Method and system for patient-specific virtual percutaneous structural heart intervention
A system and method for selecting, from a series of cardiac implants having different sizes, the cardiac implant having optimum size for implantation in a patient. The method includes obtaining data representative of a patient-specific cardiac region and predicting the optimum size of the cardiac implant best matching a predefined criterion when deployed in the cardiac region. The predicting includes querying a database; determining parameter values for a parametric model representation of the patient-specific cardiac region; and/or entering the data representative of the patient-specific cardiac region into an artificial intelligence device.
1. A computer implemented method for selecting, from a series of cardiac implants having different sizes, a cardiac implant having an optimum size, and predicting an optimum deployment position, for implantation of the cardiac implant in a patient, the method comprising:
obtaining a patient-specific three-dimensional anatomical model corresponding to data representative of a patient-specific three-dimensional image of a cardiac region, the patient-specific three-dimensional anatomical model comprising a finite element mesh;
obtaining an implant model representing a finite element representation of the cardiac implant;
virtually deploying the implant model into the patient-specific three-dimensional anatomical model;
calculating a deployed shape of the implant model at a plurality of deployment locations of the implant model; and
predicting the optimum size, and optimum position, of the cardiac implant when deployed in the cardiac region based on the patient-specific three-dimensional anatomical model and the implant model,
wherein the predicting comprises querying a database including a plurality of records, each record including data representative of a patient-specific three-dimensional image of a cardiac region and an associated size and associated position of the cardiac implant of the series.
2. The method of claim 1 , wherein the database includes records associated with respective patient-specific clinical data, and/or records associated with simulated data.
3. The method of claim 1 , wherein the database includes records obtained by applying augmentation techniques to other records, the augmentation techniques comprising scaling and/or modifying a histogram.
4. The method of claim 1 , wherein the querying of the database includes using extreme gradient boosting.
5. The method of claim 1 , wherein the predicting is further based on metadata with respect to the patient, the meta data comprising one or more of demographic data, known pathology, or medicament use.
6. The method of claim 1 , wherein a predefined criterion is a lowest risk of complications during and/or after deployment of an actual implant in an actual cardiac region of the patient.
7. A computer implemented method for estimating a risk of complications arising in and/or after structural heart intervention, the method comprising:
obtaining a patient-specific three-dimensional anatomical model corresponding to data representative of a patient-specific three-dimensional image of a cardiac region, the patient-specific three-dimensional anatomical model comprising a finite element mesh;
obtaining data representative of a size and type of a cardiac implant to be implanted in the cardiac region of the patient;
obtaining an implant model representing a finite element representation of the cardiac implant;
virtually deploying the implant model into the patient-specific three-dimensional anatomical model;
calculating a deployed shape of the implant model at a plurality of deployment locations of the implant model; and
predicting an interaction between the cardiac implant and cardiac region based on the patient-specific three-dimensional anatomical model and the implant model,
wherein the predicting comprises querying a database including a plurality of records, each record including data representative of a patient-specific three-dimensional image of a cardiac region, the size and type of the cardiac implant, and the interaction.
8. The method of claim 7 , wherein the interaction is at least one of mechanical interaction leakage, regurgitation, cardiac conduction abnormalities, or risk of implant misplacement.
9. The method of claim 1 , including:
using a neural network for generating the plurality of records; and
storing the plurality of records in a database.
10. A computer implemented method for planning structural heart intervention, the method comprising:
obtaining a patient-specific three-dimensional anatomic model corresponding to data representative of a patient-specific three-dimensional image of a cardiac region, the patient-specific three-dimensional anatomic model comprising a finite element mesh;
obtaining data representative of a cardiac implant and corresponding to a size and type of the cardiac implant, the cardiac implant configured to be implanted in the cardiac region of the patient;
obtaining an implant model representing a finite element representation of the cardiac implant;
virtually deploying the implant model into the patient-specific three-dimensional anatomical model; and
predicting a deployed shape of the cardiac implant in the cardiac region based on the implant model and the patient-specific three-dimensional anatomical model,
wherein the predicting comprises querying a database including a plurality of records, each record including data representative of a patient-specific three-dimensional image of a cardiac region, a cardiac implant, and the predicted deployed shape of the cardiac implant in the respective cardiac region.
11. The method of claim 10 , wherein the prediction of the deployed shape is presented to a user as an overlay on, a view of the patient-specific three-dimensional image of the cardiac region.
12. The method of claim 10 , further comprising determining a neo-LVOT area;
wherein the obtaining data representative of a cardiac implant includes obtaining data representative of a mitral valve implant.
13. The method of claim 12 , wherein the prediction of the deployed shape is presented to a user, as an overlay on, a view of the patient-specific three-dimensional image of the cardiac region corresponding to a mitral valve annulus region.
14. The method of claim 5 , wherein the neo-LVOT area is determined from the patient-specific three-dimensional image of the cardiac region corresponding to the mitral valve annulus region and the predicted deployed shape.
15. The method of claim 2 , further comprising
calculating an interaction between the implant model and the patient-specific three-dimensional anatomical model based on the predicted deployed shape.