Computer assisted identification of appropriate anatomical structure for medical device placement during a surgical procedure
A method for computer assisted identification of appropriate anatomical structure for placement of a medical device, comprising: receiving a 3D scan volume comprising set of medical scan images of a region of an anatomical structure where the medical device is to be placed; automatically processing the set of medical scan images to perform automatic segmentation of the anatomical structure; automatically determining a subsection of the 3D scan volume as a 3D ROI by combining the raw medical scan images and the obtained segmentation data; automatically processing the ROI to determine the preferred 3D position and orientation of the medical device to be placed with respect to the anatomical structure by identifying landmarks within the anatomical structure with a pre-trained prediction neural network; automatically determining the preferred 3D position and orientation of the medical device to be placed with respect to the 3D scan volume of the anatomical structure.
1. A method for computer assisted identification of an anatomical structure for placement of a medical device, the method comprising:
receiving a three-dimensional (3D) scan volume comprising a set of medical scan images of a region of the anatomical structure where the medical device is yet to be placed;
processing the set of medical scan images to obtain segmentation data of the anatomical structure that identifies different anatomical parts of the anatomical structure;
determining a subsection of the 3D scan volume as a 3D region of interest (ROI) where the medical device is yet to be placed, the 3D ROI including image data from the set of medical scan images and segmentation data associated with the subsection of the 3D scan volume;
processing the 3D ROI with a prediction neural network model to identify landmarks associated with a set of anatomical parts within the anatomical structure for the placement of the medical device; and
determining preferred 3D positioning, orientation, and dimensions of the medical device yet to be placed with respect to the 3D scan volume of the anatomical structure based on the landmarks.
2. The method according to claim 1 , further comprising:
identifying and storing, after the medical device has been placed by a surgeon within the anatomical structure during the surgical procedure, actual 3D positioning and orientation of the medical device; and
training the prediction neural network using the actual 3D positioning and orientation of the medical device to improve accuracy of the prediction neural network to subsequently identify preferred positioning, orientation, and dimensions of the medical device to be suggested to the surgeon in successive surgical procedures.
3. The method according to claim 1 , further comprising:
processing the medical scan images of the set of medical scan images of the set of anatomical parts associated with the landmarks; and
determining physical dimensions of the set of anatomical parts in the 3D ROI for the placement of the medical device.
4. The method according to claim 3 , wherein determining the preferred 3D dimensions includes determining at least one of size, diameter, or length of the medical device to be placed based on the physical dimensions of the set of anatomical parts.
5. The method according to claim 1 , wherein the set of medical scan images is collected using an intraoperative scanner.
6. The method according to claim 1 , wherein the set of medical scan images is collected using a presurgical stationary scanner.
7. A computer-implemented system for computer assisted identification of an anatomical structure for placement of a medical device, comprising:
at least one nontransitory processor-readable storage medium that stores at least one of processor-executable instructions or data; and
at least one processor communicably coupled to the at least one nontransitory processor-readable storage medium, wherein the at least one processor is configured to:
process a set of medical scan images to obtain segmentation data of the anatomical structure that identifies different anatomical parts of the anatomical structure, the set of medical scan images being from a three-dimensional (3D) scan volume of a region of the anatomical structure where the medical device is yet to be placed;
determine a subsection of the 3D scan volume as a 3D region of interest (ROI) where the medical device is yet to be placed, the 3D ROI including image data from the set of medical scan images and segmentation data associated with the subsection of the 3D scan volume;
process the 3D ROI with a prediction neural network model to identify landmarks associated with a set of anatomical parts within the anatomical structure for the placement of the medical device; and
determine preferred 3D positioning, orientation, and dimensions of the medical device yet to be placed with respect to the 3D scan volume of the anatomical structure based on the landmarks.
8. The method according to claim 1 , further comprising:
resizing the 3D ROI such that a ROI in each medical scan image from the set of medical scan images stacked in the subsection of the 3D scan volume has a same size,
the processing of the 3D ROI with the prediction neural network model occurring after the resizing of the 3D ROI.
9. The method according to claim 8 , further comprising:
reverting, after the processing of the 3D ROI with the prediction neural network model, the size of the ROI in each medical scan image from the set of medical scan images stacked in the subsection of the 3D scan volume to determine a position of the identified landmarks within the 3D ROI.
10. The method according to claim 1 , further comprising:
combining, after the processing of the 3D ROI with the prediction neural network model, coordinate information indicative of a position of the landmarks within a local coordinate system of the 3D ROI with coordinate information indicative of a position of the 3D ROI within a global coordinate system of the 3D scan volume, and
the determining the preferred 3D positioning, orientation, and dimensions of the medical device is based on the combined coordinate information.
11. The method according to claim 1 , further comprising:
training the prediction neural network model using a training set including previously determined ROIs with marked characteristic features associated with the landmarks.
12. The method according to claim 11 , wherein the landmarks include at least one of a pedicle center or a tip of the medical device.
13. The method according to claim 11 , further comprising validating a quality of the prediction neural network model using a validation set including 3D scan volumes of regions of anatomical structures.
14. The method according to claim 11 , further comprising augmenting the training set by transforming the previously determined ROIs using one or more of: rotation, scaling, movement, horizontal flip, additive noise of Gaussian or Poisson distributions and Gaussian blur, volumetric grid deformation, or a generative algorithm,
and the predictive neural network model being trained using the augmented training set.
15. The method according to claim 1 , wherein the processing the 3D ROI with the prediction neural network model includes processing the 3D ROI through a set of layers of the prediction neural network model in a standard forward pass to obtain outputs of the prediction neural network model.
16. The method according to claim 15 , wherein the processing the 3D ROI further comprises calculating a value of a loss function associated with the prediction neural network model based on the outputs of the standard forward pass.
17. The method according to claim 16 , wherein the processing the 3D ROI further comprises updating a set of weights of the prediction neural network model by backward-propagating the value of the loss function through the set of layers of the prediction neural network model.
18. The method according to claim 14 , wherein the augmenting the training set includes recalculating a position of the landmarks in each of the transformed 3D ROIs.
19. The method according to claim 1 , further comprising backward recalculating one or more of the preferred 3D positioning, orientation, and dimensions of the medical device yet to be placed, 3D scan volume of the anatomical structure, and the landmarks.
20. The method according to claim 1 , wherein the preferred 3D positioning, orientation, and dimensions of the medical device yet to be placed is based on two of the landmarks.
21. The method according to claim 1 , further comprising visualizing the preferred 3D positioning, orientation, and dimensions of the medical device yet to be placed with respect to the anatomical structure.