Dosimetric characteristics of the LIAC intraoperative radiotherapy beams: Assessment of sensitivity to measurement errors at the commissioning phase

Abstract

Commissioning of the LIAC intraoperative electron radiotherapy (IOERT) unit relies on the vendor's SWL-LIAC software. This software simulates the percentage depth-doses (PDDs), off-axis dose profiles and relative output factors (ROFs) for all beams based on a small set of input measurements. To date, the impact of various possible measurement errors during commissioning on LIAC dosimetric characteristics has not been studied in detail. We aimed to assess the sensitivity of the dosimetric characteristics of the ‘LIAC 12 MeV’ IOERT unit beams to various measurement errors in the commissioning phase by obtaining the magnitude of the influence of each error.The required SWL-LIAC inputs (PDDs of applicator diameters 3, 7, and 10 cm and open field) were measured with intentional applicator misalignments including 1° rotation or lateral (2 mm) or vertical (±2 mm) displacements. Also, inaccurate water surface definition for the detector was simulated by using PDDs that were shifted by ±2 mm. The errors were applied one at a time. To identify the most influential errors, the computed relative dosimetric characteristics resulting from such erroneous measurement conditions were compared to the corresponding ‘error-free’ ones. The depths of 90% and 50% doses (R90 and R50, respectively) were investigated further.All the resulting errors in R90, R50, flatness, symmetry and penumbra were within 3.9 mm, 2.8 mm, 4.3%, 4.6% and 3.5 mm, respectively. The overall averages of the absolute differences in these indices from the error-free fields were 0.8 mm, 0.8 mm, 0.6%, 0.4% and 0.5 mm, respectively. Errors in water surface definition for the detector had 4.4 to 22 times greater influence on PDDs than the other errors. The corresponding values for the profiles were 1.3–5.8 times. Among the errors in applicator alignment, PDDs and profiles were more affected by rotational and vertical ‘stand-in’ applicator misalignments, respectively. The R90 and R50 values were less affected by the errors in the small (3 and 4 cm diameter) and 45° bevelled applicators.To conclude, on average, the studied errors caused relatively small changes in beam characteristics although the maximum magnitudes of the observed errors were fairly substantial. Among the investigated errors, defining the water surface inaccurately has the largest impact on the dosimetric characteristics. Overall, sensitivity to the studied errors was found to be greatest with large-diameter flat applicators and/or lower electron energies. Although individual errors in applicator alignment have less effect on dosimetric characteristics compared to water surface error, given that these errors may occur simultaneously during dosimetry, special attention should be paid to placing the applicator correctly at the surface. Special set-up aids, such as our in-house tool for aligning applicators, are therefore indicated. The findings also further highlight the need for checking the water level periodically in long-term dosimetry procedures.