WO2019102123A1 - Method for evaluating the maximum pull-out resistance of an anchoring screw implanted in a vertebra - Google Patents

Abstract

The invention concerns a method for evaluating maximum forces of pull-out resistance (FA) for a determined model of anchoring screw, for a determined vertebra (VD) of a patient in whom the anchoring screw is able to be implanted according to the principles of a given surgical technique, the method comprising: - mechanical pull-out tests of an anchoring screw in control vertebrae (VT) determining the experimental maximum pull-out forces (FA_EX), then calculating the mean density of interest (DM) calculated from volumes of interest (VI) for each control vertebra (VT) at the same vertebral position as the determined vertebra (VD), then determining at least one conversion function (Fconv) parameterised by at least one variable of mean density of interest (DM) for all the determined vertebrae (VT), and correlating these values with the experimental forces (FA_EX), finally, applying the one or more conversion functions (Fconv) to the same values of mean density of interest (DM) for the determined vertebra (VD) determined in the same way as for the control vertebrae (VT), in order to calculate, using the one or more regression functions deduced from the one or more correlations, the maximum force or forces of pull-out resistance (FA) of the anchoring screw for the determined vertebra (VD).

Description

 Method for evaluating the maximum resistance to tearing of an anchor screw implanted in a vertebra

 The field of the invention is that of the evaluation of parameters of the planning of a spinal surgery for the implantation of an anchoring screw system in one or more vertebrae, in a manner specific to the patient.

 More specifically, the invention relates to the estimation of the maximum tensile strength of one or more anchoring screws implanted in a vertebra.

 Conventionally, anchoring screws can be implanted in anchoring sites of the vertebrae of a patient during the execution of a surgical act.

 For example, the correction of scoliosis curvatures in an adult may require surgical treatment by posterior fusion of the spine by means of instrumentation to obtain derotation of pathological curvatures and permanent stabilization of the spine. This instrumentation can comprise, inter alia, anchoring screws called "pedicle screws", as well as rods intended to connect the anchoring screws by their heads, the rods being dimensioned and bent to exert when they are secured with screws a mechanical action on the patient's column to correct the pathological spinal curvature (s) of the patient.

 In order to derotation, the anchoring screws are implanted in the anchoring sites of the vertebrae of the patient, then the rods are coupled to the screws on each side of the vertebrae, the rods being optionally coupled to each other thereafter to better rigidify the system.

 During the derotation and postoperatively, significant mechanical stress is exerted between the vertebrae of the patient and the instrumentation. Indeed, the forces required to correct scoliosis and the stability of the correction can be relatively high.

 The importance of the forces exerted on the vertebrae and on the instrumentation is such that there is a risk of instability of the assembly.

 More specifically, there is a risk of loss of significant or even complete hold anchor screws implanted in the vertebrae.

It is essential to avoid the occurrence of such loss of posture, which may compromise the correction desired peroperatively, but also lead in the longer term to non-operative pseudarthrosis or discomfort that may require surgical revision of the patient. To plan a surgical procedure, operators rely on their experience as a practitioner, a rigorous clinical examination, and additional examinations of static and postural medical imaging.

 For this purpose, the surgeons also evaluate the overall strength of the assembly by referring to the recommendations and the average values of tensile strength of anchor screws in vertebras published in the scientific medical press.

 Following these assessments and the extent of the corrections they want to apply, the surgeons plan the vertebrae to implant. Surgeons also plan the deformities they intend to apply to the stems to change the position of each pathological vertebral stage. They may possibly use vertebral osteotomy angle corrections.

 This type of evaluation, however, only provides a value indicative of the maximum force of tearing resistance that the anchor screw can withstand without risking a loss of significant strength, force obtained except exception in continuous traction in the axis insertion of the screws.

 In addition these indicative values may not reflect the local influence on the anchoring of the screws, the distribution of degradation of the mechanical properties of the vertebrae which have a decreased overall bone quality or pathological.

 Recent scientific studies have shown that a significant number of loss of holding or tearing anchor screws has been observed in elderly patients. In addition, the incidence of spinal pathologies that can be treated using screw-based systems increases with increasing patient survival, and conjugated severity factor, age is also a known factor of deterioration. mechanical properties of the vertebrae.

 The need to increase the strength of the screws used in these surgical procedures is fully documented in the scientific medical literature.

 Several additional and additional solutions are available to increase this anchoring, including:

 the extension of the assembly to other vertebral stages,

 the extension of the assembly to ilio-sacral screws;

 the use of different implantation trajectories for the pedicle screws;

 the use of new screw designs, such as:

o expansive screws, o double-screw implantations,

 o different metal alloys,

 o different thread profiles, screws with several thread types or double threads,

 o the polyaxiality of the heads,

 o new designs of the coupling of the screws with the rods, o conical screws,

 o screws coated with hydroxyapatite, calcium-phosphate or having a surface with a micrometric roughness, increasing techniques:

 o Cannulated and perforated screws that allow the injection of product increasing the bone density in the vertebra,

 o the injection of product increasing the bone density in the trajectory of the screw after tapping,

 o preoperative percutaneous vertebroplasty,

 medications for osteoporosis.

 However, sufficient assembly can not be conditioned, or even guaranteed because of their intrinsic risks, to systematic recourse to all or part of these solutions; this is especially so because some of them have potentially serious or even fatal side effects, which must be evaluated specifically for a patient under the benefit / risk balance compared to a first-line surgical treatment, called here "Standard treatment". This standard treatment includes in particular a number of screws directly related to the number of vertebrae included between the upper and lower vertebral levels healthy bordering the pathological deformations, this without presuming a possibly lower holding vertebral anchors.

 In the first place, it is essential to evaluate the feasibility of the standard treatment specifically to a patient to validate this setup with smaller margins of error, or if so, to justify the additional risk involved in using some of the additional or additional medical or surgical treatments and techniques to improve the erection

Accordingly, there is a need to evaluate more specifically for a patient and with a reduced margin of error, the force to which each of the screws implanted as part of the standard treatment can be subjected. The invention particularly aims to overcome this need of the prior art.

More specifically, the invention aims to provide a method for evaluating the maximum tear resistance without irreducible displacement that each anchor screw is capable of supporting, when it is implanted as part of a treatment standard in a particular anchoring site of a determined vertebra of a patient, according to the principles of a surgical technique of reference.

