The present invention relates to a method of constructing generic tooth geometry, in particular for a dental implant. The invention also relates to a generic dental implant formed according to the method, and to a method of forming a cutting head for a surgical procedure using the method of constructing generic tooth geometry. The invention also relates to a method of constructing a generic biomedical geometry, biomedical implant and biomedical cutting head, not necessarily solely being related to a tooth.

The process for creating bespoke biomedical patient implants for use during a surgical procedure is both time-consuming and expensive. However, generic biomedical patient implants can often be ill-fitting and inconsiderate to the patient's post-surgical needs.

One context in which generic patient implants are particularly ill-suited is in the field of dental surgery. At present, if a tooth has been removed from a patient's jaw, a dental implant can be fitted. This will typically comprise three portions: a substantially conical implant which is mechanically affixed to the mandible or jaw during surgery; an abutment element which fits into the implant in a wedging arrangement; and a crown portion, which engages with a projecting portion of the abutment element. It is feasible that the implant and abutment may be one-piece instead of being separate parts.

The implant anchors the dental implant in the jaw, but a dental surgeon must perform a great deal of surgery in order to ensure that the conical implant is accepted into the patient's mandible/maxilla and gingiva. This is traumatic to the patient undergoing surgery.

Furthermore, the dental surgeon is required to choose a dental implant based on a best fit from a collection of implants which they may have in their possession, and the dental surgeon must then create an acceptable fit between mandible/ maxilla and implant, which may involve corrective surgery to the jaw, often including bone harvesting, bone grafting drilling and reaming.

Bespoke implants are created by creating a replica of the extracted tooth which can be inserted into the gum and mandible, which can be fixed in place whilst the implant is naturally secured in the mandible by replacement osseous tissue. Such bespoke implants must be uniquely machined, and therefore the manufacturing capability to produce such implants in vast numbers is not available.

It is an object of the present invention to improve or substantially obviate the problems as presented above by providing a means of generating a generic dental implant.

The above issues are also prevalent in other biomedical implants and geometry, for example, at the ends of bones that need to be resected or rebuilt to accommodate an implant or prosthesis.

According to a first aspect of the invention there is provided a method of constructing a range of generic dental implants from a collection of tooth samples, the method comprising the steps of: a] providing a database of digitised like tooth samples from an analysis group; b] creating a baseline model from the digitised like tooth samples; c] transforming the baseline model to the digitised like tooth samples database to generate a normalised digitised like tooth sample database; d] applying principal component analysis to the normalised digitised like tooth sample database; e] extracting characteristic data relevant to one or more characteristics of the analysis group; f] determining from the characteristic data one or more generic tooth geometries; and g] constructing a range of generic dental implants in accordance with the range of generic tooth geometries.

The advantage of the present invention is that it is possible to extract generic tooth geometry from a database of tooth samples which have been analysed. This advantageously allows for the insertion of an anatomically representative dental implant into a patient, thereby avoiding the discomfort and complications of standard conical dental implants, whilst avoiding the time and expense associated with bespoke dental implants. By quantitatively providing for a range of generic tooth geometries, a dental surgeon can cater for the majority of the general population whilst only stocking a relatively small number of dental implants at any given time.

Although the terms 'dental implant', 'tooth' and 'teeth' are used herein and throughout, this is not intended to be limiting, and the methodology is applicable to other kinds of biomedical implants. As such, in additional aspects of the invention, the terms 'dental implant', 'tooth' and 'teeth' may be replaced by 'biomedical implant' and its associated terminology throughout, if required.

The application of principal component analysis to determine characteristic data in connection with the analysis group beneficially allows for the characteristics of the tooth geometry to be independently varied in order to create a nominally generic dental implant.

Optionally during step a], the database of digitised like tooth samples may be obtained from scans of physical tooth samples, for instance, via three-dimensional slice tomography, and the digitised like tooth samples may be segmented based on a density profile of each physical tooth sample.

By using data collected from physical tooth samples, a detailed statistical analysis of the teeth of a population can be determined accurately, which will help the fit of any dental implants made from the generic tooth geometry. Segmentation advantageously allows for the internal tooth structure and material profile to be determined and analysed, which can better inform the subsequent analysis.

Preferably, the digitised like tooth samples may be represented as three-dimensional meshes. Furthermore, the baseline model may be a standard mesh, the standard mesh being mapped onto the three-dimensional meshes using a node-to-node comparison, and the node-to-node comparison may be performed using elastic mesh morphing. The mesh density may be in the range 0.001 to 1 mm maximum node-to-node distance, and more preferably may be in the range 0.1 to 0.5 mm maximum node-to-node distance.

