Optimization of dual‐energy imaging systems using generalized NEQ and imaging task

Authors:
  • JSJ. H. Siewerdsen
    J. H. Siewerdsen
Publisher website
Google Scholar
Abstract
Dual‐energy (DE) imaging is a promising advanced application of flat‐panel detectors (FPDs) with a potential host of applications ranging from thoracic and cardiac imaging to interventional procedures. The performance of FPD‐based DE imaging systems is investigated in this work by incorporating the noise‐power spectrum associated with overlying anatomical structures (“anatomical noise” modeled according to acharacteristic) into descriptions of noise‐equivalent quanta (NEQ) to yield the generalized NEQ (GNEQ). Signal and noise propagation in the DE imaging chain is modeled by cascaded systems analysis. A Fourier‐based description of the imaging task is integrated with the GNEQ to yield a detectability index used as an objective function for optimizing DE image reconstruction, allocation of dose between low‐ and high‐energy images, and selection of low‐ and high‐kVp. Optimal reconstruction and acquisition parameters were found to depend on dose; for example, optimal kVp varied fromat typical radiographic dose levels (entrance surface dose, ESD) but increased toat high dose. At very low dose, detectability index indicates an optimal low‐energy technique ofbut was largely insensitive to the choice of high‐kVp in the range. Similarly, optimal dose allocation, defined as the ratio of low‐energy ESD and the total ESD, varied from 0.2 to 0.4 over the range. Furthermore, two applications of the theoretical framework were explored: (i) the increase in detectability for DE imaging compared to conventional radiography; and (ii) the performance of single‐shot vs double‐shot DE imaging, wherein the latter is found to have a DQE approximately twice that of the former. Experimental and theoretical analysis of GNEQ and task‐based detectability index provides a fundamental understanding of the factors governing DE imaging performance and offers a framework for system design and optimization.
Funding Information
  • National Institutes of Health (R01‐CA112163‐01)

This publication has 46 references indexed in Scilit: