Complexity metrics for Tomotherapy: application to the verification of transferred plans

To study the impact of transferring a treatment plan between two machines on the delivered dose, new complexity metrics based on Leaf Opening and Closing Times (LOT and LCT) were developed.

Basic tomotherapy treatment plans were created to identify the ranges of LOT that are accurately delivered and those that deviate from the planned values. Newly developed complexity metrics were derived from these ranges. These metrics from 103 clinical plans and 6 phantom plans were analyzed to identify treatment sites affected by short LCTs during plan transfer. A software tool was developed to estimate the additional dose associated with short LCTs, and these estimations were verified by measurements.

Opening time deviations were observed for short LOTs (below 50 ms) and short (but non-zero) LCTs (up to 40 ms). When a plan is created, no LOTs are generated in these regions. However, when a plan is transferred between two machines (e.g., in case of a breakdown), a fraction of LOTs equal to LOTmax (i.e., with no leaf closing) are converted into short LCTs. The additional dose due to these short LCTs could reach up to 3%. Affected treatment sites included pelvis, anal canal, and bone metastases. Measured dose discrepancies were consistent with estimations, with mean errors of 0.3% for phantom plans and 0.6% for clinical plans. Using an alternative transfer method, mean dose discrepancies decreased from 3% to less than 1% for all plans.

To date, transferring Tomotherapy plans increases LOT distribution complexity by creating short LCTs. The additional dose due to short LCTs is significant for plans with high gantry rotation speed and "hockey stick" shaped LOTs distributions. An alternative transfer method can reduce delivery errors by eliminating short LCTs.