Method and apparatus for analyzing tissue texture - Patents.com

Abstract

A method and apparatus for analyzing the texture of tissue is provided. The invention relates to a method for analysing the texture of human or animal tissue from digitised images (obtained by an X-ray based imaging system), comprising the step of calculating at least one texture score B of the image by applying an empirical variogram to the texture of the tissue (typically the grey levels of the digitised image).

Description

[Technical field]

[0001] The present invention relates to a method and apparatus for analysing the texture of human or animal tissue from digitised images. The present invention also relates to, but is not limited to, bone health, medical x-ray based imaging, primary and secondary osteoporosis, and fragility fractures. [Background technology]

[0002] In the early 1990s, the World Health Organization (WHO) conceptualized osteoporosis as a systemic skeletal disease characterized by low bone mass (reduced quantity) and microarchitectural deterioration (deteriorated quality) of bone tissue, resulting in increased bone fragility and susceptibility to fracture (Consensus development conference: diagnosis, prophylaxis, and treatment of osteoporosis. Am J Med 94, 646-650 (1993)). Furthermore, in the early 2000s, the National Institutes of Health (NIH) defined osteoporosis as a skeletal disease that reduces bone strength and increases the risk of fracture (Osteoporosis prevention, diagnosis, and therapy. Jama. 2001;285(6):785-95.). Essentially, in osteoporosis, reduced bone strength leads to a traumatic outcome: fragility fractures.

[0003] Bone strength reflects two main characteristics: bone mass (BMD) and bone quality. BMD is expressed as the amount of mineral per unit area or volume and is determined by an individual's maximum bone mass and the amount of bone loss. Bone quality refers to bone structure, turnover, accumulation of damage (e.g., microfractures), and mineralization. Bone structure is a collective term used for a variety of different entities and can be further refined. At the macroscopic level, bone structure is known as bone macroarchitecture, which describes the overall shape and geometric form of the bone, as well as its differentiation into cancellous bone (also called trabecular bone) and cortical bone. The most typical parameters describing macroarchitecture are cortical bone thickness, moment of inertia, and other geometric measurements.

[0004] At the microscopic level, i.e. when the spatial resolution of the acquired images is usually better than 100 μm, the structure of bone is known as bone microarchitecture and can be described by a combination of parameters that comprehensively describe the network structure, such as trabecular thickness and spacing, trabecular number, trabecular connectivity, and the Structural Model Index (SMI) (Macro- and Microimaging of Bone Architecture, Klaus Engelke, Sven Prevrhal, Harry K.Genant, in Principles of Bone Biology (Third Edition) Volume II, 2008, Pages 1905-1942. Editors: John Bilezikian Lawrence Raisz T. John Martin -Publisher: Academic Press. Published Date: 29th September 2008).

[0005] Osteoporosis is classified as primary or secondary based on its etiology. Primary or age-related osteoporosis is the most common type (Mirza F, Canalis E. Management of endocrine disease: Secondary osteoporosis: pathophysiology and management. European journal of endocrinology. 2015;173(3): R131-51). Physiologically, peak bone mass is usually reached by the age of 30 in both women and men. Thereafter, women lose bone mass more rapidly than men, and this bone loss becomes more pronounced after menopause due to estrogen deficiency. Therefore, this type of osteoporosis is more prevalent in women and is also called postmenopausal osteoporosis. Its main cause is the deficiency of sex hormones associated with aging.

[0006] Secondary osteoporosis occurs when underlying disease (genetic: cystic fibrosis, endocrine: diabetes, hypogonadism, hyperthyroidism, hyperparathyroidism, Cushing's disease, autoimmune: rheumatoid arthritis, etc.), lifestyle (smoking, alcohol abuse, sedentary lifestyle), and / or drug use (glucocorticoids, hypogonadotropic drugs, etc.) causes excessive bone loss. This type of osteoporosis occurs primarily in young adults and in the majority of men with osteoporosis. Approximately one-third of postmenopausal women and one-half of premenopausal women and men have secondary osteoporosis.

[0007] In any case, the hallmark of osteoporosis is fragility fractures. Fragility fractures are defined as fractures that occur due to falls from standing in response to mechanical forces that do not usually result in fracture (Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. Lancet (London, England). 2002;359(9319):1761-7. Warriner AH, Patkar NM, Yun H, Delzell E. Minor, major, low-trauma, and high-trauma fractures: what are the subsequent fracture risks and how do they vary? Current osteoporosis reports. 2011;9(3):122-8.). The proximal femur (hip), spine, humerus, and forearm are the most common skeletal sites where fragility fractures occur and are referred to as major osteoporotic fractures. It is estimated that after age 50, one in two women and one in four men will experience a major osteoporotic fracture in their remaining lives (Kanis JA, Johnell O, Oden A, Sembo I, Redlund-Johnell I, Dawson A, et al. Long-term risk of osteoporotic fracture in Malmoe. Osteoporosis international: a journal established as result of cooperation between the European Foundation for Osteoporosis and the National Osteoporosis Foundation of the USA. 2000;11(8):669-74.). In women over age 45, osteoporosis results in more hospital stays than many other diseases, including diabetes, myocardial infarction, and breast cancer. Furthermore, a prior fracture increases the risk of a subsequent fracture by 86%.Fractures are associated with high morbidity and mortality and are often a precursor to disability, loss of independence, and premature death in older adults. (Binkley N, Blank RD, Leslie WD, Lewiecki EM, Eisman JA, Bilezikian JP. Osteoporosis in Crisis: It's Time to Focus on Fracture. Journal of bone and mineral research: the official journal of the American Society for Bone and Mineral Research. 2017;32(7):1391-4.)

[0008] Once fracture risk is identified, preventive measures are taken at certain "hierarchical" levels: general lifestyle advice, appropriate calcium and / or vitamin D supplementation in addition to a healthy lifestyle, and / or pharmacotherapy. Based on the mechanism of action, there are two types of pharmacotherapy for osteoporosis: antiresorptive drugs (e.g., bisphosphonates, estrogen agonists / antagonists, estrogen, calcitonin, denosumab) that inhibit bone resorption, and osteogenic drugs (e.g., teriparatide) that stimulate bone formation. Recently, romosozumab, which has osteogenic effects, has been approved (Tu KN, Lie JD, Wan CKV, Cameron M, Austel AG, Nguyen JK, et al. Osteoporosis: A Review of Treatment Options. P & T: a peer-reviewed journal for formulary management. 2018;43(2):92-104.).

[0009] Clearly, fractures - this highly traumatic experience - and their management represent a major social, economic, and health burden that is intensified by the aging of the population and the increase in life expectancy. This is a major public health challenge and requires a special research focus. Since independence is one of the main factors that define healthy aging of the population, it is of great importance to prevent its loss by preventing fractures. The overall management and prevention of osteoporosis requires accurate clinical assessment (direct or indirect) of bone strength, bone resilience, and fracture risk.

[0010] In clinical practice, bone mineral density (BMD) measured by dual-energy x-ray absorptiometry (DXA) has become the gold standard for diagnosing osteoporosis in the absence of established fragility fractures. Although BMD is one of the primary determinants of bone strength and fracture risk, there is a significant overlap (up to 40%) in BMD values ​​between individuals who will and will not develop fractures. In essence, defining osteoporosis based solely on projected bone mineral density (BMD by DXA) is considered to have reached its limits. Indeed, the multifactorial nature of the disease encourages the current definition of osteoporosis to evolve towards a combined risk model based on clinical risk factors (CRFs) and BMD (e.g., FRAX®). Taking CRFs into account when assessing BMD increases the sensitivity of screening without sacrificing specificity. However, although some of the limitations in the current use of DXA are addressed by the concomitant use of CRFs, information on bone microarchitecture is only partially taken into account. Thus, additional information on microarchitecture should help reduce the significant overlap that exists between those truly at increased fracture risk and those who are not. However, ideally, such information should be available in the clinical setting, without significantly interfering with clinical workflow and without exposing patients to additional ionizing radiation.

[0011] EP1576526 describes a method for determining the mechanical strength of bone from a two-dimensional image of the bone. The method does not depend on BMD or only on BMD. The method allows the measurement of a parameter called α (also known as Trabecular Bone Score (TBS)).

