Determining optimal spot delivery pattern using log file Derived dose discrepancy factor in pencil beam scanning...

Spatial uncertainties in spot position during reference dosimetry measurements of pencil beam scanning system were found to cause fluctuations in the reference dosimetry curve. This study aims to develop a method to determine the optimal delivery pattern for reference dosimetry measurement using information obtained from log file and quantifying the expected dose discrepancy.

A total of 7 different delivery pattern were delivered five times each and log files for each delivery were collected. We introduced a new metric known as dose discrepancy factor (DDF) to quantify the expected dose discrepancy in the measurement arising from spot positioning error. DDF is calculated using a simple pencil beam algorithm and actual spot positions from the log file. Type A and B uncertainties of DDF are quantified to draw statistically rigorous conclusion on the most optimal delivery pattern.

Relative dose difference of up to 2 % can be observed in some of the delivery pattern when reference dosimetry curve was plotted. Linear fit of DDF calculated from log file against measured dose ratio yields an R² of 0.56 which suggest moderate correlation. Delivery patterns with repainting and X fast scan (50MU) have the lowest average uncertainty. Friedman test and post-hoc test shows that different delivery patterns were proven to be statistical different with the exception of Y fast scan (50MU) versus both X fast scan (50MU) and random scan (50MU). Y fast scan (50MU) with the highest average DDF value of 0.979 ± 0.005 was found to be the most optimal delivery pattern as it has the least discrepancy from an ideal delivery.

Delivery patterns used during reference dosimetry had a non-negligible impact on measurement. We had detailed a log file-based approach to determine the optimal delivery pattern such that dose measurement would be least affected by spot positioning error.