METHOD OF ASSESSING THE FIT OF DENTAL RESTORATION

This invention relates to a method of assessing the fit between two components, and in particular to a method of assessing the accuracy of the fit of dental restorations.

Dental prostheses are used to restore the function and appearance of teeth. Traditionally, they are individually made to fit a particular tooth or teeth. Often, an impression of a patient's mouth is taken and from this a model is produced.

One way to produce the restoration is to create a wax-up over a relevant portion of the model and to produce a mould from this. For a ceramic restoration, powder is pressed into the mould using the model thus producing a green state shell having an inner form which replicates that of the model and an outer shape based on the wax-up. The shell is subsequently sintered and coated with enamel to produce the final restoration.

Alternatively, the patient's mouth or the model are scanned and the inner form of a restoration is digitised. A wax model may again be used to produce an outer shape which can be scanned and machined or press moulded to produce the outer form of the restoration. The inner surface is then machined to produce the mating surface of the restoration.

However a restoration is manufactured, a gap is required between the restoration and the tooth or tooth part that the restoration sits on to accommodate cement that is used to fix the restoration in place.

One problem which is encountered for all (both ceramic and metallic) restorations is that errors in the reproduction process at each stage in that process can result in the restoration being a bad fit. If the restoration is too small then it may be remedied if the problem region can be identified as excess material can often be

ground away. If the restoration is too big, then it may not fit the patient at all necessitating the production of a new restoration. Both of these options are time consuming and costly and, depending on where the errors occurred in the process, may only be identifiable when the dentist tries to fit the restoration in the patient's mouth.

Even if the restoration appears to be a reasonable fit, there can be problems. It can be important that the fit of a restoration at the margin line is good all the way round the margin line for a number of reasons. The margin line is where the restoration and tooth join. Inaccuracies here can cause gum problems or ingress of matter between the restoration and remaining tooth part which can lead to tooth decay. Thus, a poor fit at the margin line can result in further dental work being required.

One way to assess the accuracy of a restoration is to apply a layer of silicon between the restoration and the model or tooth. The thickness of the silicon layer giving a measure of the gap between the restoration and model or tooth. Obviously, it is difficult to perform this test on the patient but, if the model is used then only errors after that step on the production process are identified. Additionally, although this method can identify touching points between the surfaces, a quantitative thickness of the gap cannot be obtained.

According to a first aspect of the invention, there is provided a method of assessing the fit between two components, the method comprising in any suitable order the steps of: i) obtaining a digitised representation of a surface of a first component; ii) obtaining a digitised representation of a surface of a second component which is for co-operation with the surface of the first component; and iii) determining a best fit position by comparing the digitised representation of the surface of the first component with the surface of the second component in a plurality of different relative positions taken in a first degree of freedom; in which the step of determining the best fit position is subject to the condition that there is

no interference between the surface of the first component and the surface of the second component.

Subjecting step iii) to the condition that there is no interference between the surface of the first component and the surface of the second component prevents the method from determining a physically impossible position as being the best fit position, e.g. one in which the first and second components would have to intersect each other. Accordingly, the invention ensures that a mechanical constraint is applied thus the inner form cannot pass through the co-operating surface. The method of the invention therefore provides a reliable way of determining the best fit position between two components.

Preferably the digitised representations of the surfaces of the first and second components are data obtained from scanning the surfaces. The steps of obtaining the digitised representations of the surfaces of the first and second components can comprise scanning the surfaces. Preferably, the surfaces are scanned using a contact scanning system.

Optionally, the step of scanning the surfaces can be performed prior to execution of the method of the invention. In this case, obtaining the digitised representations can comprise retrieving the digitised representations from a stored location. For instance, obtaining the digitised representations can comprise retrieving the digitised representations from a computer memory device.