[0012] It is an object of the present invention to provide a method or device for analyzing bones that is not dependent on BMD or solely on BMD, such as:

[0013] - improving the ability to identify or predict fractures; and / or - To improve the reproducibility of measurements, and / or - have a better correlation with bone microarchitecture, and / or - insensitive to the physiological characteristics of the patient, e.g. the volume and / or nature of the soft tissue surrounding the bone; and / or - Insensitive to the choice of technical parameters of image acquisition. Summary of the Invention

[0014] One aspect of the invention relates to a method or process (preferably computer-implemented) for analyzing the texture of human or animal tissue from a digitized image (preferably obtained by an X-ray based imaging system), comprising the step of calculating at least one texture score B of the image by applying an empirical variogram to the texture of the tissue, typically the grey levels of the digitized image.

[0015] The method or process according to the present invention may further include: - at least one patient factor associated with the patient on the image, and / or - At least one technical factor related to the acquisition of the image It may include applying a robustness improvement step that takes into account

[0016] The at least one patient factor can include:

[0017] - Patient morphological effects including at least one of the following: - the effect of soft tissue, and / or tissue thickness, and / or its distribution, and / or its composition, and / or indirect substitutes, and / or - The patient's weight and / or body mass index (BMI), and / or - Patient height, and / or - The effect of at least one medical condition or condition on the patient; and / or -Effect of patient positioning during image acquisition.

[0018] The at least one technical factor may include:

[0019] Possibly faulty detectors and sensors for image acquisition, and / or - the effect of scanning modes and settings for image acquisition; and / or - the technical characteristics of the imaging device used to acquire the images; - the effect of variability between imaging systems for image acquisition; and / or - the signal-to-noise ratio (SNR) of the image, and / or - Image resolution.

[0020] The tissue may be human tissue.

[0021] The tissue may be bone tissue (preferably human tissue) and the texture score B is preferably a bone texture score B. In a less preferred variant, the tissue is a soft tissue.

[0022] The digitized two-dimensional image may be selected in an area having cancellous bone structure.

[0023] A method or process according to the present invention may include:

[0024] - Each pixel Pi = (xi, yi) ∈ S has its gray level value h(P i ) to determine an optimized pixel sampling S of the image.

[0025] - for at least one region of interest (ROI) of the pixel sampling S, selecting a set of predefined directions I depending on the given region of interest (ROI). In a variant, the set of predefined directions is not selected depending on the ROI but is a default set of predefined directions I.

[0026] - For each pixel, calculate at least one variogram of the gray levels of the sampling S by moving from this pixel a distance r ∈ [1, R0] along these directions I. The variograms are calculated for each given direction or simultaneously for all given directions.

[0027] -Evaluate at least one of the following parameters on a log-log scale for each variogram of each pixel and / or for the global variogram of a sampling S combining the variograms of each pixel:

[0028] - initial gradient a, - sill b, which represents the asymptotic value of the variogram, - range c, which represents the distance at which the variogram curve transitions from a quasi-linear progression to an asymptotic behavior; - the nugget d, which represents the initial value of the variogram, and - Area under the variogram curve e.

[0029] -Combine the following:

[0030] - combining the evaluated parameters (preferably a, b, c, d and / or e) obtained for each pixel into a texture score B for each pixel; and / or - combining the evaluated parameters (preferably a, b, c, d and / or e) obtained for a sampling S into a texture score B of the sampling S, and / or - combining the estimated parameters (preferably a, b, c, d and / or e) obtained for each pixel into a texture score B of the sampling S;

[0031] This combining step further comprises applying a robustness improving step related to at least one patient factor and / or at least one technical factor to the texture score B, preferably as follows: - determining and / or correcting at least one parameter used to calculate or determine score B before or during the determination or calculation of score B as a function of patient and / or technical factors. Preferably, - at least one of the parameters R0, a, b, c, d, and / or e; and / or at least one parameter (α, β, γ, δ, ε) used to give respective weights between a, b, c, d, and / or e for calculating the texture score B; and / or - Adjust score B as a function of patient and / or technical factors.

[0032] The set of V predetermined directions I may depend on: - a bony skeletal site, if the human or animal tissue is bone, and / or Region of Interest (ROI); and / or - Image resolution, and / or - Image signal-to-noise ratio.

[0033] The step of selecting a set of predetermined directions I includes selecting a set of N direction vectors

number

[0034] The step of moving a distance r is for each pixel Pi = (xi, yi) ∈ S and for each direction

number

number

number

[0035] The variogram may be calculated simultaneously for all given directions. The step of calculating the variogram V of the grey level as a function of distance rh is preferably performed by averaging the squared difference of h over pairs of pixels separated by a distance r, according to the following formula:

number

[0036] A variogram may be calculated for each given direction, where h(O) is the grey level of an initial given pixel before displacement and h(r) is the grey level of a given new pixel after displacement a distance r along one of the given directions from the initial given pixel. The variogram is preferably calculated by the following formula: V i (r) = [h(r) - h(O)] 2 (where i∈I)

[0037] Each estimated parameter (preferably a, b, c, d and / or e) may be evaluated as a least-squares regression model of the considered variogram.

[0038] The evaluated parameters (preferably a, b, c, d and / or e) may be combined into a texture score B using linear or non-linear equations, depending on the clinical situation.

[0039] A method or process according to the invention may comprise evaluating, for each variogram of each pixel and / or for a global variogram of a sampling S combining the variograms of each pixel, at least one of the following parameters on a log-log scale:

[0040] - sill b, which represents the asymptotic value of the variogram, - range c, which represents the distance at which the variogram curve transitions from a quasi-linear progression to an asymptotic behavior; - the nugget d, which represents the initial value of the variogram, and - Area under the variogram curve e.

[0041] A method or process according to the invention may comprise evaluating, for each variogram of each pixel and / or for a global variogram of a sampling S combining the variograms of each pixel, at least two of the following parameters on a log-log scale:

[0042] - initial gradient a, - sill b, which represents the asymptotic value of the variogram, - range c, which represents the distance at which the variogram curve transitions from a quasi-linear progression to an asymptotic behavior; - the nugget d, which represents the initial value of the variogram, and - Area under the variogram curve e.

[0043] A method or process according to the invention may comprise evaluating, for each variogram of each pixel and / or for a global variogram of a sampling S combining the variograms of each pixel, an initial gradient a and at least two of the following parameters in a log-log scale:

[0044] - sill b, which represents the asymptotic value of the variogram, - range c, which represents the distance at which the variogram curve transitions from a quasi-linear progression to an asymptotic behavior; - the nugget d, which represents the initial value of the variogram, and - Area under the variogram curve e. The texture score B is preferably dimensionless.

[0045] Another aspect of the invention relates to a computer program comprising instructions for performing the steps of a process or method according to the invention when executed by a computer.

[0046] Another aspect of the invention relates to a computer program product comprising instructions which, when executed by a computer, cause the computer to carry out the steps of a process or method according to the invention.

[0047] Another aspect of the invention relates to a computer-readable storage medium containing instructions which, when executed by a computer, cause the computer to perform steps of a process or method according to the invention.

[0048] Another aspect of the invention relates to an apparatus for analysing the texture of human or animal tissue from digitised images (preferably obtained by an X-ray based imaging system), comprising means configured and / or programmed to calculate at least one texture score B of the image by applying an empirical variogram to the texture of the tissue (typically the grey levels of the digitised image).

[0049] The device of the present invention further adds the following to the texture score B: - at least one patient factor associated with the patient on the image, and / or - At least one technical factor related to the acquisition of the image The method may further comprise means configured and / or programmed to apply a robustness improvement step taking into account:

[0050] The at least one patient factor can include: - Patient morphological effects including at least one of the following: - the effect of soft tissue and / or tissue thickness and / or its distribution in the patient and / or its composition and / or indirect substitutes and / or - The patient's weight and / or body mass index (BMI), and / or - Patient height, and / or - The effect of at least one medical condition or condition on the patient; and / or -Effect of patient positioning during image acquisition.