 This objective, as well as others which will appear later, are achieved thanks to the invention which relates to a method of evaluation of maximum forces of tearing resistance (FA) of an anchor screw for a determined vertebra (VD) in which the anchoring screw is capable of being implanted according to the principles of a given insertion technique, comprising successively:

 at. acquiring an imaging reference system including unit density data (du) as well as their spatial coordinates for at least one control vertebra with at least one non-invasive medical imaging (Tl) technique;

b. at least one mechanical tensile strength test of the anchoring screw implanted in the control vertebra (s) (VT) to determine the maximum force of resistance to experimental tearing (FA _EX );

 c. defining at least one volume of interest (VI) in each control vertebra (VT) using the imaging reference acquired in step (a);

 d. calculating for each volume of interest (VI) of each control vertebra (VT) of a mean density of interest (MD);

 e. defining at least one conversion function (Fconv) comprising at least one average density of interest (MD) variable calculated for each control vertebra (VT),

 f. correlating the values of each conversion function with values of the one or more mechanical tests to establish a regression function;

 g. acquisition for a determined vertebra (DV) of a patient, an imaging repository similarly to step (a);

 h. transposition of step (c) to the determined vertebra (VD);

i. transposing step (d) to the results of step (h); j. applying a conversion function defined in step (e) to the average densities (DM) from step (i) to obtain a score; k. obtaining an estimate of a maximum pulling force (FA) per application to the score obtained in step 0 of the regression function associated with the conversion function of step 0.

 The method according to the invention makes it possible to obtain at least one estimate of the maximum force of tearing resistance (FA) and its statistical indicators of margin of error. This estimated force (FA) corresponds to the limit value beyond which the anchor screw could begin to move irreversibly out of its insertion site in a vertebra in which it would be implanted.

In step (b), it is preferentially a maximum force of resistance to experimental tearing in the axis of insertion of the screw (FAE _X ), determined on the force / displacement curve recorded during each test, in the upper limit of elastic deformation, therefore without irreversible displacement of the screw in this axis.

 The anchor screw meets the precise parameters defined by an industrial (materials, length, diameter, thread, ...). The method can be repeated for anchor screws of different design.

 The characteristics of the anchor screw and the principles of the surgical insertion technique are the same for the control vertebrae (VT) and the determined vertebra (VD).

 For a control vertebra (VT), one or more sites of implantation can be tested (for example a pedicular aiming on the right and left of the vertebra).

 With the method according to the invention, it is possible to evaluate, with a lower margin of error, the strength, estimated in the form of the maximum force of tearing resistance (FA) of an anchor screw which must be implanted in a site of insertion of a vertebra (VD), compared to the evaluation method according to the prior art.

Unit densities (du), for estimating the mechanical properties of vertebrae insertion sites determined in the patient (VD), are acquired by or less than one medical imaging modality (Tl). Then, thanks to the positioning of the volumes of interest (VI) and the determination of their average density of interest (DM), and possibly other variables such as morphological parameters or bone texture index acquired at the steps ( a) and (g). The term "density" evoked without adjective refers to vertebral bone density.

 Consequently, if the vertebra determined (VD) has, globally and locally for specific volumes (VI), densities (DM) below a normal density threshold for a healthy vertebra, then the method according to the invention allows to take into account these global or localized density deficits (DM) to a particular volume of interest (VI), an anatomical volume for example.

 The use of the densities of the volumes of interest (VI), and possibly other variables such as morphological parameters or bone texture index, reduce the margin of error of the estimation of the maximum strengths of resistance. tearing (FA) by the method according to the invention.

 The volume or volumes of interest (VI) determined for a control vertebra or vertebra determined may be anatomical, geometric, or deduced from the theoretical trajectory of the screws in the vertebra according to the surgical technique retained, or further deduced from morphological parameters of the vertebra, or be a combination of these parameters.

 For example, for a vertebral pedicle, cortical bone can be distinguished from cancellous bone (also called trabecular bone). The cortical area can also be subdivided, for example, into quadrants (lateral, medial, superior, inferior).

 An anatomical volume of interest (VI) corresponds, for example, to the anterior vertebral body or to the vertebral pedicles.

 The volumes of interest (VI) are defined by the spatial coordinates of their contours for each control vertebra or vertebra determined. Optionally, as mentioned above, other morphological variables (eg length and diameter of the pedicle) can be measured or deduced from the data acquired in step (a) as in step (g).

 The biomechanical parameters of stress resistance of a vertebra depend on several factors (geometrical, global density, cortical density or its approximation by the cortico-subcortical density, spongy density, thickness of spans and density of the trabecular network of the vertebra). spongy bone).

The method according to the invention thus has the advantage of taking into account the main parameters for the biomechanical resistance: the average density of interest (MD) of the cortical zone and / or spongy bone of one or more volumes of 'interest (VI), and possibly other variables such as morphological parameters or bone matrix texture index of the vertebra.

 The control vertebrae (VT) come from deceased donors (specimens) chosen so that their clinical characteristics are representative of the population of patients for whom implantation of anchor screws is planned. For example, control spines can be taken from specimens with non-modifiable risk factors for osteoporosis.

 Synthetic substitutes can also be used, all the more so if they succeed in reproducing the relevant parameters of vertebrae of biological origin of the targeted population (morphology of a typical vertebra, average densities of the zones reproducing the cortical bone and cancellous bone, reproduction of a trabecular structure in cancellous bone ...).

Of course, the more the number of measurements (FA _Ex ) on vertebrae controls (VT) for the same vertebral level is high, and the results obtained in step (f) are statistically representative for the same level of the vertebra determined (VD) and for the same target patient population. Similarly, for synthetic vertebrae, mechanical tearing tests must respect the validity thresholds of a repeatability protocol.

During step (f), at least one mathematical function called conversion function (Fconv) is defined, having as variable one or more average bone densities (DM) attributed to the volumes of interest (VI) of each control vertebra (VT). ), and possibly other variables such as morphological parameters or bone frame texture index available thanks to the data acquired in step (a). It is now possible to calculate a correlation between the mathematical images (I) of these conversion functions (Fconv) for each control vertebra (VT) and the maximum experimental resistance to tearing forces (FA _EX ) determined during the test. step (b). The characteristics of the regression line (s) deduced from these correlations make it possible to define as many regression functions (Freg) that can be used to estimate the maximum forces of tearing resistance (FA) for a given vertebra (VD). ). It is understood that invasive methods, such as that used in step (b), are either unthinkable for a patient, or are not recommended to evaluate the behavior of screws in the vertebra determined (VD), for example biopsies bone. The method according to the invention has the advantage of contributing to the decision-making process of the operator by providing him with encrypted, factual, relevant and patient-specific information items, derived from the biomechanical properties located for each of the vertebrae of the patient (VD). , to better estimate the risk of performing standard surgery in terms of intraoperative mechanical complications, or even failure.

 Indeed, an anchor screw implanted in a vertebra with a normal bone density can withstand greater stresses than an anchor screw implanted in a vertebra that has a decreased bone density or pathological.