A mesh is the most straightforward way in which three-dimensional tooth geometry can be digitally represented and compared, allowing for areas of curvature of the tooth geometry to be accurately mapped.

Preferably, the characteristic data relevant to one or more characteristics of the analysis group may be represented by independent modes of shape variation of the digitised like tooth samples.

By separating the characteristics of the tooth into modes which can be varied independently of the other determined modes, the user is beneficially able to individually alter the modes in order to achieve a range of realistic tooth geometries.

In a preferred embodiment, the method may further comprise a step subsequent to step d] of validating the results of the principal component analysis.

Validating the results of the principal component analysis, using one or more specific validation techniques, allows the user to determine whether erroneous or anomalous data has been incorporated from the analysis group and also if sufficient unique geometry instances have been included, thereby allowing recalibration of the analysis to generate a more accurate overall tooth geometry.

Preferably, the digitised like tooth samples may be taken from a set of tooth samples having a predetermined type, and more preferably the digitised like tooth samples may be taken from a sub-set within the set of tooth samples having a predetermined type. In order to represent the teeth digitally, it may be preferred to group the teeth, for instance, by type, that is, incisor, canine, premolar and molar, or a subset thereof, according to dentition. This allows the analysis to be conducted separately for each tooth type, increasing the accuracy of the results. Preferably, the method may further comprise a step between steps f] and g] of strategically modifying the characteristic data of the range of generic tooth geometries in order to improve a fit of the range of generic dental implants. The modification of the characteristic data may include modifying any one of: a mesio-distal aspect; a bucco-lingual aspect; and/or a superior- inferior aspect, and preferably any one of: oversizing the mesio-distal aspect by up to 20%; undersizing the bucco-lingual aspect by up to 20%; and/or undersizing or segmenting the superior-inferior aspect by up to 20%. This may allow for simpler insertion or alignment of a generic dental implant into a generic patient's oral cavity, thereby providing for a range of dental implants which are suitable for insertion across a statistically normal population.

According to a second aspect of the invention, there is provided a method of forming one or more generic dental implants comprising the steps of: a] providing a database of digitised like tooth samples from an analysis group; b] creating a baseline model from the digitised like tooth samples; c] transforming the baseline model to the digitised like tooth samples database to generate a normalised digitised like tooth sample database; d] applying principal component analysis to the normalised digitised like tooth sample database; e] extracting characteristic data relevant to one or more characteristics of the analysis group; and f] determining from the characteristic data one or more generic tooth geometries.

Constructing a generic dental implant in accordance with the determined generic tooth geometries advantageously allows for the provision of a small number of dental implants which can adequately fit the population at large. This reduces cost to the dental surgeon and patient, whilst also reducing the surgical challenges associated with the implant of cylindrical ill-fitting dental implants.

Preferably, during step g] a plurality of generic dental implants may be created having a range of geometries which vary according to the variation of the characteristic data across the analysis group. Creating a variety of generic dental implants, varying in geometry according to the statistical deviations of the determined independent modes of the generic tooth geometry ensures a wide coverage of the population for which the dental implants are effective. The method may preferably further comprise a step subsequent to step f] of parameterising the characteristic data for supplying to a CAD-CAM machine, the CAD-CAM machine constructing the one or more generic dental implant during step g].

By using the generic tooth geometry to inform design parameters for CAD-CAM manufacturing, generic dental implants can be mass-produced, thereby avoiding the time and expense associated with handmade bespoke dental implants.

According to a third aspect of the invention, there is provided a generic dental implant constructed having a generic tooth geometry formed according to a method in accordance with the second aspect of the invention.

According to a fourth aspect of the invention, there is provided a method of forming a cutting head of a dental cutting instrument, the method comprising the steps of: a] providing a database of digitised like tooth samples from an analysis group; b] creating a baseline model from the digitised like tooth samples; c] transforming the baseline model to the digitised like tooth samples database to generate a normalised digitised like tooth sample database; d] applying principal component analysis to the normalised digitised like tooth sample database; e] extracting characteristic data relevant to one or more characteristics of the analysis group; fj determining from the characteristic data one or more generic tooth geometries; and g] utilising the or each generic tooth geometry to define a geometry of a cutting head of a dental cutting instrument. According to a fifth aspect of the invention, there is provided a cutting head of a biomedical cutting instrument constructed having a generic biomedical geometry formed according to a method in accordance with the fourth aspect of the invention.

In order to insert a generic dental implant into a patient, it may be necessary to shape the bone to ensure a good acceptance of the implant. This can be readily achieved by forming an oscillatory, preferably ultrasonic, dental cutting instrument, which has a bone cutting head which is shaped so as to correspond with the root of the generic dental implant being inserted. Not only does this ensure a good fit, it also encourages the patient's natural healing mechanisms to ossify around the implant, forming a biological junction between implant and bone.

