Apparatus and method for breast cancer diagnosis

Abstract

An ultrasound apparatus (1) for performing a breast diagnostic examination for breast cancer diagnosis, comprising two ultrasound probes (10, 20), characterized in that each of the probes (10, 20) comprises a first array (11, 21) of piezoelectric transducers having a first nominal frequency (f1) and a second array (12, 22) of piezoelectric transducers having a second nominal frequency (f2), the probes (10, 20) being arranged opposite each other, configured to rotate along a circular trajectory centered on a midpoint of a segment connecting them and to slide each other along the straight line connecting them, so that they are in contact with the breast of a patient from fully opposite portions, wherein each of the arrays (11, 12, 21, 22) comprises a plurality of piezoelectric transducers positioned so that acquisitions are performed on acquisition planes orthogonal to the plane of the circular trajectory, and wherein the apparatus is configured to perform a plurality of scans with respect to a plurality of acquisition planes (P0, P1...) passing through the midpoint, orthogonal to the circular trajectory and rotated by an angle (a) from each other, by moving the probes (10, 20) along the circular trajectory.

Description

Technical Field

[0001] This invention relates to a device for performing diagnostic examinations for breast cancer. Background Technology

[0002] More specifically, the present invention relates to a device that allows for the reconstruction of a 3D ultrasound model of the breast volume in a perfectly overlapping manner and information about the spectrum and acoustic attenuation coefficients, and calculates indications of the presence of breast cancer based on these data.

[0003] Existing technology

[0004] Many problems are related to breast ultrasound examinations.

[0005] First, it is desirable to ensure the repeatability of the examination, and in particular the positioning of the ultrasound probe (or multiple probes) relative to the breast, regardless of the operator's skill and expertise; second, if 3D reconstruction of the breast is to be performed, then when moving the ultrasound probe (or multiple ultrasound probes) to scan multiple locations, it is necessary to ensure that the breast is in a defined position relative to the probe, obviously ensuring that there is a suitable acoustic coupling means (water, gel or other suitable material) between the probe and the skin.

[0006] Finally, once the aforementioned technical issues are resolved, it will be necessary to analyze the collected ultrasound data using a method that allows for high accuracy in breast cancer diagnosis.

[0007] Many devices and methods are known for ultrasound and / or ultrasound data analysis intended to diagnose breast cancer, and they attempt to address the aforementioned problem.

[0008] CN110680380 describes a device using a ring with multiple ultrasound transducers arranged thereon, located outside a cylindrical container in which the breast of a patient lying prone is immersed in water. This device is used to generate ultrasound images, which a doctor can use to diagnose breast cancer.

[0009] WO02089672 describes a device in which a linear array of piezoelectric transducers is arranged parallel to the axis of symmetry of the breast and configured to rotate about the axis of symmetry of the breast within a container providing coupling fluid.

[0010] EP2868279 describes a device for breast ultrasound that uses multiple transducers arranged on a support element.

[0011] US2013 / 0041261 describes a device using a ring transducer comprising multiple 2MHz frequency ultrasound transmitters and receivers configured to determine the spatial distribution of acoustic and mechanical parameters of the breast.

[0012] US2012029358 discloses a device using multiple ultrasound probes arranged symmetrically about the axis of symmetry of the breast, configured to image in multiple planes passing through the axis of symmetry of the breast at frequencies between 7.5 and 10 MHz.

[0013] US2004064046 describes an apparatus for breast ultrasound tomography, including a static chamber for containing fluid, within which a movable chamber is provided, the movable chamber being part of an array of ultrasound transducers and receivers.

[0014] Other examples are shown in CN111213065, US2013041260, US4222274, and US10285667.

[0015] Technical issues

[0016] To the best of the inventor’s knowledge, all the devices just described, as well as other devices known in the prior art, have limitations because they do not allow for a perfect overlap of the distribution of breast images and other characteristic parameters of the analyzed tissue (such as, for example, acoustic attenuation coefficients) during the same examination.

[0017] Moreover, no known device allows for the calculation of diagnostic parameters that explicitly take into account both of these pieces of information (information available from ultrasound imaging and information available from the analysis of acoustic attenuation coefficients).

[0018] Furthermore, no known prior art device allows for the calculation of diagnostic parameters that explicitly consider the frequency analysis of the associated raw ultrasound signal. It should also be noted that, in the following text, the raw ultrasound signal refers to the radiofrequency ultrasound signal received by the probe prior to the various filters and processing required to obtain an ultrasound image. Summary of the Invention

[0019] The object of the present invention is to provide an apparatus for performing examinations for breast cancer diagnosis that overcomes the limitations associated with known embodiments in the prior art, and in particular allows for volumetric reconstruction of perfect overlap between the distribution of characteristic parameters (such as, for example, acoustic attenuation coefficients) of the breast image and the analyzed tissue during the same examination.

[0020] According to another objective, the device object of the present invention includes a computing unit on which a computer program is loaded to calculate diagnostic parameters that explicitly take into account two pieces of information (information available from ultrasound imaging and information available from the analysis of acoustic attenuation coefficients).

[0021] However, another object of the present invention is to calculate diagnostic parameters that also explicitly take into account parameters derived from frequency analysis of ultrasound signals associated with specific portions of breast tissue. Attached Figure Description

[0022] Figure 1 , 2 Views 3 and 4 show three views of a preferred, rather than limiting, embodiment of the device; Figure 4 A schematic diagram of a breast cross-section along a universal acquisition plane passing through the device is shown; Figure 5 A schematic cross-sectional view of a portion of the tissue along a plane orthogonal to the axis of symmetry of the device is shown; Figure 6 A schematic diagram of a breast half-section along a universal acquisition plane passing through the device is shown, and Figure 7 The shapes of the signals received in transmission are schematically shown in the time domain (a) and frequency domain (b), and the shapes of the signals received in reflection are shown in the time domain (c) and frequency domain (d).