 Consequently, specifically to the patient, and according to the conventional application of a given surgical technique, the quantified estimates provided (FA) by the method allow the operator to better evaluate whether the standard number of vertebral stages to be implanted is sufficient to to guarantee the stability of the assembly, or if it is probable that we should resort to solutions of circumvention of the deficit of bone density which could involve an over-risk of morbi-mortality for this patient for the same objective of treatment.

 The method provides the practitioner with information to better secure the planning of surgical procedures by identifying with a lower margin of error the risk of intraoperative mechanical complications for a given patient, and by resorting to other treatments only if the standard treatment only seems too risky for the success of the intervention.

 Preferably, the imaging data is obtained by one of the following techniques:

 X-ray computed tomography;

 - X-ray stereo-x-ray and dual-energy x-ray absorptiometry;

 planar bi-photonic x-ray absorptiometry;

 - X-ray volumetric bi-photonic absorptiometry.

 These imaging techniques, each with their specificities (availability, irradiation dose delivered, spatial resolution, etc.), make it possible to obtain unit density data (du) for the vertebrae.

X-ray computed tomography, X-ray double-energy X-ray stereo-x-ray and absorptiometry, and X-ray volumetric bi-photon absorptiometry alone provide density measurements. units (of) as well as their spatial coordinates in vertebral volume (3D).

 Complementary to an imaging technique intended to obtain only planar density data (2D), it is possible to couple an imaging technique making it possible to obtain 3D geometry data. For example, it is possible to combine an X-ray stereo-X-ray technique and a planar bi-photon x-ray absorptiometry technique for certain areas of the spine and for the lumbar vertebrae of L1 to L4.

 Advantageously, to complete the bone density data (du), it is also possible to use a technology that makes it possible to obtain bone texture indices.

 According to a preferred embodiment, step (c) and / or step (h) comprise a substep of segmentation by clipping the bone boundaries of the vertebrae on images acquired by the one or more techniques of imaging (Tl).

 In other words, the step of positioning volumes of interest (VI) in the images of the vertebrae (VT) in step (c) (as in step (g), VD) comprises a sub-component. vertebral segmentation step (extraction of specific data (from) to a vertebra from all the data acquired for the spine) by definition of the bone contours (trimming) of each of the control vertebrae (VT) on images acquired by the modality (s) 3D imaging (Tl).

 More specifically, the vertebral segmentation (s) are performed on sectional planes acquired by the 3D (Tl) imaging technique.

 The substep of segmentation makes it possible to define the outer bony envelope of each control vertebra (VT) by recomposition from the cutouts on the cuts.

 The segmentation sub-step also makes it possible to directly cut off particular zones which, assembled on all or part of the sections, form one or more specific volumes (VI).

 This segmentation can be manual, automatic with adapted software, semi-automatic by integrating a human correction.

 This segmentation can also consist of the recording of spatial coordinates of points, lines, planes or remarkable volumes making it possible to reconstitute the volumes of interest (VI) by means of an algorithm.

For example, to obtain the segmentation of volumes (VI), remarkable morphological or geometric landmarks simplifying the subsequent constitution of a volume of interest (VI) can be declared in a generic numerical model 5 (Figure 3) of vertebra type in steps (c) and (g).

 The identification of these remarkable coordinates in the control vertebrae makes it possible to simplify the identification of the volume or volumes of interest and therefore the calculation in steps (d) and (h) of the average density of interest (MD) of each volume of interest of these vertebrae.

 Advantageously, step (c) and / or (g) comprise a sub-step of excluding non-bone tissue or pathological bone reconstructions.

 This sub-step exclusion can advantageously be performed during the segmentation of the vertebra (VT or VD) or directly volumes (VI) that we want to constitute.

 Strict trimming during segmentation eliminates pathological bone reconstructions. For example, areas of osteophytosis are excluded. These are defined quality criteria for bone densitometry by bi-photon absorptiometry.

 Advantageously, step (c) and / or step (h) are performed by adapting the shape of a generic numerical model (5), comprising predefined volumes of interest (VI), to the form or vertebrae obtained during the substep of segmentation or as a function of marks recorded directly during the substep of segmentation of the control vertebrae or vertebrae determined.

 In other words, the definition of at least one volume of interest (VI), and possibly the taking of morphological measurements, is carried out by adapting, possibly using an algorithm, a generic numerical model containing volumes of interest. (VI) types in the coordinate coordinate system of this numerical model. This model can be adapted by conforming these volumes of interest (VI) to the outer envelopes of the vertebrae (VT or VD) obtained during the substep of segmentation or according to markers recorded during this same substep.

 Advantageously, step (e) for determining at least one regression function comprises:

 a sub-step of defining at least one conversion function (Fconv) comprising as variable at least one average density of interest (MD) for each control vertebra (VT);

a sub-step of calculating mathematical images (I) using the conversion function (Fconv) applied to the values of the densities mean of interest (MD) calculated in (d) for each control vertebra (VT);

a sub-step of correlating said mathematical images (I) of each conversion function (Fconv) with the maximum tensile strengths (FAE _X ) measured in step (b);

a sub-step of literally defining a regression function (Freg) associated with the correlation;

 a substep of calculating statistical error margins and statistical significance for each correlation.

 The conversion function (Fconv) may correspond to a linear combination of one or more variables of average interest density (MD). In this same sub-step, the densities of the volumes of interest (DM) can undergo transformations to linearize them (logarithmic function, exponential, polynomial, ...) in the conversion function (Fconv).

 Optionally, other variables such as morphological parameters or bone texture index may be integrated into the conversion function (Fconv). More specifically, the conversion function (Fconv) may include morphological parameters such as the length of the anterior vertebral body or vertebral pedicles, in order to take into account the theoretical effects of a screw of the largest diameter and the greatest possible length. ; the estimation of these parameters on the forces (FA) is obtained by a complementary statistical analysis of the data obtained in the steps (a) to (d). By integrating parameters other than density (DM), the sub-step of defining the regression function (Freg) can also be obtained by using a multilinear correlation. The linear regression function can take the form "y = a. x + b "where:

 "Y" corresponds to the evaluation of the maximum force of pulling force (FA) of the anchoring screw in a determined vertebra (VD) that one wishes to estimate;

 "X" corresponds to the mathematical image (I) of the conversion function (Fconv) for the parameters of the determined vertebra (VD);

"A" corresponds to the slope of the straight linear regression;

"B" is the right-hand value of the linear regression, such as:

- FA = Freg [VD] = a. Fconv [VD] + b, with Fconv [VD] = ai. DM [VD] [1] + a ₂ . DM [VD] [2] + ... + a _w .

DM [RV] [w]

 Of course, in the context of a multilinear regression, the regression function (Freg) corresponds to a more complex polynomial than that of the preceding example.

 Preferably, the step (e) of determining at least one conversion function (Fconv) comprises a final sub-step of pre-selecting one or more conversion functions (Fconv).