What is considered to be the best fit position can depend on a number of factors, including the reason for finding the best fit position and the purpose for which the first and second components will be used. For instance, the method of the present invention could be used to determine how best to orientate the actual physical first and second components in order to achieve desired physical properties. For instance, it might be important to find the relative position between the first and second components in which the gap between them does not fall above or below

predetermined threshold levels. Optionally, the method of the present invention could be used as part of a method for determining how the actual physical components will sit together when they are brought together.

Accordingly, the best fit position could be the relative position between the first and second components at which they fit closest to each other. For example, the best fit position could be the relative position which provides the smallest mean offset between the surfaces of the first and second components. Optionally, the best fit position is the relative position which provides the smallest maximum offset between the surfaces of the first and second component. Optionally, the best fit position is the relative position at which the centre of masses of the surfaces of the first and second components are closest to each other. As will be understood, the selection of the criteria to determine the best fit position will depend on the subsequent use of the first and second components.

Comparing the digitised representation of the surface of the first component with the surface of the second component can comprise determining the alignment error between the surfaces. Determining the alignment error can comprise determining the difference between the maximum and rninimum offset distances, taken in a first dimension, between the surfaces when the surfaces substantially directly face each other. The best fit position can be considered to be the relative position between the surfaces of the first and second components which provides the smallest alignment error.

Preferably, the offsets between the surfaces of the first and second components are all measured in the same direction. This can simplify the determination of the best fit position. More preferably the offsets are all taken substantially parallel to the direction in which the surfaces of the first and second components must be moved to bring the surfaces together when the surfaces substantially directly face each other. Determining the best fit position based on the alignment error calculated using the offset taken in this direction has been found to be advantageous as it has

been found to provide an efficient way of accurately and reliably determining how the surfaces of the first and second components will co-operate with each other when the actual components are brought together.

Depending on how the surfaces are digitised, the method can preferably comprise co-aligning the digitised representation of the surfaces of the first and second components. Co-aligning the digitised representations can simplify the comparison of the surfaces of the first and second components.

Co-aligning the digitised representations can comprise orienting the digitised representations of the surfaces of the first and second components such that surface of the first component directly faces the surface of the second component. This could comprise changing the orientation of the surface of one of the first and second components such that they have substantially the same orientation. For instance, when the first and second components are substantially concave and convex, co-aligning the digitised representations preferably comprises orienting the digitised representations such that the concave surface faces the convex surface.

Co-aligning the digitised representations can comprise aligning the centre of mass of the digitised representation of the surfaces of the first components with the centre of mass of the digitised representation of the second component. This helps to bring the digitised representations into the same approximate space. In this case, co-aligning the digitised representations preferably comprises displacing one of the digitised representations of the surfaces relative to another such that there is no interference between the two data sets. Preferably, the surfaces of the first and second components are displaced translationally relative to each other. Preferably, the surfaces of the first and second components are displaced relative to each other such that surface of the first component substantially directly faces the surface of the second component.

Preferably, the first degree of freedom is a rotational degree of freedom about a first axis. Preferably, the first axis extends through the centre of mass of at least one the surfaces of the first and second components when the surfaces of the first component directly faces the surface of the second component.. More preferably, the first axis extends substantially parallel to the direction in which the surfaces of the first and second components must be moved to bring the surfaces together when the surfaces substantially directly face each other. In this case, preferably step iii) comprises rotating the digitised representation of the surface of one of the first and second components relative to the surface of the other component by a predetermined angle between each comparison. Preferably, the predetermined angle is not less than 0.5 degrees, more preferably not less than 1 degree, especially preferably not less than 5 degrees. Preferably, the predetermined angle is not more than 30 degrees, more preferably not more than 25 degrees.

Preferably, step iii) comprises: for each different relative position taken in the first degree of freedom, comparing the digitised representation of the surface of the first component with the surface of the second component in a plurality of different relative positions taken in at least a second degree of freedom. This is advantageous as it enables a more accurate determination of the best fit position to be made.

Preferably, the second degree of freedom is a first linear degree of freedom. Accordingly, when the first degree of freedom is a rotational degree of freedom about a first axis, then a plurality of comparisons can be made for a plurality of different translational positions for each rotational position. Preferably, the first linear degree of freedom extends substantially orthogonally to the first axis.

