Performance of semiconductor dosimeters with a range of radiation qualities used for mammography: A calibration laboratory study.

PURPOSE: To investigate the radiation quality dependence of the response of commercial semiconductor-based dosimeters, and to estimate potential errors and uncertainties related to different measurement and calibration scenarios. METHODS: All measurement results were compared to reference values measured at the IAEA dosimetry laboratory which is traceable to the international system of units (SI). Energy dependence of the response of eight semiconductor dosimeters were determined for five different anode-filter combinations and tube voltages from 25 to 35 kV. For systems capable of deriving half value layer (HVL) and tube voltage from measurements, calibration coefficients for these measurements were calculated. RESULTS: For six dosimeters, the maximum deviations from the reference value of the air kerma measurement were within ±5% as required by IEC 61674. Calibration coefficients for radiation qualities (anode-filter and tube voltage combinations) relative to reference radiation quality Mo-Mo 28 kV deviate up to 12%. HVL and tube voltage measurements exhibited deviations up to 11% and 10%, respectively. CONCLUSIONS: The air kerma responses of modern semiconductor dosimeters have a small energy dependence. However, no dosimeter tested complied with the accuracy limits stated by the manufacturer for tube voltage measurements, and only two dosimeters complied with the limits for HVL measurements. Absolute measurement of HVL and tube voltage with semiconductor dosimeters have to be verified for actual clinical radiation conditions on clinical mammography systems. Semiconductor dosimeters can be used for quality control measurements if individual calibration coefficients are available for the radiation condition applied. If other conditions are applied, additional uncertainty needs to be considered, particularly in the case of HVL and tube voltage measurements.

IAEA-SSDL bilateral comparisons for diagnostic level air kerma measurement standards.

Comparisons of national standard of air kerma for conventional and mammographic diagnostic X-ray radiation qualities were conducted by the IAEA. Eleven secondary standards dosimetry laboratories provided calibration data for Exradin A3 and Radcal RC6M transfer ionization chambers circulated. Each comparison result expressed as the ratio of the participant and IAEA calibration coefficient were within the acceptance limit of ±2.5%. From the 67 results of 11 participants and 10 available beam qualities, the comparison result was within its standard uncertainty in 63 cases, and within the expanded (k = 2) uncertainty in four cases. Detailed calibration uncertainty budgets from participant laboratories are presented. The relative standard calibration uncertainty of each participant was in the range of 0.5-1.3%. These results indicate that the calibration related uncertainty component is reasonable low for a clinical measurement. In addition to the calibration coefficient, other corrections should be applied for clinical measurement to achieve the recommended accuracy.

Comparison of pencil-type ionization chamber calibration results and methods between dosimetry laboratories.

A comparison of calibration results and procedures in terms of air kerma length product, PKL, and air kerma, K, was conducted between eight dosimetry laboratories. A pencil-type ionization chamber (IC), generally used for computed tomography dose measurements, was calibrated according to three calibration methods, while its residual signal and other characteristics (sensitivity profile, active length) were assessed. The results showed that the "partial irradiation method" is the preferred method for the pencil-type IC calibration in terms of PKL and it could be applied by the calibration laboratories successfully. Most of the participating laboratories achieved high level of agreement (>99%) for both dosimetry quantities (PKL and K). Estimated relative standard uncertainties of comparison results vary among laboratories from 0.34% to 2.32% depending on the quantity, beam quality and calibration method applied. Detailed analysis of the assigned uncertainties have been presented and discussed.