis systems do not use a set pixel value scale, such as the Hounsfield scale in CT, the percentage of the maximum absolute difference between uniformity regions was used for vendor comparison.
Results:
Comparison of 49 cases from six different vendors showed that the uniformity measurements can differ somewhat between scans for the same vendor and between vendors.
Conclusion:
By carefully selecting regions of interest, valuable information can be gained offering information on tomosynthesis image uniformity by using phantom that also contains other targets.
Funding Support, Disclosures, and Conflict of Interest: Research supported by Image Owl, Inc., The Phantom Laboratory, Inc. and Raforninn ehf. Authors Mallozzi and Healy employed by The Phantom Laboratory, Inc. Author Goodenough is a consultant to The Phantom Laboratory, Inc. Authors Fredriksson, Kristbjornsson, Olafsson, Oskarsdottir, and Olafsdottir are employeed by Raforninn ehf.
Contact Email: