Solid-state dosimeters: a new approach for mammography measurements

To compare responses of modern commercially available solid-state dosimeters (SStDs) used in mammography medical physics surveys for two major vendors of current digital mammography units. To compare differences in dose estimates among SStD responses with ionization chamber (IC) measurements for several target/filter (TF) combinations and report their characteristics. To review scientific bases for measurements of quantities required for mammography for traditional measurement procedures and SStDs.

SStDs designed for use with modern digital mammography units were acquired for evaluation from four manufacturers. Each instrument was evaluated under similar conditions with the available mammography beams provided by two modern full-field digital mammography units in clinical use: a GE Healthcare Senographe Essential (Essential) and a Hologic Selenia Dimensions 5000 (Dimensions), with TFs of Mo/Mo, Mo/Rh; and Rh/Rh and W/Rh, W/Ag, and W/Al, respectively. Measurements were compared among the instruments for the TFs over their respective clinical ranges of peak tube potentials for kVp and half-value layer (HVL) measurements. Comparisons for air kerma (AK) and their associated relative calculated average glandular doses (AGDs), i.e., using fixed mAs, were evaluated over the limited range of 28-30 kVp. Measurements were compared with reference IC measurements for AK, reference HVLs and calculated AGD, for two compression paddle heights for AK, to evaluate scatter effects from compression paddles. SStDs may require different positioning from current mammography measurement protocols.

Measurements of kVp were accurate in general for the SStDs (within -1.2 and +1.1 kVp) for all instruments over a wide range of set kVp's and TFs and most accurate for Mo/Mo and W/Rh. Discrepancies between measurements and reference values were greater for HVL and AK. Measured HVL values differed from reference values by -6.5% to +3.5% depending on the SStD and TF. AK measurements over limited (28-30) kVp's ranged from -6% to +7% for SStDs, compared with IC reference values. Relative AGDs for each SStD using its associated measurements of kVp, HVL and AK underestimated AGD in nearly all cases, compared with reference IC values, with discrepancies of <-1% to ∼-10%. Some differences in AGD were related to differences in contributions of compression paddle scatter to AK measurements made by ICs. Applying measured factors for scatter effects in AK measurements for three SStDs reduced discrepancies between -6.2% and +1.3%, shifting AGDs from SStDs closer to IC AGDs.

This study revealed that SStD measurements yielded good agreement with set kVp, poor agreement with standard HVL determinations, and AK measurements that were substantially different from IC measurements. Discrepancies are partly related to the scattered radiation measured by ICs in determining AK. As a result, IC measurements required for AGD, using currently accepted methodology, typically result in higher AGDs than SStDs, because current methodologies do not account for differing instrument responses to scatter. HVLs reported by SStDs contribute to discrepancies in calculated AGD that depend on kVp and TF. Medical physicists are encouraged to compare their results for SStD instruments using a similar methodology for potential discrepancies with their traditional instruments.