[0051] The at least one technical factor may include: Possibly faulty detectors and sensors for image acquisition, and / or - the effect of scanning modes and settings for image acquisition; and / or - the technical characteristics of the imaging device used to acquire the images - the effect of variability between imaging systems for image acquisition; and / or - the signal-to-noise ratio (SNR) of the image, and / or - Image resolution.

[0052] The tissue may be human tissue. The tissue may be bone tissue (preferably human tissue) and the texture score B is preferably a bone texture score B. In a less preferred variant, the tissue is a soft tissue. The digitized two-dimensional image may be an image that images cancellous bone structure.

[0053] The device of the invention may comprise: - Each pixel Pi = (xi, yi) ∈ S has its gray level value h(P i means configured and / or programmed to determine an optimized pixel sampling S of an image, such that the optimized pixel sampling S of the image has - means configured and / or programmed to select, for at least one region of interest (ROI) of the pixel sampling S, a set of predefined directions I depending on the given region of interest (ROI); this means is optional, since in a variant the set of predefined directions is not selected depending on the ROI but is a default set of predefined directions I. means configured and / or programmed to calculate, for each pixel, at least one variogram of the grey levels of the sampling S by moving from this pixel by a distance r ∈ [1,R0] along these directions I, configured and / or programmed to calculate the variograms for each predetermined direction or simultaneously for all predetermined directions, - means configured and / or programmed to evaluate, for each variogram of each pixel and / or for a global variogram of a sampling S combining the variograms of each pixel, parameters from the variograms, preferably configured and / or programmed to evaluate at least one of the following parameters on a log-log scale:

[0054] - initial gradient a, - sill b, which represents the asymptotic value of the variogram, - range c, which represents the distance at which the variogram curve transitions from a quasi-linear progression to an asymptotic behavior; - the nugget d, which represents the initial value of the variogram, and - area under the variogram curve e,

[0055] -means configured and / or programmed to combine: - combining the evaluation parameters obtained for each pixel (preferably a, b, c, d and / or e) into a texture score B for each pixel, and / or - combining the evaluation parameters (preferably a, b, c, d and / or e) obtained for a sampling S into a texture score B of the sampling S, and / or - combining the evaluation parameters obtained for each pixel (preferably a, b, c, d and / or e) into a texture score B of the sampling S

[0056] The device of the present invention may further comprise means configured and / or programmed to apply a robustness improvement step related to at least one patient factor and / or at least one technical factor to the texture score B, preferably as follows:

[0057] - by determining and / or correcting at least one parameter used to calculate or determine score B before or during the determination or calculation of score B as a function of patient and / or technical factors, preferably - at least one of the parameters R0, a, b, c, d, and / or e, and / or At least one parameter (α, β, γ, δ, ε) used to give respective weights between a, b, c, d, and / or e in order to calculate the texture score B. and / or - by correcting score B as a function of patient and / or technical factors.

[0058] The set of predefined directions I may depend on: - a bony skeletal site, if the human or animal tissue is bone, and / or Region of interest (ROI), and / or - Image resolution, and / or - Image signal-to-noise ratio.

[0059] The means for selecting a set of predetermined directions I is configured and / or programmed to: - A set of N direction vectors

number

[0060] The means configured and / or programmed to calculate at least one variogram is

number

number

number

[0061] The means configured and / or programmed to calculate the at least one variogram may be configured and / or programmed to calculate the variograms simultaneously for all given directions. The means configured and / or programmed to calculate the at least one variogram is preferably configured and / or programmed to calculate the variogram as a function of grey level h by averaging the squared difference of h over a number of pairs of pixels separated by a distance r, preferably with the following formula:

number

[0062] The means configured and / or programmed to calculate the at least one variogram may be configured and / or programmed to calculate a variogram for each predefined direction, where h(O) is the grey level of the initial predefined pixel before the movement and h(r) is the grey level of the predefined new pixel after moving a distance r from the initial predefined pixel along one of the predefined directions. The means configured and / or programmed to calculate the at least one variogram is preferably configured and / or programmed to calculate the variogram with the following formula: V i (r) = [h(r) - h(O)] 2 (where i∈I)

[0063] The means configured and / or programmed to evaluate the parameters may be configured and / or programmed to evaluate each parameter (preferably a, b, c, d, and / or e) as a least squares regression model of the considered variogram.

[0064] The means configured and / or programmed to combine the parameters may be configured and / or programmed to combine the assessment parameters (preferably a, b, c, d and / or e) into a texture score B using a linear or non-linear equation, depending on the clinical situation.

[0065] The means configured and / or programmed to evaluate the parameters may be configured and / or programmed to evaluate, for each variogram of each pixel and / or for a global variogram of a sampling S combining the variograms of each pixel, at least one of the following parameters in a log-log scale: - sill b, which represents the asymptotic value of the variogram, - range c, which represents the distance at which the variogram curve transitions from a quasi-linear progression to an asymptotic behavior; - the nugget d, which represents the initial value of the variogram, and - Area under the variogram curve e.

[0066] The means configured and / or programmed to evaluate the parameters may be configured and / or programmed to evaluate, for each variogram of each pixel and / or for a global variogram of a sampling S combining the variograms of each pixel, at least two of the following parameters in a log-log scale: - initial gradient a, - sill b, which represents the asymptotic value of the variogram, - range c, which represents the distance at which the variogram curve transitions from a quasi-linear progression to an asymptotic behavior; - the nugget d, which represents the initial value of the variogram, and - Area under the variogram curve e.

[0067] The means configured and / or programmed to evaluate the parameters may be configured and / or programmed to evaluate, for each variogram of each pixel and / or for a global variogram of a sampling S combining the variograms of each pixel, an initial gradient a and at least one of the following parameters on a log-log scale: - sill b, which represents the asymptotic value of the variogram, - range c, which represents the distance at which the variogram curve transitions from a quasi-linear progression to an asymptotic behavior; - the nugget d, which represents the initial value of the variogram, and - Area under the variogram curve e.

[0068] The texture score B is preferably dimensionless. Other advantages and features of the present invention will become apparent from a consideration of the detailed description of the non-limiting embodiments and the accompanying drawings. [Brief description of the drawings]

[0069] [Figure 1] 1 shows images reconstructed from a DXA spine scan. [Diagram 2] 1 is an embodiment of the method of the present invention showing sampling areas corresponding to measurement areas in a best realization mode; [Diagram 3] 13 shows sub-sampling areas of a bone site corresponding to measurement areas in a variant of the invention. [Figure 4] 1 shows an image reconstructed from a DXA spine scan with determination of a set of directions I in an embodiment of the method of the present invention. [Diagram 5] 1 shows an example of a variogram V and its least squares regression curve when the calculation range R0 is 20 mm in an embodiment of the method of the present invention. [Figure 6] 5 shows a curve V (corresponding to FIG. 5) on a log-log scale for an embodiment of the method of the present invention. [Figure 7] A curve V on a log-log scale for an embodiment of the method of the present invention is shown with parameters a, b, c, d, and e. [Figure 8] The composition and / or correlation of bone strength and bone texture are shown. [Figure 9] 1 shows control curves for age, ethnicity, and gender in an embodiment of the method of the present invention. [Figure 10] 1 is a table combining bone texture score B with bone mineral density status or BMD in an embodiment of the method of the present invention to obtain a bone health category (1 to 9) corresponding to fracture risk. [Figure 11] In an embodiment of the method of the present invention, the fracture risk assessment tool (FRAX®) adjusted by texture score B is shown. [Figure 12A] FIG. 2 is a block diagram illustrating steps a) to e) of an embodiment of the method of the present invention. [Figure 12B] FIG. 2 is a block diagram illustrating steps f) to i) of an embodiment of the method of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0070] Hereinafter, these embodiments are not limiting, and variations of the present invention that include only a selection of the features described below are also included within the scope of the present invention, if the selection is sufficient to provide a technical advantage or to distinguish the present invention from the prior art. This selection includes at least one feature, preferably a functional feature without structural details, or if only a part of structural details, the part is sufficient to provide a technical advantage or to distinguish the present invention from the prior art.