[0023] Figure 8 and 9 An example flowchart of a preferred embodiment of a method that can be performed by means of a device according to the invention is shown; Figure 10 A schematic cross-sectional view of the probe is shown, highlighting the mechanical coupling components; Figure 11 The equipment installed under the inspection table is shown; Figure 12 Various three-dimensional schematic diagrams of obtaining planar arrangements are shown. Detailed Implementation

[0024] Before proceeding with the following description, it should be noted that: - For tumor tissue, it refers to tissue associated with the following suspicious areas where the presence of a tumor has been confirmed by histological examination or by the method according to the invention; - For non-tumor tissue, it refers to tissue associated with a suspicious area where the presence of a tumor has been ruled out by histological examination or by the method according to the invention. In other words, non-tumor tissue involves heterogeneity detectable by ultrasound or by means of ultrasound analysis, and the result is benign; - For healthy tissue, it refers to tissue that does not have any heterogeneous features that can be detected by ultrasound or by means of ultrasound analysis, and is not classified as a suspicious area.

[0025] It should also be noted that the implementation of the device is described with reference to the use of piezoelectric transducers, but other types of transducers, such as CMUT capacitor transducers, may also be used without departing from the purpose of this invention.

[0026] refer to Figure 1 The device (1) according to the invention includes two ultrasonic probes (10, 20), each ultrasonic probe being associated with two piezoelectric transducer arrays (11, 12, 21, 22).

[0027] In the first embodiment, the array is a linear array. In the second embodiment, the array is concave, allowing for a better fit to the breast shape. It should be noted that the concavity is related to the shape of the probe.

[0028] Each of the two ultrasonic probes (10, 20) includes a first piezoelectric transducer array (11, 21) having a first nominal frequency or band center frequency (f1) and a second piezoelectric transducer array (12, 22) having a second nominal frequency or band center frequency (f2).

[0029] The first nominal frequency (f1) is selected to perform B-mode ultrasound imaging starting from the signal reflected from the breast tissue and is preferably, but not limited to, between 7 MHz and 10 MHz.

[0030] As will be clear from the description below, each probe is used for high-resolution ultrasound imaging of the breast portion, including the skin and its axis of symmetry. For this purpose, given the limited depth of the tissue being analyzed, high frequencies can be used without the associated attenuation, which becomes a problem.

[0031] The second nominal frequency (f2) is chosen to better penetrate tissue and thus allows for acoustic attenuation measurements by working in transmission mode with an array of other probes having the same nominal frequency. The second nominal frequency (f2) is preferably between 1 MHz and 3 MHz.

[0032] Two arrays (12, 22) with the lowest nominal frequency (f2) are equipped with the same number of piezoelectric transducers, such that the corresponding transducer (221) of the array (22) of the second probe (20) corresponds to each transducer (121) of the array (12) of the first probe (10).

[0033] The implementation of each of the two probes ensures that all piezoelectric transducers associated with it are rigidly fixed to each other. In other words, the relative positions of the two piezoelectric transducer arrays are fixed. Specifically, each probe comprises two piezoelectric transducer arrays, linearly or concavely arranged adjacent to each other.

[0034] Each probe also includes a coupling component (16, 26) to a control device (not shown in the figure), which, according to the following detailed description, is configured to guide the probe and detect and analyze the signals acquired therefrom.

[0035] In a first preferred embodiment, each probe (10, 20) further includes a retaining rod (13, 23). The two retaining rods (13, 23) are associated with the annular support (30) in completely opposite positions. In other words, the two probes (10, 20) are positioned along the same direction, opposite each other, and the array (11, 12, 21, 22) is positioned in the same plane as and symmetrical to the axis of symmetry (a) of the annular support (30).

[0036] Each of the two probes (10, 20) is positioned along the radial direction of the annular support (30) and associated in the same way, such that the fixing rod (13, 23) can slide, thereby allowing the probes (10, 20) to be positioned at different distances from the axis of symmetry (a) of the annular support (30).

[0037] Preferably, the device further includes a propulsion component configured to advance the two probes (10, 20) radially toward the center. Figure 1 (Not shown in the image). Preferably, the propulsion component includes at least a spring element configured to propel the respective probes (10, 20) toward the center.

[0038] The device also includes a movable component (40) configured to rotate the annular support (30) about its own axis of symmetry (a).

[0039] The moving part (40) includes one or more electric motors and corresponding control components. These are moving parts known in the prior art.

[0040] Regardless of the embodiments described above, the embodiments are preferred rather than limiting variations. The device (1) according to the invention comprises two ultrasonic probes (10, 20), each ultrasonic probe comprising two arrays (11, 12, 21, 22) of piezoelectric transducers having two different nominal frequencies (f1, f2), the probes being arranged opposite each other on the same straight line and configured such that they can rotate along a circular trajectory centered on the midpoint of the segment connecting them, so that they can slide relative to each other along the straight line connecting them.

[0041] like Figure 10As shown, and preferably each probe also includes an acoustic coupling element (14, 15) to the breast skin. In a first preferred embodiment, the acoustic coupling element comprises a flexible membrane (14) filled with gel (15). The combination of membrane and gel is configured to be compressed as the probe is advanced toward the breast, thereby conforming to the shape of the breast and thus ensuring mechanical coupling with the latter for ultrasound transmission.

[0042] Preferably, for this purpose, the breast is covered with a gel, liquid, or any other material suitable for ultrasound transmission.

