Distinguishing the indistinguishable : Identification of signal aliasing with a novel x - ray detector design for improving image signal - to - noise ratio

Introduction: Breast cancer is the most prevalent type of cancer and the second leading cause of mortality amongst Canadian women [1]. Regular mammography screening is currently the best method of early detection. Recently, a 10-year study of the Ontario Breast Screening Program revealed that not all screening centers are the same [2]. In particular, sites with digital x-ray detectors that acquired images with higher signal-to-noise ratios (SNR) had significantly higher cancer detection rates (44% increase) [3]. The ability of an x-ray detector to produce high SNR images for a given amount of radiation is described by the detective quantum efficiency (DQE). DQE is a Fourier-based metric and function of spatial-frequency, where low frequencies correspond to large image features and high frequencies to fine detail. The Apodized Aperture Pixel (AAP) is a novel detector design proposed by our group [4] that more than doubles DQE at high frequencies by increasing contrast and eliminating aliasing. However, while reducing noise aliasing always improves DQE, it is not obvious whether aliasing helps or hinders visualization of fine structures and classification of micro-calcification morphology, an indicator of cell grade and malignancy risk [5]. In this presentation, I describe a method of controlling aliasing with the AAP design and investigate aliasing effects on detectability in x-ray images.