# Radiation treatment planning for IMRT

#### Inverse treatment planning

The contouring of planning target volumes and the region of interests (ROI) quite similar as in the case of conformal therapy, but very often we create the higher number of ROI, sometimes only for dose optimisation purpose. The beam number is between 5 and 9 with uniform equal arrangement around the patient.

Figure 1. : Workflow for IMRT treatment planning

We have to determine only the beam energy and the field geometry. The dose intensity profile and weight of the fields will be determined by optimization algorithms. On the base of conformal treatment planning and clinical experience we have to apply for planning target volumes and critical structures several dose-volume constraints. In general we are prescribing the dose (minimum, maximum or mean) of different target volumes or structures. In fact, the shape of dose-volume-histogram (DVH) is determined by these parameters. The priority of various conditions can be taken in account by relative weight factors.

Figure2. Dose-volume constraints for planning target volumes and critical structures for IMRT.

At the end of optimization we receive the intensity modulated fields in 2D, the dose distribution, which can be evaluated with conventional methods (on the base of DVH and isodose). If it is not convenient the whole procedure should be repeated with new dose volume limitation. Initially this is a complex process, but when we have a relevant experience, we can apply the treatment planning protocols for IMRT.
In inverse treatment planning of IMRT, the clinical objectives are specified mathematically in the form of an objective function. The objective function measures the goodness of a treatment plan, so the choice of the objective function is crucial for the optimization of a treatment plan. Usually, the objective function is a function of the beamlet intensities. Two types of objective functions are used: physical models and radiobiological models. Physical models are solely based on dose, while biological models argue that optimization should be based on the biological effects produced by the underlying dose distributions. A common method to express radiobiological objective functions is based on tumour control probabilities (TCP) and normal-tissue complication probabilities (NTCP).
Two type of optimisation algorithm can be applied: conjugated gradient approach and stochastic methods. In IMRT, an advanced conformal radiotherapy method, each beam is divided in a number of small beamlets (bixels). The intensity of each beamlet can individually be adjusted. A sparse dose matrix is precalculated and contains the dose value at each sampling point from each bixel with unit radiation intensity. The intensity (weight) of each beamlet has to be determined such that the produced dose distribution is “optimal”. This gradient method can use for the simple case. Stochastic optimization algorithms offer the advantage that they can find the optimal treatment parameters even for complex objective functions with potential local minima. The most commonly used method is the "simulated annealing”.