Differential atlas for identifying sites of recurrence (disrn) in predicting atrial fibrillation recurrence

An apparatus comprising:

• a memory configured to store a set of computed tomography (CT) images acquired from a population of subjects, where a member of the set of CT images includes a left atrium (LA) region, where the set of CT images is acquired from a population of patients, where the population includes an atrial fibrillation (AF) recurrence positive (AF+) subpopulation, and an AF recurrence negative (AF−) subpopulation;
• where each member of the set of CT images includes an atrial mask that defines an LA represented in each member of the set of CT images, respectively;
• one or more processors configured to:
• access the set of CT images;
• construct a statistical shape differential atlas from the set of CT images;
• generate a template LA model from the statistical shape differential atlas, where the template LA model includes a site of interest (SOI);
• acquire a pre-ablation CT image of a region of tissue in a patient demonstrating AF pathology;
• generate a patient LA model from the pre-ablation CT image;
• compute an optimal deformation field that registers the SOI to the patient LA model using a deformable registration approach;
• compute a patient feature vector based on the optimal deformation field;
• generate an AF probability score for the patient based, at least in part, on the patient feature vector;
• generate a classification of the patient demonstrating AF pathology based, at least in part, on the AF probability score;
• and
• display the classification or the AF probability score.

A non-transitory computer-readable storage device storing executable instructions that, in response to execution, cause an atrial fibrillation (AF) recurrence prediction device to perform operations for producing a quantification of differences between a pre-ablation image of a left atrium (LA) and a statistical shape atlas, the operations comprising:

• accessing a set of computed tomography (CT) images acquired from a population of subjects, where a member of the set of CT images includes a left atrium (LA) region;
• constructing a statistical shape differential atlas from the set of CT images;
• generating a template LA model from the statistical shape differential atlas, where the template LA model includes a site of interest (SOI);
• acquiring a pre-ablation CT image of a region of tissue in a patient demonstrating AF pathology;
• generating a patient LA model from the pre-ablation CT image;
• computing an optimal deformation field that registers the SOI to the patient LA model using a deformable registration approach;
• computing a patient feature vector based on the optimal deformation field;
• generating an AF probability score for the patient based, at least in part, on the patient feature vector;
• generating a classification of the patient based, at least in part, on the AF probability score;
• generating a personalized AF treatment plan based on the classification, the AF probability score, and the pre-ablation CT image;
• and
• displaying the personalized AF treatment plan, the classification, the AF probability score, or the pre-ablation CT image.