Graphical representation of a dynamic knee score for a knee surgery

Abstract

This disclosure relates to a computer assistant for a surgeon with a graphical representation of a dynamic knee score for a knee surgery. A processor receives computer tomography data of a current patient's knee and user input from the surgeon, the user input comprises an identifier of a knee implant. The processor then retrieves multiple machine learning model parameters indicative of a machine learning performed on historical patient records. For multiple values of a rotation of the tibial component and a slope of the tibial component the processor configures a post-operative kinematic model performs a kinematic simulation and estimates a current patient outcome by applying the machine learning model. Finally, the processor generates a shaded surface spanning the multiple values of a rotation of the tibial component and a slope of the tibial component on a user interface to graphically represent the estimated current patient outcome for each of the rotation and slope.

What is claimed is: 
     
         1 . A method for assisting a surgeon with a graphical representation of a dynamic knee score for a knee surgery, the method comprising:
 retrieving a 3D model of a current patient's knee, the 3D model representing computer tomography data of the current patient's knee;   receiving user input comprising an identifier of a knee implant, the knee implant comprising a component;   retrieving multiple machine learning model parameters indicative of machine learning performed on historical patient records, the historical patient records comprising multiple historical kinematic parameters of each of multiple historical patients as inputs and a reported patient outcome for each historical patient as output, the machine learning model parameters being indicative of a relationship between the multiple historical kinematic parameters and the reported patient outcome;   for each of multiple values of rotation of the component and slope of the component:
 configuring a post-operative kinematic model of the current patient's knee based on the computer tomography data, the user input, and the value of the rotation and the slope, wherein configuring the post-operative kinematic model comprises virtually performing the surgery by introducing cut surfaces to change a shape of bones in the 3D model and adding a shape of the knee implant to determine the post-operative kinematic model; 
 performing a kinematic simulation based on the post-operative kinematic model to determine multiple simulated kinematic parameters; and 
 estimating a current patient outcome by applying the multiple machine learning model parameters to the multiple simulated kinematic parameters of the current patient; and 
   representing the estimated current patient outcome for each of the multiple values of rotation of the component and slope of the component.   
     
     
         2 . The method of  claim 1 , wherein the multiple historical kinematic parameters are indicative of a kinematic simulation of the historical patients' knees. 
     
     
         3 . The method of  claim 1 , wherein the method further comprises performing the machine learning on the historical patient records. 
     
     
         4 . The method of  claim 3 , wherein performing the machine learning comprises selecting the multiple historical kinematic parameters from a larger set of potential kinematic parameters. 
     
     
         5 . The method of  claim 3 , further comprising receiving from each of the multiple historical patients the reported patient outcome as user input via a user interface. 
     
     
         6 . The method of  claim 1 , wherein the multiple historical kinematic parameters and the multiple simulated kinematic parameters are independent of a knee implant geometry. 
     
     
         7 . The method of  claim 1  further comprising determining one or more simulated kinematic parameters that are most responsible for the current patient outcome. 
     
     
         8 . The method of  claim 1 , further comprising generating a shaded surface spanning the multiple values of rotation of the component and slope of the component on a user interface to graphically represent the estimated current patient outcome for each of the multiple values of rotation of the component and slope of the component. 
     
     
         9 . The method of  claim 8 , further comprising generating one shaded surface for each of multiple surgery parameters by repeating the following steps (a) and (b) for each of the multiple surgery parameters:
 (a) for each of multiple values of rotation of the component and slope of the component:
 (a1) configuring a post-operative kinematic model of the current patient's knee based on the computer tomography data, the user input, and the value of the rotation and the slope and that surgery parameter; 
 (a2) performing a kinematic simulation based on the post-operative kinematic model to determine multiple simulated kinematic parameters; and 
 (a3) estimating a current patient outcome by applying the multiple machine learning model parameters to the multiple simulated kinematic parameters of the current patient; and 
   (b) generating a shaded surface for that surgery parameter, spanning the multiple values of rotation of the component and slope of the component on a user interface to graphically represent the estimated current patient outcome for each of the multiple values of rotation of the component and slope of the component.   
     
     
         10 . The method of  claim 9 , further comprising arranging the shaded surface for each of the multiple surgery parameters in a grid on a user interface to indicate combinations of surgery parameters. 
     
     
         11 . The method of  claim 1 , further comprising receiving computer tomography data of the current patient's knee and building the 3D model to represent the computer tomography data. 
     
     
         12 . The method of  claim 1 , wherein determining the post-operative kinematic model comprises simplifying the 3D model. 
     
