Multileaf collimator tongue-and-groove effect on depth and off-axis doses: A comparison of treatment planning data with measurements and Monte Carlo calculations

Verification of the junctional dose for irradiation of the chest wall and supraclavicular regions under the circumstances of advanced technologies

A single-isocenter half-beam technique is commonly used when irradiating the chest wall and supraclavicular regions in patients with high-risk breast cancer. However, several studies have reported that underdosage can occur at the junction of the chest wall and supraclavicular regions due to a "tongue-and-groove" effect. This study verified the efficacy of an open leaf technique (OL-tech) that involves placing a multileaf collimator 5 mm outside from the beam central axis to remove the effect of the multileaf collimator in a single-isocenter half-beam technique. We compared the junction doses of the OL-tech with those of a conventional technique (C-tech) in square and clinical plans, using 4 and 10 MV x-rays in the Clinac iX and 6 and 10 MV x-rays in the Trilogy accelerators (Varian Medical Systems, Palo Alto, CA). EBT3 radiochromic films were used for measurements. Measurements were performed at a depth of 3 cm when verifying field matching. The EBT3 films in the square plan indicated junction doses for the C-tech of 78.3% with the Clinac iX accelerator and 73.6% with the Trilogy accelerator. By contrast, the corresponding doses for the OL-tech were 107.2% and 99.8%, respectively. In the clinical plan, the junction doses for the C-tech were 76.5% with the Clinac iX accelerator and 72.6% with the Trilogy accelerator; the corresponding doses for the OL-tech were 108.3% and 101.7%, respectively. As with the square plan, variations in the junction dose were much smaller using the OL-tech than using the C-tech. Our results suggest that the OL-tech can be useful for improving dose homogeneity at the junction of the chest wall and supraclavicular regions.

Commissioning of the tongue-and-groove modelling in treatment planning systems: from static fields to VMAT treatments

Adequate modelling of the multi-leaf collimator (MLC) by treatment planning systems (TPS) is essential for accurate dose calculations in intensity-modulated radiation-therapy. For this reason modern TPSs incorporate MLC characteristics such as the leaf end curvature, MLC transmission and the tongue-and-groove. However, the modelling of the tongue-and-groove is often neglected during TPS commissioning and it is not known how accurate it is. This study evaluates the dosimetric consequences of the tongue-and-groove effect for two different MLC models using both film dosimetry and ionisation chambers. A set of comprehensive tests are presented that evaluate the ability of TPSs to accurately model this effect in (a) static fields, (b) sliding window beams and (c) VMAT arcs. The tests proposed are useful for the commissioning of TPSs and for the validation of major upgrades. With the ECLIPSE TPS, relevant differences were found between calculations and measurements for beams with dynamic MLCs in the presence of the TG effect, especially for the High Definition MLC, small gap sizes and the 1 mm calculation grid. For this combination, dose differences as high as 10% and 7% were obtained for dynamic MLC gaps of 5 mm and 10 mm, respectively. These differences indicate inadequate modelling of the tongue-and-groove effect, which might not be identified without the proposed tests. In particular, the TPS tended to underestimate the calculated dose, which may require tuning of other configuration parameters in the TPS (such as the dosimetric leaf gap) in order to maximise the agreement between calculations and measurements in clinical plans. In conclusion, a need for better modelling of the MLC by TPSs is demonstrated, one of the relevant aspects being the tongue-and-groove effect. This would improve the accuracy of TPS calculations, especially for plans using small MLC gaps, such as plans with small target volumes or high complexities. Improved modelling of the MLC would also reduce the need for tuning parameters in the TPS, facilitating a more comprehensive configuration and commissioning of TPSs.