Compositions and methods for monitoring the treatment of breast disorders
A breast, algorithmic technology, applied in the direction of medicine or prescription, application, mammography, etc.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0205] Database construction; feature classification
[0206] Dataset (KARMA)
[0207] In order to build the largest possible pool of potential training cases for predictor construction, mammographic images of 41,353 breast cancer-free women were sampled from KARMA, from which raw and processed images were available (Available from General Electrics, Philips, Sectra, Hologic, SIEMENS, FUJI, Agfa, Array Group and Vidar).
[0208] software application
[0209] Modified from the Java-based software ImageJ programming framework developed by the National Institutes of Health (http: / / rsb.info.nih.gov / ij / index.html, accessed March 30, 2017) To develop one aspect of the present invention, the computer program (hereinafter "STRATUS").
[0210] data input
[0211] Raw and processed images of the same mammograms from 41,353 women were analyzed and used to measure breast density using the VuComp M-Vu tool, an FDA-approved density measurement tool. This serves as a reference measure f...
Embodiment 2
[0223] Predictor construction and density measurement
[0224] For predictor construction using machine learning to measure density measurements from breast images and image metadata, STRATUS was configured to compare the 1,027 feature variables of Example 1 (including the breast compression score) with the FDA-approved VuComp M-Vu tool in Raw image-reference measurements of density obtained on the same mammogram were correlated.
[0225]The learning steps use the R programming framework developed by r-project.org and run as programs within the STRATUS programming framework. The STRATUS 1,027 image analysis variables of Example 1 were loaded as an R dataset. The (raw) image reference measurements of breast density were matched to the corresponding processed 1,027 image feature variables generated by STRATUS in Example 1. Perform a scaled principal component analysis (PCA) on the feature variables and create a predictive dataset based on the PCA data.
[0226] STRATUS was co...
Embodiment 3
[0232] image comparison
[0233] Alignment algorithms were developed and incorporated into STRATUS for aligning multiple breast images. The comparison tool can be a stand-alone computer program or an add-on. as in Figure 7 As observed in , alignment of breast images provides reduced non-biological variability between images over time.
[0234] Several techniques are available for image comparison (Thévenaz, U.E. Ruttimann, M. Unser, A Pyramid Approach to Subpixel Registration Based on Intensity, IEEE Transactions on Image Processing, vol. 7, no. 1, pp. 27-41, January 1998 ; US20160019690; US20150023576; US20060245629; US20090060300).
[0235] The transformation registration method used here preserved the largest fraction of the original image area and was used for the final analysis below (Example 4). This alignment scheme was developed for analyzing several mammograms in time series and is insensitive to differences in pixel intensity that can be seen between processed a...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