According to a seventh aspect of the invention, there is provided a generic biomedical implant constructed having a generic biomedical geometry formed according to a method in accordance with the sixth aspect of the invention. According to an eighth aspect of the invention, there is provided a cutting head of a biomedical cutting instrument having a generic biomedical geometry formed according to a method in accordance with the sixth aspect of the invention.

According to a sixth aspect of the invention, there is provided a method of constructing generic biomedical geometry from a collection of biomedical samples, the method comprising the steps of: a] providing a database of digitised like biomedical samples from an analysis group; b] creating a baseline model from the digitised like biomedical samples; c] transforming the baseline model to the digitised like biomedical samples database to generate a normalised digitised like biomedical sample database; d] applying principal component analysis to the normalised digitised like biomedical sample database; e] extracting characteristic data relevant to one or more characteristics of the analysis group; and f] determining from the characteristic data one or more generic biomedical geometries.

According to a seventh aspect of the invention, there is provided a method of forming a generic biometric implant comprising the steps of: a] providing a database of digitised like biomedical samples from an analysis group; b] creating a baseline model from the digitised like biomedical samples; c] transforming the baseline model to the digitised like biomedical samples database to generate a normalised digitised like biomedical sample database; d] applying principal component analysis to the normalised digitised like biomedical sample database; e] extracting characteristic data relevant to one or more characteristics of the analysis group; f] determining from the characteristic data one or more generic biomedical geometries; and g] constructing one or more generic biometric implant in accordance with the one or more biomedical geometries.

According to an eighth aspect of the invention, there is provided a generic biomedical implant constructed having a generic biomedical geometry formed according to a method in accordance with the sixth aspect of the invention. According to a ninth aspect of the invention, there is provided a method of forming a cutting head of a biomedical cutting instrument, the method comprising the steps of: a] providing a database of digitised like biomedical samples from an analysis group; b] creating a baseline model from the digitised like biomedical samples; c] transforming the baseline model to the digitised like biomedical samples database to generate a normalised digitised like biomedical sample database; d] applying principal component analysis to the normalised digitised like biomedical sample database; e] extracting characteristic data relevant to one or more characteristics of the analysis group; f] determining from the characteristic data one or more generic biomedical geometries; and g] utilising the or each generic biomedical geometry to define a geometry of a cutting head of a biomedical cutting instrument.

According to a tenth aspect of the invention, there is provided a cutting head of a biomedical cutting instrument having a generic biomedical geometry formed according to a method in accordance with the ninth aspect of the invention.

According to an eleventh aspect of the invention, there is provided a method of constructing generic biomedical geometry from a collection of biomedical samples, the method comprising the steps of: a] providing a database of digitised like biomedical samples from an analysis group; b] creating a baseline model from the digitised like biomedical samples; c] transforming the baseline model to the digitised like biomedical samples database to generate a normalised digitised like biomedical sample database; d] analysing the normalised digitised like biomedical sample database, and extracting characteristic data relevant to one or more characteristics of the analysis group; and e] determining from the characteristic data one or more generic biomedical geometries. According to a twelfth aspect of the invention, there is provided a method of forming a generic biometric implant comprising the steps of: a] providing a database of digitised like biomedical samples from an analysis group; b] creating a baseline model from the digitised like biomedical samples; c] transforming the baseline model to the digitised like biomedical samples database to generate a normalised digitised like biomedical sample database; d] analysing the normalised digitised like biomedical sample database, and extracting characteristic data relevant to one or more characteristics of the analysis group; e] determining from the characteristic data one or more generic biomedical geometries; and f] constructing one or more generic biometric implant in accordance with the one or more said generic biomedical geometries.

According to a thirteenth aspect of the invention, there is provided a generic biomedical implant constructed having a generic biomedical geometry formed according to a method according to the eleventh aspect of the invention.

According to a fourteenth aspect of the invention, there is provided a method of forming a cutting head of a biomedical cutting instrument, the method comprising the steps of: a] providing a database of digitised like biomedical samples from an analysis group; b] creating a baseline model from the digitised like biomedical samples; c] transforming the baseline model to the digitised like biomedical samples database to generate a normalised digitised like biomedical sample database; d] analysing the normalised digitised like biomedical sample database, and extracting characteristic data relevant to one or more characteristics of the analysis group; e] determining from the characteristic data one or more generic biomedical geometries; and fj utilising the or each generic biomedical geometry to define a geometry of a cutting head of a biomedical cutting instrument. According to a fifteenth aspect of the invention a cutting head of a biomedical cutting instrument having a generic biomedical geometry formed according to a method in accordance with the fourteenth aspect of the invention.