[0071] In this specification, the following definitions are used (see FIG. 8).

[0072] Bone strength: defined as the ultimate stress that a bone structure can withstand before fracture, the maximum stress on the stress-strain curve. There are many determinants of bone strength, including anatomical properties such as mineral density, micro- and macro-architecture, and geometric morphology. Includes bone mass and quality.

[0073] Bone mass or bone density: corresponds to the total amount of mineralized tissue in the bone. It is evaluated by bone density, which is evaluated in grams per unit volume or unit area. Bone density can be assessed by DXA bone densitometers, quantitative computed tomography (QCT), peripheral quantitative computed tomography (pQCT) or high-resolution peripheral quantitative computed tomography (HR-pQCT) scanners, or dedicated ultrasound devices.

[0074] Bone quality: corresponds to the overall qualitative state of bone. According to the NIH definition, it is composed of bone turnover, bone mineralization, accumulation of bone lesions, and bone structure or bone architecture. "Bone structure" is divided into macrostructure and microstructure.

[0075] Macrostructure: Primarily refers to the overall geometry of the bone, its shape and distribution, including its differentiation into cortical and cancellous bone (also called trabecular bone).

[0076] Microstructure: This corresponds to the internal structure of trabecular bone and can be described by a combination of parameters that comprehensively describe the network structure, such as trabecular thickness and spacing, trabecular number, trabecular connectivity, and the structure model index (SMI).

[0077] Bone mineralization: corresponds to the quantity and specific quality of the organic matrix of collagen and hydroxyapatite crystals.

[0078] Bone turnover: Bone turnover includes the processes of bone resorption and bone formation. Bone turnover can be assessed indirectly by measurement of biochemical markers (blood and / or urine) and by analysis of biopsy samples (e.g., by bone morphometry).

[0079] Accumulation of bone damage: This corresponds to the gradual accumulation of damage in the bone.

[0080] Bone texture: A statistical property of bone that can be related to the macro- and microstructure of bone.

[0081] Bone resilience index: An index of resistance to fracture based on BMD value or a combination of BMD T-score and texture score.

[0082] Fracture Risk: Related to conditions and medical conditions involving a disruption of bone continuity. Related to bone strength and bone repair.

[0083] Embodiments of an apparatus according to the invention for implementing embodiments of a method according to the invention will now be described with reference to figures 1 to 12.

[0084] Based on prior art methods described in patent documents (CA2549281A1, EP1576526B1, FR_2848694, US_7609867), the analysis of two-dimensional medical X-ray images is enhanced according to the present invention to determine certain aspects of bone quality, e.g. the bone texture score B, which relates to bone macro- and microstructure.

[0085] The bone texture score B in an embodiment of the present invention can predict osteoporotic fractures in both men and women, independent of bone mineral density.

[0086] In an embodiment of this invention, the bone texture score B is interpreted in relation to age- and sex-matched controls and the patient's risk of fragility fracture (location relative to different states of bone microarchitecture based on quartiles: a high bone texture score B is consistent with normal microarchitecture and reduced risk of fragility fracture, a low bone texture score B is consistent with poor microarchitecture and increased risk of fragility fracture).

[0087] A method embodiment of the invention is a method for analysing the texture of human or animal tissue from digitised images (preferably obtained by an X-ray based imaging system), comprising the step of calculating at least one texture score B of the images by applying an empirical variogram to the texture of the tissue (typically the grey levels of the digitised images). As explained in step i) below, method embodiments of the invention preferably apply to the texture score B (preferably to the calculation of the texture score B) a robustness improvement step that takes into account:

[0088] - at least one patient factor associated with the patient on the image, and / or - At least one technical factor related to the acquisition of the image.

[0089] By "variogram" (unless specified as being theoretical or experimental or empirical) herein is meant "empirical variogram" (also referred to by those skilled in the art as "empirical variogram").

[0090] In statistics, a theoretical variogram is a function that describes the degree of spatial dependence of a spatial random field or stochastic process.

[0091] A theoretical variogram can be defined as the mean squared difference of values ​​between two points (in space, in an image, etc.) separated by the distance between the two points.

[0092] The theoretical variogram can be equivalently defined as the variance of the difference in field values ​​at two locations.

[0093] The empirical variogram is an estimator of the theoretical variogram from data. Empirical variograms are used for measured data where sample information is not available at all positions. For example, the variation in intensity of the pixels of a radiological image acquired by the sensor of an X-ray machine, where each pixel Pi = (xi,yi) ∈ S has its gray level value h(P i ).

[0094] The concept of a variogram is clear to those skilled in the art, see for example https: / / fr.wikipedia.org / wiki / Variogramme and https: / / en.wikipedia.org / wiki / Variogram.

[0095] Also see the following documents:

[0096] -Matheron, Georges (1963). “Principles of geostatistics”. Economic Geology. 58 (8): 1246-1266. doi:10.2113 / gsecongeo.58.8.1246. ISSN 1554-0774. -Ford, David. “The Empirical Variogram” -Bachmaier, Martin; Backes, Matthias (2008-02-24). “Variogram or semivariogram? Understanding the variances in a variogram”. Precision Agriculture. Springer Science and Business Media LLC. 9 (3): 173-175. doi:10.1007 / s11119-008-9056-2. ISSN 1385-2256. -Nguyen, H.; Osterman, G.; Wunch, D.; O’Dell, C.; Mandrake, L.; Wennberg, P.; Fisher, B.; Castano, R. (2014). “A method for colocating satellite XCO2 data to ground-based data and its application to ACOS-GOSAT and TCCON”. Atmospheric Measurement Techniques. 7 (8): 2631-2644. Bibcode:2014AMT.....7.2631N. doi:10.5194 / amt-7-2631-2014. ISSN 1867-8548. -Arregui Mena, J.D.; et al. (2018). “Characterisation of the spatial variability of material properties of Gilsocarbon and NBG-18 using random fields”. Journal of Nuclear Materials. 511: 91-108. Bibcode:2018JNuM..511...91A. doi:10.1016 / j.jnucmat.2018.09.008. -Schiappapietra, Erika; Douglas, John (April 2020). “Modelling the spatial correlation of earthquake ground motion: Insights from the literature, data from the 2016-2017 Central Italy earthquake sequence and ground-motion simulations”. Earth-Science Reviews. 203: 103139. Bibcode:2020ESRv..20303139S. doi:10.1016 / j.earscirev.2020.103139. -Sokolov, Vladimir; Wenzel, Friedemann (2011-07-25). “Influence of spatial correlation of strong ground motion on uncertainty in earthquake loss estimation”. Earthquake Engineering & Structural Dynamics. 40 (9): 993-1009. doi:10.1002 / eqe.1074. -Olea, Ricardo A. (1991). Geostatistical Glossary and Multilingual Dictionary. Oxford University Press. pp. 47, 67, 81. ISBN 9780195066890. -Cressie, N., 1993, Statistics for spatial data, Wiley Interscience. -Chiles, J. P., P. Delfiner, 1999, Geostatistics, Modelling Spatial Uncertainty, Wiley-Interscience. -Wackernagel, H., 2003, Multivariate Geostatistics, Springer. -Burrough, PA and McDonnell, RA, 1998, Principles of Geographical Information Systems. -Isobel Clark, 1979, Practical Geostatistics, Applied Science Publishers. -Clark, I., 1979, Practical Geostatistics, Applied Science Publishers. -David, M., 1978, Geostatistical Ore Reserve Estimation, Elsevier Publishing. -Hald, A., 1952, Statistical Theory with Engineering Applications, John Wiley & Sons, New York. -Journel, AG and Huijbregts, Ch. J., 1978 Mining Geostatistics, Academic Press. The texture score B is dimensionless.

[0097] An embodiment of the method of the present invention will now be described in detail.

[0098] Method embodiments of the present invention include obtaining a digitized image, preferably obtained by an X-ray based imaging system.

[0099] An example of an acquired two-dimensional image is shown in Figure 1. Figure 1 shows an image reconstructed from a DXA spine scan.

[0100] This digitized image may be a two-dimensional image or a projected three-dimensional image.