     
         13 . The method of  claim 1 , wherein the component is a tibial component. 
     
     
         14 . The method of  claim 1 , wherein
 the historical patient records further comprise historical anatomical measurements;   the multiple machine learning model parameters are indicative of a relationship between the historical anatomical measurements and the reported patient outcome; and   estimating the current patient outcome comprises applying the multiple machine learning model parameters to anatomical measurements of the current patient's knee.   
     
     
         15 . The method of  claim 1 , wherein
 the historical patient records further comprise historical demographic parameters and patient questionnaire data capture parameters;   the machine learning model parameters are indicative of a relationship between (i) the historical demographic parameters and patient questionnaire data capture parameters and (ii) the reported patient outcome; and   estimating the current patient outcome comprises applying the multiple machine learning model parameters to a current patient's demographic and patient questionnaire data capture parameters.   
     
     
         16 . The method of  claim 1 , wherein estimating the current patient outcome is based on kinematic expert knowledge to either modify or reweight penalty factors from the kinematic simulation or is based on new penalty factors from the kinematic simulation. 
     
     
         17 . The method of  claim 16 , wherein the kinematic expert knowledge is different for different users of a computer system performing the method. 
     
     
         18 . The method of  claim 1 , further comprising determining further placement input parameters of the knee implant based on simulated kinematic parameters other than slope and rotation of the component in order to optimise the estimated current patient outcome. 
     
     
         19 . A non-transitory computer readable medium with program code stored thereon that, when executed by a computer, is configured to cause the computer to perform one or more operations comprising:
 retrieving a 3D model of a current patient's knee, the 3D model representing computer tomography data of the current patient's knee;   receiving user input comprising an identifier of a knee implant, the knee implant comprising a component;   retrieving multiple machine learning model parameters indicative of machine learning performed on historical patient records, the historical patient records comprising multiple historical kinematic parameters of each of multiple historical patients as inputs and a reported patient outcome for each historical patient as output, the machine learning model parameters being indicative of a relationship between the multiple historical kinematic parameters and the reported patient outcome;   for each of multiple values of rotation of the component and slope of the component:
 configuring a post-operative kinematic model of the current patient's knee based on the computer tomography data, the user input, and the value of the rotation and the slope, wherein configuring the post-operative kinematic model comprises virtually performing a surgery by introducing cut surfaces to change a shape of bones in the 3D model and adding a shape of the knee implant to determine the post-operative kinematic model; 
 performing a kinematic simulation based on the post-operative kinematic model to determine multiple simulated kinematic parameters; and 
 estimating a current patient outcome by applying the multiple machine learning model parameters to the multiple simulated kinematic parameters of the current patient; and 
   representing the estimated current patient outcome for each of the multiple values of rotation of the component and slope of the component.   
     
     
         20 . A computer system for assisting a surgeon with a graphical representation of a dynamic knee score for a knee surgery, the computer system comprising:
 a data source connection configured to:
 retrieve a 3D model of a current patient's knee, the 3D model representing computer tomography data of the current patient's knee; 
 receive user input comprising an identifier of a knee implant, the knee implant comprising a component; and 
 retrieve multiple machine learning model parameters indicative of a machine learning performed on historical patient records, the historical patient records comprising multiple historical kinematic parameters of each of multiple historical patients as inputs and a reported patient outcome for each historical patient as output, the machine learning model parameters being indicative of a relationship between the multiple historical kinematic parameters and the reported patient outcome; 
   a data storage device with program code stored thereon; and   a processor configured to execute the program code stored on the data storage device to perform the steps of:   for each of multiple values of rotation of the component and slope of the component:
 configuring a post-operative kinematic model of the current patient's knee based on the computer tomography data, the user input, and the value of the rotation and the slope, wherein configuring the post-operative kinematic model comprises:
 virtually performing the surgery by introducing cut surfaces to change a shape of bones in the 3D model and adding a shape of the knee implant to determine the post-operative kinematic model; 
 
 performing a kinematic simulation based on the post-operative kinematic model to determine multiple simulated kinematic parameters; and 
 estimating a current patient outcome by applying the multiple machine learning model parameters to the multiple simulated kinematic parameters of the current patient; and 
   represent the estimated current patient outcome for each of the multiple values of rotation of the component and slope of the component.   
     
     
         21 . The computer system of  claim 20 , further comprising a display device to display a shaded surface that represents the estimated current patient outcome or each of the multiple values of rotation of the component and slope of the component to a surgeon.