Maximum-Likelihood Estimation of Glandular Fraction for Mammography and its Effect on Microcalcification Detection

TL;DR

This paper presents a maximum likelihood algorithm to estimate pixel-wise glandular fraction in mammography, improving breast density localization and microcalcification detection, with potential to enhance breast cancer screening accuracy.

Contribution

The study introduces a novel maximum likelihood method for pixel-wise glandular fraction estimation, validated on simulated and clinical images, enhancing microcalcification detection.

Findings

Glandular fraction estimation achieved RMSE of 2.5-3.2%

Contrast-to-noise ratio for microcalcifications improved by up to 548%

Algorithm provides accurate breast density localization

Abstract

Objective: Breast tissue is mainly a mixture of adipose and fibro-glandular tissue. Cancer risk and risk of undetected breast cancer increases with the amount of glandular tissue in the breast. Therefore, radiologists must report the total volume glandular fraction or a BI-RADS classification in screening and diagnostic mammography. A Maximum Likelihood algorithm is shown to estimate the pixel-wise glandular fraction from mammographic images. The pixel-wise glandular fraction provides information that helps localize dense tissue. The total volume glandular fraction can be calculated from pixel-wise glandular fraction. The algorithm was implemented for images acquired with an anti-scatter grid, and without using the anti-scatter grid but followed by software scatter removal. The work also studied if presenting the pixel-wise glandular fraction image alongside the usual mammographic image has the potential to improve the contrast-to-noise ratio on micro-calcifications in the breast. Results: For the TOPAS simulated images, the glandular fraction was estimated with a root mean squared error of 3.2% and 2.5% for the without and with anti-scatter grid cases. Average absolute errors were (3.7 +/- 2.4)% and (3.6 +/- 0.9)%, respectively. Results from DICOM clinical images (where the true glandular fraction is unknown) show that the algorithm gives a glandular fraction within the average range expected from the literature. For microcalcification detection, the contrast-to-noise ratio improved by 17.5-548% in DICOM images and 5.1-88% in TOPAS images. Conclusion: We show a method of accurate estimation of pixel-wise glandular fraction image, providing localization information of breast density. The glandular fraction images also showed an improvement in contrast to noise ratio for detecting microcalcifications, a risk factor in breast cancer.