Quantitative characterization of ex vivo breast tissue via x-ray phase-contrast tomography

J Xray Sci Technol. 2019;27(3):503-516. doi: 10.3233/XST-180453.

Abstract

Background: Grating-based X-ray phase-contrast imaging (GPCI) has received growing interests in recent years due to its high capability of visualizing soft tissue. Breast imaging is one of the most promising candidates for the first clinical application of this imaging modality.

Objective: In this work, quantitative breast tissue characterization based on GPCI computed tomography (CT) is investigated with a laboratory X-ray tube through a comparison between attenuation-based CT images and phase-contrast CT images.

Methods: The Hounsfield units (HU) scale was introduced to phase-contrast images due to its wide application in clinical medicine. In this work, instead of water, plastic cylinders composed of polyethylene terephthalate (PET) was treated as the calibration material. An alternative test-retest reliability (TRR) was presented to evaluate the repeatability of GPCI. Comparison between attenuation-based CT imaging and GPCI CT imaging was operated with the use of statistical analysis methods like histograms and receiver operating characteristic (ROC) curves.

Results: The determined mean TRR related to cylinders is slightly larger in phase-contrast imaging (0.93) than that in attenuation-based imaging (0.89). With respect to distinguishing breast tissues, the AUC (area under curve) values of ROC curves of phase-contrast images are higher than that of attenuation-based images.

Conclusions: An ex vivo study of GPCI shows that it is a stable imaging modality for visualizing the breast tissue with good repeatability, and that it could be of potential for the diagnosis of breast cancer as well.

Keywords: Breast imaging; Hounsfield units; phase-contrast tomography.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Breast / diagnostic imaging*
  • Calibration
  • Female
  • Humans
  • Mammography / methods*
  • ROC Curve
  • Reproducibility of Results
  • Tomography, X-Ray Computed / methods*