The invention will be appreciated as preferably not only being relevant in the field of dental surgery, but also potentially to more generic biomedical implants, for example, arthroplasty or skin grafting. The invention permits the user to compare the geometry of the biomedical sample from a population, so as to inform the generic geometry that suits the majority, preferably being 75 percent or more, of the population and which can then be utilised during a surgical procedure. The invention could also feasibly be used to inform digital reconstructions of forensic medical and/or historical archaeological biomedical samples. The invention will now be more particularly described, by way of example only, with reference to the accompanying drawings, in which:

Figure 1 shows a pictorial representation of one embodiment of a method of constructing generic tooth geometry, in accordance with the first and sixth aspects of the invention;

Figure 2 shows a diagrammatic representation of the method of Figure 1; Figure 3 shows a perspective representation of a physical tooth sample and a stack of scans through the physical tooth sample, used as input into the method shown in Figure 1;

Figure 4 shows a perspective representation of a digitised tooth sample of the physical tooth sample of Figure 3;

Figure 5a shows the variation of a plurality of digitised normalised tooth samples according to a first independent mode of shape variation, as determined during the method of Figure 1;

Figure 5b shows the variation of a plurality of digitised normalised tooth samples according to a second independent mode of shape variation;

Figure 5c shows the variation of a plurality of digitised normalised tooth samples according to a third independent mode of shape variation;

Figure 6 shows a [dental implant] formed in accordance with the third or seventh aspects of the invention; and

Figure 7 shows a pictorial side view of one embodiment of a cutting head of a dental cutting instrument, in accordance with the fourth or eighth aspects of the invention and for use in conjunction with the said methods.

Referring firstly to Figures 1 and 2 of the drawings, there is shown a generalised embodiment of a method of producing a generic biomedical geometry, in particular for the generation of a patient implant, shown globally as 1000. In the described embodiment, the generic biomedical geometry is shown in the context of a dental implant for a patient, but the person skilled in the art will appreciate that the method may be applied generally to any biomedical geometry. It will therefore be understood that the following references to teeth may be readily substituted for any biomedical sample, biomedical device, or biomedical implant to which the method is applied.

The generic biomedical geometry, and therefore generic biomedical implant, is designed to be applicable across a population, and therefore a range of generic biomedical geometries and generic biomedical implants will be generated in accordance with the following invention.

In order to generate a generic biomedical geometry, reference data is provided from a population SI 00. A plurality of physical tooth samples 10 from a population or analysis group may be provided, such as the physical tooth sample 10 shown in Figure 3. These physical tooth samples 10, comprising a root portion 12 and a crown portion 14, may be provided from any number of sources, for example: teeth extracted from patients during dental procedures; historical or archaeological tooth samples; or in vivo tooth samples which have been scanned to reveal the dental anatomy. It has been found that an analysis group of around thirty is sufficiently large to accurately capture the variance of the population, but any group size may be utilised. Biomedical samples are grouped such that the reference data only refers to like biomedical samples. For example, in the illustrated embodiment, only physical tooth samples 10 are shown. However, this could be readily streamlined, such that all tooth samples were specific tooth types, for instance, an incisor, or a maxillary right-side first molar. The scope of the categorisation of tooth type will naturally alter the nature of any generic biomedical geometry output therefrom. It will also be appreciated that physical tooth samples 10 may vary across the analysis group, and therefore the particular population chosen for the analysis group may be specifically filtered. For example, physical tooth samples 10 can be selected based on one or more characteristics of the person from which it was taken, such as age, gender and/or genetic or cultural characteristics.

In order to proceed with the method 1000, the collection of biomedical samples must be provided in a digitised database, ready for processing. The physical tooth samples 10 may be scanned, for example, in order to digitise at step SI 10 the collection, thereby forming a plurality of digital tooth samples 16, illustrated in part A of Figure 1, stored on a computing device 18 having at least a processor 20 and a memory storage device 22.

This digitisation step SI 10 can be performed in a number of ways. If the physical tooth samples 10 are provided from unextracted patient teeth, then the process of scanning the teeth using, for example, a CT scanner, will result in a plurality of digital CT scans. Alternatively, extracted or archaeological teeth may be scanned ex vivo, which may provide a more accurate representation of the structure of the physical tooth sample 10.

