Cross-modality planning using feature detection
Systems and methods for planning the position of surgical hardware to be robotically implanted in a subject. The system extracts information about the planned position of hardware from an operative plan based on preoperative images, and converts this information into mathematical vectors. Intraoperatively, at least one three-dimensional scan of the operative site is obtained. The intraoperative images are processed by image analysis, to which are applied artificial intelligence algorithms for feature identification. The vectors derived from the preoperative plan are superimposed on identified anatomical features from the processed intraoperative images. The surgical plan can then be updated intraoperatively, taking into account any shift in position of the anatomical features between the preoperative images and the intraoperative images, prior to robotic insertion of the hardware.
1 . A system, comprising:
at least one processor executing instructions stored on at least one non-transitory storage medium, to cause the at least one processor to:
a) based on a surgical plan derived from preoperative images, define planned poses for implantable hardware elements comprising at least one three-dimensional implantable hardware element, or at least two two-dimensional implantable hardware elements, to be attached to or inserted into an anatomical part;
b) convert the planned poses into a three-dimensional geometric function for the anatomical part;
c) process at least one intraoperative three-dimensional image of the anatomical part to identify anatomical features to which the three-dimensional geometric function is to be aligned;
d) on the at least one processed intraoperative three-dimensional image, define a range of possible positions for the three-dimensional geometric function; and
e) using the identified anatomical features, virtually align the three-dimensional geometric function in the at least one processed intraoperative three-dimensional image, the alignment achieving a position of the implantable hardware elements that is compatible with the surgical plan and has greater accuracy than aligning each implantable hardware element individually on the at least one processed intraoperative three-dimensional image, without the need to perform registration between the preoperative images and the at least one processed intraoperative three-dimensional image.
2 . The system according to claim 1 , wherein steps a) to e) are performed for a plurality of anatomical parts.
3 . The system according to claim 2 , further comprising steps:
f) repeat steps a) to e) on each anatomical part, such that a plurality of three-dimensional geometric functions is generated;
g) compare virtual alignments in the at least one processed intraoperative three-dimensional image of all three-dimensional geometric functions with the surgical plan; and
h) if the virtual alignments in the at least one processed intraoperative three-dimensional image are inconsistent with the surgical plan, repeat step e), such that the positioning of the implantable hardware elements has greater accuracy than aligning each implantable hardware element individually on the at least one processed intraoperative three-dimensional image for a complete set of implantable hardware elements for the plurality of anatomical parts.
4 . The system according to claim 1 , wherein a robotic surgical system enabled to carry out the surgical plan uses at least one aligned three-dimensional geometric function in the at least one processed intraoperative three-dimensional image to update the surgical plan.
5 . The system according to claim 1 , wherein the preoperative images comprise at least one three-dimensional MRI or CT image.
6 . The system according to claim 1 , wherein virtually aligning the three-dimensional geometric function is based on one or more predetermined constraints corresponding to the anatomical features.
7 . The system according to claim 1 , wherein the surgical plan is for a spinal fusion, the implantable hardware elements are pedicle screws and intervertebral rods, and the anatomical part is a vertebra.
8 . The system according to claim 1 , wherein the three-dimensional geometric function defines a fixed angle between the at least two two-dimensional implantable hardware elements in reference to each other.
9 . The system according to claim 1 , wherein at least one form of artificial intelligence and anatomical images are used to identify a range of possible positions for implantable hardware elements within the at least one intraoperative three-dimensional image.
10 . The system according to claim 1 , further comprising the step of determining a mismatch in the planned poses of implantable hardware elements between the preoperative images and the at least one processed intraoperative three-dimensional image, wherein the determined mismatch is used to adjust the planned poses of the implantable hardware elements.
11 . The system according to claim 1 , wherein the virtual alignment reduces a stress cost function using an algorithm that identifies a stress minimum between one of the implantable hardware elements and the anatomical part.
12 . The system according to claim 11 , wherein the algorithm uses artificial intelligence applied to at least one of feature detection, intensity detection, finite element analysis with meshing, and image segmentation to reduce the stress cost function.
13 . The system according to claim 12 , wherein the three-dimensional geometric functions for all implantable hardware elements are positioned in combination with analysis of preoperative images showing motion analysis of the anatomical part.
14 . The system according to claim 11 , wherein the stress cost function further takes into account predicted stress of all implantable hardware elements on the anatomical part.
15 . The system according to claim 1 , wherein the surgical plan is for a spinal fusion, and the position of the implantable hardware elements is determined by analysis of measured spinal mobility limitations over substantial lengths of the patient's spine in order to plan a correction procedure with minimal surgical corrective steps.
16 . The system according to claim 1 , wherein the three-dimensional geometric function is defined as a mathematical quantity having four points, the four points representing the beginning and ending of each of a pair of pedicle screws for a given vertebra in three-dimensional space.
17 . The system according to claim 16 , wherein a first of the pair of pedicle screws is used after implantation in combination with the at least one processed intraoperative three-dimensional image to accomplish the virtual alignment with a second of the pair of pedicle screws.
18 . The system according to claim 1 , wherein the three-dimensional geometric function comprises the position, length, and angle of a right pedicle screw and a left pedicle screw having a fixed angle between them, for implantation into a single vertebra.
19 . A system, comprising:
a memory for storing a surgical plan, which is based on preoperative images, for a region of interest, the surgical plan comprising planned poses of implantable hardware elements to be inserted into anatomical parts in the region of interest; and
at least one processor having a controller to;
convert the planned poses into a three-dimensional geometric function for an anatomical part;
process at least one intraoperative three-dimensional image of the anatomical part to identify anatomical features to which the three-dimensional geometric function is to be aligned;
on the at least one processed intraoperative three-dimensional image, define a range of possible positions for the three-dimensional geometric function; and
using the identified anatomical features, virtually align the three-dimensional geometric function in the at least one processed intraoperative three-dimensional image, the virtual alignment achieving a position of the implantable hardware elements that is compatible with the surgical plan and has greater accuracy than aligning each implantable hardware element individually on the at least one processed intraoperative three-dimensional image, without the need to perform registration between the preoperative images and the at least one processed intraoperative three-dimensional image.
20 . A system, comprising:
at least one processor executing instructions stored on at least one non-transitory storage medium, to cause the at least one processor to:
a) based on a surgical plan derived from a preoperative images, define planned poses of a pair of pedicle screws for a vertebra;
b) convert the planned poses of the pair of screws into a three-dimensional geometric function for the pair of screws;
c) process at least one intraoperative three-dimensional image of the vertebra to identify surface features of the vertebra;
d) on the at least one processed intraoperative three-dimensional image, define a range of possible positions for each pedicle screw through vertebral pedicles; and
e) using the identified surface features, virtually align the three-dimensional geometric function in the at least one processed intraoperative three-dimensional image, the virtual alignment achieving a position of the pedicle screws that is compatible with the surgical plan and reduces a cost function to a greater degree than aligning each pedicle screw individually on the at least one processed intraoperative three-dimensional image, without the need to perform registration between the preoperative images and the at least one processed intraoperative three-dimensional image.