AUTOMATED PERCENTAGE OF BREAST DENSITY MEASUREMENTS FOR FULL FIELD DIGITAL MAMMOGRAPHY

• US 20160110863A1
• Filed: 05/30/2014
• Published: 04/21/2016
• Est. Priority Date: 05/30/2013
• Status: Active Grant

• Current Assignee: Mayo Foundation For Medical Education And Research (Mayo Clinic), H Lee Moffitt Cancer Center & Research Institute (State University System of Florida)
• Original Assignee: Mayo Foundation For Medical Education And Research (Mayo Clinic), H Lee Moffitt Cancer Center & Research Institute (State University System of Florida)
• Inventors: Heine, John J., Sellers, Thomas A., Vachon, Celine M., Fowler, Erin E.
• Granted Patent: US 10,007,982 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: A61B 6/502 for diagnosis of breast, i.e. mammography
• CPC Class Codes: G06F 18/2113 by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation
• CPC Class Codes: G06F 2218/04 Denoising
• CPC Class Codes: G06T 2207/10116 X-ray image
• CPC Class Codes: G06T 2207/20076 Probabilistic image processing
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/41 based on statistical description of texture
• CPC Class Codes: G06V 10/52 Scale-space analysis, e.g. wavelet analysis multi-scale boundary representations G06V10/42
• CPC Class Codes: G06V 2201/03 Recognition of patterns in medical or anatomical images

Abstract

An automated percentage of breast density measure (PDa) that analyzes signal dependent noise (SDN) based on a wavelet expansion using full field digital mammography (FFDM). A matched case-control dataset is used with images acquired from a specific direct x-ray capture FFDM system. PDa is applied to the raw and clinical display representation images. Transforming (pixel mapping) of the raw image to another representation (raw-transformed) is performed using differential evolution optimization and investigated the influence of lowering the native spatial resolution of the image by a one-half. When controlling for body mass index, the quartile specific ORs for the associations of PDa with breast cancer varied with representation and resolution. PDa is a valid automated breast density measurement for a specific FFDM technology and compares well against PD (operator-assisted or the standard) when applied to either the raw-transformed or clinical display images from this FFDM technology.

Claims

1. A method of assessing breast density for breast cancer risk applications, comprising:

receiving digital image data including a plurality of pixels;

performing a statistical analysis on the digital image data; and

associating the statistically analyzed digital image data with a measure of risk for breast cancer.

2. The method of claim 1, performing the statistical analysis comprising:

applying a logistic regression to model and compare the association between quartile measures of bread density (PD), and breast cancer status, with a lowest quartile measured PD.

3. The method of claim 2, further comprising summarizing determining magnitudes of the associations by odds ratios (ORs) having predetermined confidence intervals (CIs).

4. The method of claim 2, further comprising adjusting breast density models for body mass index (BMI).

5. The method of claim 1, further comprising automating a determination of breast density (PD_a).

6. The method of claim 5, further comprising:

using a wavelet expansion to determine PD_a application by lowering a spatial resolution; and

applying a signal dependent noise (SDN) analysis.

7. The method of claim 6, further comprising:

estimating a global reference variance signal by using all pixels in a d₁ image corresponding to a breast region by moving a n×

n pixel search window across the d₁ image constrained to breast area;

at each window location, determining a local variance;

comparing the global reference variance signal is compared with the local variance;

if the local variance deviates too far above the reference using a chi-square test, a respective n×

n region is labeled as dense; and

producing a binary mage with each pixel labeled as dense or labeled as other.

8. The method of claim 7, further comprising refining the global reference variance by restricting its second estimation in the d₁ image to those locations that were initially labeled as other.

9. The method of claim 8, wherein the pixels labeled as other correspond to adipose image regions.

10. The method of claim 7, wherein PD_a is calculated as a ratio.

11. The method of claim 1, further comprising preprocessing the digital image data using an empirical based mapping.

12. The method of claim 11, wherein the empirical based mapping is determined by:

13. The method of claim 1, further comprising determining a signal dependent noise (SDN) characteristic of the digital image data.

14. The method of claim 13, wherein the SDN is determined in accordance with:

y=c₀+c₁x+c₂x²+c₃x³+c₄exp(−

z)²,

where y is a local noise variance, x is a corresponding local signal,

15. A method of automatically measuring breast density (PD_a), comprising:

modeling a Fourier spectral density of digitized film mammograms as an inverse power-law;

estimating a power-law behavior for a given mammogram;

removing a shape of the mammogram using a deconvolution process to a filtered image; and

determining a density of area in the filtered image.

16. The method of claim 15, wherein areas of higher spatial variation in the filtered image correspond with areas of increased mammographic density in the digitized mammogram.

17. The method of claim 16, wherein the areas of higher spatial variation are a signature of signal dependent noise (SDN).

18. The method of claim 16, further comprising applying a statistical method to detect these areas of increased variation in the filtered image.

19. A computer implemented method of determining breast density, comprising:

receiving digital image data including a plurality of pixels;

filtering the digital image data;

estimating a global reference variance signal in the digital image data;

performing a statistical analysis on the digital image data; and

associating the statistically analyzed digital image data with a measure of risk for breast cancer.

20. The computer implemented method of claim 19, wherein the statistical analysis determines areas of higher density tissue in the digital image data.