Energy-sensitive scatter estimation and correction for spectral x-ray imaging with photon-counting detectors

TL;DR

This paper introduces an energy-dependent scatter estimation and correction model for spectral x-ray imaging with photon-counting detectors, improving quantitative accuracy in applications like spectral mammography without altering acquisition protocols.

Contribution

It proposes a novel scatter estimation model that accounts for energy-dependent scatter characteristics, enhancing spectral x-ray imaging accuracy with photon-counting detectors.

Findings

Energy-sensitive scatter correction improves iodine density estimation.

The model accurately accounts for energy-dependent scatter distribution.

No modifications to acquisition parameters are needed.

Abstract

As photon counting detectors are being explored for medical and industrial imaging applications, there is a critical need to understand spectral characteristics of scattered x-ray photons. Scattered radiation is detrimental to x-ray imaging by reducing image quality and quantitative accuracy. While various scatter correction techniques have been proposed for x-ray imaging with conventional energy-integrating detectors, additional efforts are required to develop approaches for spectral x-ray imaging with energy-resolving PCDs. We show the benefits of accurate scatter estimation and correction for each energy bin when using a photon counting detector. We propose a scatter estimation model that accounts for the energy-dependent scatter characteristics in projection x-ray imaging. This can then be used to restore quantitative accuracy for spectral x-ray imaging with PCDs. Results are shown in the context of contrast-enhanced spectral mammography using dual-energy subtraction to digitally isolate iodine targets. In the presence of scatter, the projected iodine densities are increasingly underestimated as the object thickness increases. The energysensitive scatter correction improves the iodine density estimation. These results suggest that our scatter estimation model can accurately account for the energy-dependent scatter distribution, which can be an effective tool for scatter compensation in spectral x-ray imaging. Implementing this scatter estimation model does not require any modifications to the acquisition parameters and is transferable to other x-ray imaging applications such as tomosynthesis and CT.