Target Volume Predictive Model for Adaptive Radiation Therapy

Given an input of a time series of dimensional points P, the PredictiveModel function performs linear quadratic estimation (Kalman filtering) on each point, one dimension at a time, through all samples. Assumes input arguments maintains row correspondence for each point.

The output produced is:

• Python: Set of points in a .roi file defining a Kalman filtered volume
• MATLAB implementation: raw_data cell array defining a Kalman filtered volume

Mathematically:

Usage

MATLAB:

Runtime example can be executed with runPredictiveModel script from command.

PYTHON:

• Define your series of volumes similar to plan.roi.
• Place your plan.roi and predModel.py in the same directory.
• Run predModel.py.
• See output.txt or newartplan.roi for output.

Clinical application

This algorithm was designed as part of adaptive radiation therapy for clinical target volume determination. Here are some images of it in use:

License

Copyright 2015, 2016 Yazan Obeidi.

Use for educational purposes is permitted.