Optionally, a scanning technique may be used which scans more than merely the external surface of each physical tooth sample 10, and permits the internal structure of each physical tooth sample 10 to be digitised as well. This will yield the enamel-dentine boundary 18, shown in Figure 3; the shape of the physical tooth sample 10 inclusive of the enamel 24, shown in white, can then be subsequently modelled, in addition to the dentine core 26, shown in grey. Airspace in the physical tooth sample 10 is shown in black. Computationally, this may be achieved by associating a binary operator with the components of the relevant model. Such a scanning technique may utilise X-ray slice imaging to generate a series, known as an image stack 28, of transverse slices through the physical tooth sample 10, from which the three- dimensional structure of the physical tooth sample 10 may be digitally generated in a process known as reconstruction.

The image stack 28 may be reformed into the three-dimensional structure, either as a complete digital tooth sample 16, or via a process known as segmentation, step SI 20, into a digital tooth sample 16 having a digital dentine portion 30 and a digital enamel portion 32.

Regardless of the form of the digital tooth sample 16, the data will be formed in step SI 30 into a digital tooth mesh 34, or a plurality thereof if segmentation in step S120 has occurred, which is a description of the three-dimensional digital tooth sample 16 using a connected point cloud to form nodes 36 and elements 38.

A digital tooth mesh 34 may be comprised from a predetermined number of nodes 36, the number defining the coarseness of the mesh, and therefore the accuracy of the digital tooth mesh 34 at representing the digital tooth sample 16. The node-to-node separation will be greater in regions of minimal geometry of the digital tooth sample 16, with the nodes 36 being relatively close-packed in regions of increased geometric variation, for instance, in areas of curvature of the digital tooth sample 16. Evidently, a greater number of nodes 36 will increase the accuracy of the digital tooth mesh 34, but will also increase the computational requirement for generation and analysis of the digital tooth mesh 34. Typical mesh densities might be in the range of 0.01 to 1 mm maximum node-to-node distances, and more preferably in the range of 0.1 to 0.5 mm maximum node-to-node distances; though any mesh density could feasibly be provided for. The mesh generation in step SI 30 is applied to every digital tooth sample 16 to produce a database 40 of digital tooth meshes 34. A baseline model 42 is also created in step S140 which is nominally representative of an average tooth. This baseline model 42 is also a mesh.

To determine the geometry of the baseline model 42, the user could enter some nominal parameters to define the shape of the baseline model 42, with the said parameters likely being based on historical data or prior knowledge on the user's part. However, it will be apparent that parameters could be generated based on some initial processing of the data at hand, from any of the physical tooth samples 10, the digital tooth samples 16, or the database 40 of digital tooth meshes 34. Alternatively, for simplicity, the baseline model 42 could be chosen to be one of the raw digital tooth samples 16, either by the user or automatically. If required, the chosen baseline model 42 may then be re-meshed in order to apply a desired mesh density.

Once the baseline model 42 has been generated, it may be transformed to each digital tooth mesh 34 in the database 40, which is known as mapping or registration at step SI 50. The registration process in step SI 50 matches the nodes 36 of the baseline model 42 to the nearest point on the surface of each digital tooth mesh to thereby determine an overall deviation of the nodes of the baseline model 42 to each digital tooth mesh 34.

To perform the transformation in step SI 50, one digital tooth mesh 34 is selected from the database 40 and designated as the target geometry. This target geometry is then spatially aligned to the baseline model 42. This may, for example, be performed, using an iterative closest point algorithm. The transformation process is conducted using a registration algorithm that can be represented mathematically as follows:

The baseline geometry (5 _X ) = {{xw yii' ^z }, {A _c }}, where 1 < i≤ N _lt 1 < c < T _t The target geometry (½) = {x _2j , y _2j , ¾,}, , where 1≤ j≤ N ₂ , l≤ d≤ T ₂

Where (Χϋ _, Υϋ,Ζΐί) is vertex i and A _c is triangle patch c for the baseline model 42 geometry and (x ₂ j, y ₂ j, z ₂ j) is vertex j and A _d is triangle patch d for the target geometry 34. N _x and N ₂ are the number of vertices and T _x and T ₂ are the number of triangles of S _x and S ₂ respectively.

A K-dimensional tree, KDT^ is constructed between the nodes of the baseline model 42 geometry and the centroids of the target geometry 34 surface triangulations.

For each vertex of the baseline model 42 geometry 5 _X the nearest m surface triangles are found using KDT^ where m is a constant dependent on mesh density, typically 50.

The intersection point G between a perpendicular line drawn from the plane defined by the target triangle patch A _d and the interrogated vertex is used to define a distance measure, y, of the closest point on the target surface A _d to

If G lies inside A _d (case 1) then y(i, d) = \ G— P ₁ (i) | or if G lies outside A _d (case 2) then y(i, d) = \ G— | + \ G — P\ where P is the closest triangle vertex to G . Only the nearest m surface triangles are interrogated for P^i) for efficiency.