[0101] From X-ray imaging modalities (including but not limited to digital X-ray, projection computed tomography (CT), projection QCT, DXA, conventional X-ray images, etc.), the available images are not always suitable for texture analysis. To improve the content of the input images to the method, the images are reconstructed using raw data whenever possible. This raw data includes, for example, raw detector data, scan parameters, and analysis data.

[0102] A two-dimensional x-ray-based image is used as input for the calculation. This image is reconstructed from data acquired from the imaging system, including but not limited to sensor data, scan parameters, and scan analysis data.

[0103] The human tissue is bone tissue, but preferably is bone tissue, and the texture score B is preferably a bone texture score B (in variations of this embodiment, the tissue may be animal tissue and / or the tissue may be soft tissue).

[0104] In this embodiment, the digitized two-dimensional image is selected in an area having cancellous bone structure.

[0105] On such a reconstructed image, the calculation processing steps are steps a) to i) and j) shown in FIG.

[0106] a) Determine the optimized pixel sampling S of the image. Each pixel Pi = (xi, yi) ∈ S is assigned its gray level value h(P i ).

[0107] As shown in Figure 2, from the reconstructed input image, the variogram V i A pixel sampling S is determined (in step d)) to select the locations for evaluating In FIG. 2, the sampling S is indicated by 1 and enclosed between the lines indicated by 2.

[0108] These locations are optimized for a given clinical situation and anatomical site such that the final bone texture score B is appropriate and efficient. In the most basic sense, the sampling region corresponds to the area in the image where bone is present.

[0109] Depending on the bone site, a subsample 3 of the area where bone is present in this image (see FIG. 3) may be used to improve performance.

[0110] b) for at least one region of interest (ROI) of pixel sampling S, preferably for a plurality of ROIs, -Determine the calculation distance R0, which depends on the given region of interest (ROI). The range R0 is determined depending on the skeletal site and the image resolution. In DXA systems, the value R0 is between 1 cm and 2 cm. - Select a set of predefined directions I that depends on a given region of interest (ROI). This step of selecting a set of predefined directions I is done by dividing N directional unit vectors

number

number

[0111] The maximum value of N depends on the complexity of the bone structure of the imaged bone considered in the ROI and its image resolution. Typically, for bones with non-complex bone structure such as the spine or lumbar vertebrae, N<9, but for complex structures such as the proximal femur, N may be significantly increased.

[0112] For example, N=3, 4, 6, or 8.

[0113] The N direction vectors are preferably uniformly distributed over an angle 2π / N around the pixel under consideration.

[0114] Set of given directions I

number

[0115] - Skeletal regions and / or ROIs on the images, if the human or animal tissue is bone, and / or the region of interest (ROI) being considered, and / or - Image resolution, and / or - Image signal-to-noise ratio.

[0116] This determination of the set of directions I is - Improve the ability to identify or predict fractures; and / or allowing for better reproducibility (or precision) of the measurements, and / or -allowing for a better correlation with bone microstructure.

[0117] Therefore, in the next step d), the variogram V Pi To calculate the value h(P i ) are compared to pixels located along a line in a particular direction.

[0118] These directions depend on both the type of bone (ie, skeletal location) and the region of interest selected for measurement on this bone in order to optimize the texture measurements.

[0119] Indeed, the chosen direction is related to the bone morphology, in particular to the trabecular direction of cancellous bone: for example, in the spine, the preferred direction of the trabecular bone is vertical, so the chosen directions are vertical and horizontal [-π / 2, 0, π / 2, π] (parallel and perpendicular to the trabecular direction), as shown by the four arrows in Figure 4.

[0120] c) For each pixel Pi = (xi, yi) ∈ S and for each direction

number

number

number

[0121] d) For each pixel, calculate at least one variogram of the grey levels of the sampling S as a function of the distance r along these directions I. That is, by moving at least a distance r ∈ [1, R0] from this pixel; as mentioned above, moving a distance r i =(x i ,y i )∈S and each direction

number

number

number

[0122] The variogram is -One variogram for each given direction (i.e., one variogram for each pixel in S and one for each direction in I). If a variogram is calculated for each given direction, h(O) is the gray level of the initial given pixel before the shift, and h(r) is the gray level of the given new pixel after shifting a distance r from the initial given pixel along one of the given directions. The variogram is calculated by the following formula: V i (r) = [h(r) - h(O)] 2 (where i∈I) or - calculated simultaneously for all given directions (i.e. one variogram for each pixel of S). When variograms are calculated simultaneously for all given directions, the step of calculating the variogram V as a function of the grey level h is done by averaging the squared difference of h over pairs of pixels separated by a distance r, preferably according to the following formula:

number

[0123] V Pi is applied to each Pi∈S for a given range of values ​​R∈[1,R0]. This particular range of values ​​of r is the variogram V Pi To evaluate all the parameters of V Pi :r / →V Pi is chosen to allow (r) to converge.

[0124] The range R0 is determined depending on the skeletal location and image resolution. In DXA systems, the value R0 is between 1 cm and 2 cm.

[0125] The computational range R0 should not be confused with the range parameter c of the variogram model.

[0126] e) For every pixel Pi ∈ S, V Pi (r) and trace or calculate or determine the associated curve on a log-log scale. An example of a variogram V(r) on a log-log scale for one pixel is shown in FIG.

[0127] Variogram Curve V Pi Representing it on a log-log scale means that the values ​​along each axis become unitless.

[0128] f) V Pi The full model V is estimated as a least-squares regression model of the representation.

[0129] g) Evaluate the parameters of this model including, but not limited to, the initial slope a, the sill b, the range c, the nugget d, and / or the area under the curve e.

[0130] Each variogram V Pi For the global variogram V of the sampling S combining the variograms of each pixel, we evaluate at least one of the following parameters on a log-log scale:

[0131] - the initial slope of the variogram a (shown as 4 in Figure 7), - sil b, which represents the asymptotic value of the variogram (denoted as 5 in Figure 7); - range c (indicated as 6 in Figure 7), which represents the distance at which the variogram curve transitions from a quasi-linear progression to an asymptotic behavior; - the nugget d (indicated by 7 in Figure 7), which represents the initial value of the variogram, and - the area under the variogram curve e (shown as 8 in Figure 7).

[0132] V Pi Or a log-log plot of V, a mathematical model is fitted. The parameters (i.e., coefficients) a, b, c, d, and e of this model are combined to create a bone texture score, B.

[0133] The parameters a, b, c, d, and / or e are estimated from a least-squares regression model of the variogram considered.

[0134] Depending on the content of the image, the selected coefficients of the model may vary, since they may not be well-defined (e.g., the variogram curves may not converge to an asymptote and therefore the "range" may not be defined).

[0135] Preferably, the method comprises evaluating at least one of the parameters sill b, range c, nugget d and area e on a log-log scale for each variogram of each pixel and / or for a global variogram of a sampling S combining the variograms of each pixel.

[0136] Preferably, the method comprises evaluating at least two of the parameters slope a, sill b, range c, nugget d and area e on a log-log scale for each variogram at each pixel and / or for a global variogram of a sampling S combining the variograms of each pixel.

[0137] Preferably, the method comprises evaluating, on a log-log scale, the initial slope a and at least one of the sill b, range c, nugget d and area e for each variogram of each pixel and / or for a global variogram of a sampling S combining the variograms of each pixel.

[0138] Preferably, the method includes evaluating the slope a, sill b, range c, nugget d, and area e, all on a log-log scale, for each variogram at each pixel and / or for a global variogram of a sampling S combining the variograms at each pixel.

[0139] h) Combine the parameters (preferably a, b, c, d and / or e) into a bone texture score B (dimensionless), for example using a linear or non-linear equation depending on the clinical situation.