[0043] In a second embodiment, the device is immersed in a container filled with water or any other liquid suitable for use as an acoustic coupling component, and has an opening at the top so that the breast can be introduced from above.

[0044] The device (1) is configured to be positioned at an opening (51) below the examination table (50) into which the breast of a woman lying prone on the examination table can be introduced.

[0045] Having described the device, we can now describe its functions.

[0046] Once the patient lies prone on the examination table (50), the device (1) positions its own axis of symmetry (a) on the axis of symmetry of the breast, and the two probes (10, 20) advance toward the axis of symmetry until they contact the patient’s breast from completely opposite parts at the first corner position (P0).

[0047] According to another embodiment, the device can be integrally fixed to the lower part of the inspection table (50) at a suitable location through a hole obtained on its surface.

[0048] Now, the first scan can be performed, and according to the detailed description below, the annular support (30) rotates at an angle ( ), until the two ultrasound probes are brought to position (P1), and the scanning and rotation sequence is repeated until the annular support (30) is rotated 180°, and then returned, with the two probes aligned at the first corner position (P0), but in the inverted position.

[0049] In this way, the entire breast volume is scanned.

[0050] It should be noted that the probe's control device and the acquisition of the detected ultrasonic signals are configured as follows: - Individually guide each piezoelectric transducer in each array of each probe; - Detect the signal acquired by each piezoelectric transducer of each probe; - Store the "raw" (also known as "radio frequency") ultrasonic signals acquired by each transducer, making them available for subsequent processing; - Process radio frequency signals to obtain B-mode ultrasound images; - The acquired relative positions are associated with each detected signal and with each image.

[0051] It is clear that, for the process just described and all other processes described in the document, the device according to the invention includes an electronic computing component on which a computer program is loaded and configured to perform relative movement of the device, data acquisition, data storage, data processing, and diagnostic parameter calculation.

[0052] The acquisition process at each position (P0, P1, ...) is as follows: 100) By sending multiple ultrasound signals and detecting relevant ultrasound signals reflected by breast tissue, a first B-mode ultrasound image related to a portion of the tissue included between the first probe (10) and the axis of symmetry is acquired by means of a first array (11) of a first probe (10); 110) By sending multiple ultrasound signals and detecting relevant ultrasound signals reflected from breast tissue, a second B-mode ultrasound image related to a portion of the tissue included between the second probe (10) and the axis of symmetry is acquired by means of the first array (21) of the second probe (20); 120) Store the “raw” (or “radio frequency” ultrasound signal, as detected by each probe (10, 20) and before any subsequent processing, especially before the processing used for the creation of the associated B-mode image; 130) An ultrasonic pulse is emitted by means of the first piezoelectric transducer (121) of the second array (12) of the first probe (10), and the transmitted ultrasonic signal is detected by means of the first piezoelectric transducer (221) of the second array (22) of the second probe (20) and the reflected ultrasonic signal is detected by means of the first piezoelectric transducer (121) of the second array (12) of the first probe (10); 140) Repeat step 130 for each pair of piezoelectric transducers in the second array (12, 22) of the first and second probes.

[0053] At the end of the acquisition process, for each angle position (P0, P1, ...), the following processing can be performed: - Two ultrasound images, acquired in the same plane and correlated with tissue portions (P0_1, P0_2) included between the axis of symmetry and each probe, are obtained by analyzing the reflected signals acquired by each of the two arrays having a first nominal frequency; - Multiple pairs of signals, consisting of transmitted signals (known based on the structural characteristics of the probes used and the associated guidance system, often referred to as "beamformers") and received signals in transmission mode, each pair of signals being opposite to a pair of transducers in an array of first and second probes having a second nominal frequency; - Multiple ultrasonic signals reflected in the time domain, acquired by an array of probes, each with a second nominal frequency; - Multiple ultrasound signals (110S) in the time domain are reflected by breast tissue on pulses transmitted by an array having a first nominal frequency and are associated with corresponding propagation segments.

[0054] exist Figure 6 The image shows the path (111) of the ultrasonic reflected signal detected by the first transducer of the first array (11) of the first ultrasonic probe (10). For illustrative purposes, Figure 6 Only the portion of the tissue contained between the axis of symmetry and the probe (10) is shown.

[0055] As from Figure 6 The analysis clearly shows that the figure also shows that the arrival time (tA) of the ultrasound reflected signal to point A, which is located far from the probe, is greater than the arrival time (tB) of the signal reflected from point B, which is basically placed on the skin and in contact with the probe.

[0056] Based on this known consideration, the value of the original (i.e., unprocessed) ultrasound signal corresponding to the reflection can be associated with each point in segment AB.

[0057] The combination of acquisitions performed defines the volume grid for acquiring the volume, schematically shown in Figure 5 As shown in the figure, this diagram illustrates the planar projection of the i-th (Vi) acquired volume associated with the general point C.

[0058] Therefore, the angular amplitude (alfa) of the i-th volume is equal to the angle between the two subsequently acquired positions, and the radial amplitude (dr) is equal to the amplitude of the ultrasound imaging system in the radial direction (i.e., in the height corresponding to the depth of the ultrasound probe and the "lateral spatial resolution" of the ultrasound image). Figure 5 Spatial resolution (not shown in the text) in a meaningful sense.

[0059] As is known, the depth resolution of an ultrasound imaging system theoretically corresponds to half the wavelength of the incident ultrasound pulse, which is inversely proportional to the frequency.

[0060] It should also be noted that the angular amplitude of the acquired volume depends on the focusing of the ultrasonic beam used. Conveniently, the angle between the two subsequent acquisition positions (P0, P1) is selected. (in order to obtain the entire volume and ensure that no area is left uncovered).

[0061] Therefore, for each i-th volume, the following will be obtained.

