Method for assessing breast density
Breast density is a significant breast cancer risk factor measured from mammograms. Evidence suggests that the spatial variation in mammograms may also be associated with risk. The variation in calibrated mammograms as a breast cancer risk factor was investigated and its relationship with other measures of breast density was explored using full field digital mammography (FFDM) as described herein. A matched case-control analysis was used to assess a spatial variation breast density measure in calibrated FFDM images, normalized for the image acquisition technique variation. The findings indicate the variation measure is a viable automated method for assessing breast density. Insights gained by this work may be used to develop a standard for measuring breast density.
1. A method of assessing breast density for breast cancer risk applications, comprising:
receiving digital image data including a plurality of pixels;
calibrating the digital image data;
measuring a variation of pixel values of the calibrated digital image data, wherein measuring a variation of pixel values of the calibrated digital image data further comprises at least one of calculating an l 2 norm or order derived therefrom or calculating an l 1 norm or order derived therefrom; and
associating the variation of pixel values with a measure of risk for breast cancer, wherein the variation of pixel values correlates with at least one of a relative risk for breast cancer, an odds ratio for breast cancer, or an absolute risk prediction for breast cancer.
2. The method of claim 1 , wherein measuring a variation of pixel values of the calibrated digital image data further comprises:
calculating an l 2 norm or order derived therefrom;
calculating an l 1 norm or order derived therefrom; and
calculating a combination of measures based on results of the l 2 norm or order derived therefrom and the l 2 norm or order derived therefrom.
3. The method of claim 2 , wherein calculating a combination of measures further comprises at least one of using a linear method, using a non-linear method, or using a Gram-Schmidt orthogonalization process, Principal Component Analysis, partial least squares or kernel-based method.
4. The method of claim 1 , wherein calibrating the digital image data further comprises adjusting for image acquisition technique parameters by adjusting for at least one of variation in target/filter combination, x-ray tube voltage, radiation exposure, or compressed breast thickness.
5. The method of claim 4 , wherein calibrating the digital image data is performed pixel-by-pixel.
6. The method of claim 4 , wherein calibrating the digital image data further comprises:
calculating an average pixel value of an n×n pixel region; and
calibrating the average pixel value.
7. The method of claim 1 , wherein the digital image data comprises an image having a breast tissue area and a background area, the method further comprising:
segmenting the breast tissue area from the background area of the image.
8. The method of claim 7 , further comprising:
assigning pixel values within the breast tissue area a first value; and
assigning pixel values within the background area a second value.
9. The method of claim 7 , further comprising positioning a radial coordinate system origin at a side of the image at a first direction centroid position estimated from the segmented image.
10. The method of claim 9 , further comprising eroding a percentage of the image between the radial coordinate system origin and a perimeter of the breast area along a radial direction.
11. A method of assessing breast density for breast cancer risk applications, comprising:
receiving digital image data including a plurality of pixels;
calibrating the digital image data;
measuring a variation of pixel values of the calibrated digital image data, wherein measuring a variation of pixel values of the calibrated digital image data further comprises calculating the variation using an n th central or non-central moment of an integer or a fractional order or any real number order; and
associating the variation of pixel values with a measure of risk for breast cancer, wherein the variation of pixel values correlates with at least one of a relative risk for breast cancer, an odds ratio for breast cancer, or an absolute risk prediction for breast cancer.
12. The method of claim 11 , wherein measuring a variation of pixel values of the calibrated digital image data further comprises:
calculating a first n th central or non-central moment;
calculating a second n th central or non-central moment, the second n th central or non-central moment being different than the first n th central or non-central moment; and
calculating a combination of measures based on results of the first and second n th central or non-central moments using at least one of a linear method or a non-linear method.
13. The method of claim 11 , wherein calibrating the digital image data further comprises adjusting for image acquisition technique parameters by adjusting for at least one of variation in target/filter combination, x-ray tube voltage, radiation exposure, or compressed breast thickness.
14. A method of assessing breast density for breast cancer risk applications, comprising:
receiving digital image data including a plurality of pixels;
measuring a variation of pixel values of the digital image data, wherein measuring a variation of pixel values of the digital image data further comprises at least one of calculating an l 2 norm or order derived therefrom or calculating an l 1 norm or order derived therefrom: and
associating the variation of pixel values with a measure of risk for breast cancer, wherein the variation of pixel values correlates with at least one of a relative risk for breast cancer, an odds ratio for breast cancer, or an absolute risk prediction for breast cancer.
15. The method of claim 14 , wherein measuring a variation of pixel values of the digital image data further comprises:
calculating an l 2 norm or order derived therefrom;
calculating an l 1 norm or order derived therefrom; and
calculating a combination of measures based on results of the l 2 norm or order derived therefrom and the l 1 norm or order derived therefrom.
16. The method of claim 15 , wherein calculating a combination of measures further comprises at least one of using a linear method, using a non-linear method, or using a Gram-Schmidt orthogonalization process, Principal Component Analysis, partial least squares or kernel-based method.
17. The method of claim 14 , wherein the digital image data comprises an image having a breast tissue area and a background area, the method further comprising:
segmenting the breast tissue area from the background area of the image.
18. The method of claim 17 , further comprising:
assigning pixel values within the breast tissue area a first value; and
assigning pixel values within the background area a second value.
19. A method of assessing breast density for breast cancer risk applications, comprising:
receiving digital image data including a plurality of pixels;
measuring a variation of pixel values of the digital image data, wherein measuring a variation of pixel values of the digital image data further comprises calculating the variation using an n th central or non-central moment of an integer or a fractional order or any real number order; and
associating the variation of pixel values with a measure of risk for breast cancer, wherein the variation of pixel values correlates with at least one of a relative risk for breast cancer, an odds ratio for breast cancer, or an absolute risk prediction for breast cancer.
20. The method of claim 19 , wherein calculating a variation of pixel values of the digital image data further comprises:
calculating a first n th central or non-central moment;
calculating a second n th central or non-central moment, the second n th central or non-central moment being different than the first n th central or non-central moment; and
calculating a combination of measures based on results of the first and second n th central or non-central moments using at least one of a linear method and a non-linear method.
21. The method of claim 19 , wherein the digital image data comprises an image having a breast tissue area and a background area, the method further comprising:
segmenting the breast tissue area from the background area of the image.