The minimum y(i, d) is used to establish the closed location on A _d to From this information the displacement vectors D^i. d = G— for case 1 or D^i. d = P— for case 2 are calculated.

The quality of the resulting displaced baseline model 42 mesh W ^k is improved by iterating the above process by n number of iterations, where k = 0,1,2, ... , n. The displacement vectors are calculated according to D ₁ (i, d) _1+k = D ₁ (i, d) _k - (n— k) ^'1 . After each iterative step the resulting shape is smoothed using Laplacian smoothing and the above-described steps may be repeated.

The registration process in step S I 50 results in a normalised digital sample database 44 of normalised digital tooth mesh data.

It will be apparent to the skilled person that such elastic mesh morphing represents only a single means of arriving at the normalised digital sample database 44, and other morphing or mapping algorithms are available.

Once a normalised digital sample database 44 has been generated in step S I 50, it is possible to analyse in step S I 60 the normalised digital sample database 44 completely. The preferred method of statistical analysis of the normalised digital sample database 44 is by using principal component analysis (hereinafter referred to as 'PCA'). PCA enables the statistical deconstruction into an output dataset 46 of the normalised digital tooth mesh data, from which can be derived a plurality of independent modes 48 of shape variation, such as those as illustrated in Figure 5a, 5b and 5c as independent modes 48a, 48b, 48c. PCA can be conducted in accordance with the following algorithms:

Each normalised digital tooth mesh is re-defined by x _t xi— [ ^χ ΐί' ΐί' ^ζ ι, l' ... , x _ni , y _ni , z _ni ] ^T , l≤ i≤ N

Or, if material information is included from the segmentation in step S 120, xi ⁼ i ^x ii> yii> ^z ii> En — · ^χ ηί· Ύηί· 1 < i < N , where F is a numerical identifier relating to a material property, for example: density; modulus; greyscale; Hounsfield value; or similar, where N is the number of geometries and n is the number of nodes in each mesh.

The mean shape is defined by the geometric database as

The correlations are established by the covariance matrix

PCA of the covariance matrix gives a set of i = N— 1 eigenvalues Λ _έ and eigenvectors e _t . The variation from the average shape by uncorrelated components is described as Where ω _; are the weights associated to the eigenvectors e _t . Selection of individual eigenvectors and manipulation of weightings allows the creation of a generic tooth geometry based on the desired characteristics. The resultant eigenvectors e _t are then representative of the modes. The modes 48 which result from the PCA are independent and orthogonal, and this is critical to the present method 1000. The modes 48 are non-interacting; varying the defining parameter of one mode will not affect the other modes 48 determined by PCA. This may result in the modes 48 being similar yet different for each normalised digital sample database. The number of modes 48 which is required in order to capture the majority of the shape variation of each digital tooth sample 16 is generally between around three and ten, but any number of modes 48 could feasibly be determined from the analysis.

Examples of modes of shape variation 48 are shown in Figures 5a, 5b and 5c, each showing three indicative examples of normalised digital tooth samples. In Figure 5a, each of the three normalised digital tooth samples 16a', 16a", 16a'" vary primarily in vertical scale, and this would typically be the primary mode determined from the PCA. Figure 5b shows variation of the normalised digital tooth samples 16b', 16b", 16b'" according to a combination of the horizontal scale and the root curvature. Finally, Figure 5c shows the variation of the normalised digital tooth samples 16c', 16c", 16c'" according to an inclination of the lowermost portion of the root of each tooth. The modes shown in Figures 5b and 5c might be typical secondary independent modes of shape variation 48, which might inform the and/or range of generic implant geometries to a lesser degree, but potentially seen as being a refinement or increasing refinement.

This independence between modes 48 allows each mode 48 to be parameterised in step SI 70 and plotted, as shown in Figure 2, and formed into a characteristic output dataset 50, which shows the variation of the mode across the entire range of normalised digital tooth mesh data. The modes 48 in question could be anything related to the geometry or material distribution of the digital tooth samples 16, but the primary modes 48 might be readily associated, for instance: tooth scale; root length; dentine width; crown angular inclination; and so on.

Not only does the PCA result in the parametrisation in step SI 70 of the modes 48, the output dataset 46 also illustrates the relative importance of the modes 48 to the overall geometry of the digital tooth sample 16. This advantageously enables the modes 48 to be ranked in terms of importance, which can help to instruct the final generic biomedical geometry.

