Neural network-based radiation treatment planning
A neural network is trained using a training corpus having a plurality of information features, each of the information features including both a reference radiation treatment dose and at least one corresponding post-treatment patient datum. By one approach, the at least one corresponding post-treatment patient datum comprises patient imagery such as, but not limited to, one or more computed tomography images. That trained neural network facilitates radiation treatment planning by generating resultant treatment efficacy probability information and resultant treatment complications probability information.
1 . A method to train a neural network configured to facilitate radiation treatment planning by generating resultant treatment efficacy probability information and resultant treatment complications probability information, the method comprising:
accessing a training corpus comprising a plurality of information features, each of the information features comprising both a reference radiation treatment dose and at least one corresponding post-treatment patient datum;
training the neural network using the training corpus;
wherein the neural network comprises a probabilistic model that uses probability theory to make predictions and to classify inputs.
2 . The method of claim 1 wherein the probabilistic model includes a variational latent space in a radiation dose prediction pipeline.
3 . The method of claim 1 wherein the neural network includes an encoder and a decoder that are combined as a U-Net shape structure which receives multi-channel radiation treatment input for a particular patient and outputs corresponding feature maps.
4 . The method of claim 1 wherein the neural network is configured to generate a prior latent space, wherein at least some positions in the prior latent space encode a corresponding radiation dose variant to provide a plurality of radiation dose variant codes.
5 . The method of claim 4 wherein the neural network is configured to output predicted radiation dose results from the plurality of radiation dose variant codes and feature maps.
6 . The method of claim 1 wherein the at least one corresponding post-treatment patient datum comprises patient imagery.
7 . The method of claim 6 wherein the patient imagery comprises at least one patient computed tomography image.
8 . The method of claim 1 wherein the plurality of information features includes information representing a plurality of different patients.
9 . The method of claim 1 wherein the plurality of information features includes information representing a plurality of different radiation treatment approaches.
10 . An apparatus comprising:
a control circuit configured as a neural network that facilitates radiation treatment planning by generating resultant treatment efficacy probability information and resultant treatment complications probability information, wherein the neural network was trained with a training corpus comprising a plurality of information features, each of the information features comprising both a reference radiation treatment dose and at least one corresponding post-treatment patient datum, and wherein the neural network comprises a probabilistic model that uses probability theory to make predictions and to classify inputs.
11 . The apparatus of claim 10 wherein the probabilistic model includes a variational latent space in a radiation dose prediction pipeline.
12 . The apparatus of claim 10 wherein the neural network includes an encoder and a decoder that are combined as a U-Net shape structure which receives multi-channel radiation treatment input for a particular patient and outputs corresponding feature maps.
13 . The apparatus of claim 10 wherein the neural network is configured to generate a prior latent space, wherein at least some positions in the prior latent space encode a corresponding radiation dose variant to provide a plurality of radiation dose variant codes.
14 . The apparatus of claim 13 wherein the neural network is configured to output predicted radiation dose results from the plurality of radiation dose variant codes and feature maps.
15 . The apparatus of claim 10 wherein the neural network is configured to receive input comprising at least one of computed tomography content, contouring/segmenting content, and radiation treatment platform geometry content.
16 . The apparatus of claim 15 wherein the neural network is configured to receive input comprising each of computed tomography content, contouring/segmenting content, and radiation treatment platform geometry content.
17 . The apparatus of claim 10 wherein the plurality of information features includes information representing a plurality of different patients.