Method for determining mammary gland tissue composition from an x-ray image

Abstract

The invention relates to the field of medical X-ray technology and can be used in conducting X-ray examinations. To determine the tissue composition of a mammary gland from an X-ray image, at least one X-ray image of the mammary gland and at least one image of thicknesses of the mammary gland are obtained. Using a previously obtained array of flat-field calibration data, stored in a memory, at least one flat-field image is obtained. At least one image of X-ray attenuation and at least one image of the tissue composition of the mammary gland are obtained. The invention increases the diagnostic value of X-ray examination without the need for additional manipulation of the patient or an increase in the radiation burden.

Description

METHOD FOR DETERMINING THE TISSUE COMPOSITION OF THE MAMMARY GLAND FROM AN X-RAY IMAGE The invention relates to the field of medical X-ray technology and can be used in conducting X-ray examinations. As a first approximation, the mammary gland (hereinafter referred to as MG) can be thought of as consisting of loose adipose tissue and dense tissue, including fibrous and glandular tissue, blood, water, skin, and other structures. With this representation, the tissue composition of the MG can be characterized by volumetric density, i.e., the volumetric proportion of dense tissue within the total volume of the MG being examined. Volumetric density can be represented either as an integral or pixel-by-pixel representation, i.e., as an image of the tissue composition. Increased breast density is a strong predictor of cancer development, increases the risk of pathologies being missed by a radiologist, and may in itself be a pathological condition requiring diagnosis and treatment (mastopathy). A method for determining breast density from an X-ray image is known (see John A. Shepherd et al. “Novel use of single X-ray absorptiometry for measuring breast density”, Technol Cancer Res Treat, 4 (2), 2005, pp. 173-182 [1]), based on comparing the pixel brightness of the tissue of the breast being studied with the pixel brightness of a tissue-equivalent phantom compressed together with the breast being studied and imaged together with it. The disadvantage of the known method is the need to place a reference object in the field of view when obtaining each image, which complicates the work of medical personnel, reduces the size of the effective field of view and increases the cost of the equipment used. A method is known for determining the tissue composition (content of fatty and dense tissue) in mammary glands from an X-ray image (see US 10733724, 04.08.2020 [2]), which consists of finding a reference value on the X-ray image, on the basis of which a quantitative analysis of the area of ​​interest in the image is performed, including determination of the density of mammary glands. The disadvantage of the known method is the use of the energy measure from the area of ​​the X-ray image that has been directly exposed to determine the measure of radiation attenuation in the area of ​​the object of study, while, firstly, such an energy measure is not always adequate due to the upper limitation of the dynamic range of the X-ray receiver (saturation of the receiver), and secondly, the accuracy of the extrapolation of the energy measure to the area of ​​the object of study is limited by the combined influence of the anode heel effect and additional filtering materials on the spatial distributions of the spectral composition and energy of X-ray radiation. The technical problem solved by the claimed invention consists in creating a method for determining the tissue composition of mammary glands that is free from the aforementioned disadvantages. The technical problem is solved and the technical result is achieved as a result of the creation of a method for determining the tissue composition of mammary glands using an X-ray image, which includes: - obtaining at least one X-ray image of the mammary gland, - obtaining at least one image of the thickness of the mammary glands, - obtaining at least one flat field image corresponding to the conditions for obtaining said at least one X-ray image of the breast, - obtaining at least one X-ray attenuation image using said at least one X-ray image of the breast and said at least one flat field image - and obtaining at least one image of the tissue composition of the mammary gland using said at least one image of X-ray attenuation and said at least one image of the thickness of the mammary gland. The said at least one flat field image is obtained using a flat field calibration data array previously obtained and stored in memory. In one particular embodiment, X-ray images of the mammary gland are obtained in craniocaudal and mediolateral projections. In another particular embodiment, X-ray images of the mammary gland are obtained from several angles, each of which is characterized by a corresponding angle of passage of the X-ray beam through the mammary gland. In another particular embodiment, X-ray images of the breast are obtained in at least two different ranges of the X-ray spectrum. In another particular embodiment, said at least one image of the thickness of the mammary gland is obtained using said at least one X-ray image of the mammary gland and a compression thickness value obtained from data containing conditions for obtaining said at least one X-ray image of the mammary gland. In yet another particular embodiment, said at least one image of the tissue composition of the mammary gland is obtained using data containing the conditions for obtaining said at least one X-ray image of the mammary gland, said at least one X-ray attenuation image, said at least one image of the thickness of the mammary gland, as well as a calibration data array previously obtained and stored in memory. In a preferred embodiment, the calibration data array is preliminarily obtained in the form of a dependence of the image of the thicknesses of the base materials on the image of the attenuation of the X-ray radiation and on the spectral composition of the X-ray radiation. In another particular embodiment, said at least one image of the tissue composition of the mammary gland is corrected using information about the presence or location on said original image of at least one area with a given tissue composition. In yet another particular embodiment, at least one mask of the region of interest is obtained using the at least one image of the breast or the at least one image of the breast tissue composition, at least one image of the specific volumes of the breast is obtained, and the values ​​of the total volume of the breast, the total volumes of individual tissues in the breast volume, and the total percentage shares of the volumes of individual tissues in the breast volume are obtained using the at least one image of the specific volumes of the breast, the at least one image of the breast tissue composition, and the at least one mask of the region of interest. In a preferred embodiment, said at least one region of interest mask is obtained using one of the operations selected from the group including image thresholding, image boundary detection, and the use of a pre-trained deep neural network. In another particular embodiment, the results obtained for breast images obtained in different projections, or from different angles, or in different spectral ranges are analyzed together. This achieves a technical result that increases the information content of standard X-ray examination of the mammary gland.

