Monte Carlo Model Tuning for an Independent Dose Calculation Software: Testing and Validation of Optimal Parameters in Clinical Radiotherapy Treatment Plans

The present thesis work was carried out at the AUSL-IRCCS Medical Physics Unit, Reggio Emilia, Italy, and focuses on activities related to Patient-Specific Quality Assurance (PSQA), particularly on dosimetric verification through independent dose calculation in radiotherapy. The dosimetric check plays a crucial role in this process by comparing the dose calculated by the Treatment Planning System (TPS) with that recalculated independently by software such as RadCalc, in order to detect clinically relevant discrepancies. The purpose of this study was to validate the RadCalc’s Monte Carlo calculation module (RCMC), with a specific focus on the Additional Radiation to Light Field Offset (ARLFO) parameter, which characterizes the multileaf collimator (MLC) dosimetric parameter in the clinical planning and corresponds to the Dosimetric Leaf Gap parameter (DLG)used in TPS. The optimal ARLFO value was derived through an analytical method, starting from the standard baseline of 0 cm, in order to maximize the agreement with phantom-based PSQA measurements. The percentage dose difference between the point dose measured at the central axis position (CAX) inside the dedicated water phantom and that calculated by RCMC was evaluated for several fields with varying MLC apertures, ranging from 2 mm to 20 mm. The procedure has been conducted for 6 MV (6X) beams energy and results demonstrated a linear dependence between the dose difference at the CXS and the ARLFO value for each field. From this linear relationship, the ARLFO value that theoretically yields a 0% dose difference was extrapolated and considered optimal. Since the sensitivity to ARLFO increases as the field size decreases, a weighted average of the optimal ARLFO value for each field was calculated, with weights based on the MLC field aperture. The resulting weighted value was–0.053 cm for 6X energy. Subsequently, a dataset of 26 clinical treatment plans was analyzed, involving different tumor sites (breast, lung, head and neck, brain). For each plan, measurements and RCMC dose calculations were compared using both standard ARLFO = 0.000 cm and the optimal ARLFO =-0.053 cm configurations. Agreement was assessed using the gamma index analysis, applying both the 3%/2 mm criterion with a 10% global dose threshold and 3%/3 mm with a 50% local dose threshold. The gamma index evaluates the level of agreement between two dose distributions by combining spatial and dosimetric tolerances. Results showed an average improvement of 6 percentage points in gamma passing rates when using the optimal ARLFO value, confirming the field-size sensitivity of the parameter. The analytical method developed in this study to determine the optimal ARLFO value was proved to be both feasible and straightforward to apply. It led to a notice able improvement in the agreement between the measured dose and that based on RCMC calculation. In the near future, this approach can be extended to other beam energies, with the specific aim to identify the most appropriate parameter values to integrate into routine clinical practice.