Modeling & optimization

External beam radiation treatment is a common approach to treat cancer with radiation. With development in technology especially with computer controlled multileaf collimators (MLC) era of Intensity Modulated Radiation Therapy (IMRT) dawned. IMRT distinguishes itself from the traditional 3D radiation therapy in that the intensity of the beam was modulated over a given aperture/ opening.

The intensity variation as a function of position (beamlets or bixels) in a given beam is in fact accomplished by changing the fluence or number of photons through the bixel. The energy of the beam remains the same. One can argue that the intensity change from bixel to bixel is achieved only by the fluence – the beam energy remains fixed. 

To a critical reader, it is not the optimization of intensity but optimization of fluence, therefore the term intensity modulation is a misnomer.

XMRT uses a simple formalism where two discrete megavoltage clinical photon energies 6 MV and 18 MV are used as part of decision making by the computer algorithm to determine which one is appropriate from a particular location for a custom patient, and tumor and normal tissue geometry. The investigation still continues with application from phantom geometries to real world scenarios. 

Low-dose-rate prostate brachytherapy is a common procedure for treating prostate cancer through interstitial implantation of small radioactive seeds throughout the prostate. Numerous regimens exist where radioactive seed implant has been used as either a monotherapy or in conjunction with external beam radiotherapy for the treatment of various stages of prostate cancer. Distribution of about 60-100 radioactive iodine seeds in prostate can be efficiently planned by using optimization algorithms with no user intervention.

Genetic algorithms are from the class of stochastic algorithm. They use concepts based on principles of Darwinian Natural Selection. Using probabilities on an initial population of n random seed distributions, operators such as elitism, roulette wheel selection, singlepoint crossover, and single-point mutation are applied.

A custom designed optimization algorithm developed by our group uses an objective function which calculates radiation dose and a fitness score for each of the seed distributions on a generation by generation basis.

This study from our group introduced mathematical algorithm for optimizing volumetric modulated arc therapy (VMAT) using mixed photon beams (MP-VMAT), rather than the traditional single photon beam. The algorithm uses a two-step approach: first, it optimizes dual energy intensity maps (DEIMs) for beam segments, and then it disperses these maps across the beam segments using a mixed integer linear programming (MILP) model.

The results showed that MP-VMAT could achieve better dose distribution and reduce radiation exposure to surrounding tissues like the rectum and bladder, compared to single energy VMAT, especially in prostate cancer cases. The study demonstrated the potential benefits of MP-VMAT compared to single photon energy beams. More details in Momin et al.