Program Information
Quantifying the Amount of Missing Tissue in a Digital Breast Tomosynthesis
D Goodenough1 , H Olafsdottir2 , I Olafsson3*, J Fredriksson4 , R Mallozzi5 , S Kristinsson6 , A Healy7 , G Oskarsdottir8 , A Kristbjornsson9 , J Levy10 , (1) George Washington University, Washington Dc, Washington, (2) Raforninn Ehf., Reykjavik, Gullbringusysla, (3) Raforninn Ehf., Reykjavik, Gullbringusysla, (4) Raforninn Ehf., Reykjavik, Gullbringusysla, (5) The Phantom Laboratory, Salem, New York, (6) Raforninn Ehf., Reykjavik, Gullbringusysla, (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-31 (Sunday, July 12, 2015) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose:
To automatically quantify the amount of missing tissue in a digital breast tomosynthesis system using four stair-stepped chest wall missing tissue gauges in the Tomophan™ from the Phantom Laboratory and image processing from Image Owl.
Methods:
The Tomophan™ phantom incorporates four stair-stepped missing tissue gauges by the chest wall, allowing measurement of missing chest wall in two different locations along the chest wall at two different heights. Each of the four gauges has 12 steps in 0.5 mm increments rising from the chest wall. An image processing algorithm was developed by Image Owl that first finds the two slices containing the steps then finds the signal through the highest step in all four gauges. Using the signal drop at the beginning of each gauge the distance to the end of the image gives the length of the missing tissue gauge in millimeters.
Results:
The Tomophan™ was imaged in digital breast tomosynthesis (DBT) systems from various vendors resulting in 46 cases used for testing. The results showed that on average 1.9 mm of 6 mm of the gauges are visible. A small focus group was asked to count the number of visible steps for each case which resulted in a good agreement between observer counts and computed data.
Conclusion:
First, the results indicate that the amount of missing chest wall can differ between vendors. Secondly it was shown that an automated method to estimate the amount of missing chest wall gauges agreed well with observer assessments. This finding indicates that consistency testing may be simplified using the Tomophan™ phantom and analysis by an automated image processing named Tomo QA. In general the reason for missing chest wall may be due to a function of the beam profile at the chest wall as DBT projects through the angular sampling.
Funding Support, Disclosures, and Conflict of Interest: Research supported by Image Owl, Inc., The Phantom Laboratory, Inc. and Raforninn ehf. Mallozzi and Healy employed by The Phantom Laboratory, Inc. Goodenough is a consultant to The Phantom Laboratory, Inc. Fredriksson, Kristbjornsson, Olafsson, Oskarsdottir and Olafsdottir are employed by Raforninn, Ehf
Contact Email: