Systems and methods for predicting surgical outcomes

Patent metadata

Field	Value
Publication number	US-12484965-B2
Application number	US-202117163924-A
Country	US
Kind code	B2
Filing date	Feb 1, 2021
Priority date	Feb 1, 2021
Publication date	Dec 2, 2025
Grant date	Dec 2, 2025

Abstract

Systems and methods for predicting surgical outcomes are provided. A surgical plan comprising information about a planned surgery and at least one preoperative image depicting a planned surgical result and at least one postoperative image depicting an actual surgical result resulting from execution of the planned surgery may be received. The postoperative image may be registered to the preoperative image. One or more features may be automatically identified in each of the postoperative image and the preoperative image. A difference may be automatically measured in at least one parameter of each of the one or more features to yield training data. A function for predicting the difference may be generated using artificial intelligence and based on the training data. The function may be applied to an unexecuted surgical plan.

First claim

What is claimed is:

1 . A method for predicting surgical outcomes, comprising:
receiving, by a processor, a surgical plan comprising information about a planned surgery and at least one preoperative image depicting a planned surgical result; receiving, by the processor, at least one postoperative image depicting an actual surgical result resulting from execution of the planned surgery; registering, by the processor, the postoperative image to the preoperative image; automatically identifying, by the processor using feature recognition, one or more features in each of the postoperative image and the preoperative image; automatically measuring, by the processor, a difference between a value of a parameter of a feature in the postoperative image and the value of the parameter of the feature in the preoperative image to yield training data, wherein the difference comprises at least one of a distance and an angle; generating, by the processor, using artificial intelligence and based on the training data, a function for predicting the difference, wherein generating the function comprises fitting a transfer function to the training data using regression analysis; determining, using artificial intelligence and the training data, at least one unessential parameter to remove from a plurality of parameters, wherein each of the at least one unessential parameter does not affect a result of the function, and wherein the at least one unessential parameter comprises at least one of the distance and the angle; predicting, by the processor using the function and based on an unexecuted surgical plan, an expected difference between the value of the parameter of the feature within a planned surgical result described in the unexecuted surgical plan and the value of the parameter of the feature in an actual surgical result resulting from execution of the unexecuted surgical plan; and updating, by the processor, the unexecuted surgical plan to apply at least one change based on the predicted expected difference to reduce the predicted expected difference for the value of the parameter of the feature within the actual surgical result, wherein the at least one change comprises at least one of a change in a tool to use in the unexecuted surgical plan, a change in a tool trajectory, a change in an insertion point of a tool, and/or a change in insertion depth of a tool.

2 . The method of claim 1 , further comprising:
generating, by the processor, a notification based on the predicted expected difference, the notification being at least one of an audible notification or a visual notification.

3 . The method of claim 1 , further comprising:
generating, by the processor, a notification based on the updated unexecuted surgical plan, the notification including a prompt to accept or decline the at least one change in the updated unexecuted surgical plan.

4 . The method of claim 1 , wherein the unexecuted surgical plan is automatically updated based on pre-authorized historical plans, the historical plans having at least one surgical step substantially similar to the at least one change in the updated unexecuted surgical plan.

5 . The method of claim 1 , wherein the measuring includes automatically determining the value of the parameter of the feature based on the surgical plan.

6 . The method of claim 1 , wherein the measuring includes automatically identifying at least one location to measure the difference between the value of the parameter of the feature in the postoperative image and the value of the parameter of the feature in the preoperative image.

7 . The method of claim 1 , wherein the identifying further uses segmentation.

8 . The method of claim 1 , wherein the feature comprises at least one of one or more implants or one or more anatomical elements.

9 . The method of claim 1 , wherein the value of the parameter of the feature comprises at least one of a position or an orientation of at least one of one or more implants or one or more tools.

10 . The method of claim 1 , further comprising:
determining, using artificial intelligence and the training data, a weight for the value of the parameter of the feature.

11 . The method of claim 1 , wherein the function is based on a plurality of inputs.

12 . The method of claim 1 , wherein the registering includes overlaying a depiction of the feature from the at least one postoperative image over a depiction of the feature from the at least one preoperative image.

13 . A method for predicting surgical outcomes comprising:
receiving, by a processor, at least one preoperative image depicting a planned surgical result and at least one postoperative image depicting an actual surgical result; identifying, by the processor using feature recognition, one or more features in each of the postoperative image and the preoperative image; overlaying, by the processor, the one or more features of the at least one preoperative image over the one or more features of the at least one postoperative image; measuring, by the processor, a difference between a value of a parameter of a feature in the postoperative image and the value of the parameter of the feature in the preoperative image to yield training data, wherein the difference comprises at least one of a distance and an angle; generating, by the processor using artificial intelligence and based on the training data, a function for predicting the difference, wherein generating the function comprises fitting a transfer function to the training data using regression analysis; determining, using artificial intelligence and the training data, at least one unessential parameter to remove from a plurality of parameters, wherein each of the at least one unessential parameter does not affect a result of the function, and wherein the at least one unessential parameter comprises at least one of the distance and the angle; predicting, by the processor using the function and based on an unexecuted surgical plan, an expected difference between the value of the parameter of the feature within a planned surgical result described in the unexecuted surgical plan and the value of the parameter of the feature in an actual surgical result resulting from execution of the unexecuted surgical plan; and updating, by the processor, the unexecuted surgical plan to apply at least one change based on the predicted expected difference to reduce the predicted expected difference for the value of the parameter of the feature within the actual surgical result, wherein the at least one change comprises at least one of a change in a tool to use in the unexecuted surgical plan, a change in a tool trajectory, a change in an insertion point of a tool, and/or a change in insertion depth of a tool.

14 . The method of claim 13 , further comprising:
generating, by the processor, a notification based on the updated unexecuted surgical plan, the notification including a prompt to accept or decline the at least one change in the updated unexecuted surgical plan.

15 . The method of claim 13 , wherein the unexecuted surgical plan is automatically updated based on pre-authorized historical plans, the historical plans having at least one surgical step substantially similar to the at least one change in the updated unexecuted surgical plan.

16 . A system for predicting surgical outcomes comprising:
at least one processor; and at least one memory storing instructions for execution by the at least one processor that, when executed, cause the at least one processor to:
receive a surgical plan comprising information about a planned surger

Assignees

• Mazor Robotics Ltd

Inventors

• Kleyman Leonid
• Regev Ofer
• Sandelson Adi
• Glaser Adam D
• Riding Shane Lars

Classifications

• G06N3/08

  Learning methods · CPC title

• A61B2034/102

  Modelling of surgical devices, implants or prosthesis · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title