 According to a preferred solution, the step (e) of determining one or more conversion functions (Fconv) that can be used in clinical routine requires a substep of selection of conversion functions (Fconv) that lead to statistically acceptable results. This can also reduce the list of explored interest volumes (VIs) to only those volumes (VIs) necessary and sufficient to obtain the pre-selected conversion functions (Fconv).

This substep of selection of relevant conversion functions (Fconv) is advantageously performed according to statistical criteria: the final sub-step of selection is performed according to selection criteria which comprise, for each regression function, a coefficient of determination (R ² ), and a standard error estimation value (SEE) of the linear correlation of each score with the maximum experimental tearing resistance forces (FA _EX ) on the control vertebrae (VT).

In statistics, the coefficient of determination (R ² ) is used to measure the performance of the prediction allowed by the linear fit and the standard error of estimate (SEE) is a measure of the variability of the data points of both sides. other of the regression line.

 The selection criteria may also incorporate other calculations of the performance of the prediction allowed by the adjustment.

Finally, when applied to a patient likely to benefit from the surgical treatment with a method according to the invention, the application criteria can also include clinical, practical, financial implementation constraints, due to the time required to the realization of the method (VI), due to the technical limitations of the medical imaging systems used (Tl). These criteria must also take into account the case where particular areas of a specific vertebra (VD) could be unusable. For example, if one of the pedicles has impairments such that it is not implantable, the conversion functions (Fconv) having as variable the density of the two pedicles (DM) no longer make sense for this determined vertebra (VD). It will then be necessary to retain only the conversion functions (Fconv) having as variable the density of only one of the pedicles (DM).

 Other features and advantages of the invention will appear more clearly on reading the description hereafter of a preferred embodiment of the invention given by way of illustrative and non-limiting example, as well as the appended drawings among which :

 Figure 1 is a schematic representation of the evaluation method according to the invention;

 FIG. 2 is an illustration according to an axial sectional view of a vertebra with a pedicle screw inserted to the left in the pedicle and the anterior body, and a delimitation on the section of a volume of interest (VI) in the pedicle. to the right ;

 Figure 3 is an illustration of a generic digital vertebra model.

 As shown in FIG. 1, the subject of the invention is a method for estimating maximum pulling force (FA) forces of a particular anchor screw model in an implantation site chosen for a determined vertebra (VD) according to the principles of a given insertion technique.

 The anchoring screw is in particular a pedicle screw. FIG. 2 shows the implantation of an anchoring screw 3 into a vertebra 1 by a path passing through the middle of the vertical and horizontal sections of the left vertebral pedicle 4 to terminate in the anterior body of the vertebra 2. The trajectory is oblique. If two screws are implanted, their respective longitudinal axes converge towards the points of the screws.

 The method according to the invention comprises a succession of steps, of which a first part is carried out on control vertebrae (VT), and a second part is performed for at least one determined vertebra of a patient (VD) for at least one Insertion site given according to the principles of a given surgical technique.

 These control vertebrae (VT) are preferably cadaveric, as detailed later, or synthetic.

In this example two volumes (VI) will be described in control vertebrae (VT) of a sufficient number so that the determined vertebra (VD) benefits from statistically significant and representative calculations for the clinical category of the population to which the patient and the patients belong. donors (in practice at least 6 vertebrae witnesses by vertebral level). The vertebral level (VD) to be implanted is L4. This significance is established by a value of p-value <0.01. The two volumes are: Vl [1], constituted by the volume of the anterior vertebral body, and Vl [2], constituted by the volumes of the two pedicles. Similarly for the example, three conversion functions [Fconv] are defined literally:

 Fconv [1] = 1. DM [Anterior Vertebral Body] = 1. DM [1] = DM [1]; Fconv [2] = 1. DM [Anterior Vertebral Body] + 1. DM [Pedicles] = 1. DM [1] + 1. DM [2] = DM [1] + DM [2]);

 Fconv [3] = 0.3. DM [Anterior Vertebral Body] + 1. DM [Pedicles] = 0.3. DM [1] + 1. DM [2] = 0.3. DM [1] + DM [2]

 The steps of the first part of the process according to the invention are as follows:

 at. acquisition of unit density data (du) as well as their spatial coordinates for control spines by at least one non-invasive medical imaging technique (T1) for identifying a subset of at least one unit density (of the ) and its spatial coordinates for each control vertebra (VT), the set of unit density data (du) measured for each control spine is [1], [2], ..., [z] , "Z" being the maximum number of measurements allowed by the resolution of the imaging system, a subset of these data (du) is occupied by the bone volume of each control vertebra {VT [1], VT [2] , ..., VT [6]};

b. mechanical tensile strength tests of the anchoring screw implanted in one or more insertion sites of each of the control vertebrae (VT) with determination of the maximum force of resistance to experimental tearing (FA _EX ) on the Force / Displacement curve recorded during each test: here a single value per control vertebra (VT) from 1 to 6 is 6 values: {FA _EX [1], FA _EX [2], ..., FA _EX [6] };

 c. delimiting here volumes of interest VI [1] and VI [2] for each vertebra of control {VT [1], VT [2], ..., VT [6]} of lumbar level L4, ie 12 volumes: { Vl [1] [1] Vl [2] [1], ..., Vl [6] [1]} + {Vl [1] [2] Vl [2] [2], ..., Vl [ 6] [2]} from:

[5] [1] and [5] [z]] donor unit spine unit data, the subset of the data unit (s) of [1] [ai], [1] [a ₂ ], ..., [1] [a _n ]} being assigned to volume Vl [1] of control vertebra VT [1] ] of lumbar level L4 of the donor d, ie Vl [1] [1],

- the same for the donor e and the subset of unit data {[2] [bi], [2] [b ₂ ], ..., [2] [b _m ]} being attributed to the volume VI [1] of the control vertebra L4 VT [2], ie Vl [2] [1];

- finally, for donor z, the subset of unit data {of [6] [zi], [6] [z ₂ ], ..., [6] [z ₀ ]} being attributed to volume VI [1] of the control vertebra L4 VT [6], ie Vl [6] [1];

 Same reasoning with other subsets of unit data (du) for each control vertebra (VT) for DM [2]:

 - the subset of the unit data of [1] [a1], [1] [a2], ..., [1] [ap]} of vertebra L4 of spine d is assigned to volume Vl [ 1], [2]

 - the subset of the unit data [of [2] ^ 1], [2] ^ 2], ..., [2] ^ q]} of vertebra L4 of rachis e is attributed to volume Vl [ 2] [2],

- the subset of the unit data of [6] [oo1], [6] [oo2], ..., [6] [ooo] of the vertebra L4 of the rachis z is attributed to the volume Vl [6] ] [2]

