Method to track three-dimensional target motion with a dynamical multi-leaf collimator

• US 7,469,035 B2
• Filed: 12/11/2007
• Issued: 12/23/2008
• Est. Priority Date: 12/11/2006
• Status: Expired due to Fees

• Current Assignee: Varian Medical Systems Incorporated (Siemens AG), Board of Trustees of the Leland Stanford Junior University (Stanford Management Co.)
• Original Assignee: Varian Medical Systems Incorporated (Siemens AG), Board of Trustees of the Leland Stanford Junior University (Stanford Management Co.)
• Inventors: Keall, Paul J., Suh, Yelin, Povzner, Sergey, Sawant, Amit, Cattell, Herbert
• Primary Examiner(s): SONG, HOON K
• Application Number:
  US12/001,762
• Publication Number:
  US 20080159478A1
• Time in Patent Office: 378 Days
• Field of Search: 378/65, 378/205, 378/152
• US Class Current: 378/65
• CPC Class Codes: A61N 2005/1059 using cameras imaging the patient
• CPC Class Codes: A61N 2005/1061 using an x-ray imaging system having a separate imaging source
• CPC Class Codes: A61N 5/1037 taking into account the movement of the target, e.g. 4D-image based planning
• CPC Class Codes: A61N 5/1042 with spatial modulation of the radiation beam within the treatment head
• CPC Class Codes: A61N 5/1067 in real time, i.e. during treatment

Abstract

A method of continuous real-time monitoring and positioning of multi-leaf collimators during on and off radiation exposure conditions of radiation therapy to account for target motion relative to a radiation beam is provided. A prediction algorithm estimates future positions of a target relative to the radiation source. Target geometry and orientation are determined relative to the radiation source. Target, treatment plan, and leaf width data, and temporal interpolations of radiation doses are sent to the controller. Coordinates having an origin at an isocenter of the isocentric plane establish initial aperture end positions of the leaves that is provided to the controller, where motors to position the MLC midpoint aperture ends according to the position and target information. Each aperture end intersects a single point of a convolution of the target and the isocenter of the isocentric plane. Radiation source hold-conditions are provided according to predetermined undesirable operational and/or treatment states.

Claims

1. A method of continuous real-time monitoring of anatomy and continuous real-time beam positioning using multi-leaf collimators (MLC) during on and off radiation exposure conditions of radiation therapy treatment in order to account for target and anatomical motion relative to a radiation beam comprising:

a. using a prediction algorithm within a controller to estimate a future position of a target in relation to a radiation source reference frame;

b. determining a position, a size, a shape and an orientation of said target according to a frame of reference of said radiation source;

c. providing said determined target information to a controller;

d. determine an initial position of each leaf of said MLC;

e. provide a treatment plan to said controller from a treatment plan data file;

f. provide temporal interpolations of radiation doses and a respiratory phase according to said treatment plan, whereby predetermined positions of said leaves are provided that correlate to radiation dose segments along a total duration of a total irradiation of said target;

g. using Cartesian coordinates having an origin at an isocenter of an isocentric plane of said target to establish an initial position of aperture ends of said leaves;

h. providing said initial position information of said aperture ends to said controller;

i. determining an aperture end width and determining a midpoint of said aperture end for each said leaf;

j. providing said width and midpoint information to said controller, wherein said controller directs motors to position said MLC midpoint aperture ends according to said initial position information and said provided target information;

k. positioning each said aperture end midpoint to intersect a single point of a convolution of said target and said isocenter of said isocentric plane; and

l. providing radiation source hold conditions, wherein said hold condition is according to predetermined undesirable operational and/or treatment states.

2. The method of claim 1, further comprising calculating optimal radiation source jaw positions, wherein said jaw position optimization comprises;

a. using Cartesian coordinates in a plane that is transverse to said radiation source having an origin concentric to said source to determining initial positions of said jaw;

b. providing said jaw positions to said controller;

c. dynamically positioning exposure ends of said jaws, wherein said jaw ends are positioned proximal to said aperture ends of said MLC.

3. The method of claim 2, wherein said proximal position is from 0 to 20 said leaf widths along a first axis of said jaw coordinate system and is from 0 to a maximum travel range of said leaf along a second axis of said jaw coordinate system.

4. The method of claim 1, wherein said prediction algorithm is selected from a group comprising neural networks, autoregressive moving average models, Kalman filters, adaptive filters, modified linear adaptive filter, adaptive and Kalman filters.

5. The method of claim 1, wherein said target position, size and shape determination comprises:

a. determine an initial position of a target by monitoring at least one external target marker and/or at least one internal marker;

b. using a position prediction algorithm to estimate a future position of said target;

c. using a rotation matrix to calculate rotational changes in said position of said target relative to said radiation apertures;

d. using Cartesian coordinates along an MLC plane to calculate an X-offset and a Y-offset of said target, wherein said offsets result from movement of said target along a path of said radiation source; and

e. using Cartesian coordinates along an isocentric plane of said target to calculate an X-translation and a Y-translation of said target.

6. The method of claim 1, wherein said predetermined undesirable operational and/or treatment state is selected from a group comprising when said target moves below one or more said jaws, when said leaf aperture end intersects more than a single point of said convolution of said target and said isocenter of said isocentric plane, when no said apertures exist in said MLC plane, when any said leaves in a leaf pair overlap, and when said leaves are separated more that a maximum allowable distance.

7. The method of claim 1 further comprises dynamically moving MLC carriages in a direction perpendicular to leaf motion in order to track a principal component of motion along said direction perpendicular to said MLC leaf motion, whereby moving a plurality of MLC leaves in a manner to account for components of said target and anatomical motion along a direction of leaf travel.

8. The method of claim 1, wherein said radiation beam is a particle beam, whereas said continuous real-time beam positioning is performed using a particle actuator, whereby said particle actuator shapes said particle beam according to said convolution of said target and said isocenter of said isocentric plane.