Method to estimate 3D abdominal and thoracic tumor position to submillimeter accuracy using sequential x-ray imaging and respiratory monitoring
A method of estimating target motion for image guided radiotherapy (IGRT) systems is provided. The method includes acquiring by a kV imaging system sequential images of a target motion, computing by the kV imaging system from the sequential images an image-based estimation of the target motion expressed in a patient coordinate system, transforming by the kV imaging system the image-based estimation in the patient coordinate system to an estimate in a projection coordinate system, reformulating by the kV imaging system the projection coordinate system in a converging iterative form to force a convergence of the projection coordinate system to output a resolved estimation of the target motion, and displaying by the kV imaging system the resolved estimation of the target motion.
1. A method of estimating target motion for image guided radiotherapy (IGRT) systems, comprising:
a. acquiring by a kV imaging system sequential images of a target motion;
b. computing by said kV imaging system from said sequential images an image-based estimation of said target motion expressed in a patient coordinate system;
c. transforming by said kV imaging system said image-based estimation in said patient coordinate system to an estimate in a projection coordinate system;
d. reformulating by said kV imaging system said projection coordinate system in a converging iterative form to force a convergence of said projection coordinate system to output a resolved estimation of said target motion; and
e. displaying by said kV imaging system said resolved estimation of said target motion.
2. The method of claim 1 , wherein said target position T(x, y, z) comprises an x-direction corresponding to a left-right (LR) orientation in said patient coordinate system, a y-direction corresponding to a superior-inferior (SI) orientation in said patient coordinate system, and a z-direction corresponding to an anterior-posterior (AP) orientation in said patient coordinate system.
3. The method of claim 1 , wherein said imaging coordinate system comprises a kV source located at a kV source to axis distance (SAD), wherein said axis is a target isocenter, and an imager located at said SAD minus a source to imager distance (SID), wherein an origin of said imaging coordinate system is coincident with said target isocenter.
4. The method of claim 1 , wherein said kV imaging system is selected from the group consisting of an alternate stereoscopic sequential imaging system, a rotational monoscopic sequential imaging system, and a synchronous stereoscopic imaging system.
5. The method of claim 1 , wherein said transforming said image-based estimation in said patient coordinate system to an estimate in a projection coordinate system comprises:
a. sequentially applying a counter-clock wise rotation φ about the z-axis in said patient coordinate system and rotation of θ around the y-axis in said patient coordinate system;
b. applying a projection into the z′ direction of the imager coordinate system; and
c. performing a least squares estimate in said projection coordinate system using said suitably programmed computer.
6. The method of claim 1 , transforming said image-based estimation in said patient coordinate system to an estimate in a projection coordinate system comprises reformulating a least-squares estimation of target-surrogate model parameters from expressions in said patient coordinates to expressions in said projection coordinates.
7. The method of claim 1 , wherein said target position T(x, y, z) is continuously estimated using a signal R through the correlation between sequential signals R i and projected marker positions p i on intermittently acquired kV images, wherein said signal R is generated by optical respiratory sensors or accelerometers.