Step iii) can further comprise: for each different relative position taken in the first degree of freedom, comparing the digitised representation of the surface of the first component with the surface of the second component in a plurality of different relative positions taken in at least a third degree of freedom. Preferably,

the third degree of freedom is a second linear degree of freedom. Preferably, the second linear degree of freedom extends substantially orthogonally to the first axis and to the first linear degree of freedom. Accordingly, for each rotational position, a plurality of comparisons can be made for different relative positions between the surfaces of the first and second component taken in two substantially orthogonal dimensions.

Preferably, the method further comprises: iv) fine-tuning the best fit position. Preferably step iv) comprises, starting from the best fit position determined in step iii), comparing the digitised representation of the surface of the first component and the surface of the second component in a plurality of relative positions taken in at least one degree of freedom. Accordingly, once an initial best fit position has been found between the two components, then a more precise best fit position can be found. This is advantageous as it enables a more accurate determination of the best fit position to be made. Furthermore, the determination of an initial best fit position and then fine-tuning the best fit position has been found to provide a particularly efficient way of determining an accurate and reliable best fit position. Preferably, step iv) is also subject to the condition that there is no interference between the surface of the first component and the surface of the second component.

The preferred number of degrees of freedom in which the surfaces of the first and second components can be moved relative to each other in order to fine-tune the best fit position will depend on a number of factors, such as the accuracy to which the fit needs to be made. Preferably, the first and second surfaces can be moved relative to each other in at least two degrees of freedom, more preferably at least three degrees of freedom, especially preferably at least four degrees of freedom, for example in all six degrees of freedom. Preferably, the surfaces of the first and second components can be moved relative to each other in at least one rotational degree of freedom. Preferably, the at least one rotational degree of freedom is taken about at least a second axis that extends orthogonal to the first axis.

The method could provide an output based on the best fit position determined in either or both of step iii) or iv). Preferably, the method further comprises: v) evaluating the quality of the fit between the surface of the first component and the surface of the second component at the best fit position determined in step iii) or iv). Preferably, the method further comprises v) providing an output based on the evaluation. Evaluating the quality of the fit can comprise performing an analysis of the gap between the surface of the first component and the surface of the second component. The gap between the surface of the first and second component can be the distance between them taken perpendicularly to one of the surfaces of the first and second components.

The analysis of the gap can comprise determining the mean gap between the surface of the first component and the surface of the second component. Optionally, the analysis can comprise determining the standard deviation of the gap between the surface of the first component and the surface of the second component. The analysis can comprise determining the maximum gap between the surface of the first component and the surface of the second component. The analysis can comprise determining the minimum gap between the surface of the first component and the surface of the second component.

The evaluation could comprise comparing the result of the analysis with a predetermined threshold level. In this case, the output could be a pass or fail signal.

At least part of the gap between the surfaces of the first and second components could be analysed. Optionally, the entire gap between the surfaces of the first and second components could be analysed. Preferably, step v) comprises: evaluating the quality of fit between the surface of the first component and the surface of the second component around the margin line region of the digitised representation of the surface of the first component. The margin line region can be a band

extending annularly around at least one of the surfaces of the first or second components, at its perimetral edge. For instance, if the surface of the first component is concave in shape, then the margin line region can be a band extending annularly around the surface at its open end. The preferred width of the band can depend on a number of factors, such as the purpose for which the first and second components will be used and the size of the surfaces of the first and second components.

As the fit at the margin line region can be paramount, it can be preferred that a tolerance is applied to this region. Optionally, the method further comprises determining if the gap between the surface of the first component and the surface at any point around the margin line region does not comply with a predetermined threshold tolerance. Preferably, the method further comprises determining if the mean gap between the surfaces of the first and second components at the margin line region is greater than a predetermined threshold. Optionally, the method further comprises determining if the mean gap between the surfaces of the first and second components at the margin line region is less than a predetermined threshold. In this case, the output could be a pass or fail signal based on the comparison of the properties of the gap at the margin line region with a predetermined threshold.