The person performing the PCA can optionally validate in step SI 80 the results of the parametrisation in step S I 70 to ensure that the characteristic output dataset 50 captures the tooth shape variation as accurately as possible. Possible methods of validating the characteristic output dataset 50 include variance analysis, eigenmode analysis and deviation capturing. Additionally or alternatively, the PCA can be re-run for the entire range of normalised digital tooth mesh data, but with a single digital tooth mesh 34 being left out each time. This allows for the identification of any anomalous physical tooth samples 10 which may be otherwise skewing the results.

Based on the characteristic output dataset 50 it is possible to extract characteristic data about a nominally average and therefore generic tooth which is based on the characteristics of the population or analysis group. This characteristic data can then be used to determine, in step S I 90, generic tooth geometry 52.

However, as a single generic tooth geometry 52 has now been generated in step S I 90, we can use it to create a range of generic tooth geometries which may be usable by the majority of patients across a population. This can be achieved by creating a plurality or range of generic tooth geometries which vary according to at least one of the modes 48 as determined during the analysis in step S I 60.

To provide some context, it is possible that, following analysis in step S I 60, the most important modes 48 are root length and root width for an incisor, for example. The generic tooth geometry 52 has a root length of x cm and a root width of y cm. If we vary the modes associated with root length and root width each by one standard deviation, that is, a root length of χ±σ _χ cm and a root width of y±o _y cm, then create a total of nine generic tooth geometries by crossing the two ranges for root size, a large proportion of the general patient population will find one of the geometries to be suitable.

The generic tooth geometries 52 can now be parameterised geometrically to provide input data for manufacturing. This would typically involve the generation of a non-uniform rational basis spline (NURBS) geometry which can be supplied for use in CAD-CAM manufacturing. This allows for the rapid manufacture of generic tooth implants 54 which will be usable by a large proportion of the population, such as that shown in Figure 6.

A typical generic tooth implant 54 might comprise a root portion 54a, and an abutment 54b extending upwardly from the root portion 54a, to which a crown portion may be affixed. Such a two-part implant 54 allows for different generic root portions 54a and crown portions to be used for a single patient, to best fit their oral geometry. However, a unitary or multi-part implant is equally feasible. A greater proportion of the patient population may be serviced by the provision of generic tooth implants 54 which comply with, for example, two standard deviations from a particular mode 48. Accommodation may also be made for unusual modal shapes; the prior description assumes a Gaussian modal distribution, but bimodal distributions are possible, for instance, in the distribution of geometries between male and female teeth.

The above-described method 1000 specifically relates to the creation of generic tooth implants 54, and it will be appreciated that this may not be specifically limited to human teeth. Indeed, the present method could feasibly be used in a veterinary context. Additionally, and as stated throughout, whilst the creation of dental implants has driven the present invention, all or some further biomedical implants may be subject to the same geometric variations throughout a patient population, and the invention could readily be applied to other biomedical fields where generic patient implants might be useful. This could apply to both internal prosthetics, for instance, in the replacement of bone, or external prosthetics, such as for facial prosthetics.

The characteristic data of the generic tooth geometries 52 derived from the above-described method 1000 can also be utilised for purposes other than the manufacture of dental implants 54.

One ancillary use of the generic tooth geometries 52 is in the construction of the associated surgical tools which are required during the implant surgery. For example, a cutting tool 56 can be created, such as that shown in Figure 7, which has a head 58 which is shaped in accordance with the root portion of a generic dental implant 54. This can be formed using the characteristic data of the generic tooth geometries 52.

The cutting tool 56 is, in this embodiment, an ultrasonic cutter, which can safely cut osseous tissue in the jaw without causing damage to the gum. This allows the dental surgeon to shape the tooth-accepting portion of the mandible to the exact shape of the generic dental implant 54, ensuring an optimal fit. The method 1000 of determining generic biomedical geometries is primarily applicable to surgical uses. However, it is also entirely possible to utilise the method to determine a probable geometry based upon partial information.

Considering the physical tooth samples 10 as previously described, if historical archaeological teeth are used, it is entirely possible that part of the physical tooth sample has degraded or been ground down over time, particularly the crown. This can lead to misidentification of the tooth, which can hamper academic studies of the physical tooth sample 10. However, based on the characteristic data of the generic tooth geometries 52, which may, for example, have been determined for a plurality of different teeth types. If the root 12 of the physical tooth sample 10 is well preserved, then an optimisation of the root shape from the statistical model to the preserved root 12 can give the user a probabilistic indication of what type of tooth the complete physical tooth sample 10 is likely to be. Such reconstruction may also have use in a forensic medical context.