[0062] - Values ​​related to the intensity of reflected ultrasound images (S)i ), - The grayscale value associated with the B-mode ultrasound image (B i ).

[0063] Furthermore, the following content is available: - The value of the acoustic attenuation coefficient calculated for each propagation line of the ultrasonic signal emitted by the second array (12) of the first probe (10) and received by the second array (22) of the second probe (20); - Acquisition of raw radio frequency ultrasound signals, which involves all propagation lines of the ultrasound signals for each array, for all locations acquired.

[0064] A method for analyzing previously collected data to calculate diagnostic parameters indicating the presence of breast cancer includes the following steps: 200) Segment the 3D model of the breast volume to distinguish the presence of one or more “suspicious areas” characterized by being brighter (hyperechoic) or darker (hypoechoic) in color relative to the surrounding tissue, or in any other way characterized by any other form of “non-uniformity” that can be detected by means of grayscale analysis, which makes these areas “enclosed”.

[0065] In the case of individualizing one or more suspicious regions, a diagnostic parameter (Dj) is calculated for each region, indicating the probability that the region can be classified as breast cancer. The diagnostic parameter (Dj) depends on one or more of the following factors: (i) Morphological information derived from B-mode imaging of points of the 3D model within the j-th region (Zj); (ii) Information regarding the average acoustic attenuation coefficient relative to at least a portion of the propagation line of the ultrasonic signal contained in the j-th region (Zj); (iii) Information about the spectrum of the signal associated with the j-th region (Zj), or information obtained in any way through analysis thereof, may be performed by means of a quantitative comparison with a “model spectrum” obtained in patients with a known diagnosis.

[0066] It should be noted that, for reference Figure 4 (Where, for simplicity, the propagation line of the ultrasonic signal is shown, but the corresponding probe is not shown.) The average acoustic attenuation coefficient relative to at least a portion of the propagation line of the ultrasonic signal contained in the j-th region (Zj) can be calculated as follows: - The acoustic attenuation coefficient (Ats) of healthy tissue is calculated, which is the propagation line (of length L1) of the ultrasound signal outside any suspicious area in the ultrasound image. Preferably, the acoustic attenuation coefficient (Ats) of healthy tissue is calculated as the average of the acoustic attenuation coefficients calculated for multiple propagation lines of the ultrasound signal outside any suspicious area detected during the ultrasound examination.

[0067] Therefore, for each propagation line (total length Li) passing through the j-th region (Zj), the average acoustic attenuation coefficient (Aj) relative to the internal tissue of the suspected region can be calculated according to the following relationship, considering that the total acoustic attenuation (Ai) calculated for the i-th line is given by the sum of the acoustic attenuation along the healthy portion (lengths Li' and Li'') and the acoustic attenuation along the portion contained within the suspected region (length Li'''), where the only unknown factor is the value of the average acoustic attenuation coefficient (Aj) relative to the internal tissue of the suspected region, because the average acoustic attenuation coefficient (Ai) along the entire i-th propagation line can be calculated considering the propagation line, as it passes through homogeneous tissue: Ats x (Li' + Li'') / (Li) + Aj x Li''' / Li = Ai It should be clarified that the lengths (Li', Li'', Li''') of the outer and inner segments of the suspicious region can only be measured by acquiring ultrasound images performed at the same acquisition location. The frequency of the ultrasound image is greater than that of the ultrasound signal in order to improve the resolution of the ultrasound image and thus improve the accuracy of the calculation of the lengths (Li', Li'', Li''') of the outer and inner segments of the suspicious region.

[0068] In other words, such measurements are only possible with the configuration of the piezoelectric transducer array of the device according to the invention, which includes two ultrasonic probes facing each other, each ultrasonic probe comprising two arrays with different nominal frequencies, one array optimized for acquiring high-resolution ultrasonic images and the other array optimized for calculating acoustic attenuation coefficients, and generally used to perform acoustic measurements in transmission at frequencies lower than those required to obtain good ultrasonic images. In a preferred embodiment, the calculation of diagnostic parameters can be performed according to the following steps.

[0069] 210) Calculate multiple morphological and intensity properties derived from ultrasound imaging of the j-th region (Zj), for example: - Volume of the region; - The ratio between surface area and volume; - Maximum dimension; - Shape parameters, such as eccentricity; - The ratio of the average gray value of the j-th region (Zj) to the average gray value of the surrounding regions.

[0070] Preferably, the surrounding area is defined as a combination of all breast tissue portions that do not belong to any suspicious area, or as a set of all breast tissue portions that do not belong to any suspicious area and are traversed by the same propagation line in the j-th region (Zj) under consideration.

[0071] 220) Calculate the average acoustic attenuation coefficient (Aj) relative to at least a portion of the propagation line of the ultrasonic signal contained in the j-th region (Zj); 230) Calculate the ratio between the average acoustic attenuation coefficient (Aj) of the j-th region (Zj) and the average acoustic attenuation coefficient (Ats) from the outer region to the j-th region. 235) Calculate the acoustic propagation speed of the ultrasound signal associated with at least one propagation line of the ultrasound signal passing through the j-th region (Zj) and at least one propagation line of the ultrasound signal not passing through any suspected region.

[0072] Conveniently, the propagation speed that gives an indication of tissue density can be calculated as a function of the reception time of the emitted ultrasound signal and the relative distance between the probes (both of which are automatically measured by the device according to the invention).

[0073] 240) Calculate BUB (broadband ultrasound backscattering) relative to the set of 3d model points in the j-th region (Zj); It should be noted that BUB is the average value of the backscattering coefficients (which are frequency-dependent parameters) over a defined frequency interval, preferably between 0.2 and 0.6 MHz, or in any way within a frequency interval where the spectrum of the reflected signal shows a decreasing trend. The backscattering coefficient can be calculated as the ratio between the intensity of the ultrasound signal reflected by the portion of the tissue under consideration and a reference ultrasound signal. In the examination procedure, the signal reflected by the air-water interface positioned at the same distance from the probe relative to the tissue portion can be considered the reference ultrasound signal. Therefore, it is clear that once the parameters of the transmitted signal (frequency, power, etc.) and the reference distance are fixed, the reference ultrasound signal is the parameter characterizing the probe.

[0074] 250) Calculate the IRC (Integrated Reflection Coefficient) of the interface between the j-th region (Zj) and the surrounding tissue.

[0075] IRC can be calculated as the average value of the energy reflection coefficient over a defined frequency interval, preferably between 0.2 and 0.6 MHz, or in any way within a frequency interval where the spectrum of the reflected signal has a decreasing trend, and is obtained as the logarithmic difference between the spectrum of the signal reflected from the interface (between the j-th region and the surrounding tissue) and the spectrum of the signal reflected from the air-water reference interface (stored in the system during the device construction step and available for calculation). Following point 250, a set of values ​​for each parameter relative to each region is thus defined.

[0076] 260) Calculate at least the spectrum of the original radio frequency ultrasound signal emitted by one of the first arrays (11, 21) and reflected by a segment of the propagation line of the ultrasound signal contained in the j-th region.

[0077] 270) Calculate at least the quantitative parameters extracted from the spectrum of point 260), including them in a list containing the following parameters: - Average spectrum value; - The region to which the spectrum corresponds at defined frequency intervals; - The spectral width at a predetermined intensity level (intended as the difference between the maximum and minimum frequencies), particularly at a level defined by an intensity value lower than the maximum value of the spectrum for the predetermined amount, especially lower than 1 dB or 3 dB; - The frequency value corresponding to the maximum value of the spectrum; - The slope of the line that interpolates multiple points of the spectrum at predetermined frequency intervals; - The coefficients of a polynomial interpolated at frequency intervals containing the maximum value of the spectrum.

[0078] Figure 6 A portion of the propagation line of an ultrasonic signal emitted by a piezoelectric crystal of a first array (11) of a first probe (10) included in the j-th region (Zj) is shown. Figure 7 The relevant ultrasound signal obtained from the reflection in the time domain is shown, and in particular the portion (A'-B') associated with the portion of the propagation line contained in the j-th region (Zj). Figure 7 The spectrum (S1) is also shown as a frequency transformation of the ultrasound signal reflected by tissue included in the j-th region (Zj).

[0079] Preferably, the spectrum associated with the j-th region (Zj) is calculated as the average spectrum of multiple spectra associated with multiple propagation segments of the signal contained within the j-th region. Preferably, in order to calculate the average spectrum, an ultrasonic signal selection step is also performed for calculating the average spectrum relative to the j-th region; - Calculate the correlation coefficient between the average spectrum and all spectra used in its calculation; - Exclude spectra with correlation coefficients below a predetermined threshold (e.g., 0.9); - Calculate the new average spectrum relative to the j-th region; - Repeat the calculation process of spectral correlation and exclusion until all remaining spectra have a correlation coefficient of average spectrum greater than or equal to the predetermined threshold.

[0080] 280) Calculate diagnostic parameters, which is a function comparing at least one of the parameters calculated in steps 210, 220, 230, 235, 240, 250, 260, and 270 with the following: - Relevant parameters related to tumors detected during ultrasound examinations performed on patients subsequently confirmed to have breast cancer by histological examination; - Relevant parameters associated with suspicious tumors detected during ultrasound examinations of patients who were subsequently ruled out for breast cancer by means of biopsy or other equivalent reliable techniques; - Relevant parameters related to tissues of no interest in cancer, detected during ultrasound examinations performed on patients who have subsequently been ruled out for breast cancer.

[0081] Therefore, it is clear that comparisons can only be made after examining a statistically significant number of patients, for whom the possible tumors have subsequently been confirmed by other tests, and thus a dataset is available that is obtained based on the just described and has been confirmed relative to the nature of the tumor (Zt), relative to the nature of the tumor (Znt), and relative to the tissue portion not of interest to the ultrasound-visible suspicious area (Tnz).

[0082] Therefore, the comparison occurs after the following process is defined by the reference value of the parameter.

[0083] 300) For each parameter calculated at points 210 to 250, define the mean tumor value (VMt) and the relative confidence intervals of 75% and 95%, starting from the statistical distribution of the values ​​of each parameter in all suspicious regions belonging to the set (Zt) of suspicious regions whose tumor nature has been confirmed; 310) For each parameter calculated at points 210 to 250, define the non-tumor mean (VMnt) and the 75% and 95% relative confidence intervals, starting from the statistical distribution of the values ​​of each parameter in all suspicious regions belonging to the set (Znt) of suspicious regions whose tumor nature has been excluded; 320) For each parameter calculated at points 210 to 250, define the healthy mean (VM) and the 75% and 95% relative confidence intervals, starting from the statistical distribution of the values ​​of each parameter for all parts of the tissue belonging to the set (Tnz) of tissues not of interest in the ultrasound-visible suspicious area.

[0084] After point 320, three sets of reference values ​​are defined for each parameter: the first set is associated with the tumor region, the second set is associated with suspicious regions where the result is not a tumor, and the third set is associated with tissues that do not belong to suspicious regions (healthy tissues).