 According to a particular embodiment, the first part of the method further comprises the following steps:

• determination of the average densities of interest {DM [1] [1], DM [2] [1], ..., DM [6] [1]} and {DM [1] [2], DM [2] ] [2], ..., DM [6] [2]} for the two volumes of interest Vl [1] and VI [2] defined in (c) from the total volume of unit volumes of each unit data (from) each subset, ie 12 average densities of interest (it should be noted that each subset of unit data (du) assigned to each volume of interest (VI) for each control vertebra (VT) can to be directly in the referential image of each spine); • arbitrary definition of three conversion functions (Fconv):

 - Fconv [1] = DM [1];

 - Fconv [2] = DM [1] + DM [2];

 - Fconv [3] = 0.3. DM [1] + DM [2]

 Then we calculate, by the three conversion functions (Fconv [1], Fconv [2], Fconv [3]), thanks to the values of the average densities of interest {DM [1] [1], ..., DM [6] [2]}, mathematical images l [1] = {Fconv [1] [1], ..., Fconv [6] [1]}, l [2] = {Fconv [2] [1 ], ..., Fconv [6] [2]} and l [3] = {Fconv [1] [3], ..., Fconv [6] [3]} (Example: Fconv [6] [3 ] = 0.3 DM [6] [1] + DM [6] [2]), ie 6 values per conversion function (Fconv).

 It is then proceeded to a computation of three correlations of sets of mathematical images l [1], l [2] and l [3] with FAEx = {FAEx [1], ..., FAEx [6]}, and to a literal definition of their respective regression function (Frég [1], Frég [2] and Frég [3]) from the parameters of definition of the regression lines:

 - Freg [1] = a. Fconv [1] + b,

 - Freg [2] = c. Fconv [2] + d,

 - Freg [3] = e. Fconv [3] + f,

where "a", "c" and "e" are the slopes of the regression lines and "b", "d" and "f" are their original values.

 It is assumed that only two correlations have acceptable statistical error margins and therefore only the conversion functions (Fconv [2] and Fconv [3]) are retained at the end of this pre-selection substep.

The control vertebrae (VT) are obtained by sampling from deceased donors (specimens). These specimens and their vertebrae must have clinically relevant characteristics to establish maximum experimental tearing strengths (FAE _X ) of the anchor screw.

For example, one way to obtain consistent information on bone densities is to select donors with no apparent surgical history (or dissection) of the spine, no spinal deformities, and excluding imaging and surgical dissection vertebrae presenting signs of pathologies such as cancers, bone metastases, angiomas and those with signs or sequelae of fractures. The use of cadaveric vertebrae makes it possible to have donor Vertebrae (VT) which have the most clinical similarities for their bone quality with the patient population for which the vertebrae (VD) are likely to be implanted by screws. anchor. Advantageously, the cadaveric vertebrae (VT) also contain a trabecular micro-structure not available in commercial synthetic foams.

 For the step (a) of acquisition of at least one unitary density data (du) by imaging (Tl), an acquisition of a clinical imaging (Tl) producing a 3D geometry of the unit bone densities (of the ) control vertebrae (VT).

 The imaging data (s) are obtained by one of the following clinical imaging techniques (Tl):

 X-ray computed tomography;

 X-ray stereo-radiography and dual-energy absorptiometry;

 planar bi-photonic x-ray absorptiometry;

 - X-ray volumetric bi-photonic absorptiometry.

With the exception of planar bi-photonic x-ray absorptiometry, these acquisition methods make it possible to obtain unitary data sets (du) for imaging all the vertebrae (VT) by projecting the planes of the unit volumes determining unit densities (du), these sets are called slices. These unit volumes, according to the modalities used preferentially (Tl), called voxels, are set during the acquisition of the data (of) in step (a). Each voxel is assigned a gray level value according to the tissue densities and a calibration of the imaging system by a calibration phantom or by direct equivalence by imaging a calibration phantom at the same time as the spines. The unit densities (of) measured are convertible into weight equivalence expressed in grams of hydroxyapatite equivalent per mm ³ (g (HA) .mnr ³ ).

 Advantageously, the step (c) of definition for each control vertebra (VT) of the volumes of interest (VI) comprises:

a sub-step of segmentation of the vertebra (VT), or part of the vertebra, and / or directly of clipping of the volumes of interest (VI) or the location thereof, on images acquired by the or imaging techniques acquired in step (a); a sub-step of excluding areas of pathological bone reconstructions and non-bone areas of the vertebra.

 The substep of segmentation makes it possible to delimit the external bone envelope of the vertebra (VT), or directly of the volumes of interest (VI) in the vertebra (VT).

 The sub-step of clipping also makes it possible to perform the sub-step of excluding zones of pathological bone reconstructions. It excludes osteophytes, osteoarthritic excrescences peripheral to the cortical bone which do not contribute to the mechanical qualities of the bone of the vertebra (VT).

This exclusion also makes it possible not to distort bone densities (MD) by excluding non-bone areas that are of lower densities (soft tissues, intervertebral disks, cartilage) and artificially reduce the average density of interest (MD) of the zone. that one studies without contributing to the biomechanical properties of the vertebral bone (FA _EX ).

 More specifically, and with reference to FIG. 1, step (b) of the mechanical tearing tests comprises:

a sub-step of implantation of the anchoring screws in an insertion site defines in an identical manner for all the control vertebrae (VT) and according to the principles of the selected insertion surgical technique; - a substep of tearing all the anchor screws in the control vertebrae (VT), and obtaining the maximum tensile strengths experimental experimental FAE _X [1], FAE _X [2], ..., FAE _X [6]} on the Force / Displacement curves of each test.

 Prior to the implantation sub-step, a surgical preparation of each control vertebra (VT) is carried out in order to eliminate all the soft tissues and the intervertebral discs and the anchoring screws are inserted according to the principles of the surgical technique. retained insertion (oblique transpedicular trajectory, bicortical ...). The vertebrae (VT) thus prepared are included in a synthetic cement and presented in the desired tear axis on the test bench.

 X-rays can be taken to check the positional compliance of the anchor screws in the vertebrae (VT) against the principles of the selected reference surgical technique.

During the sub-step of tearing the screws of anchoring of the control vertebrae (VT), one obtains the maximum forces of resistance to the experimental tearing (FA _EX ) by the exploitation of the curves Force / displacement obtained during mechanical tensile tests in continuous traction along the axis of insertion of the anchoring screws (pedicle screw) in the control vertebrae (VT).

 Of course, the mechanical test can be performed with a combination of traction axes, and may optionally include mechanical fatigue cycles to simulate longer-term postoperative degradation of the deterioration of the screw strength if one wants to obtain longer-term evaluations.

