Breast Density Quantification Using Liquid Phantoms with Dual-Energy Mammography
A Lam Ng*, h ding, S Molloi, University of California, Irvine, CASU-C-116-7 Sunday 1:00PM - 1:55PM Room: 116
Purpose:Breast density has shown to be an important indicator for breast cancer risk. A computational calibration study using breast-equivalent liquid phantoms in dual-energy mammography is investigated to determine its feasibility and improvement in accuracy for breast density quantification.
Methods: A computational model simulates a dual energy mammography system which is equipped with a tungsten anode x-ray tube and acquires dual energy images at 28 kVp with a 50 μm rhodium filter and 49 kVp with a 300 μm copper filter. Liquid phantoms, consisting of mixtures of water (H2O), glycerol (C3H6O3) and isopropyl alcohol (C3H6O) were used to simulate the X-ray attenuation properties of breast tissues. Calibration was accomplished with the simulated dual energy signals from the liquid phantoms using a rational polynomial function. The accuracy of liquid phantom calibration was tested with different phantom configurations of various thicknesses (2-8 cm) and densities (0%-100%), along with simulated glandular and adipose tissues. The performance of the liquid phantom calibration was compared to standard CIRS plastic phantom calibration by evaluating the errors in breast density quantification in simulated breast tissues.
Results:Simulation results indicate that a mixture of liquid materials based on hydrogen, carbon and oxygen can be used to create glandular and adipose equivalent phantoms. For glandular and adipose tissues H2O, C3H6O3 and C3H6O percentages were estimated to be 42.68%, 22.62%, 34.69% and 0.11%, 16.37%, 83.51%, respectively. RMS errors for thicknesses were smaller (0.063 mm for glandular and 0.072 mm for adipose) using liquid phantoms than the results reported with CIRS phantoms. RMS errors for variable glandular densities were 1.93%.
Conclusion:The results of this work indicate that a proper selection of liquid mixtures can be used as breast equivalent calibration phantoms in dual-energy mammography, which may potentially improve the accuracy in dual energy breast density quantification.