[0140] For example, the method may include combining the following:

[0141] - combining the evaluation parameters obtained for each pixel (preferably a, b, c, d and / or e) into a texture score B for each pixel, and / or - combining the evaluation parameters (preferably a, b, c, d and / or e) obtained for a sampling S into a global texture score B of the sampling S, and / or - combining the evaluation parameters obtained for each pixel (preferably a, b, c, d and / or e) into a global texture score B of the sampling S;

[0142] At least two parameters a, b, c, d, e are preferably combined into a texture score B using a linear or non-linear equation depending on the clinical situation. In practice, the parameters of the variogram model are combined into a bone texture parameter B using a combination equation. By way of example, such combination equations include, but are not limited to, multiple linear models for a given clinical situation and anatomical site. In such a case scenario, B is defined as follows: B = αa + βb + γc + δd + εe where the coefficients α, β, γ, δ, and ε are related to the slope, sill, nugget, and area under the curve, respectively. These coefficients are carefully defined during a pre-clinical performance optimization stage.

[0143] These coefficients are typically obtained from different experimental analyses.

[0144] The selection of the best coefficients is obtained, for example, from a combination of large-scale clinical studies and images, followed by a grid search optimization step.

[0145] More generally, α, β, γ, δ, ε can be obtained in different ways, including:

[0146] - Correction abacus based on experimental measurements, and / or - mathematical models and / or simulations of corrections based on theory, and / or -Machine learning or artificial intelligence (AI) algorithms, and / or - Selection of the best coefficients obtained from an extensive clinical study and image combination, and / or obtained from a grid search optimization step with clinical performance results; - A combination of the above methods.

[0147] This combination of at least two parameters improves the ability to identify or predict fractures.

[0148] i) applying a robustness improvement step related to at least one patient factor and / or at least one technical factor, preferably a robustness improvement step related to both patient and technical factors, to the texture score B.

[0149] The robustness step may be implemented as follows.

[0150] - before or during the previous step h) (by determining and / or correcting at least one parameter used to calculate or determine score B before or during the determination or calculation of score B as a function of patient factors and / or technical factors, preferably by determining and / or correcting R0, a, b, c, d and / or e before or during the determination or calculation of score B as a function of patient factors and / or technical factors); and / or - after the previous step h) (by correcting the score B (obtained from at least parameters R0, α, β, γ, δ, ε, a, b, c, d, and / or e) as a function of patient and / or technical factors).

[0151] The robustness step may be implemented in different ways, including by determining and / or correcting R0, α, β, γ, δ, ε, a, b, c, d, e, and / or B.

[0152] - a corrected abacus based on experimental measurements taking into account patient and / or technical factors and their observed influence on the variogram and / or texture score B, and / or - mathematical models and / or simulations of theory-based corrections taking into account patient and / or technical factors and their predicted influence on the variogram and / or texture score B, and / or - Machine learning or artificial intelligence (AI) algorithms that learn to minimize the impact of patient and / or technical factors on the variogram and / or texture score B, and / or - Selection of the best coefficients obtained from an extensive clinical study and combination of images, and / or from a grid search optimization step with the results of clinical performance until the influence of patient and / or technical factors on the variogram and / or texture score B is minimized; and / or - A combination of the above methods.

[0153] For example, R0 may be determined and / or corrected as a function of the image resolution of the x-ray acquired image.

[0154] The at least one patient factor includes:

[0155] - Patient morphological effects including at least one of the following:

[0156] - the effect of soft tissue, and / or tissue thickness, and / or its distribution, and / or its composition, and / or indirect substitutes, and / or - The patient's weight and / or body mass index (BMI) and / or waist circumference, and / or - Patient height, and / or - Affected by at least one patient disease or condition (e.g., arthropathy, ascites, aortic calcification, gas, etc.); and / or -Effect of patient positioning during image acquisition.

[0157] The at least one technical factor includes:

[0158] Possibly defective detectors and sensors for image acquisition, and / or - the effect of scanning modes and settings for image acquisition; and / or - the technical characteristics of the imaging device used to acquire the images - the effect of variability between imaging systems for image acquisition; and / or - the signal-to-noise ratio (SNR) of the image, and / or - Image resolution.

[0159] The robustness improvement step allows:

[0160] - Less sensitive to the physiological characteristics of the patient, e.g. the volume and / or nature of the soft tissue surrounding the bone; and / or - Less sensitive to the choice of technical parameters of image acquisition.

[0161] Typically, - The coefficient α is -Affects reproducibility. -Optimize fracture prediction.

[0162] - The coefficient β is -Affects the overall mineralization contribution of (bone) tissue.

[0163] - The coefficient γ is - Influences the contribution of the overall geometric properties of the (bone) tissue.

[0164] - The coefficient δ is - Related to the signal-to-noise ratio of an image.

[0165] - The coefficient ε is -Affects reproducibility.

[0166] j) Finally, at the end of the method, the bone texture score B is interpreted by comparison with age (Figure 9), ethnicity and sex matched controls.

[0167] A patient's fragility fracture risk profile is assessed either alone or incorporated into a probabilistic model such as FRAX (FIG. 11), or in combination with the minimum BMD T-score for both spine and hip in a given clinical situation (FIG. 10), or as a combination of BMD T-score and TBS category to define a bone resilience index.

[0168] For FIG. 9, the following parameters were used or determined:

[0169] -Bone images correspond to Fig. 1.

[0170] - The variogram corresponds to Figure 7.

[0171] -Calculated a=0.73 -Calculated b=1.35 -Calculated c=2.0 -Calculated d=0.08 -Calculated e=1.30

[0172] B = αa + βb + γc + δd + εe = 1.1 Where: α=1.6 β=0.08 γ=0.12 δ=-0.97 ε = -0.26

[0173] Technical and patient factors considered: -Age = 73 -Weight = 65 kg -Height = 1.66 m -BMI = 23.59 -Tissue thickness = 19 cm α, β, γ, δ, ε are parameters used to give respective weights between a, b, c, d, and / or e to calculate or determine the texture score B.

[0174] An embodiment of the method of the present invention further comprises the step of displaying (on the screen):

[0175] - a sampling S determined on the image, and / or - the direction I selected on the image, and / or - the calculated variogram, and / or - the evaluated parameters a, b, c, d and / or e, and / or the calculated texture score B, and / or - Patient and / or technical factors taken into account in the robustness improvement step.

[0176] Typically, at least one step of the method of the invention described above, or more precisely each step of the method of the invention described above, is not carried out purely abstractly or purely intellectually, but involves the use of technical means.

[0177] Typically, each of steps a) to j) of the method of the invention described above is implemented by technical means, preferably at least one computer, one central processing unit or computing device, one analogue electronic circuit (preferably dedicated), one digital electronic circuit (preferably dedicated) and / or one microprocessor (preferably dedicated) and / or software means.

[0178] Tables 1 and 2 below are examples of metric results of univariate tests and logistic regression models, respectively, that can best obtain the optimization of distance R0.

[0179] Univariate - Discrimination of fracture groups [Table 1]

[0180] Table 1: Univariate tests for optimization

[0181] Optimization of technical factors (logistic model): An example of compensating low image resolution with high R0 [Table 2]

[0182] Table 2: Examples of metric tests for logistic models to obtain best significance

[0183] Optimizing Patient Factors: Soft Tissue Examples: Logistic regression: INC_FX~B(INC_FX:incident fracture) [Table 3]

[0184] Table 3: Summary of logistic regression results for raw and adjusted data

[0185] Table 3: In a given cohort, the logistic model allows comparison of raw and patient factor-adjusted Texture Score B measurements.

[0186] Such models aim to predict new fractures, i.e., INC_FX, given score measurements of either the raw or corrected data.

[0187] - Significance of the test statistic for patient-adjusted B (p-value = 0.02) versus 0.12 for raw B - Odds ratios showing an increase (from 35% to 58%) in the odds of incident fracture (per increase in B)

[0188] Table 1: Calculated distances: in the same cohort, 1) Case-control group - fracture group vs control group: a. Increasing the distance from 2px to 10px increases the p-value for univariate analysis by Student's test of cases vs controls.

[0189] Table 2: Calculated distances: in the same cohort, 1) Case-control group #2 - New fracture group vs. control group, logistic model, new fracture ~ B (distance) a. As the distance increases from 2 to 10, the p-value increases. b. Increasing distance from 2 to 10 shows an improvement in the odds of a new fracture per increase in B.