Where the term generic is used above to refer to a biomedical implant, in particular a dental implant, which can be used as a substitute for a missing tooth in a patient, it will be appreciated that a generic geometry is not necessarily identical to the expected average geometry. In order to improve a fit of a dental implant into a patient's mouth, strategic modifications from the anatomic shape may be provided, in order to improve the ease or acceptance of the implant by the patient. For instance, for a tooth, an increase in the mesiodistal direction of up to 20%, preferably between 2% and 10% may be useful, and additionally or alternatively, a reduction in the buccal-lingual direction of up to 20%, preferably in the range of 2% to 10% may also be of assistance. Furthermore, an undersizing or segmentation in the superior-inferior aspect may also be useful, preferably by up to 20%. This alteration will clearly be applicable for the dental cutting tool, and can be applied to other biomedical implants and/or cutting tools as well.

It is therefore possible to provide a method of constructing generic biomedical implant geometry, and more particularly but not necessarily exclusively tooth geometry from a collection of associated biomedical samples, in this case tooth samples. By creating a database of digitised biomedical samples, which in the embodiments described above are tooth samples, which can be transformed to a baseline model to normalise for analysis yields a plurality of independent modes of shape variation which can be varied in order to arrive at generic implant, in this case tooth, geometry. The determination of such generic tooth geometry can be used to inform the manufacture of both dental implants and the associated tools, such dental implants being superior to standard conical implants which are currently available, as they will be of suitable dimensions so as to fit a statistically large proportion of the general population.

The term dental implant is used in context in the described embodiments to refer to an implanted and direct replacement for a missing tooth within a patient, with the dental implant having the visual appearance which is close to, similar or generally resembling that of the tooth which has been lost. However, it will be appreciated that a wide range of dental implants could be produced using the present invention, including, but not necessarily limited to: dentures; one-piece, two- piece, three-piece or multi-piece dental implants; restorative crowns; dental bridges; root canal; or veneers. The invention could also be used more generally and be applied to surgical planning prior to the insertion of a dental implant into a patient's oral cavity.

In the field of other biomedical implants, the generation of a range of generic implants could be used across a range of contexts where there is a statistical variation of sizes and or shapes across the population. Examples of such biomedical implants include: joint prostheses, such as primary implants, revision implants and/or megaprostheses, for instance, in the case of tumour-related bone loss; trauma devices, such as intermedullary nails or fracture plates; endoprosthetic components, such as above-knee amputation socket ischial brims; and cochlear implants. The biomedical implants therefore may not be strictly applicable solely for the replacement of existing tissue of a patient, but for any context in which the variation in size of the implant will be statistically varied across the population in a standardisable manner.

The words 'comprises/comprising' and the words 'having/including' when used herein with reference to the present invention are used to specify the presence of stated features, integers, steps or components, but do not preclude the presence or addition of one or more other features, integers, steps, components or groups thereof.

It is appreciated that certain features of the invention, which are, for clarity, described in the context of separate embodiments, may also be provided in combination in a single embodiment. Conversely, various features of the invention which are, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable sub-combination.

The embodiments described above are provided by way of examples only, and various other modifications will be apparent to persons skilled in the field without departing from the scope of the invention herein described and defined.

Claims

1. A method of constructing a range of generic dental implants from a collection of tooth samples, the method comprising the steps of: a] providing a database (40) of digitised like tooth samples (16) from an analysis group; b] creating a baseline model (42) from the digitised like tooth samples (16); c] transforming the baseline model (42) to the digitised like tooth samples database (40) to generate a normalised digitised like tooth sample database (44); d] applying principal component analysis to the normalised digitised like tooth sample database (44); e] extracting characteristic data relevant to one or more characteristics of the analysis group; f] determining from the characteristic data a range of generic tooth geometries (52); and g] constructing a range of generic dental implants (54) in accordance with the range of generic tooth geometries (52).

2. A method as claimed in claim 1, wherein during step a], the database of digitised like tooth samples (16) is obtained from scans of physical tooth samples (10).

3. A method as claimed in claim 2, wherein the scans of physical tooth samples (10) are obtained via three-dimensional slice tomography.

4. A method as claimed in claim 2 or claim 3, wherein the digitised like tooth samples (16) are segmented based on a density profile of each physical tooth sample (10).

5. A method as claimed in any one of the preceding claims, wherein the digitised like tooth samples (16) are represented as three-dimensional meshes.

6. A method as claimed in claim 5, wherein the baseline model (42) is a standard mesh, the standard mesh being mapped onto the three-dimensional meshes using a node-to-node comparison.

7. A method as claimed in claim 6, wherein the node-to-node comparison is performed using elastic mesh morphing.