Optimization of breast cancer detection in Dual Energy X-ray Mammography using a CMOS imaging detector

Dual energy mammography has the ability to improve the detection of microcalcifications leading to early diagnosis of breast cancer. In this simulation study, a prototype dual energy mammography system, using a CMOS based imaging detector with different X-ray spectra, was modeled. The device consists of a 33.91 mg/cm(2) Gd2O2S:Tb scintillator screen, placed in direct contact with the sensor, with a pixel size of 22.5 mu m. Various filter materials and tube voltages of a Tungsten (W) anode for both the low and high energy were examined. The selection of the filters applied to W spectra was based on their K-edges (K-edge filtering). Hydroxyapatite (HAp) was used to simulate microcalcifications. Calcification signal-to-noise ratio (SNRtc) was calculated for entrance surface dose within the acceptable levels of conventional mammography. Optimization was based on the maximization of SNRtc while minimizing the entrance dose. The best compromise between SNRtc value and dose was provided by a 35kVp X-ray spectrum with added beam filtration of 100 mu m Pd and a 70kVp Yb filtered spectrum of 800 mu m for the low and high energy, respectively. Computer simulation results show that a SNRtc value of 3.6 can be achieved for a calcification size of 200 mu m. Compared with previous studies, this method can improve detectability of microcalcifications.