Construction of the medical linear accelerator model

The MCNP 5 Monte Carlo (MC) code was employed for the simulation of a 6 MV (Philips SL 75/5, Philips/Elekta, The Netherlands) medical LINAC. The simulation incorporated all the main beam modifying components of the LINAC head. In order to save computer time the simulation was performed in two steps [6]. Figure 1 shows the relative positions of the modeled components.

In the first step a 6 MeV electron beam impinged on a heavy metal target. The generated bremsstrahlung photons crossed the flattening filter and its holder and were tallied just under the flattening filter holder. Figure 2 shows the annular tallying surfaces and their position under the flattening filter holder. The number of photons crossing the tallying surface was recorded separately for thirty-two 250 KeV energy bins spanning an energy range from 0 to 8 MeV.

The spectra that were calculated in the first step of the simulation were re-emitted from a point source. The distance separating the point source from the simulated secondary collimator was the same with the distance between the actual secondary collimator and the upper surface of the heavy metal target. Each spectrum was emitted in the same angular aperture that it was measured by the annular tally rings. Figure 3 shows the setup used in the second step of the simulation.

Verification of the simulated beam's dosimetric properties

Percentage depth dose (PDD) and dose profiles measured on a water phantom (RFA-300, Scanditronix Wellhofer, Uppsala, Sweden) using an ion chamber (CC13-S, Scanditronix Wellhofer, Uppsala, Sweden) were compared to calculations obtained in simulated geometries including a water phantom. Figure 4 depicts the setup of the simulated detectors used inside the water phantom for the calculation of PDDs and dose profiles. Calculations were performed for a square 20 x 20 cm² field. Source to surface distance (SSD) was set to 100 cm for all actual measurements and simulations. Dose profiles were obtained at the depth of 10 cm (d₁₀) in the actual and the simulated water phantom as well.

The mathematical phantom

The BodyBuilder^TM software (White Rock Science, Los Alamos, New Mexico, USA) was employed for the generation of a mathematical phantom representing an average female patient. The phantom was modified in order to include all twenty-eight radiosensitive organs recently defined by the International Commission on Radiological Protection (ICRP) [7]. Lymph nodes [8] and salivary glands [9] were added to the phantom geometry. The dose imparted to the red bone marrow, bone surface, extrathoracic tissue, muscle and oral mucosa was approximated suitably by the dose imparted to nearby organs.

Breast cancer radiotherapy simulation

The second part of the LINAC geometry in conjunction with the mathematical phantom, were used for the simulation of breast radiation therapy. Each therapy consisted of irradiation with a lateral and a medial field of the same size. Medial and lateral fields with a dimension of 16 x 10 cm² and 20 x 10 cm² were defined in cooperation with an experienced radiation oncologist. The gantry angle was 286^o and 115^o for the medial and lateral field respectively, so that the inner field edges matched in the transverse plane of the phantom. The target tumor was assumed to lie at the center of the left breast and the prescribed tumor dose was set to 50.4 Gy for each therapy. Any lateral or medial projection contributed 50% to the total tumor dose. SSD was selected to be 100 cm for all projections. Figure 5 shows the therapy setup for the 10 x 16 cm² fields. Mean doses to all radiosensitive organs were calculated.