[0190] The inventive apparatus comprises technical means (in particular means configured and / or programmed for calculating, determining, selecting, calculating, evaluating, combining and applying an improvement step, respectively) configured and / or programmed to implement all the steps mentioned above (in particular means configured and / or programmed for calculating, determining, selecting, calculating, evaluating, combining and applying an improvement step, respectively).

[0191] Typically, at least one of the means of the device of the invention as described above, preferably each of the means of the device of the invention (in particular the means configured and / or programmed to calculate, determine, select, compute, evaluate, combine, apply an improvement step) is a technical means.

[0192] Typically, each means of the inventive apparatus implementing the above mentioned steps a) to j) (in particular means configured and / or programmed to calculate, determine, select, compute, evaluate, combine, apply improvement steps) comprises at least one computer, one central processing unit or computing device, one analogue electronic circuit (preferably dedicated), one digital electronic circuit (preferably dedicated) and / or one microprocessor (preferably dedicated) and / or software means.

[0193] The apparatus of the present invention comprises:

[0194] - means for implementing the step of obtaining a digitized image; The means for acquiring a digitized image typically comprises: - Traditional X-ray imaging systems, or -Digital X-ray imaging system, or -Dual-energy X-ray absorptiometry (DXA) imaging system, or -Projection computed tomography (CT) imaging systems, or - a quantitative computed tomography (QCT) imaging system, or - Projection quantitative computed tomography imaging system, or - A peripheral quantitative computed tomography (pQCT) imaging system, or - a high-resolution peripheral quantitative computed tomography (HR-pQCT) imaging system; and

[0195] - means for implementing all of the above steps a) to j), which means are typically integrated in a single computer;

[0196] a screen (configured to display the determined sampling S on the image, and / or the selected direction I, and / or the calculated variogram, and / or the evaluated parameters a, b, c, d and / or e, and / or the calculated texture score B, and / or the patient factors and / or technical factors taken into account in the robustness improvement step).

[0197] Of course, different features, aspects, variations, and embodiments of the invention can be combined with each other in various combinations, unless they are mutually incompatible or mutually exclusive.

[0198] In the Summary of the Invention and the Detailed Description of the Drawings and the Embodiments, the word "method" is used, which may be replaced in an equivalent manner with the word "process."

[0199] Of course, the invention is not limited to the examples described herein, and numerous modifications can be made to these examples without exceeding the scope of the invention, provided that they fall within the scope of the claims.

Claims

1. A method for analyzing the texture of human or animal tissue from a digitized image obtained by an X-ray based imaging system, comprising: calculating, by technical means, a texture score B of at least one image by applying an experimental variogram to the texture of the tissue; - at least one patient factor related to the patient depicted in the image, and / or - at least one technical factor related to the acquisition of the image applying, by technical means, a robustness improvement step to the texture score B taking into account A method comprising.

2. At least one patient factor is - soft tissue, and / or tissue thickness, and / or its distribution, and / or the influence of its composition, and / or indirect surrogates, and / or - patient weight and / or body mass index (BMI), and / or - patient height, the influence of the patient's morphology including at least one of, and / or the influence of at least one medical condition or state of the patient, and / or the influence of the patient's positioning during image acquisition, The method according to claim 1, comprising.

3. At least one technical factor is - a detector and sensor that may be defective for image acquisition, and / or - the influence of scan mode and settings for image acquisition, and / or - the technical characteristics of the imaging device used for image acquisition, - the influence of variations between imaging systems for image acquisition, and / or - the signal-to-noise ratio (SNR) of the image, and / or - the resolution of the image The method according to claim 1 or 2, comprising.

4. The method according to claim 1 or 2, wherein the tissue is bone tissue and the texture score B is a bone texture score B.

5. The method according to claim 4, wherein the digitized two-dimensional image is selected in a region having a cancellous bone structure.

6. Steps implemented by the technical means, comprising: selecting, for at least one region of interest (ROI) of pixel sampling S, a set of predetermined directions I according to the given region of interest (ROI); Determining an optimized pixel sampling S of the image such that each pixel Pi = (xi, yi) ∈ S has its gray level value h(P ｉ ) for each experimental variogram of each pixel, and / or for the global experimental variogram of sampling S combining the experimental variograms of each pixel, on a log-log scale, For each pixel, by moving a distance r ∈ [1, R ０ along these directions I from this pixel, calculating at least one experimental variogram of the gray levels of the sampling S, the experimental variogram being calculated for each predetermined direction or simultaneously for all predetermined directions; - an initial gradient a, - a sill b representing the asymptote of the experimental variogram, ​ - A range c representing the distance at which the experimental variogram curve transitions from a quasi-linear progression to an asymptotic behavior, - A nugget d representing the initial value of the experimental variogram, and - An area e under the experimental variogram curve, The step of evaluating at least one of the parameters; - Combining the parameters obtained for each pixel into the texture score B of each pixel, and / or - Combining the parameters obtained for the sampling S into the texture score B of the sampling S, and / or - Combining the parameters obtained for each pixel into the texture score B of the sampling S, A step, further including a robustness improvement step regarding at least one patient factor and / or at least one technical factor, - As a function of the patient factor and / or the technical factor, before or during the determination or calculation of the score B, preferably used for calculating or determining the score B, -R ０ at least one parameter of a, b, c, d, and / or e, and / or - By determining and / or correcting at least one parameter (α, β, γ, δ, ε) of at least one parameter used to give each weight between a, b, c, d, and / or e for calculating the texture score B, And / or - By correcting the score B as a function of the patient factor and / or the technical factor, The step applied to the texture score B, The method according to claim 1.

7. The set of predetermined directions I is - When the human or animal tissue is bone, the skeletal site of the bone, and / or - The region of interest (ROI), and / or - The resolution of the image, and / or - The signal-to-noise ratio of the image The method according to claim 6, which depends on.

8. The step of selecting the set of predetermined directions I is performed by determining a set of N direction vectors 【Number 1】 Here, θ ｋ ∈ [−π, π], ∀k ∈ 1, …, N The method according to claim 6 or 7.

9. The step of moving by a distance r is for each pixel Pi = (xi, yi) ∈ S and each direction 【Number 2】 For, 【Number 3】 along the distance r ∈ [1, R in pixel units ０ (where 【Number 4】 Is the gray value of such a pixel) The method according to claim 8, which is performed by moving only by.

10. The experimental variogram is calculated simultaneously for all predetermined directions, and the step of calculating the experimental variogram of the gray level as a function of the distance r is to average the squared difference of h over a plurality of pixel pairs, the following formula: 【Number 5】 At each distance r, V Ｐｉ (r) is calculated for all pixels P ｉ ∈ S, the method according to claim 8, which is performed by such calculation.

11. The experimental variogram is calculated for each predetermined direction, h(O) is the gray level of an initial predetermined pixel before movement, h(r) is the gray level of a new predetermined pixel after moving a distance r along one of the predetermined directions from the initial predetermined pixel, The experimental variogram is given by the following equation V ｉ V(r)=[h(r)-h(0)] ２ (where i ∈ I) The method according to claim 6 or 7, calculated by

12. The method according to claim 6 or 7, wherein each parameter a, b, c, d, and / or e is evaluated as a least squares regression model of the considered experimental variogram.

13. The method according to claim 6 or 7, wherein the parameters are combined into the texture score B using linear or non-linear equations depending on the clinical situation.

14. For each experimental variogram of each pixel and / or for the global experimental variogram of the sampling S combining the experimental variograms of each pixel, - silb representing the asymptote of the experimental variogram, - range c representing the distance at which the experimental variogram curve transitions from a quasi-linear progression to an asymptotic behavior, - nugget d representing the initial value of the experimental variogram, and - area e under the experimental variogram curve The method according to claim 6 or 7, comprising evaluating at least one of the parameters on a log-log scale.

15. For each experimental variogram of each pixel and / or for the global experimental variogram of the sampling S combining the experimental variograms of each pixel, - initial gradient a, - silb representing the asymptote of the experimental variogram, - range c representing the distance at which the experimental variogram curve transitions from a quasi-linear progression to an asymptotic behavior, - nugget d representing the initial value of the experimental variogram, and - area e under the experimental variogram curve The method according to claim 6 or 7, comprising evaluating at least two of the parameters on a log-log scale.

