Apparatus and method for generating clinical decision support
An apparatus and method for generating clinical decision support is disclosed. The apparatus includes at least a processor and a computer-readable storage medium communicatively connected to the at least a processor, wherein the computer-readable storage medium contains instructions configuring the at least processor to receive user data, generate a fused feature vector correlating the user data to a plurality of clinical outcomes by training a plurality of deep neural networks (DNNs) to output a first set of feature vectors, a second set of feature vectors and a third set of feature vectors, fusing the first, second, and third set of features vectors to form the fused feature vector, generate a procedural output using the fused feature vector, and display the procedural output through a user interface.
1. An apparatus for generating clinical decision support, wherein
the apparatus comprises:
at least a processor;
a computer-readable storage medium communicatively connected to the at least a processor, wherein the computer-readable storage medium contains instructions configuring the at least processor to:
receive user data associated with a subject, the user data comprising:
electrocardiogram (ECG) data comprising a plurality of signals representative of an electrical activity of a heart of the subject;
image data comprising a plurality of images associated with the heart of the subject; and
user historical data comprising a plurality of records;
generating, using a first deep neural network (DNN), a first set of feature vectors representative of the ECG data;
generating, using a second DNN, a second set of feature vectors representative of the image data;
generating, using a third DNN, a third set of feature vectors representative of the user historical data; and
generate a set of dimensionally reduced feature vectors for each of the first set, the second set, and the third set of feature vectors; fusing the first, second, and third set of dimensionally reduced features vectors to form the fused feature vector;
generate a procedural output using the fused feature vector; and
display the procedural output through a user interface.
2. The apparatus of claim 1 , wherein the first DNN is trained with time series data of exemplary user data correlated to clinical outcomes to output the first set of feature vectors.
3. The apparatus of claim 1 , wherein the second DNN is trained with image data of exemplary user data correlated to clinical outcomes to output the second set of feature vectors.
4. The apparatus of claim 1 , wherein the third DNN is trained with natural language data of exemplary user data correlated to clinical outcomes to output the third set of feature vectors.
5. The apparatus of claim 1 , wherein the procedural output comprises a pre-procedure output, a during-procedure output, and a post-procedure output.
6. The apparatus of claim 5 , wherein the pre-procedure output comprises an ablation strategy outlining focus on key aspects of cardiac ablation prior to an ablation procedure.
7. The apparatus of claim 5 , wherein the during-procedure output comprises an integrative assessment data assessing lesion quality and contiguity implemented in an intraprocedural visualization stage of an ablation procedure.
8. The apparatus of claim 5 , wherein the post-procedure output comprises a post-ablation summative assessment of durability of ablation estimating a likelihood of atrial fibrillation recurrence after an ablation procedure.
9. The apparatus of claim 1 , wherein generating the procedural output comprises training a fused feature classifier based on the fused feature vector to output a pre-procedure output and a post-procedure output.
10. The apparatus of claim 1 , wherein generating the procedural output comprises training a fused object detection model based on the fused feature vector to output a during-procedure output.
11. A method for generating clinical decision support, wherein the method comprises:
receiving, by a computing device, user data associated with a subject, the user data comprising:
electrocardiogram (ECG) data comprising a plurality of signals representative of an electrical activity of a heart of the subject;
image data comprising a plurality of images associated with the heart of the subject; and
user historical data comprising a plurality of records;
generating, using a first deep neural network (DNN), a first set of feature vectors representative of the ECG data;
generating, using a second DNN, a second set of feature vectors representative of the image data;
generating, using a third DNN, a third set of feature vectors representative of the user historical data; and
generate a set of dimensionally reduced feature vectors for each of the first set, the second set, and the third set of feature vectors; fusing the first, second, and third set of dimensionally reduced features vectors to form the fused feature vector;
generating, by the computing device, a procedural output using the fused feature vector; and
displaying, by the computing device, the procedural output through a user interface.
12. The method of claim 11 , wherein the first DNN is trained with time series data of exemplary user data correlated to clinical outcomes to output the first set of feature vectors.
13. The method of claim 11 , wherein the second DNN is trained with image data of exemplary user data correlated to clinical outcomes to output the second set of feature vectors.
14. The method of claim 11 , wherein the third DNN is trained with natural language data of exemplary user data correlated to clinical outcomes to output the third set of feature vectors.
15. The method of claim 11 , wherein the procedural output comprises a pre-procedure output, a during-procedure output, and a post-procedure output.
16. The method of claim 15 , wherein the pre-procedure output comprises an ablation strategy outlining focus on key aspects of cardiac ablation prior to an ablation procedure.
17. The method of claim 15 , wherein the during-procedure output comprises an integrative assessment data assessing lesion quality and contiguity implemented in an intraprocedural visualization stage of an ablation procedure.
18. The method of claim 15 , wherein the post-procedure output comprises a post-ablation summative assessment of durability of ablation estimating a likelihood of atrial fibrillation recurrence after an ablation procedure.
19. The method of claim 11 , wherein generating the procedural output comprises training a fused feature classifier based on the fused feature vector to output a pre-procedure output and a post-procedure output.
20. The method of claim 11 , wherein generating the procedural output comprises training a fused object detection model based on the fused feature vector to output a during-procedure output.