[0085] (Referring to the calculation of tumor reference values, step 310) Figure 9 The flowchart shows that, for all tumor spots (XT1, ... ZTn), all parameters (morphological, acoustic attenuation, relative to BUB and IRC, and relative to the spectrum) are calculated, and then the average of all values ​​calculated for a single tumor region is calculated for each parameter to obtain a set of tumor reference values ​​including the average for each parameter.

[0086] The procedure is the same for non-tumor suspicious areas and for tissues of interest that are not visible on ultrasound.

[0087] This method provides steps for selecting the following important parameters.

[0088] 330) Select the reference set parameters for both the 75% confidence interval between the mean tumor value (VMt) and the 75% confidence interval between the mean healthy value (VMs) and the mean non-tumor value (VMnt) in the reference set parameters.

[0089] The process of defining the reference value also includes a series of calculations of the reference spectrum.

[0090] 340) Select multiple segments of ultrasound propagation from data obtained from ultrasound examinations performed on patients with a confirmed diagnosis of breast cancer (presence / absence): - Contained within the tumor area; - Contained within a suspicious area that was not subsequently found to be a tumor; - Contained within an external organization in the suspected area; 350) Calculate the frequency transformation of the original radio frequency ultrasound signal associated with each segment in each segment of ultrasound propagation; 360) Calculation: -Tumor reference spectrum, obtained as the average spectrum relative to all ultrasound signals associated with the tumor region; - Non-tumor reference spectrum, obtained as the average spectrum of all ultrasound signals associated with non-tumor regions; - The reference spectrum relative to the external tissue of the suspected area is obtained as the average spectrum of all ultrasound signals associated with the external tissue of the suspected area.

[0091] Furthermore, preferably, the method includes the step of selecting the ultrasound signal to be used for calculating the reference spectrum, as defined below: 370) For each reference spectrum calculated at point 360: - Calculate the correlation coefficient between the reference spectrum and all the spectra used in its calculation; - Exclude spectra with correlation coefficients below a predetermined threshold (e.g., 0.9); - Calculate the new reference spectrum; - Repeat the calculation process of spectral correlation and exclusion until the correlation coefficients of all remaining spectra with the reference spectrum are greater than or equal to the predetermined threshold.

[0092] Once the reference values ​​of the parameters and the reference spectrum are known, a diagnostic parameter indicating the probability that the j-th region (Zj) is a tumor is calculated.

[0093] In the first embodiment, the diagnostic parameter is a classification of the j-th region (Zj) as either tumor or non-tumor. The process for classifying the j-th region is as follows; 400) Calculate the correlation coefficients between the parameter set relative to the j-th region and the following items: -A reference set relative to the tumor region -A reference set relative to the non-tumor suspicious areas that subsequently turned out; - A reference set relative to organizations that do not belong to the suspicious area.

[0094] 410) If one of the three coefficients calculated at point 400) is greater than a first predetermined threshold (e.g., r > 0.9) and the remaining coefficients are less than a second predetermined threshold (e.g., r < 0.7), then if the correlation coefficient is greater than a predetermined threshold relative to a reference set of tumor regions, then the j-th region is classified as a tumor, and if the correlation coefficient is one of the other two, then it is classified as a non-tumor region. 420) If no correlation coefficient is greater than the first predetermined threshold, or if both lower correlation coefficients are not lower than the second predetermined threshold, then the j-th region is not classified.

[0095] In another embodiment, the classification process for the j-th region is as follows: 500) Calculate the correlation coefficient between the average spectrum relative to the j-th region and the following items. - Reference spectrum of the tumor region; -The subsequent results were not a reference spectrum of the suspicious area of ​​the tumor; - Reference spectrum of external organizations relative to the suspected area.

[0096] 510) If one of the three coefficients calculated at point 500 is greater than a first predetermined threshold (e.g., r > 0.9) and the remaining coefficients are less than a second predetermined threshold (e.g., r < 0.7), then if the coefficient is greater than a predetermined threshold relative to the tumor reference spectrum, then the j-th region is classified as a tumor, and if the coefficient is one of the other two, then it is classified as a non-tumor region. 520) If no correlation coefficient is greater than the first predetermined threshold, or if both lower correlation coefficients are not lower than the second predetermined threshold, then the j-th region is not classified.

[0097] In another embodiment, the classification process for the j-th region is as follows: 600) Subdivide the available reference datasets in the training and validation datasets. 610) The classification neural network is trained using the training set to classify the dataset relative to tumor or non-tumor suspicious regions. 620) Present the dataset relative to the j-th region to the trained network. 630) Classify the j-th region as tumor or non-tumor based on the network output.

[0098] In the second embodiment, the diagnostic parameter is a numerical value representing the probability that the j-th region (Zj) is a tumor or a non-tumor.

[0099] The probability that the j-th region (Zj) is a tumor can be calculated as a percentage by converting the correlation coefficient between the average spectrum of the j-th region (calculated at point 500) and the reference spectrum of the tumor region.

