Well calculated is better than quickly calculated: Comparison of Pencil beam and Monte Carlo algorithms according to the number of lesions and fractionation in radiosurgery of multiple brain metastases

Abstract

PurposeIn this work we compared pencil beam (PB) and Monte Carlo (MC) algorithms in single isocenter plans of multiple brain metastases radiosurgery (SIMM-SRS) plans using the quality indices reported for SRS. MethodThe plans were evaluated concerning the prescribed dose, fractions and the number of metastases. The quality indices studied were mean dose (Dmean), D95, Paddick conformity index (PCI), Radiation Therapy Oncology Group (RTOG) homogeneity (HIRTOG) and quality of coverage indices (QRTOG), gradient index (GI), efficiency index for targets (Gη12Gy) and organs at risk (OARη12Gy) and V12-V18 for brain. ResultsThe D95 for plans calculated with PB algorithm increased and differences were statistically significant (p < 0.001). For Dmean no differences were observed (p > 0.194). The PCI for the single-fraction cases showed statistical significant differences (p < 0.039). The PCI for the three-fraction cases did not show statistical significant difference (p < 0.569). However, the mean value of the index for all cases did not differ significantly between PB (0.84) and MC (0.81). The GI showed statistically significant differences, only for the plans with more than 10 metastases for a single-fraction (p = 0.0001). The Gη12Gy values reported are within the interval of 0.26–0.80, and for all cases, there were no statistically significant differences. ConclusionConsidering that MC is more accurate for small volumes and heterogeneities, and computational time is reasonable for clinical use, it should be selected in all cases for SIMM-SRS plans. We introduced the potential of novel indices as Gη12Gy, and OARη12Gy for clinical evaluation that potentially serve as optimization factor.