Method and system for improved hemodynamic computation in coronary arteries

1. A method for non-invasive assessment of an arterial stenosis, comprising:

segmenting a plurality of mesh candidates for an anatomical model of an artery including a stenosis region of a patient from medical imaging data, the plurality of mesh candidates representing different anatomical models of the artery resulting from variations in segmenting techniques, wherein segmenting the plurality of mesh candidates comprises:

determining one or more candidate locations for each vertex of the anatomical model,

assigning a confidence score to each of the plurality of mesh candidates by averaging a probability associated with each vertex in a respective mesh candidate, and

applying cluster analysis to reduce a number of the plurality of mesh candidates based on the confidence score;

computing a hemodynamic index for the stenosis region in each of the plurality of mesh candidates; and

determining whether a clinical decision for treatment of the stenosis region of the patient is ambiguous by determining whether a variation among values of the hemodynamic index for the stenosis region in each of the plurality of mesh candidates is entirely below a threshold for the clinical decision or is entirely above the threshold for the clinical decision.

2. The method as recited in claim 1 , wherein segmenting a plurality of mesh candidates for an anatomical model of an artery including a stenosis region of a patient from medical imaging data comprises:

generating the plurality of mesh candidates based on the one or more candidate locations.

3. The method as recited in claim 1 , wherein determining one or more candidate locations for each vertex of the anatomical model comprises:

determining the probability that a respective voxel is a boundary point for each voxel along a surface normal within a predetermined distance at each vertex.

4. The method as recited in claim 3 , wherein determining one or more candidate locations for each vertex of the anatomical model further comprises:

determining voxels associated with a probability above a threshold value as the one or more candidate locations for each vertex.

5. The method as recited in claim 2 , wherein generating the plurality of mesh candidates based on the one or more candidate locations comprises:

generating the plurality of mesh candidates by enforcing a plurality of connection rules for connecting the one or more candidate locations of neighboring vertices of the anatomical model, wherein each of the plurality of connection rules results in a corresponding mesh candidate.

6. The method as recited in claim 5 , wherein generating the plurality of mesh candidates based on the one or more candidate locations further comprises:

projecting each of the plurality of mesh candidates onto a learned shape space of the artery using an active shape model.

7. The method as recited in claim 1 , further comprising:

in response to determining that the clinical decision for treatment of the stenosis region of the patient is not ambiguous, displaying results of the hemodynamic index without receiving user input.

8. The method as recited in claim 1 , further comprising:

in response to determining that the clinical decision for treatment of the stenosis region of the patient is ambiguous:

displaying at least one of the plurality of mesh candidates; and

receiving user input to select and/or edit the at least one of the plurality of mesh candidates.

9. The method as recited in claim 8 , wherein displaying at least one of the plurality of mesh candidates comprises:

displaying each of the plurality of mesh candidates overlaid on a same multiplanar reformatted planes of the artery.

10. The method as recited in claim 8 , wherein displaying at least one of the plurality of mesh candidates comprises:

displaying each of the plurality of mesh candidates overlaid on an instance of a same multiplanar reformatted planes of the artery in a side-by-side configuration.

11. The method as recited in claim 8 , wherein displaying at least one of the plurality of mesh candidates comprises:

displaying mesh candidates of the plurality of mesh candidates having a value of the hemodynamic index for the stenosis region above the threshold on a first portion of a display; and

displaying mesh candidates of the plurality of mesh candidates having the value of the hemodynamic index for the stenosis region below the threshold on a second portion of the display.

12. The method as recited in claim 1 , wherein computing a hemodynamic index for the stenosis region in each of the plurality of mesh candidates comprises:

simulating blood flow and pressure in each of the plurality of mesh candidates for the artery of the patient; and

computing a fractional flow reserve value for the stenosis region in each of the plurality of mesh candidates based on the blood flow and pressure simulations.

13. An apparatus for non-invasive assessment of an arterial stenosis, comprising:

means for segmenting a plurality of mesh candidates for an anatomical model of an artery including a stenosis region of a patient from medical imaging data, the plurality of mesh candidates representing different anatomical models of the artery resulting from variations in segmenting techniques, wherein the means for segmenting the plurality of mesh candidates comprises:

means for determining one or more candidate locations for each vertex of the anatomical model,

means for assigning a confidence score to each of the plurality of mesh candidates by averaging a probability associated with each vertex in a respective mesh candidate, and

means for applying cluster analysis to reduce a number of the plurality of mesh candidates based on the confidence score;

means for computing a hemodynamic index for the stenosis region in each of the plurality of mesh candidates; and

means for determining whether a clinical decision for treatment of the stenosis region of the patient is ambiguous by determining whether a variation among values of the hemodynamic index for the stenosis region in each of the plurality of mesh candidates is entirely below a threshold for the clinical decision or is entirely above the threshold for the clinical decision.

14. The apparatus as recited in claim 13 , wherein the means for segmenting a plurality of mesh candidates for an anatomical model of an artery including a stenosis region of a patient from medical imaging data comprises:

means for generating the plurality of mesh candidates based on the one or more candidate locations.

15. The apparatus as recited in claim 13 , wherein the means for determining one or more candidate locations for each vertex of the anatomical model comprises:

means for determining the probability that a respective voxel is a boundary point for each voxel along a surface normal within a predetermined distance at each vertex.

16. The apparatus as recited in claim 15 , wherein the means for determining one or more candidate locations for each vertex of the anatomical model further comprises:

means for determining voxels associated with a probability above a threshold value as the one or more candidate locations for each vertex.

17. The apparatus as recited in claim 14 , wherein the means for generating the plurality of mesh candidates based on the one or more candidate locations comprises:

means for generating the plurality of mesh candidates by enforcing a plurality of connection rules for connecting the one or more candidate locations of neighboring vertices of the anatomical model, wherein each of the plurality of connection rules results in a corresponding mesh candidate.

18. The apparatus as recited in claim 17 , wherein the means for generating the plurality of mesh candidates based on the one or more candidate locations further comprises:

means for projecting each of the plurality of mesh candidates onto a learned shape space of the artery using an active shape model.

19. The apparatus as recited in claim 13 , further comprising:

means, responsive to determining that the clinical decision for treatment of the stenosis region of the patient is not ambiguous, for displaying results of the hemodynamic index without receiving user input.

20. The apparatus as recited in claim 13 , further comprising:

means, responsive to determining that the clinical decision for treatment of the stenosis region of the patient is ambiguous, for:

displaying at least one of the plurality of mesh candidates; and

receiving user input to select and/or edit the at feast one of the plurality of mesh candidates.

21. A non-transitory computer readable medium storing computer program instructions for non-invasive assessment of an arterial stenosis, the computer program instructions when executed by a processor cause the processor to perform operations comprising:

segmenting a plurality of mesh candidates for an anatomical model of an artery including a stenosis region of a patient from medical imaging data, the plurality of mesh candidates representing different anatomical models of the artery resulting from variations in segmenting techniques, wherein segmenting the plurality of mesh candidates comprises:

determining one or more candidate locations for each vertex of the anatomical model,

assigning a confidence score to each of the plurality of mesh candidates by averaging a probability associated with each vertex in a respective mesh candidate, and

applying cluster analysis to reduce a number of the plurality of mesh candidates based on the confidence score;

computing a hemodynamic index for the stenosis region in each of the plurality of mesh candidates; and

determining whether a clinical decision for treatment of the stenosis region of the patient is ambiguous by determining whether a variation among values of the hemodynamic index for the stenosis region in each of the plurality of mesh candidates is entirely below a threshold for the clinical decision or is entirely above the threshold for the clinical decision.

22. The non-transitory computer readable medium as recited in claim 21 , wherein segmenting a plurality of mesh candidates for an anatomical model of an artery including a stenosis region of a patient from medical imaging data comprises:

generating the plurality of mesh candidates based on the one or more candidate locations.

23. The non-transitory computer readable medium as recited in claim 21 , the operations further comprising:

in response to determining that the clinical decision for treatment of the stenosis region of the patient is not ambiguous, displaying results of the hemodynamic index without receiving user input.

24. The non-transitory computer readable medium as recited in claim 21 , the operations further comprising:

in response to determining that the clinical decision for treatment of the stenosis region of the patient is ambiguous:

displaying at least one of the plurality of mesh candidates; and

receiving user input to select and/or edit the at least one of the plurality of mesh candidates.

25. The non-transitory computer readable medium as recited in claim 24 , wherein displaying at least one of the plurality of mesh candidates comprises:

displaying each of the plurality of mesh candidates overlaid on a same multiplanar reformatted planes of the artery.

26. A method for supporting a clinical decision regarding an anatomical object of interest, comprising:

segmenting a plurality of mesh candidates for an anatomical model of an anatomical object of interest of a patient from medical imaging data, the plurality of mesh candidates representing different anatomical models of the artery resulting from variations in segmenting techniques, wherein segmenting the plurality of mesh candidates comprises:

determining one or more candidate locations for each vertex of the anatomical model,

assigning a confidence score to each of the plurality of mesh candidates by averaging a probability associated with each vertex in a respective mesh candidate, and

applying cluster analysis to reduce a number of the plurality of mesh candidates based on the confidence score;

computing a quantitative metric for each of the plurality of mesh candidates; and

determining whether a clinical decision is ambiguous by determining whether a variation among values of the quantitative metric for each of the plurality of mesh candidates is entirely below a threshold for the clinical decision or is entirely above the threshold for the clinical decision.