Claims

1. An ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis, comprising two ultrasound probes, Its features Each of the probes comprises a first array (11, 21) of piezoelectric or CMUT transducers having a first nominal frequency (f1) and a second array (12, 22) of piezoelectric or CMUT transducers having a second nominal frequency (f2). The probes are arranged opposite to each other and configured to rotate along a circular trajectory centered on the midpoint of the segments connecting them and slide relative to each other along the straight lines connecting them, such that they come into contact with the patient's breast from completely opposite portions. in Each of the arrays (11, 12, 21, 22) includes a plurality of piezoelectric or CMUT transducers positioned such that acquisition is performed on an acquisition plane orthogonal to the plane of the circular trajectory. And among them The device is configured to perform a rotation angle relative to the midpoint, orthogonal to the circular trajectory, and relative to each other, by moving the probe along the circular trajectory. Multiple scans of multiple acquisition planes (P0, P1...) The device further includes an electronic computing unit loaded with a computer program configured to execute a method according to the following steps for calculating diagnostic parameters indicating the presence or absence of breast cancer: 210) Calculate multiple morphological properties derived from ultrasound imaging of each suspicious region, the multiple morphological properties being selected from the following list: - Volume of the region; - The ratio between surface area and volume; - Maximum dimension; -Eccentricity; - The ratio of the average gray value of each suspicious area to the average gray value of the surrounding area; 220) Calculate the average acoustic attenuation coefficient (Aj) relative to at least a portion of the propagation line of the ultrasound signal contained in each suspected region; 230) Calculate the ratio between the average acoustic attenuation coefficient and the average acoustic attenuation coefficient (Ats) from the outer region to each suspected region; 235) Calculate the acoustic propagation speed of the ultrasound signal associated with at least one propagation line of the ultrasound signal passing through each suspected region and at least one propagation line of the ultrasound signal not passing through any suspected region; 260) Calculate at least one spectrum of the original radio frequency ultrasound signal emitted by one of the first arrays (11, 21) and reflected by tissue portions corresponding to segments of the propagation lines of the ultrasound signals contained in each suspicious region; 270) Calculate at least one quantitative parameter extracted from the spectrum of step 260, said at least one quantitative parameter being selected from a list including the following parameters: - Average spectrum value; - The region to which the spectrum is directed at a defined frequency interval and / or a defined intensity interval; - Spectral width at predetermined frequency intervals; - The frequency value corresponding to the maximum value of the spectrum; - The slope of the line that interpolates multiple points of the spectrum at predetermined frequency intervals; - The coefficients of a polynomial interpolated at frequency intervals containing the maximum value of the spectrum; 280) Calculate diagnostic parameters, which are functions of comparing at least one of the parameters calculated in steps 210 to 270 with corresponding parameters obtained by means of an ultrasound examination performed by the device and referencing the following: - Subsequently, histological examination was used to confirm the presence of suspicious areas for pathological examination; - Subsequently, a biopsy or other equivalent and reliable techniques are used to rule out areas suspected of containing breast cancer; - Suspicious areas are parts of organizations that do not pay attention to.

2. The ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 1, characterized in that... The device also includes a ring-shaped support (30), and wherein Each probe also includes a fixing rod (13, 23) associated with the annular support (30), such that two probes are positioned in the same direction and opposite each other, and the array (11, 12, 21, 22) is positioned in the same plane of the axis of symmetry (a) of the annular support (30).

3. The ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 1 or 2, characterized in that... The head of each probe in the probe is concave, which allows for better breast shape attachment.

4. The ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 1 or 2, characterized in that... Select the first nominal frequency (f1) to perform B-mode ultrasound imaging starting from the signal reflected from the breast tissue. And select the second nominal frequency (f2) to perform acoustic attenuation measurements, working in transmission mode with an array of other probes having the same nominal frequency.

5. The ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 1 or 2, characterized in that... The first nominal frequency (f1) is between 7 MHz and 10 MHz and the second nominal frequency (f2) is between 1 MHz and 3 MHz.

6. The ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 1 or 2, characterized in that... Each probe also includes an acoustic coupling component with the breast skin, comprising a flexible membrane (14) filled with gel (15). The combination of the membrane (14) and gel (15) is configured to be squeezed as the probe is advanced toward the breast, thereby adapting to the shape of the breast and thus ensuring mechanical coupling for ultrasound transmission.

7. The ultrasound apparatus (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 2, wherein the computer program is further configured to perform the following acquisition process for each acquired location: 100) A first B-mode ultrasound image relating to a portion of breast tissue included between the first probe (10) and the axis of symmetry is acquired by means of the first array (11) of the first probe (10) of the two ultrasound probes; 110) A second mode B ultrasound image related to a portion of the tissue included between the second probe (20) and the axis of symmetry is acquired by means of the first array (21) of the second probe (20) of the two ultrasound probes; 120) Store the raw reflected ultrasound signals, which are those detected by each probe before any further processing; 130) Ultrasonic pulses are emitted by means of the first piezoelectric or CMUT transducer (121) of the second array (12) of the first probe (10), and: - The transmitted ultrasonic signal is detected by means of the first piezoelectric or CMUT transducer (221) of the second array (22) of the second probe (20), and - The reflected ultrasonic signal is detected by means of the first piezoelectric or CMUT transducer (121) of the second array (12) of the first probe (10). 140) Repeat step 130 for each pair of piezoelectric or CMUT transducers in the second array (12, 22) of the first probe (10) and the second probe.

8. The ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 7, characterized in that... The ultrasonic devices are configured to rotate relative to each other by a predetermined angle ( A series of acquisitions are performed on planes (P0, P1) to perform multiple acquisitions, such that for each point on each acquired plane, the association is: - Value related to the intensity of the reflected ultrasound signal (S) i ), - The grayscale value associated with the B-mode ultrasound image (B i The value (B) i The group thus constitutes a 3D ultrasound model of breast volume. And among them For each location retrieved, store: - The original radio frequency reflected ultrasound signals associated with all propagation lines of the ultrasound signals of each of the first arrays (11, 21) and the second arrays (12, 22). - The original radio frequency transmitted ultrasound signal associated with all propagation lines of the ultrasound signal of the second array (12, 22).

9. The ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 8, characterized in that... The method for calculating diagnostic parameters indicating the presence or absence of breast cancer further includes the following steps before calculating the diagnostic parameters: 200) The 3D ultrasound model of the segmented breast volume to distinguish the presence of one or more suspicious areas, characterized by a brighter or darker color relative to the surrounding tissue, or in any other way characterized by any other non-uniformity detectable by means of grayscale analysis, wherein a brighter color corresponds to hyperechoicity and a darker color corresponds to hypoechoicity.

