LAUSR

Purpose To implement a Quality Assurance program for mammography integrating the dose-tracking software Radimetrics™ (Bayer) in a workflow for a quick assessment of equipment performance and patient dose reduction. Methods A setting was created in Radimetrics™ [1] to collect the acquisition parameters and dosimetric data of all the mammographies performed. However, to implement an optimization program, a deeper level of data stratification was needed. A tool was developed in Matlab®, to separately store the data exported from Radimetrics™ according to the acquisition type (2D/3D), control mode (manual/automatic), equipment, operator. The tool automatically produces reports and graphs synthesizing acquisition and dosimetric parameters (anode-filter coupling, kV, mAs, compression thickness, Entrance Dose ED, Average Glandular Dose AGD) for each mammography unit, highlighting possible outliers and/or changes of the equipment performance. Data of 70806 acquisitions were collected. Dosimetric data of manual acquisitions of prothesis using producer indications were compared to those obtained with parameters chosen by the technicians according to experience. Data from automatic acquisition with three different SenoEssential equipments (GE Healthcare) were compared in terms of acquisition settings (% of Mo-Mo, Mo-Rh and Rh-Rh, minimum–maximum values of kV and mAs) and dose, also comparing AGD with the achievable and acceptable levels reported in the European Guidelines [2] . Results For prothesis acquisitions, lower AGD values were delivered using parameters suggested by manufacturer for compression thickness During automatic acquisitions of 2D projections, the use of the bucky dedicated to tomosynthesis lead to thickness underestimation, and to a higher proportion of Mo-Rh acquisitions delivering AGD above the achievable threshold (Figure). Conclusions For prothesis acquisitions, new indications were provided to the technicians suggesting the optimal settings resulted for each compression thickness interval. According to our results, use of tomosynthesis bucky for 2D acquisitions should be discouraged.