 For the step (c) of defining at least one volume of interest (VI), a volume (VI) is identified that can be anatomical, geometric, derived from morphological parameters of the vertebra, or be a combination of these definitions.

 For example, a volume of interest (VI) may correspond to the trabecular area of the anterior body of a vertebra or to a cylinder contained in this zone.

 With reference to FIG. 2, a volume of interest (VI) approximating the entire pedicle (4) is represented on the image of the vertebra 1.

 By iteration of the method of steps (c) to (f), the definition of a volume of interest (VI) can be approximated or refined by another new volume (VI). This new volume of interest (VI) could consist of a cylindrical or oval volume included within the limits of the pedicle, passing right through it into the largest possible diameter, then developing longitudinally along the major axis of the pedicle, and circular shape, in the anterior vertebral body, defining a trajectory called "oblique trans-pedicular pathway". This kind of volume (VI) makes it possible to better reflect the densities of the various anatomical zones (VI) traversed by the screws inserted in step (b) and makes it possible to better determine in step (g) the parameters of a screw maximum diameter and maximum length, without causing cortical break-in at the insertion, likely to get the best hold (FA).

 Volume (VI) definitions can be introduced that reduce segmentation efforts, which can be compared with the performance of their conversion function (Fconv) against defined volumes (VIs) with complete rigorous segmentation and clipping.

According to an advantageous characteristic of the evaluation method, during step (c) of definition of at least one volume of interest (VI), volumes of interest (VI) are selected that are whole (cortical and cancellous bone). ), or partial (cortical bone, cancellous or cortical-subcortical), or volumes (VI) whose cortical-subcortical zone is excluded. The cortical-subcortical is composed of the cortical bone and an additional margin of spongy bone in continuity of a fixed thickness. This thickness can be defined separately for the cortex of the pedicle or of another part such as the anterior vertebral body.

 Although having a higher density, the cortical bone tissue is sometimes less than the spatial resolution of the clinical imaging systems (T1), which is a source of error. Consequently, based more on the cortical bone but on a cortico-subcortical zone, the systematic error of the determination of this cortical zone due to the artificial decrease of the unit densities of the corticals whose Unit volumes may include some of the lower density tissue lining the outside of the cortex on one side (soft tissue, osteophytes) and / or cancellous bone on the inner side of the cortex, with spongy bone being less dense than the cortical bone in the healthy subject. This artifact due to the resolution of imaging systems (Tl) is known in clinical radiology as "partial volumes".

 This step (c) of delimitation of at least one volume of interest (VI) can be performed by adapting the definition parameters of a generic digital model comprising typical interest volumes (VI) in this generic reference frame, for conform to the limits of the outer envelope of the vertebrae (VT or VD) and / or marks obtained during the substep segmentation. This model makes it possible to directly transpose predefined interest volumes (VI), identified on the generic numerical model and to obtain morphological measurements, for example by an algorithm.

 Such a generic numerical model representing an axial section of a typical vertebra is represented by FIG.

 During the deformation of the model, for example by means of an algorithm, so that it conforms to the limits of the external envelope of the determined vertebra, or other markers, the volumes of interest (VI) predefined types in this model are also deformed by fitting on predetermined morphological variables of a vertebra type to the actual variables measured, or thanks to markings defined on the vertebrae (VT or VD).

For the step (d) of determining the average density of interest (MD) of each volume of interest (VI), the weight equivalents of each voxel in a volume of interest (VI) are calculated (by to its value of gray level which can be transformed into density and weight according to a method well known in the art previous), then the average density of interest (MD) is calculated from the volume of interest (VI) and from the sum of the weight equivalents of each voxel.

 This step (d) of determining the average density of interest (MD) can for example be performed using one of the methods derived from the mathematical theory of Gaussian quadrature to pass from the volume of the subset of the voxels to that of the volume of interest (VI).

 Finally, the determination step (e) preferably comprises, for at least one regression function (Frég):

 a sub-step of literally defining at least one conversion function (Fconv) comprising as variable at least one average density of interest (MD) for each control vertebra (VT); a sub-step of calculating the mathematical images (I) by this conversion function (Fconv) applied to the values of the average densities of interest (MD) calculated in (d) for each control vertebra (VT);

a sub-step of correlating these mathematical images (I) of each conversion function (Fconv) with the maximum forces of experimental tear resistance (FA _EX ) measured in step (b); a sub-step of literal definition of its regression function (frig) associated with the correlation;

 a substep of calculating statistical error margins and statistical significance for each correlation.

 The use of the regression function or functions (Freg) associated with each conversion function (Fconv) including as variables one or more average interest densities (MD) and possibly other morphological parameters and index parameters. bone structure texture, allows for the same parameters calculated in step (g) for a determined vertebra (VD), and transformed by the same conversion functions (Fconv), to obtain the estimate of maximum forces of Custom tear resistance (FA).

 In other words, the average interest densities (MD) of the selected interest volumes (VI), and possibly other morphological parameters or bone frame texture indices are combined, which may correspond, for example , to the following formula:

"Fconv [n] = a. DM [1] + b. DM [2] + ... + x. DM [w] ". For example, with a single average density of interest (MD), one can have a regression function "m" (Freg) formulated as follows:

 "Fconv [m] = 1. DM [Anterior Vertebral Body]", ie "Fconv [m] = DM [Anterior Vertebral Body]".

 Finally, the definition parameters of the regression line make it possible to obtain, from the images (I) of the conversion function (Fconv) applied to the parameters (DM) of the determined vertebra (VD), an estimation of the force maximum tear resistance (FA).

 The definition of the parameters of the regression function (Freg) corresponds to the values (slope + value at the origin) of the linear regression line.

 The regression function can take the form "y = a. x + b "where:

 "Y" is the estimate of the maximum pull-out force (FA) of the anchor screw in a determined vertebra (VD) that is desired;

 "X" is the value of the conversion function (Fconv) for the determined vertebra data (VD);

 "A" corresponds to the slope of the straight linear regression;

"B" is the right-hand value of the linear regression. Let: FA [n] (VD) = a. Fconv [n] (VD) + b, with Fconv [n] (VD) = ai. DM [VD] [1] + a ₂ . DM [VD] [2] + ... + a _w . DM [RV] [w]

 To improve the composition of the conversion function (Fconv), a study of variation of the weighting coefficients of the average interest densities (MD) can be performed.

 Similarly, the average interest densities (MD) can undergo transforms to linearize (logarithmic function, exponential, polynomial, ...).

 Since it is possible to calculate as many estimates of maximum tearing forces estimated (FA) per vertebra (VD) as there are regression functions (Freg) and thus conversion functions (Fconv), according to the present mode embodiment, step (e) comprises a final sub-step of selecting the maximum pull-out force (FA) estimates for a specific vertebra insertion site (VD).

