Customised surgical apparatus

1. A method for designing a customised surgical instrument or prosthesis for a body part of a specific patient, the body part having a body part surface, the method comprising the step of:

processing at least one x-ray image of the body part to generate a processed patient image, wherein processing the at least one x-ray image includes filtering the at least one x-ray image, creating a differential image corresponding to differences in intensity of the filtered at least one x-ray image and normalizing peaks in the differential image by applying a normalizer function to the differential image;

instantiating a statistical model generated using a minimum description length approach and having a set of anatomical correspondence points across the statistical shape model using image data derived from the processed patient image to generate a patient specific model of the body part having a modeled surface, the instantiating step comprising:

(i) generating a pseudo x-ray image of the body part from a CT image reconstructed from the statistical model using parameters of a model of the body;

(ii) processing the pseudo x-ray image of the body part to generate a processed pseudo x-ray image, wherein processing the pseudo x-ray image includes filtering the pseudo x-ray image, creating a differential image corresponding to differences in intensity of the filtered pseudo x-ray image, and normalizing peaks in the differential image by applying a sigmoid function to the differential image;

(iii) comparing the processed patient image and processed pseudo x-ray images to determine the differences between the processed patient image and the processed pseudo x-ray image; and

(iv) repeating steps (i) to (iii) using different parameters of the model of the body part to minimize a cost function indicating the differences between the processed patient image and the processed pseudo x-ray image and generate the patient specific model; and

generating a customised surgical instrument or prosthesis having a contact derived from the modeled surface of the patient specific model.

2. The method of claim 1 , wherein the patient specific model has a surface shape that varies by less than approximately 1 to 2 mm from the surface shape of the body part.

3. The method of claim 1 , wherein the patient specific model includes bone and soft tissue.

4. The method of claim 3 , wherein the step of generating the customised surgical instrument or prosthesis is based on patient specific data relating to both bone and soft tissue.

5. The method of claim 1 , wherein the contact surface is a portion of the surgical instrument that is shaped to fit on the body part and/or is a portion of the surgical instrument that is shaped to fit into a space around the body part.

6. The method of claim 1 , wherein demographic data about the patient is supplied to the statistical shape model and wherein the statistical shape model instantiates a model from a sub-population matching the demographic data of the patient.

7. The method of claim 1 , wherein the processing step includes applying a high pass filter to the pseudo x-ray image.

8. The method of claim 1 , wherein the processing step includes generating a differential image.

9. The method of claim 8 , wherein the processing step includes separating the pseudo x-ray image into a positive features image and a negative features image.

10. The method of claim 8 , wherein the processing step includes applying a broadening function to features of the differential image.

11. The method of claim 8 , wherein the processing step includes applying a normalising function to the differential image features.

12. The method of claim 1 , wherein the statistical shape model is a surface model and the correspondence are confined to the surface.

13. The method of claim 1 , wherein the statistical shape model is a volume model and the correspondences are explicit across the entire volume of interest.

14. The method of claim 1 , wherein instantiating the patient specific model includes using a quasi-Newton optimisation method.

15. The method of claim 1 further comprising the step of using patient specific data from the patient specific model in a kinematic model.

16. The method of claim 15 , wherein the kinematic data is also used to generate the design of the customised surgical instrument or prosthesis.

17. The method of claim 1 , wherein the generating step comprises the step of manufacturing the surgical instrument or prosthesis.

18. The method of claim 1 , wherein the generating step comprises the step of manufacturing the surgical instrument or prosthesis having a contact surface that is the negative of the modeled surface.

19. A computer implemented method for generating a patient specific model of a body part, comprising the steps of:

processing an x-ray image of a body part of a patient to produce a processed x-ray image, wherein processing the x-ray image includes filtering the x-ray image, generating a differential image corresponding to differences in intensity of the filtered x-ray image and normalizing peaks in the differential image by applying a normalizer function to peaks in the differential image;

reconstructing a CT scan type image using parameters of a model of the body part from a statistical shape model generated using a minimum description length approach and having anatomical correspondences;

generating a pseudo x-ray image corresponding to the x-ray image of the body part from the CT scan type image;

processing the pseudo x-ray image of the body part to generate a processed pseudo x-ray image, wherein processing the pseudo x-ray image includes filtering the pseudo x-ray image, generating a differential image corresponding to differences in intensity of the filtered pseudo x-ray image and normalizing peaks in the differential image by applying a sigmoid function to the differential image; and

using a quasi-optimisation method to minimize a cost function indicating a residual between the processed x-ray image and processed pseudo x-ray image by varying the parameters of the model of the body part to generate a patient specific model of the body part.