Claims

A method for determining tissue composition of a mammary gland from an X-ray image, comprising obtaining at least one X-ray image of a mammary gland, obtaining at least one image of the thickness of a mammary gland, obtaining at least one flat field image corresponding to the conditions for obtaining said at least one X-ray image of a mammary gland, obtaining at least one X-ray attenuation image using said at least one X-ray image of a mammary gland and said at least one flat field image, and obtaining at least one image of tissue composition of a mammary gland using said at least one X-ray attenuation image and said at least one image of the thickness of a mammary gland, wherein said at leastone flat field image is obtained using a previously acquired and stored flat field calibration data array. The method according to paragraph 1, characterized in that X-ray images of the mammary gland are obtained in craniocaudal and mediolateral projections. The method according to paragraph 1, characterized in that X-ray images of the mammary gland are obtained from several angles, each of which is characterized by a corresponding angle of passage of the X-ray beam through the mammary gland. The method according to claim 1, characterized in that X-ray images of the mammary gland are obtained in at least two different ranges of the X-ray radiation spectrum. The method according to claim 1, characterized in that said at least one image of the thickness of the mammary gland is obtained using said at least one X-ray image of the mammary gland and a compression thickness value obtained from data containing conditions for obtaining said at least one X-ray image of the mammary gland. The method according to claim 1, characterized in that said at least one image of the tissue composition of the mammary gland is obtained using data containing the conditions for obtaining said at least one X-ray image of the mammary gland, said at least one image of X-ray attenuation, said at least one image of the thickness of the mammary gland, as well as a previously obtained and stored in memory array of calibration data. The method according to paragraph 6, characterized in that the calibration data array is preliminarily obtained in the form of a dependence of the image of the thicknesses of the base materials on the image of the attenuation of X-ray radiation and on the spectral composition of the X-ray radiation. The method according to claim 1, characterized in that said at least one image of the tissue composition of the mammary gland is corrected using information about the presence or location on said original image of at least one area with a given tissue composition. The method according to claim 1, characterized in that at least one mask of the region of interest is obtained using the said at least one image of the mammary gland or the said at least one image of the tissue composition of the mammary gland, at least one image of the specific volumes of the mammary gland is obtained, and the values ​​of the total volume of the mammary gland, the total volumes of individual tissues in the volume of the mammary gland, and the total percentage shares of the volumes of individual tissues in the volume of the mammary gland are obtained using the said at least one image of the specific volumes of the mammary gland, the said at least one image of the tissue composition of the mammary gland, and the said at least one mask of the region of interest. The method according to claim 9, characterized in that said at least one mask of the region of interest is obtained using one of the operations selected from the group including threshold processing of the image, detection of boundaries in the image and the use of a pre-trained deep neural network. The method according to paragraph 1 or paragraph 9, characterized in that the results obtained for images of the mammary gland obtained in different projections, or from different angles, or in different spectral ranges, are analyzed together.

Images