Evaluating the Figure of Merit in mammography utilizing Monte Carlo simulation

Mammography represents the most powerful screening tool available for the early detection of breast cancer. As a screening technique, it addresses to asymptomatic women and therefore the implementation of the 'As Low As Reasonably Achievable' (ALARA) principle is of increased importance. The overall optimization of any radiographic procedure requires the definition of a suitable Figure of Merit (FOM), based on a representative combination of image quality and dose characteristics. For the present study, Monte Carlo simulation of the mammographic process was utilized to derive the energy deposition inside the breast phantom and the signal beneath it. Certain signal characteristics (Subject Contrast (SC), Contrast to Noise Ratio (CNR), squared CNR) were studied (normalized to the output signal) and combined with appropriate dose indices (Entrance Surface Dose (ESD), Average Glandular Dose (AGD), Mid Plane Dose (MPD)) to extract an appropriate FOM index, suitable for spectrum based optimization. Results, for different breast glandularities and lesion sizes and compositions, demonstrate that the choice of the optimum spectrum strongly depends not only on the breast and lesion characteristics, but also on the selection of the utilized FOM index. For screen-film mammography, where the primary image quality index is considered the SC, the traditionally used Mo and Rh-anode spectra demonstrate improved overall performance, despite their inferior dosimetric characteristics compared to W-anode spectra. If digital mammography is considered, where the CNR is of primary importance, W-anode spectra demonstrate a noticeable improved performance