The task of correctly identifying suspicious masses in the breast is often complicated by the confusing overlap of anatomical features present in the projection images provided by mammography. For that reason, there has been growing interest in obtaining volumetric information through a technique called digital breast tomosynthesis (DBT). DBT has shown promise for improving sensitivity and specificity in breast imaging compared to mammography by providing depth information that helps to distinguish overlapping breast tissue. This technique is most commonly performed using the technology of active matrix, flat-panel imagers (AMFPIs) which incorporate an x-ray converter (in the form of a layer of cesium iodide (CsI:Tl) or amorphous selenium (a-Se) deposited on a pixelated array) to detect the incident radiation. A single DBT view typically involves acquisition of 9 to 25 projection images and it is clinically desirable that the cumulative dose per view be no more than that of a single-view mammogram. However, the relatively modest amount of signal generated by each X ray in CsI:Tl and a-Se leads to technical limitations that constrain the conditions under which the projection images are acquired ? which can significantly affect image quality as well as impede efforts to minimize dose. To overcome this problem, the proposed research seeks to develop an alternative form of x- ray converter based on polycrystalline mercuric iodide and fabricated using a screen-print method ? referred to as SP HgI2. SP HgI2 has been shown to be capable of providing at least three times more signal per X ray than a-Se or CsI:Tl and, as a result, would help to address technical limitations in AMFPIs used for DBT. However, further development of this novel converter is required to improve other important converter and imager-related properties (including the reduction of charge trapping effects and non-uniformity in pixel-to-pixel signal response).
The specific aims of the project therefore focus on achieving a high level of performance for each of these properties.
These aims will be accomplished through systematic investigation of a number of strategies chosen to improve different aspects of converter and imager performance. The strategies consist of modifications to the methods used to fabricate the SP HgI2 converters combined with the introduction of a grid structure into those converters. The methodology of the research will be based on iterative design, fabrication and evaluation of prototype converters employing these strategies ? with the information obtained from a given set of prototypes used to improve subsequent prototypes. The converters will be deposited on simple detector substrates (which facilitate relatively rapid evaluation of a limited number of properties) as well as on AMFPI arrays (which allow complete performance characterization). The expected impact of the new converter technology made possible by the research will be a significant improvement in the effectiveness of digital breast tomosynthesis, including possible dose reduction ? along with the potential for facilitating dose reduction in other applications such as breast CT, fluoroscopy and cone-beam CT in radiation therapy.
Digital breast tomosynthesis (DBT) provides useful depth information that can alleviate the problem of tissue overlap in conventional mammography. However, the modest signal performance of many DBT imagers constrains their operation. The proposed development of a novel form of x-ray converter would overcome those limitations, leading to better image quality (which could improve specificity and sensitivity) as well as the possibility of lower dose (without impeding diagnosis) ? thereby improving the effectiveness and reducing the radiation risk of breast screening.
