Encrypted login | home

Program Information

Comparisons and Limitations of Physics Instrumentation in Mammography


L Brateman

L Brateman1*, P Heintz2, (1) University of Florida, Gainesville, FL, (2) Univ New Mexico Radiology Dept., Albuquerque, NM

SU-C-116-3 Sunday 1:00PM - 1:55PM Room: 116

Purpose: To evaluate discrepancies in measurements for two major vendors of mammography equipment with a few models of physics instruments, and to make known potential errors in best use of these instruments for ACR and MQSA/FDA documentation. We show variability in responses under different beam conditions and report problems encountered. A separate study shows effects on calculated mean glandular dose.

Methods: Recent mammography equipment evaluations for new installations show discrepancies in measurements of kVp and HVL with different instrumentation. This study tested well-behaved GE Essential and Hologic Dimensions units (6 target-filter combinations) with 3 vendors and 4 instruments (3 with solid-state detectors, SStDs) using recommended techniques, which sometime conflict. HVLs were measured two ways: (a) single-point measurements for the SStDs, and for some instruments, (b) with standardized mammography-grade aluminum filtration. Particular attention was paid to HVLs near values required for mammography phantom mean glandular dose calculations (28-30 kVp). Ionization chamber measurements with Al filtration, according to ACR recommendations, were used as reference HVLs for 28, 29 and 30 kVp.

Results: Measurements of kVp with SStD instruments were most consistent with Mo/Mo, with a 0.18 kV average discrepancy for all instruments (discrepancy range (-0.3,+0.4). For Mo/Mo, one vendor was consistently high and 2 low, but this trend was inconsistent among other combinations. The range of discrepancies was within 1 kV for all except Mo/Rh for one instrument and W/Ag for another. Unlike kVp, HVL measurements showed large variations. Over 28-30 kVp, discrepancies of single-point HVLs ranged from approximately zero to nearly 15% from reference values. It was impossible to follow manufacturers or ACR HVL measurement recommendations with aluminum for 2 of the 3 SStDs.

Conclusion: : All instruments appear to function well, but data show biases which affect accuracy. Users must be careful how they interpret their data, particularly for HVL.

Contact Email: