Physically-constrained modeling of a heart in medical imaging
Physically-constrained modeling of a heart is provided. Patient-specific data may be used to estimate heart anatomy locations. A model is applied to the data for estimation. For increased accuracy of estimation, the biomechanics of the heart, such as the valve, may be used to constrain the estimation. By applying a dynamic system between estimated anatomy locations of different times, the locations may be deformed or refined. The modeled heart and/or valve may be used to estimate hemodynamics. The resulting velocities or other motion information may be used to emulate ultrasound Doppler imaging for comparing with acquired ultrasound Doppler data. The comparison may validate the modeling.
1. A method for physically-constrained imaging a heart, the method comprising:
acquiring first and second medical diagnostic data representing a patient with a medical scanner;
estimating, with a processor, first anatomy locations of a valve of the patient from the first medical diagnostic data of the patient at a first time;
estimating, with the processor, second anatomy locations of the valve of the patient from the second medical diagnostic data of the patient at a second time different than the first time;
altering the first anatomy locations for the first time as a function of the second anatomy locations and a biomechanical model relating physical mechanics of the valve from the first and second times;
altering the second anatomy locations for the second time as a function of the first anatomy locations and the biomechanical model; and
generating an image as a function of the altered first or second anatomy locations.
2. The method of claim 1 wherein estimating the first and second anatomy locations comprises estimating from ultrasound data representing a volume including the valve over time.
3. The method of claim 1 wherein estimating the first and second anatomy locations comprises estimating by application of the first and second medical diagnostic data, respectively, as input features to a machine-learned matrix.
4. The method of claim 1 wherein estimating the first and second anatomy locations comprises estimating as a function of a discriminative probabilistic model.
5. The method of claim 1 wherein estimating the first and second anatomy locations comprises estimating locations for anterior and posterior papillary tips, mitral annulus, and anterior and posterior leaflets at the first and second times and estimating a mesh for the valve at the first and second times.
6. The method of claim 1 wherein altering the first and second anatomical locations comprises solving with the biomechanical model comprising a finite element model.
7. The method of claim 1 wherein altering the first and second anatomical locations comprises applying a force emulating a spring along a normal direction to the first and second anatomical locations and weighted by an amount of altering.
8. The method of claim 1 wherein altering the first and second anatomical locations comprises altering with the biomechanical model comprising a dynamic system having mass, damping, stiffness, displacement, velocity, and acceleration terms.
9. The method of claim 1 further comprising:
repeating the estimating and altering until convergence; and
repeating the estimating, altering and repeating of the estimating and altering for a third time.
10. The method of claim 1 further comprising:
modeling the heart, including the valve;
simulating hemodynamics from the modeling of the heart;
acquiring Doppler ultrasound data from the heart; and
emulating the Doppler ultrasound data from the simulated hemodynamics.