The output could be the result of the evaluation step. For instance, if the evaluation comprises determining the mean gap, then the output could be the value of the mean gap. Optionally, the output could be a graph illustrating the result of the evaluation step. For instance, the graph could be a plot of the size of the gap for a plurality of positions between the surface of the first component and the surface of the second component. Optionally, the output could be an image illustrating the result of the evaluation step. For instance, the output could be an image of the gap determined in the evaluation step. For instance, the output could be a 3-dimensional image of the gap determined in the evaluation step. The image could illustrate the shape of the gap between the surface of the first component

and the surface of the second component. Optionally, the image could illustrate the size of the gap between the surface of the first component and the surface of the second component. The 3 -dimensional image could be configured to use different colours to illustrate the different sizes of the gap at different points between the surfaces. The 3 -dimensional image could be configured to use different grades of shading of a single colour to illustrate the different sizes of the gap at different points between the surfaces.

The output could comprise a combination of different types of outputs. For instance, the output could comprise the result of the evaluation step as well as the result of a comparison of the properties of the gap with a predetermined threshold.

At least one of the first and second components could be a medical prosthesis. Preferably, one of the first and second components is a dental restoration. The other of the first and second components can be a tooth on which the dental restoration is to be fitted. Optionally the other of the first and second components can be a model of a tooth on which the dental restoration is to be fitted. Accordingly, the method of the invention can be used to assess the fit between the dental restoration and the tooth on which the dental restoration is to be fitted. In particular, the invention provides a repeatable and reliable method for assessing the fit of a restoration for example a coping, crown or framework. In preferred embodiments, the method pays particular attention to the margin line.

Accordingly, there is described herein a method of assessing the accuracy of a dental restoration comprising in any suitable order the steps of: digitising an inner form of the dental restoration; digitising a co-operating surface for the dental restoration; comparing the digitised inner form of the dental restoration with the digitised co-operating surface; and producing an output based on that comparison.

The digitised representation of the tooth could be obtained via scanning the tooth.

Optionally, the digitised representation of the tooth could be obtained via scanning

a model of the tooth.

The comparison can include fitting and aligning the inner form and co-operating surface.

The comparison step is preferably carried out by a computer program.

According to a second aspect of the invention there is provided, computer program code, comprising instructions which, when executed by a computer, causes the computer to perform the method of the invention.

According to a third aspect of the invention, there is provided a computer readable medium, bearing computer program code according to the second aspect of the invention.

The invention will now be described by way of example, with reference to the accompanying drawings, of which:

Fig Ia shows a cross-section through a restoration;

Fig Ib shows an enlargement of region A of Fig Ia;

Fig 2 is a flow diagram detailing steps of the invention;

Fig 3 shows diagrammatically the steps involved in a comparison;

Figs 4a and 4b show schematically outputs for the fit of a restoration; and

Fig 4c shows schematically a cross-section through a restoration and tooth/model.

Fig 1 shows a restoration 10, in this case a coping, and a corresponding tooth or model of a tooth 12 which the restoration 10 is designed to cover. A gap 14 is provided between the restoration 10 and tooth 12 which will be filled with cement when the restoration is secured to the tooth.

The size of the cement gap 14 is not critical except for in the region of the margin

line 20. At all other areas, there is a minimum desired cement gap to ensure that the cement cures properly and a maximum desired cement gap to ensure that the restoration fits correctly with opposing teeth in the mouth but, anywhere within this range is acceptable. However, at the margin line 20, it is important to have tighter control on the cement gap. One reason is that the gap around the margin line 20 is where excess cement exits the gap. Thus, too small a gap may hinder this exit causing the restoration to sit proud of the tooth. Another reason is that cement will erode from a large gap over time potentially allowing debris to enter the gap which could compromise the integrity of the tooth requiring further remedial work at a later date. If the erosion is severe, the restoration may come off the tooth, also necessitating further dental work.

