The cause of large differences in calculated beam monitor units between a treatment planning system and a QA software

Background/Introduction In radiotherapy, beam monitor units (MUs) calculated by a treatment planning system (TPS) are independently verified with an alternative calculation method, either manually calculated or with a separate commercial dose calculation algorithm product (henceforth called SDC for Secondary Dose Calculation). When there is a large discrepancy in MUs (>5%) between the TPS and SDC, the accurate MUs are unknown. This study aims to find the source of discrepancy and determine the accurate MUs for patient treatment. Methods We compared calculated MUs by Varian Eclipse (eMC, V.16), a Monte Carlo EGSnrc system, and an SDC on both a flat and curved 1 cm surface. The curved surface had a dip or rise relative to a flat surface ranging from 0.5 cm, to 2 cm over 4 cm width. Electron beam energies ranged from 6 to 18 MeV for 6×6 cm 2 , 10×10 cm 2 and 15×15 cm 2 fields. Results For both the dip and rise geometries, we found EGSnrc Monte Carlo calculated MUs generally agreed with Eclipse within 1-2%. For the 0.5 cm, 1 cm and 2 cm dip (rise) cases, the SDC overestimates (underestimates) MUs by 5-7% (3-8%), 11-15% (3-16%), and 25-31% (9-33%), respectively, depending on beam energy. Larger discrepancies occurred for lower energy beams. The amount of over- and under-estimation was independent of the field size when the field was larger than curvature width. Conclusions The investigation found that MUs calculated by the SDC are unreliable when the body surface is not flat and that SDCs can over- or under-estimated MUs by > 5% even at 0.5 cm dip or rise curvatures. The Eclipse calculated MUs are accurate for curvatures up to 2 cm dip (rise) over 4 cm width. These findings bring attention to practitioners to make informed decisions in choosing accurate MUs for patient treatment.