10. The ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 1, characterized in that... At step 260), the spectrum associated with each suspicious region is calculated as the average of multiple spectra relative to the corresponding propagation segments of the signal contained within each suspicious region. And among them To calculate the average spectrum, the following spectrum selection procedure is performed to calculate the average spectrum relative to each suspected region: - Calculate the correlation coefficient between the average spectrum and all spectra used in its calculation; - Exclude spectra with correlation coefficients below a predetermined threshold; - Calculate the new average spectrum relative to each suspicious region; - Repeat the calculation process of spectral correlation and exclusion until all remaining spectra have a correlation coefficient with the average spectrum that is greater than or equal to the predetermined threshold.

11. The ultrasound apparatus (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 10, characterized in that... The comparison in step 280) occurs after the following definition process of a set of reference values ​​for tumors, non-tumor cells, and healthy individuals, which are detected by means of the device in ultrasound examinations performed on patients for whom tumor pathology has been confirmed or ruled out by histological examination: 300) For each parameter calculated in steps 210 to 250, define the mean tumor value (VMt) and the 75% and 95% relative confidence intervals, starting from the statistical distribution of the values ​​of each parameter in all suspicious regions with confirmed tumor nature; 310) For each parameter calculated in steps 210 to 250, define the non-tumor mean (VMnt) and the 75% and 95% relative confidence intervals, starting from the statistical distribution of the values ​​of each parameter in all suspicious regions where tumor nature (Znt) has been excluded; 320) For each parameter calculated in steps 210 to 250, define the healthy mean (VM) and the 75% and 95% relative confidence intervals, starting from the statistical distribution of the values ​​of each parameter for all parts of the tissue belonging to the set of healthy tissue parts not of interest in the ultrasound-visible suspicious area (Tnz); 340) Select multiple segments for the propagation of the ultrasound signal: - Contained within the tumor area; - Contained within a suspicious area that was not subsequently found to be a tumor; - Included within organizations located outside the suspected area; 350) Calculate the frequency transformation of the radio frequency raw ultrasound signal associated with each segment of the segment in which the ultrasound signal propagates; 360) Calculation: -Tumor reference spectrum, obtained as the average spectrum relative to all ultrasound signals associated with the tumor region; - Non-tumor reference spectrum, obtained as the average spectrum of all ultrasound signals relative to suspicious areas that were subsequently found not to be tumors; - The reference spectrum relative to the external tissue of the suspected area is obtained as the average spectrum of all ultrasound signals associated with the external tissue of the suspected area.

12. The ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 11, characterized in that... The process of defining the reference set for tumors includes the following steps: 330) Select parameters at a 75% confidence interval for the mean tumor value (VMt) to exceed the relative healthy mean (VMs) and a 75% confidence interval for the relative non-tumor mean (VMnt).

13. The ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 1, characterized in that... The diagnostic parameters are used to classify each suspicious area as "tumor" or "non-tumor," and are obtained through the following operations: 400) Calculate the correlation coefficients between the parameter set relative to the j-th suspicious region and the following items. -A reference set relative to the tumor region -A reference set relative to suspicious regions that are not tumors in subsequent analyses; -A reference set relative to organizations that do not belong to the suspicious area 410) If one of the three coefficients calculated in step 400) is greater than a first predetermined threshold and the other coefficients are lower than a second predetermined threshold, then if the correlation coefficient is greater than a predetermined threshold relative to the reference set of the tumor region, then the j-th suspicious region is classified as a tumor, and if the correlation coefficient is one of the other two, then the j-th suspicious region is classified as a non-tumor. 420) If no correlation coefficient is greater than the first predetermined threshold, or if both lower correlation coefficients are not lower than the second predetermined threshold, then the j-th suspicious region is not classified and further diagnostic examination is recommended.

14. The ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 1, characterized in that... The diagnostic parameters are used to classify each suspicious area as either tumor or non-tumor, and are obtained through the following operations: 500) Calculate the correlation coefficient between the average spectrum relative to the j-th suspected region and the following items. -Tumor reference spectrum; -The subsequent results were not a reference spectrum of the suspicious area of ​​the tumor; - The reference spectrum of the external tissue relative to the suspicious region calculated at step 360); 510) If one of the three coefficients calculated in step 500 is greater than a first predetermined threshold and the other coefficients are lower than a second predetermined threshold, then if the correlation coefficient is greater than a predetermined threshold with respect to the tumor reference spectrum, then the j-th suspicious region is classified as a tumor, and if the correlation coefficient is one of the other two, then the j-th suspicious region is classified as a non-tumor. 520) If no correlation coefficient is greater than the first predetermined threshold, or if both lower correlation coefficients are not lower than the second predetermined threshold, then the j-th suspicious region is not classified.

15. The ultrasound device (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 1, characterized in that... The diagnostic parameters are used to classify each suspicious area as either tumor or non-tumor, and are obtained through the following operations: 600) Subdivide the available reference datasets in the training and validation datasets. 610) The classification neural network is trained using the training dataset to classify the dataset relative to tumor or non-tumor suspicious regions. 620) Present the dataset relative to the j-th suspicious region to the trained network. 630) Classify the j-th suspicious region as tumor or non-tumor based on the network output.

16. The ultrasound apparatus (1) for performing a breast diagnostic examination for breast cancer diagnosis according to claim 14, characterized in that... The diagnostic parameter is a numerical value representing the probability that each suspicious region is a tumor or not, calculated by converting the correlation coefficient between the average spectrum of each suspicious region (calculated at step 500) and the reference spectrum of the tumor region into a percentage value.

Images