This final sub-step of selection is performed according to criteria which comprise, for each correlation, a coefficient of determination (R ² ) and / or a standard error estimation value (SEE). These conditions make it possible to eliminate conversion functions [Fconv] that do not result in sufficient correlation or SEE too high, and thus eliminate volumes of interest (VI) that are never contributory to statistically efficient conversion functions (Fconv), this is based on the collection of vertebrae tested (VT).

 These selection criteria may also include other process use constraints: technological, practical, financial, time spent on the process, limitations of the clinical imaging system used, and ease or reproducibility. the placement of the volumes of interest (VI) or other criteria measured in a clinical study.

The estimates of maximum pull-out forces obtained by decreasing R ² and / or decreasing SEE are classified. With these rankings, it is possible to restrict the pre-selection of conversion functions (Fconv) to the ease of determining them or because they respond more realistically to implementation constraints, while approaching in an acceptable manner. other conversion functions (Fconv) whose error margins are the lowest.

In conclusion of the steps of the first part of the method, it has thus defined volumes of interest (VI), their respective average densities of interest (DM), conversion functions (Fconv) linking one or more of these densities (DM ), the correlation between the mathematical images (I) of each of these conversion functions (Fconv) and the experimental forces (FAE _x ), as well as the statistical error margins, and the determination of the definition parameters of the straight line of regression and thus the regression function (Freg) for each conversion function (Fconv).

 For the second part of the process according to the invention, the steps are as follows:

 acquisition for a given vertebra (VD), here L4, of at least one unit density data (du) as well as their spatial coordinates in the spine of a patient by the same imaging technique (Tl) that in step (a);

 positioning in the determined vertebra (VD) of the volumes of interest V1 [1] and V1 [2] to obtain volumes similar to those defined in (c);

calculating for the volume of interest Vl [1] and Vl [2] of the determined vertebra (VD) from the imaging unit data (du) acquired in step (f), the subset of the data units of [vi], [v2], ..., [v _p ]} being assigned to volume Vl [1] for the calculation of DM [1] of the vertebra determined (VD) and the subset {of [wi], [w2], ..., of [w _p ]} being attributed to the volume of interest Vl [2] for the calculation of DM [2], the two subsets being assigned respectively to the two volumes of interest (VI) by their spatial coordinates during the operation of segmentation of the determined vertebra (VD) L4 (manual, automated or semi-automated), and then deduced therefrom their average densities of interest DM [1] and DM [2] by determining the weight of each of the two volumes;

application of the two regression functions (Frég) resulting from the pre-selection in step (e), to the average interest densities DM [1] and DM [2] resulting from step (h), to determine estimates of the two maximum pull-out forces (FA _EX [2] and FA _EX [3]) of the anchor screw for the determined vertebra (VD) L4.

 In this second part of the method, the imaging technique (s) (Tl) are the same for the determined vertebra (VD) as those used for the imaging data acquisition step for the control vertebrae (VT).

 According to the present embodiment, the transposition step for the determined vertebra (VD) also comprises:

 a substep of vertebral segmentation (VD), or directly volumes of interest (VI) on the images acquired by the one or more imaging techniques (T1);

 a substep of exclusion of pathological bone reconstructions or non-bone tissue by vertebra clipping or directly of the volumes of interest (VI) during this segmentation.

 Still according to the present embodiment and as for the step (c) of definition of the two volumes of interest (VI), the step of transposition of the second part of the process can be carried out by deforming a generic numerical model describing these two typical interest volumes (VI) for conforming to the determined vertebra (VD);

 The use of the generic digital model is carried out in the same manner and with the same advantages as for step (c) previously described.

In the second part of the method, we apply the two regression functions (Freg [2] and Freg [3]) resulting from step (e) of the first part of the method to the parameters obtained for the determined vertebra (VD). to compute three estimates of maximum forces of pullout resistance (FA [2] and FA [3]). In the application example, it is assumed that the conversion function Fconv [3] is the best in terms of margin of error (R ² = 0.78 and SEE = 12.7%) compared to Fconv [2] also preselected, and that it can be retained since it is also posited that the vertebra determined (VD) seems intact in medical imaging (Tl) and the clinical examination for the volumes of interest Vl [1] and Vl [2] We assume that the slope of the regression line is "e" is 2437 and that the value at the origin of "f" is - 841. The result of Fconv [3] ] for the determined vertebra (Fconv [3] (VD)) is 0.3451 g (HA) .mm 3. The application of the regression function Freg [3] to the parameters of the determined vertebra (VD) (FA [ 3] = 2437. Fconv [3] (VD) - 841) results in the following result:

- FA [3] = 582 N ± 74 N, R ² = 0.78, SEE = 12.7%, p-value <0.01, where 582 N is the estimate of the maximum force of tearing resistance ( FA) with a margin of error of ± 74 N and for a sufficient number of 6 measurements of the maximum force of resistance to experimental tearing on control vertebrae (FAE _X ) of the same level L4 as the determined vertebra (VD) to satisfy the condition of p-value <0.01.

The evaluation of a maximum force of tearing resistance (FA) is specific to the parameters defined by the manufacturer for the manufacture of the anchor screw tested on the control vertebrae (VT), and to the principles of the surgical technique of implantation chosen (implantation site and trajectory), and is worth, without correction, only for these same conditions for the determined vertebra (VD).

 Then, the operator obtains an evaluation of the force corresponding to the maximum force of tearing resistance (FA) most relevant for an insertion site in this determined vertebra (VD), assuming each time that the same vertebral level was validated during steps (a) to (e).

 The clinician can repeat the steps of the second part of the procedure for all vertebrae (VD) he wants to implant for this patient.

 With these assessments of maximum pull-out strength (FA), a surgeon can estimate whether holding (in Newton) is sufficient for each of the screws to be implanted for the patient's vertebrae (VD) in order to withstand the mechanical stresses of the patient. moments of force transmitted to all the screws during the coupling of the screws to the rods to obtain the correction or corrections envisaged.

The practitioner can then optionally use the values of the estimates (FA) in a numerical simulation software designed to evaluate whether the maneuvers that the surgeon wishes to achieve comply with the recommendations of the scientific medical literature.

 Following these evaluations, possibly with the same simulation tools, the surgeon can consider several options:

 validate the implantation of anchoring screws envisaged according to the conventional surgical technique;

 to decrease or increase the number of screws or vertebral levels envisaged for the implantation of the anchoring screws;

 or use other additional treatments or additional medications, surgical or radio-interventional.

 Thus, after taking into account correction factors in resistance of holding forces (FA) promised by the bypass solutions, he can stop his surgical strategy for this patient.

 The evaluation method according to the invention thus makes it possible to obtain information for a specific surgical planning for a patient, making it possible to evaluate the need for the use of complementary or additional medical, surgical or radio-interventional treatments.