It is preferred that a tolerance is applied to a region of the margin line 20 which limits the acceptable mean gap size. For example a lmm wide band 22 is selected around the circumference at the margin line and a maximum mean gap size tolerance of lOOμm is applied to this region, as described in more detail below.

Fig 2 is a flow diagram detailing steps of the invention. The inner form or surface of a dental restoration, for example a coping, crown, or bridge, is digitised 100. A co-operating surface for the dental restoration is also digitised 110. The inner form and the co-operating surface are compared 120 and an output 130 produced based on the comparison.

One way to digitise a surface is to scan the surface using either a contact or a non- contact scanning system which maps the surface profile or obtains sufficient data points of the surface to enable construction of a triangulated surface model thereof. A suitable scanner is described in International Patent Application No. WO03/46412.

The co-operating surface is either the surface of a model of the tooth or tooth part which will be covered by the restoration or, it can be the actual tooth or tooth part

itself. A model is typically produced directly from an impression of the tooth or teeth being restored although it could be produced by, for example, rapid prototyping using data from an intra-oral scan. Data which is obtained directly from the tooth or tooth part is generally obtained by taking an intra-oral scan. Other imaging techniques such as x-ray, MRI, CAT scanning currently do not have the capabilities to provide suitable data but developments in these areas may change this. Once the data is digitised, it can be stored in a machine controller, interface or computer until the comparison is made.

The digitising of the two matching surfaces can be carried out in either order.

Referring now to Figs 2 and 3, the surface maps or triangulated surface models 221 are then compared. Firstly one of the data sets is flipped 122 so that both data sets are either concave 222a or convex 222b in orientation. Next, a translation step 124 occurs whereby the restoration is moved with respect to that of the cooperating surface to align the centre of mass of one to the other. In addition, the two data sets are vertically displaced along the Z-axis such that there is no interference between the two data sets and so that the surfaces directly face each other. This satisfies the mechanical constraints that occur in real life i.e. that the restoration must sit over the co-operating surface and cannot pass through that surface.

An initial best fit position is then determined at step 126. In the described embodiment, the initial best fit position is the relative rotational position about the Z-axis and the relative translational position along the X and Y axes at which there is the smallest alignment error between the digitised surfaces of the restoration 10 and tooth 12. The initial best fit position is determined by performing the following steps: for each of a plurality of relative rotational positions taken about the Z-axis, the data sets are shifted translationally relative to each other along the X and Y axes until the smallest alignment error is found. In a preferred embodiment, the rotation is carried out incrementally, say by 10°

W

14 intervals. However, as will be understood, the preferred angle between each rotational position can vary on a number of factors, such as the accuracy of fit required. For instance, in another preferred embodiment, the relative rotation can be carried out through 48 increments or substantially equal size.

The lowest alignment error for each relative rotational position is stored and, after the lowest alignment error has been determined for each of the relative rotational positions, the relative position with the lowest alignment error is chosen as the initial best fit position.

10

There are many suitable ways of calculating alignment error for each relative position of the surfaces the data sets represent. In the particular embodiment described, the alignment error for each relative position is determined by finding the maximum vertical (i.e. Z axis) offset distance between the two surfaces, and

15 then subtracting the minimum vertical offset distance between the two surfaces. Example offset distances used in the calculation of the alignment error in the embodiment described are illustrated by lines 226 and 228 in Figure 3. The calculated difference provides an indication of the actual vertical offset between the two surfaces when the coping 10 and model 12 are brought together. It has

20 been found that finding the relative position which provides the smallest vertical offset between the two surfaces is particularly advantageous as it has been found that such a position provides the best fit between the surfaces at the margin line region 22, described in more detail below.

