MO‐A‐141‐02: Session In Memory of Fearghus O't Foghludha – Virtual Tools for Validation of X‐Ray Breast Imaging Systems

Abstract

Given the large number of design parameters available to an x‐ray imaging system innovator (x‐ray spectrum, number of projection angles, angular range, dose and reconstruction algorithm, to name a few), simulation methods offer significant advantages over clinical studies for evaluation of the alternatives in terms of reproducibility, reduced radiation exposures of patients and volunteers, a known reference standard, and the capability for studying patient and disease subpopulations through appropriate choice of breast and patient model parameters. For these reasons, the emergent use of Virtual Clinical Trials for preclinical assessment of x‐ray breast imaging methods has introduced a demand to optimize protocols for simulation studies. The protocols should comprise of specifications for breast phantoms, the simulated data representation, methods for incorporating phantoms into a model for the imaging process, and appropriate statistical assessment methods applied to the resulting simulated images. This symposium explores simulation methods for evaluation of novel x‐ray breast imaging systems — the subject of recently established AAPM taskgroup TG234. The focus is on the various approaches of developing software anthropomorphic phantoms and their use in the statistical assessment of novel imaging systems along with computational models for the x‐ray image formation process. The symposium will review and discuss the state of the science of Virtual Clinical Trials for novel x‐ray breast imaging systems. Presentations will discuss real‐time simulation of breast anatomy, simulation of small‐scale tissue properties, generation of breast phantoms based upon clinical datasets, and task‐based evaluation of x‐ray breast imaging systems using software phantoms.Learning Objectives:1. Review current approaches for simulation of breast phantoms for x‐ray imaging, along with their strengths and limitations;2. Understand the validation of software phantoms using object and image statistics;3. Review methods for statistical and task‐based image quality assessment and discuss their applications in x‐ray breast imaging systems;4. Illustrate methods for performing virtual clinical trials of emerging 3D x‐ray breast imaging systems.P.R.B. is a PI on NSF grant IIS0916690 and co‐investigator on NIH R01 CA154444, NIH R21 CA155906‐01A1, and DoD HBCU PTA BC083639, and is also involved in research collaboration with Barco, Inc. and Hologic, Inc.N.K. is funded by NIH R01 CA134658 and R01 CA11243, and is involved in resaerch collaboration with Siement Medical Solution.