View radiation-modeling on GitHub

Project plan:

home

Started: 5/22/2023

Finished:

Contents

Long term goals

  • 400 total hours
  • 30 weeks
  • 12.5 hours a week

Project plan 1

(planned 6/22)

Phase 1 plan (10 weeks):

  • Create a model of a stationary x-ray LINAC and the radiation activity at different distances, amounts of time, and voltages.
  • Implement a method to convert that activity into exposure

Phase 2 (10 weeks):

  • Create a model of a person (for now, just one circular cross-section) with a tumor in the center
  • Modify the linac model to be able to rotate around the person
  • Record the exposure in each layer of the body

Phase 3 (10 weeks):

  • Add attenuation to the model. The beam will lose energy as it passes through layers.

Further work (the rest of the time):

  • Add functionality for the tumor to be anywhere in the cross-section, not just the center.
  • Make the person model a bit more realistic. Mark out the cross section with organs, bones, etc.
  • Implement a method to convert the exposure into effective dose. Record the effective dose in each of the different organs.
  • Repeat everything with proton therapy instead of x-ray therapy

Summary of phase 1 upon finishing:

Phase 1 summary

Project Plan 2

(planned 7/17)

This is the plan now that I’m basically finished with Phase 1:

Phase 2: Create 1D model of a radiation treatment (6 weeks)

  • Have the model shoot one beam of the correct energy of photons into one detector
  • Display the amount of dose the detector receives.
  • Create a graph of the dose profile, and the PDD.
  • Extra: flattening filter in the beam!

Phase 3: Start work on a 3DCRT model (6 weeks)

Basic requirements:

  • Model a realistic body part (pelvis, breast, etc) with multiple detectors
  • Have multiple beams of photons in different locations with the correct energies
  • Display data for how much dose each detector receives (tumor, important organs, etc)

Phase 4: Make the model better (10 weeks)

Ideas:

  • Allow for user to specify a prescription (e.g. 500 cGy) and have the model automatically stop when the prescription has been hit
  • Have the model shut off a beam when it gives too much to a sensitive organ like the heart
  • Allow to save the settings after a run including the beam locations and how much dose each beam distributes
  • Optimize for speed
  • Have the model do a couple of runs and keep the one that distributes the least dose to the sensitive organs
  • Or have the model shoot test particles all around and start the beam in the locations where the particles do the least damage.
  • Maybe do a run called, “calibration” that decides the best locations for beams, and saves the settings for future runs.
  • Allow for the model to work on multiple body parts
  • Maybe the user provides a certain file containing the information about the body part and the tumor, and then the program constructs it into a 3D model.

Project Plan 3

7/28

  • Phase 1: preparation and learning Geant4 (10 weeks)
  • Phase 2: Creating 1d-crt (10 weeks)
  • Phase 3: Creating 3d-crt (10 weeks)

Phase 2 in more detail:

3 weeks developing G4 Brems

  • G4 Brems is a Geant4 Bremsstrahlung simulation.
  • The purpose of this is to accurately model how a linac machine works
  • We will shoot a 1 mm wide beam of electrons into a tungsten target and utilize bremsstrahlung physics to create an accurate photon beam.
  • This will accurately produce data on energy distribution of photons, and might be used in the final g43dcrt project. 2 weeks researching

3 weeks developing 1D-CRT

  • 1D-CRT is a version of 3D-CRT but with one beam instead of multiple
  • The purpose is to accurately model the effect of one beam on a patient
  • We will either use G4 Brems for our beam, or just use the energy distribution found in G4 Brems

2 weeks finishing up everything

Records of time spent