25 It is preferred that the relative rotation, either incremental or not, is carried out about the Z, or vertical, axis.

Once the initial best position has been determined, the final best position is then determined at step 127. In the described embodiment this is done by performing a

30 multi-dimensional downhill simplex (MDDS) minimisation process 127 between the two sets of data, for the initial best position determined at step 126. This

involves determining, starting from, the determined initial best fit position, the relative position, taken in three rotational degrees of freedom about the X, Y and Z axes, and three translation degrees of freedom along the X, Y and Z axes, at which there is the smallest alignment error.

Once the final best fit position has been determined, the quality of the fit between the coping 10 and model 12 is evaluated at step 128. The evaluation can be based on one, or a plurality of different statistical analyses. For instance, the evaluation can be based on a number of parameters or criteria such as: a minimum gap between the two data sets; a maximum gap between the two data sets; a range within which the mean gap size over the margin line region must fall . These can be standard values or can be adjusted for different types of restoration or based on knowledge concerning the dentist who will fit the restoration.

In one preferred embodiment, the evaluation is performed by determining the gap size between the coping 10 and the model 12 taken normally to the surface of the model 12 (as illustrated by line 224) for a plurality of points between the surfaces of the coping 10 and model 12. Furthermore, the mean gap size is calculated, and evaluated, for the margin line region 22 only. In order that the cement gap fulfils its requirements i.e. is large enough to allow excess cement to exit, the preferred mean gap is not more than lOOμm in size. If it is important to ensure that the gap at the margin line is small enough to minimise the chance of wash out or erosion of the cement, then a minimum preferred mean margin line region gap threshold can also be applied.

There are a number of different ways in which the output 130 can be communicated to, for example, an operator. The simplest for an operator is to output a pass or fail 132 based on the threshold level applied to the margin line region 22. Accordingly, if the mean gap size for the margin line region 22 is above 1 OOμm, then the output can be a fail signal.

Alternatively, or indeed in addition, a graph 134 of the cement gap around the tooth and/or at the margin line or, an image 136 of the gap between the two data sets or, statistics 138 on the difference between the data can be produced. One or more of these outputs can be used or selected.

Referring now to Fig 4, outputs indicating the gap at the margin line, and elsewhere in the matching of the tooth and restoration are produced.

Fig 4a shows a line graph 134 of the gap at the margin line. The measurements to determine the gap size are taken radially 260 at the margin line, between the outer surface of tooth/model 12a and the inner surface or the restoration 10a (see Fig 4c), in this example the measurement is taken every 0.9° of relative rotation.

As shown in Fig 4a the gap at the margin line is zero 252 periodically. Clinically acceptable marginal fits are currently defined in terms of a mean marginal opening. Therefore it is acceptable that areas exist where the gap is non-zero. Further development in the scientific understanding of the effects of marginal opening may result in different specifications, such as a mean or maximum gap over a certain distance or angle or proportion of the circumference or some other statistical calculation.

Referring now to Fig 4b, an image 270 of the cement gap over the entire restoration is shown. The gap size is coded either by colour, shading or as degrees of grey shading (as shown). The gap at the margin line 254 is given as well as the gap elsewhere on the restoration. This type of imaging can be useful if a tailored service is provided to customers. Some dentists like a consistent cement gap size whereas others prefer a larger gap towards the apex allowing more cement to be used. By coding the gap size, and providing this information to the customer, they can clearly see that the gap that has been produced is to their preferred dimensions.

An advantage of providing a detailed output of the cement gap, such as a 3d image as shown, in Fig 4b is that errors in the production process can be identified and sourced. For example, if the restoration fits the model of the tooth or teeth being restored, but does not fit the patient, then the error either lies in the taking of the impression or the model of the impression. This can be easily checked by comparing the model to the impression i.e. placing them together and manipulating them to assess the fit. In a preferred embodiment, the impression shape can also be digitised and accuracy outputs given for comparisons of the impression and model and the impression and restoration.

The comparison and output steps 120, 130 are preferably controlled by computer software or a computer program. This software or program may be located in the computer in which the comparison process is carried out, or a computer, controller or interface associated with the digitising step.