16. For each experimental variogram of each pixel and / or for the global experimental variogram of the sampling S combining the experimental variograms of each pixel, the initial gradient a and - silb representing the asymptote of the experimental variogram, - range c representing the distance at which the experimental variogram curve transitions from a quasi-linear progression to an asymptotic behavior, - nugget d representing the initial value of the experimental variogram, and - area e under the curve The method according to claim 6 or 7, comprising evaluating at least one of the parameters on a log-log scale.

17. The method according to claim 1 or 2, wherein the texture score B is dimensionless.

18. An apparatus for analyzing the texture of human or animal tissue from a digitized image obtained by an X-ray based imaging system, means configured and / or programmed to calculate at least one texture score B of the image by applying an experimental variogram to the texture of the tissue; - at least one patient factor related to the patient imaged in the image, and / or - at least one technical factor related to the acquisition of the image means configured and / or programmed to apply a robustness improvement step to the texture score B taking into account; An apparatus comprising.

19. At least one patient factor is - soft tissue, and / or the thickness of the tissue, and / or its distribution, and / or the influence of its composition, and / or an indirect surrogate, and / or - the patient's weight and / or body mass index (BMI), and / or - the patient's height, the influence of the patient's morphology including at least one of, and / or the influence of at least one of the patient's medical condition or state, and / or the influence of the patient's positioning at the time of image acquisition The apparatus according to claim 18, comprising.

20. At least one technical factor is - detectors and sensors that may be defective for image acquisition, and / or - the influence of the scan mode and settings for image acquisition, and / or - the technical characteristics of the imaging device used for image acquisition, - the influence of variations between imaging systems for image acquisition, and / or - the signal-to-noise ratio (SNR) of the image, and / or - the resolution of the image The apparatus according to claim 18 or 19, comprising.

21. The apparatus according to claim 18 or 19, wherein the human tissue is bone tissue and the texture score B is a bone texture score B.

22. The apparatus according to claim 21, wherein the digitized two-dimensional image is an image that images a cancellous bone structure.

23. - Optimized pixel sampling S of an image in which each pixel Pi = (xi, yi) ∈ S has a gray level value h(P ｉ ) is determined by means configured and / or programmed to do so; - means configured and / or programmed to select a set of predetermined directions I for at least one region of interest (ROI) of pixel sampling S, depending on the given region of interest (ROI); - For each pixel, moving a distance r ∈ [1, R ０ from this pixel along these directions I, means configured and / or programmed to calculate at least one experimental variogram of the gray levels of the sampling S, and configured and / or programmed to calculate the experimental variogram simultaneously for each predetermined direction or for all predetermined directions; - for each experimental variogram of each pixel, and / or for the global experimental variogram of sampling S that combines the experimental variograms of each pixel, - an initial gradient a, - A silb representing the asymptote of the experimental variogram, - A range c representing the distance at which the experimental variogram curve transitions from a quasi-linear progression to an asymptotic behavior, - A nugget d representing the initial value of the experimental variogram, and - The area e under the experimental variogram curve Means configured and / or programmed to evaluate at least one of the parameters on a log-log scale; - Combining the parameters obtained for each pixel with the texture score B of each pixel, and / or - Combining the parameters obtained for the sampling S with the texture score B of the sampling S, and / or - Combining the parameters obtained for each pixel with the texture score B of the sampling S, Means configured and / or programmed as such; Comprising; Furthermore, -R ０ at least one parameter of a, b, c, d, e, and / or - At least one parameter (α, β, γ, δ, ε) used to give each weight between a, b, c, d, and / or e for calculating the texture score B By determining and / or correcting at least one of the parameters, And / or - By correcting the score B as a function of patient factors and / or technical factors, Means configured and / or programmed to apply a robustness improvement step regarding at least one patient factor and / or at least one technical factor to the texture score B, according to claim 18 or 19.

24. The set of predetermined directions I is - When the human or animal tissue is bone, the skeletal site of the bone, and / or - The region of interest (ROI), and / or - The resolution of the image, and / or - The signal-to-noise ratio of the image Depending on, according to claim 23.

25. Means configured and / or programmed to select a set of predetermined directions I is A set of N direction vectors 【Number 6】 Here, θ ｋ ∈ [−π, π], ∀k ∈ 1, …, N Configured and / or programmed to select a set of predetermined directions I by determining, according to claim 23.

26. Means configured and / or programmed to calculate at least one experimental variogram, for each pixel Pi = (xi, yi) ∈ S and each direction 【Number 7】 For, 【Number 8】 Moving by a distance r ∈ [1, R ０ in pixel units along, it is configured and / or programmed to calculate at least one experimental variogram of the gray levels of the sampling S, 【Number 9】 Which is the gray value of such a pixel, According to claim 25.

27. Means configured and / or programmed to calculate at least one experimental variogram are configured and / or programmed to calculate experimental variograms simultaneously for all predetermined directions, Means configured and / or programmed to calculate at least one experimental variogram, By averaging the squared differences of h over pairs of pixels separated by a distance r, the following equation: 【Number 10】 (VP ｉ (r) is calculated for all pixels Pi ∈ S) The apparatus according to claim 25, configured and / or programmed to calculate a variogram as a function of the gray level h.

28. Means configured and / or programmed to calculate at least one experimental variogram are configured and / or programmed to calculate an experimental variogram for each predetermined direction, h(O) is the gray level of an initial predetermined pixel before movement, h(r) is the gray level of a new predetermined pixel after moving a distance r along one of the predetermined directions from the initial predetermined pixel, Means configured and / or programmed to calculate at least one experimental variogram is configured and / or programmed to calculate an experimental variogram by the following equation V ｉ V(r)=[h(r) - h(0)] ２ (where i ∈ I) configured and / or programmed to calculate an experimental variogram. The apparatus according to claim 23.

29. Means configured and / or programmed to evaluate parameters are configured and / or programmed to evaluate each parameter a, b, c, d, and / or e as a least squares regression model of the considered experimental variogram, of the apparatus according to claim 23.

30. Means configured and / or programmed to combine parameters are configured and / or programmed to combine the parameters into a texture score B using a linear or non-linear equation, depending on the clinical situation, of the apparatus according to claim 23.

31. Means configured and / or programmed to evaluate parameters are for each experimental variogram of each pixel and / or for the global experimental variogram of a sampling S combining the experimental variograms of each pixel, - silb representing the asymptote of the experimental variogram, - range c representing the distance at which the experimental variogram curve transitions from a quasi-linear progression to an asymptotic behavior, - nugget d representing the initial value of the experimental variogram, and - area e under the experimental variogram curve The apparatus according to claim 23, configured and / or programmed to evaluate at least one of the parameters on a log-log scale.

32. Means configured and / or programmed to evaluate a parameter, for each experimental variogram of each pixel and / or for the global experimental variogram of a sampling S combining the experimental variograms of each pixel, - an initial gradient a, - a sill b representing the asymptotic value of the experimental variogram, - a range c representing the distance at which the experimental variogram curve transitions from a quasi-linear progression to an asymptotic behavior, - a nugget d representing the initial value of the experimental variogram, and - an area e under the experimental variogram curve The apparatus according to claim 23, configured and / or programmed to evaluate at least two of the parameters on a log-log scale.

33. Means configured and / or programmed to evaluate a parameter, for each experimental variogram of each pixel and / or for the global experimental variogram of a sampling S combining the experimental variograms of each pixel, an initial gradient a and - a sill b representing the asymptotic value of the experimental variogram, - a range c representing the distance at which the experimental variogram curve transitions from a quasi-linear progression to an asymptotic behavior, - a nugget d representing the initial value of the experimental variogram, and - an area e under the experimental variogram curve The apparatus according to claim 23, configured and / or programmed to evaluate at least one of them on a log-log scale.

34. The apparatus according to claim 18 or 19, wherein the texture score B is dimensionless.

35. A computer program comprising instructions that, when executed on a computer, perform the steps of the method according to claim 1 or 2.

36. A computer-readable storage medium comprising instructions that, when executed by a computer, direct the computer to perform the steps of the method according to claim 1 or 2.

Images