Anatomical noise in contrast-enhanced digital mammography. Part II. Dual-energy imaging

• Published in: Medical physics (Lancaster) 2013-08, Vol.40 (8), p.081907-n/a
• Main Authors: Hill, Melissa L., Mainprize, James G., Carton, Ann-Katherine, Saab-Puong, Sylvie, Iordache, Răzvan, Muller, Serge, Jong, Roberta A., Dromain, Clarisse, Yaffe, Martin J.
• Format: Article
• Language: English
• Subjects: Adult; Aged; Aged, 80 and over; anatomical noise; ANIMAL TISSUES; beta; BIOMEDICAL RADIOGRAPHY; Cancer; CONTRAST MEDIA; contrast‐enhanced; DATA ACQUISITION; Data acquisition and logging; diagnostic radiography; Digital computing or data processing equipment or methods, specially adapted for specific applications; Digital mammography; Digital radiography; dual‐energy; Edge enhancement; Female; Humans; Image data processing or generation, in general; image denoising; image enhancement; Image enhancement or restoration, e.g. from bit‐mapped to bit‐mapped creating a similar image; IMAGE PROCESSING; IODINE; MAMMARY GLANDS; mammography; Mammography - methods; Medical image artifacts; Medical image contrast; Medical image noise; medical image processing; Medical imaging; Medical X‐ray imaging; Middle Aged; Modulation transfer functions; NEOPLASMS; Noise; power‐law; Quantum noise; Radiographic Image Enhancement - methods; RADIOLOGY AND NUCLEAR MEDICINE; Signal-To-Noise Ratio; tumours; X‐ray imaging; X‐ray spectra
• ISSN: 0094-2405; 2473-4209
• DOI: 10.1118/1.4812681

Purpose: Dual-energy (DE) contrast-enhanced digital mammography (CEDM) uses an iodinated contrast agent in combination with digital mammography (DM) to evaluate lesions on the basis of tumor angiogenesis. In DE imaging, low-energy (LE) and high-energy (HE) images are acquired after contrast administration and their logarithms are subtracted to cancel the appearance of normal breast tissue. Often there is incomplete signal cancellation in the subtracted images, creating a background “clutter” that can impair lesion detection. This is the second component of a two-part report on anatomical noise in CEDM. In Part I the authors characterized the anatomical noise for single-energy (SE) temporal subtraction CEDM by a power law, with model parametersα and β. In this work the authors quantify the anatomical noise in DE CEDM clinical images and compare this with the noise in SE CEDM. The influence on the anatomical noise of the presence of iodine in the breast, the timing of imaging postcontrast administration, and the x-ray energy used for acquisition are each evaluated. Methods: The power law parameters,α and β, were measured from unprocessed LE and HE images and from DE subtracted images to quantify the anatomical noise. A total of 98 DE CEDM cases acquired in a previous clinical pilot study were assessed. Conventional DM images from 75 of the women were evaluated for comparison with DE CEDM. The influence of the imaging technique on anatomical noise was determined from an analysis of differences between the power law parameters as measured in DM, LE, HE, and DE subtracted images for each subject. Results: In DE CEDM, weighted image subtraction lowers β to about 1.1 from 3.2 and 3.1 in LE and HE unprocessed images, respectively. The presence of iodine has a small but significant effect in LE images, reducing β by about 0.07 compared to DM, with α unchanged. Increasing the x-ray energy, from that typical in DM to a HE beam, significantly decreases α by about 2 × 10−5 mm2, and lowers β by about 0.14 compared to LE images. A comparison of SE and DE CEDM at 4 min postcontrast shows equivalent power law parameters in unprocessed images, and lower α and β by about 3 × 10−5 mm2 and 0.50, respectively, in DE versus SE subtracted images. Conclusions: Image subtraction in both SE and DE CEDM reduces β by over a factor of 2, while maintaining α below that in DM. Given the equivalent α between SE and DE unprocessed CEDM images, and the smaller anatomical noise in the DE sub