 The evaluation method according to the invention also makes it easy to deduce the impact on the estimation of the maximum pull-out force (FA) of a longer anchoring screw or the largest possible diameter. for the same vertebra (VD).

 For this purpose, the insertion lengths of the anchoring screws are measured in the control vertebrae (VT), for example in the pedicle and in the anterior body, and also the shortest distance from the outer edge of the screw to the outer edge. of the pedicle cortex.

 Indeed, it is interesting to take into account the shortest distance between the outer limits of the body of the screw and the cortex of the pedicle that the screw passes right through to end in the anterior body, distance measured on the perpendicular line to the longitudinal axis of the screw. The closer the anchoring screw is to the cortex, and if it respects the entire cortex, the better the anchoring is by increasing the compaction of the cancellous bone between the body of the screw and the cortex.

In the same way, one can measure the residual volume that would persist between the outer surface of life and the cortical pedicle. It may also be interesting to analyze by micro-X-ray tomography for the control vertebrae (VT) before and after insertion of a screw, the impact of the insertion of the studied screw on the compaction of the trabecular bone. in the vicinity of the screw and its influence on the maximum force of resistance to experimental tearing (FAE _X ) and to integrate the personalized modeling of the compaction gradient at the determined vertebra (VD) according to the parameters of the screw.

 One can also compose a function linking two or more of these three parameters to assign to each volume crossed (VI) by the screw a participation in the maximum force of tearing resistance (FA).

 For the same range of screws, it is thus possible to evaluate the implantation of screws having a different diameter and / or length.

 One can also readjust the estimates of maximum pull-out forces (FA) obtained with the parameters of another given surgical technique, or a technique of increasing bone density, or any other workaround mentioned. For example, scientific publications estimate at about 2.5 the coefficient to be applied to the forces (FA) obtained by a conventional method of insertion, compared to those obtained by an increase in the density of the anterior vertebral body by infiltration of polymethacrylate. Methyl (PMMA) using cannulated pedicle screws and perforated compared to screws of identical design without perforations and without infiltration of PMMA.

 Estimates of maximum tensile strengths (FA) can be compared with postoperative clinical follow-up of patients to evaluate them as predictors in the short, medium and long term complications of surgical procedures (screw movements, non-union, handicrafts, ...).

 Variations in the density distribution (DM) between two populations whose unit data (du) are acquired in the same way in steps (a) and / or (f), as well as other parameters, can be compared to calculating the force estimates (FA) of a patient of a first population with the conversion functions (Fconv) determined for another population by adjusting for these variations of these conversion functions (Fconv).

Obtaining intraoperative data during the actual implantation of the pedicle screws may make it possible to improve the estimates (FA) both in terms of value and in terms of statistical margins of error. For example, a clinical study of practitioners' assessments of the estimates given (FA), a post-clinical follow-up operative, or measuring the torque of force applied by the screwdriver to each of the screws inserted into the vertebrae of the patients.

Claims

1. A method of evaluating maximum tensile strengths (FA) of an anchoring screw for a specific vertebra (VD) in which the anchor screw is capable of being implanted according to the principles of a given insertion technique, comprising successively:  at. acquiring an imaging repository including unit density data (du) as well as their spatial coordinates for at least one control vertebra with at least one non-invasive medical imaging (Tl) technique; b. at least one mechanical tensile strength test of the anchoring screw implanted in the control vertebra (s) (VT) to determine the maximum force of resistance to experimental tearing (FAE _X );  c. defining at least one volume of interest (VI) in each control vertebra (VT) using the imaging reference acquired in step (a); d. calculating for each volume of interest (VI) of each control vertebra (VT) of a mean density of interest (MD);  e. defining at least one conversion function (Fconv) comprising at least one average density of interest (MD) variable calculated for each control vertebra (VT),  f. correlating the values of each conversion function with values of the one or more mechanical tests to establish a regression function;  g. acquisition for a determined vertebra (DV) of a patient, an imaging repository similarly to step (a);  h. transposition of step (c) to the determined vertebra (VD);  i. transposing step (d) to the results of step (h); j. applying a conversion function defined in step (e) to the average densities (DM) from step (i) to obtain a score; k. obtaining an estimate of a maximum pulling force (FA) per application to the score obtained in step 0 of the regression function associated with the conversion function of step 0. 2. Evaluation method according to claim 1, characterized in that the imaging data is obtained from at least one of the following imaging techniques:  X-ray computed tomography;  - X-ray stereo-x-ray and dual-energy x-ray absorptiometry;  stereo x-ray bi-planar X-ray;  planar bi-photonic x-ray absorptiometry;  - X-ray volumetric bi-photonic absorptiometry;  - vertebral bone matrix texture index. 3. Evaluation method according to any one of the preceding claims, characterized in that step (c) and / or step (h) comprise a substep of segmentation by trimming of the bone boundaries of the vertebrae on images acquired by the one or more imaging techniques (Tl). 4. Evaluation method according to any one of the preceding claims, characterized in that step (c) and / or step (h) comprise a substep of exclusion of non-bone areas or bone reconstructions. pathological. 5. Evaluation method according to any one of claims (3) and (4), characterized in that step (c) and / or step (h) are performed by adapting the shape of a model generic digit (2), comprising predefined volumes of interest (VI), in the shape or vertebrae obtained during the substep of segmentation or as a function of marks recorded directly during the substep of segmentation of the control vertebrae or determined vertebrae. 6. Evaluation method according to any one of the preceding claims, characterized in that step (e) for determining at least one regression function comprises: a sub-step of defining at least one conversion function (Fconv) comprising as variable at least one average density of interest (MD) for each control vertebra (VT); a sub-step of calculating mathematical images (I) using the conversion function (Fconv) applied to the values of the average densities of interest (MD) calculated in (d) for each control vertebra (VT) ; a sub-step of correlating said mathematical images (I) of each conversion function (Fconv) with the maximum forces of experimental tear resistance (FA _EX ) measured in step (b); a sub-step of literally defining a regression function (Freg) associated with the correlation;  a substep of calculating statistical error margins and statistical significance for each correlation. 7. Evaluation method according to any one of the preceding claims, characterized in that the step (e) of determining at least one conversion function (Fconv) comprises a final substep of pre-selection of functions. of conversion (Fconv). 8. Evaluation method according to claim 7, characterized in that the final pre-selection sub-step is performed according to selection criteria which comprise, for each regression function, a coefficient of determination (R ² ), and a standard error estimation (SEE) value of the correlation of each of the sets of images (I) obtained by the application of each of the conversion functions (Fconv) with the maximum strengths of resistance. Experimental tearing out (FA _EX ) obtained on control vertebrae (VT).