Adjoint transport for dose in beam angle optimization for external beam radiation therapy

1. A method of beam angle optimization for an IMRT radiotherapy treatment, the method comprising:

providing a patient model including one or more regions of interest (ROls) for the IMRT radiotherapy treatment;

defining a delivery coordinate space (DCS) having a set of candidate vertices, each respective candidate vertex defining a respective beam's eye view (BEV) plane;

for each respective ROI of the one or more ROls:

solving an adjoint transport to obtain an adjoint solution field from the respective ROI; and

for each respective candidate vertex in the DCS:

for each respective pixel of the respective BEV plane defined by the respective candidate vertex:

evaluating an adjoint photon fluence originating from a respective beamlet incident from the respective candidate vertex and passing through the respective pixel by performing ray tracing of the adjoint solution field; and

evaluating a respective dose of the respective ROI from the respective beamlet using the adjoint photon fluence;

for each respective candidate vertex in the DCS:

for each respective pixel of the respective BEV plane defined by the respective candidate vertex:

evaluating a respective BEV score of the respective pixel using the doses of the one or more ROls evaluated for the respective beamlet incident from the respective candidate vertex and passing through the respective pixel; and

determining one or more BEV regions in the respective BEV plane based on the BEV scores of the pixels of the BEV planes corresponding to the set of candidate vertices in the DCS;

determining a BEV region connectivity manifold based on the BEV regions of the BEV planes of the set of candidate vertices in the DCS, the BEV region connectivity manifold representing connections between contiguous BEV regions between adjacent vertices; and

determining a set of IMRT fields based on the BEV region connectivity manifold, each respective IMRT field of the set of IMRT fields defining a beam angle corresponding to a respective vertex in the DCS,

wherein the IMRT fields are represented as a set of nodes, and the beam angle optimization includes finding the set of nodes that have optimal max-distance,

wherein the method further comprises selecting order of beam delivery in the set of IMRT fields, the selecting including determining a total time the beam is off during transitioning from one IMRT field in the set to another IMRT field in the set, and determining the order in which the beam is delivered based on minimum total beam-off time,

wherein the total beam-off time includes a time for acquiring one or more images for guidance of treatment between beam delivery, and

wherein the method further comprises limiting beam delivery in the set of IMRT fields to beams that can be delivered in a time-efficient order.

2. The method of claim 1 , wherein each candidate vertex in the DCS comprises an isocenter position, a gantry angle, a couch angle, and a collimator angle.

3. The method of claim 1 , wherein determining the one or more BEV regions in the respective BEV plane defined by the respective candidate vertex comprises:

determining a threshold BEV score based on the BEV scores of the pixels of the BEV planes corresponding to the set of candidate vertices in the DCS; and

determining the one or more BEV regions in the respective BEV plane by comparing each respective BEV score of a respective pixel of the respective BEV plane to the threshold BEV score, wherein pixels within the one or more BEV regions have BEV scores greater than or equal to the threshold BEV score.

4. The method of claim 1 , wherein the one or more ROls include a plurality of ROls, and wherein evaluating the BEV score for each respective pixel of the BEV plane comprises evaluating a linear combination of the doses of the plurality of ROls.

5. The method of claim 4 , wherein each respective dose of a respective ROI of the plurality of ROls is given a respective weight in the linear combination.

6. The method of claim 5 , wherein the plurality of ROls includes one or more planning target volumes (PTVs) and one or more organs at risk (OARs).

7. The method of claim 6 , wherein the respective dose of a respective ROI that is one of the one or more PTVs has a positive weight, and the respective dose of a respective ROI that is one of the one or more OARs has a negative weight.

8. The method of claim 1 , wherein determining the set of IMRT fields comprises performing optimization using a max-distance function.

9. The method of claim 8 , wherein the optimization is performed using a local gradient descent algorithm based on information contained in the BEV region connectivity manifold.